id	sid	tid	token	lemma	pos
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	1	Building	build	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	2	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	3	State	state	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	4	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	6	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	8	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	9	Art	Art	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	10	Grammatical	Grammatical	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	11	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	12	Correction	Correction	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	13	System	System	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	14	Alla	Alla	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	15	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	16	Center	Center	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	17	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	18	Computational	Computational	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	19	Learning	Learning	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	20	Systems	Systems	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	21	Columbia	Columbia	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	22	University	University	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	23	New	New	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	24	York	York	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	26	NY	NY	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	27	10115	10115	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	28	alla@ccls.columbia.edu	alla@ccls.columbia.edu	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	29	Dan	Dan	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	30	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	31	Department	Department	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	32	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	33	Computer	Computer	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	34	Science	Science	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	35	University	University	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	36	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	37	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	38	Urbana	Urbana	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	39	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	40	IL	IL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	41	61801	61801	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	42	danr@illinois.edu	danr@illinois.edu	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	43	Abstract	Abstract	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	44	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	45	paper	paper	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	46	identifies	identifie	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	47	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	48	examines	examine	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	49	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	50	key	key	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	51	principles	principle	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	52	underlying	underlie	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	53	building	build	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	54	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	55	state	state	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	56	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	57	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	58	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	59	the-	the-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	60	art	art	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	61	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	62	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	63	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	64	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	1	65	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	2	do	do	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	4	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	5	analyzing	analyze	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	8	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	9	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	10	placed	place	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	11	first	first	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	12	among	among	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	13	seventeen	seventeen	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	14	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	15	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	17	re-	re-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	18	cent	cent	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	19	CoNLL-2013	conll-2013	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	20	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	21	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	22	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	23	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	24	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	25	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	2	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	2	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	3	focuses	focus	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	5	five	five	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	6	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	7	types	type	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	8	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	9	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	10	common	common	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	11	among	among	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	12	non	non	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	13	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	14	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	15	English	english	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	16	writers	writer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	3	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	2	describe	describe	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	3	four	four	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	4	design	design	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	5	principles	principle	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	6	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	7	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	8	relevant	relevant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	9	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	10	correcting	correct	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	11	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	12	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	13	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	14	er-	er-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	15	rors	ror	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	17	analyze	analyze	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	19	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	20	along	along	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	21	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	22	dimen-	dimen-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	23	sions	sion	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	24	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	25	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	26	show	show	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	27	how	how	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	28	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	29	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	30	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	31	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	32	contributes	contribute	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	33	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	34	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	35	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	4	36	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	1	1	1	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	2	Introduction	introduction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	3	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	4	field	field	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	6	text	text	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	7	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	8	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	9	seen	see	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	10	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	11	increased	increase	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	12	interest	interest	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	13	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	15	past	past	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	16	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	17	years	year	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	19	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	20	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	21	focus	focus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	22	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	23	correcting	correct	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	24	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	25	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	26	made	make	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	27	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	28	English	English	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	29	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	30	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	31	Second	Second	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	32	Language	Language	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	33	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	34	ESL	ESL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	35	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	36	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	5	37	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	1	Three	three	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	2	competi-	competi-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	3	tions	tion	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	4	devoted	devote	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	5	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	6	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	7	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	8	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	9	non	non	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	10	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	11	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	12	writ-	writ-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	13	ers	ers	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	14	took	take	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	15	place	place	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	16	recently	recently	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	17	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	18	HOO-2011	HOO-2011	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	19	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	20	Dale	Dale	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	21	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	22	Kil-	Kil-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	23	garriff	garriff	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	24	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	25	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	26	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	28	HOO-2012	HOO-2012	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	29	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	30	Dale	Dale	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	31	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	32	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	33	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	34	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	35	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	36	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	37	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	38	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	39	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	40	CoNLL-2013	conll-2013	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	41	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	42	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	43	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	44	Ng	Ng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	45	et	et	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	46	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	47	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	48	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	49	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	50	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	6	51	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	2	most	most	RBS
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	3	recent	recent	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	4	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	5	most	most	RBS
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	6	prominent	prominent	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	7	among	among	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	8	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	11	CoNLL-2013	conll-2013	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	12	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	13	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	15	covers	cover	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	16	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	17	common	common	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	18	ESL	ESL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	19	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	21	including	include	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	22	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	23	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	24	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	25	usage	usage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	26	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	28	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	29	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	30	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	31	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	32	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	33	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	34	various	various	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	35	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	36	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	37	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	38	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	39	illustrated	illustrate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	40	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	41	Fig	Fig	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	7	42	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	1	1.1	1.1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	2	Seventeen	seventeen	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	3	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	4	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	5	1The	1the	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	6	CoNLL-2014	CoNLL-2014	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	7	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	8	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	9	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	10	completed	complete	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	11	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	13	time	time	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	14	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	15	writing	write	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	16	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	17	paper	paper	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	18	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	19	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	20	extension	extension	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	21	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	22	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	23	CoNLL-2013	conll-2013	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	24	com-	com-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	25	petition	petition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	26	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	27	Ng	Ng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	28	et	et	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	29	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	30	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	31	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	32	2014	2014	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	33	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	34	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	35	addressed	address	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	36	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	37	types	type	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	38	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	39	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	8	40	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	3	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	4	Columbia	Columbia	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	5	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	7	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	8	slightly	slightly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	9	extended	extend	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	10	version	version	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	13	Nowadays	Nowadays	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	14	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	15	phone	phone	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	16	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	17	phones	phone	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	18	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	19	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	20	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	21	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	22	many	many	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	23	functionalities	functionality	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	24	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	25	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	26	included	include	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	27	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	28	including	include	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	29	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	30	∅/a	∅/a	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	31	camera	camera	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	32	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	33	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	34	∅/a	∅/a	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	35	Wi	Wi	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	36	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	37	Fi	Fi	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	38	receiver	receiver	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	9	39	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	10	1	Figure	figure	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	10	2	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	10	3	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	10	4	Examples	example	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	10	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	10	6	representative	representative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	10	7	ESL	ESL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	10	8	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	10	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	1	participated	participate	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	2	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	4	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	5	developed	develop	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	6	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	7	wide	wide	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	8	array	array	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	10	ap-	ap-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	11	proaches	proache	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	12	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	13	include	include	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	14	discriminative	discriminative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	15	classifiers	classifier	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	17	lan-	lan-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	18	guage	guage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	19	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	21	statistical	statistical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	22	machine	machine	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	23	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	24	translation	translation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	25	sys-	sys-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	26	tems	tem	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	28	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	29	rule	rule	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	30	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	31	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	32	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	11	33	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	1	Many	many	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	2	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	4	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	5	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	6	made	make	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	7	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	8	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	9	linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	10	resources	resource	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	11	such	such	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	12	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	13	addi-	addi-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	14	tional	tional	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	15	annotated	annotate	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	16	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	17	corpora	corpora	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	19	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	20	defined	define	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	21	high-	high-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	22	level	level	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	23	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	24	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	25	take	take	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	26	into	into	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	27	account	account	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	28	syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	29	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	30	se-	se-	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	31	mantic	mantic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	32	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	12	33	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	1	Even	even	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	2	though	though	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	4	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	5	incorporated	incorporate	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	6	similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	7	re-	re-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	8	sources	source	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	11	scores	score	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	12	varied	vary	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	13	widely	widely	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	13	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	2	top	top	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	3	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	5	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	7	University	University	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	8	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	9	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	11	obtained	obtain	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	12	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	13	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	14	score	score	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	15	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	16	31.202	31.202	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	18	while	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	20	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	21	team	team	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	22	scored	score	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	23	25.01	25.01	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	24	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	25	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	26	median	median	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	27	result	result	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	28	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	29	8.48	8.48	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	30	points.3	points.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	31	These	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	32	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	33	suggest	suggest	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	34	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	35	there	there	EX
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	36	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	37	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	38	enough	enough	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	39	understanding	understanding	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	40	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	41	what	what	WP
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	42	works	work	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	43	best	well	RBS
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	44	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	45	what	what	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	46	elements	element	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	47	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	48	essential	essential	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	49	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	50	building	build	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	51	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	52	state	state	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	53	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	54	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	55	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	56	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	57	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	58	art	art	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	59	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	60	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	61	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	14	62	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	2	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	3	paper	paper	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	5	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	6	identify	identify	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	7	key	key	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	8	principles	principle	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	9	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	10	build-	build-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	11	ing	ing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	12	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	13	robust	robust	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	14	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	15	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	16	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	17	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	18	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	19	show	show	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	20	their	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	21	importance	importance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	22	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	23	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	24	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	25	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	26	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	27	shared	shared	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	28	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	15	29	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	2	do	do	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	4	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	5	analyzing	analyze	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	8	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	9	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	10	evaluating	evaluate	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	11	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	12	along	along	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	13	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	14	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	15	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	16	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	17	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	18	CoNLL-2013	CoNLL-2013	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	19	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	21	ranked	rank	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	22	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	23	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	24	top	top	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	16	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	2	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	3	descrip-	descrip-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	4	tion	tion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	8	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	9	Columbia	Columbia	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	10	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	12	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	13	refer	refer	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	15	reader	reader	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	16	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	17	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	18	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	19	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	17	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	18	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	18	2	2014a	2014a	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	18	3	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	18	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	1	2The	2the	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	2	state	state	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	3	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	5	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	7	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	8	art	art	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	9	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	12	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	13	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	14	dis-	dis-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	15	cussed	cuss	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	16	here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	17	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	18	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	19	respect	respect	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	20	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	21	individual	individual	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	22	components	component	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	23	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	24	differ-	differ-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	25	ent	ent	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	26	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	19	27	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	1	Improvements	Improvements	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	2	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	3	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	4	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	5	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	6	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	7	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	8	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	9	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	11	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	12	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	13	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	14	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	15	due	due	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	16	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	17	joint	joint	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	18	learning	learning	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	19	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	20	inference	inference	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	21	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	22	orthogonal	orthogonal	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	24	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	25	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	26	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	27	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	28	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	29	paper	paper	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	30	still	still	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	31	applies	apply	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	32	there	there	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	20	33	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	1	3F1	3F1	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	2	might	may	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	3	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	4	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	6	ideal	ideal	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	7	metric	metric	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	8	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	9	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	10	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	11	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	12	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	13	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	15	one	one	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	16	chosen	choose	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	17	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	19	evaluation	evaluation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	21	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	22	1	See	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	22	2	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	22	3	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	22	4	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	22	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	23	1	6	6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	23	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	1	419	419	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	2	Transactions	transaction	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	5	Association	Association	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	6	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	7	Computational	Computational	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	8	Linguistics	Linguistics	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	10	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	11	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	12	2014	2014	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	13	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	14	419–434	419–434	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	24	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	25	1	Action	Action	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	25	2	Editor	Editor	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	25	3	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	25	4	Alexander	Alexander	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	25	5	Koller	Koller	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	25	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	26	1	Submitted	submit	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	26	2	10/2013	10/2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	26	3	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	26	4	Revised	revise	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	26	5	6/2014	6/2014	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	26	6	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	26	7	Published	publish	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	26	8	10/2014	10/2014	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	26	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	27	1	c	c	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	27	2	©	©	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	27	3	2014	2014	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	27	4	Association	Association	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	27	5	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	27	6	Computational	Computational	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	27	7	Linguistics	Linguistics	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	27	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	1	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	2	learning	learning	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	3	algorithm	algorithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	4	;	;	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	5	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	7	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	8	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	9	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	10	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	11	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	12	annotated	annotated	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	13	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	14	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	15	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	16	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	17	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	18	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	19	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	20	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	21	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	22	made	make	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	23	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	24	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	25	writers	writer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	26	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	27	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	28	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	29	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	30	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	31	linguis-	linguis-	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	32	tic	tic	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	33	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	28	34	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	2	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	3	dimension	dimension	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	5	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	6	imple-	imple-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	7	mentations	mentation	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	8	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	9	compared	compare	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	11	including	include	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	13	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	14	possible	possible	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	16	approaches	approach	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	17	chosen	choose	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	18	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	19	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	20	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	29	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	2	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	3	vali-	vali-	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	4	date	date	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	6	obtained	obtain	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	7	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	8	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	9	another	another	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	10	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	11	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	30	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	1	Overall	overall	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	4	paper	paper	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	5	makes	make	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	6	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	7	contributions	contribution	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	8	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	9	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	10	1	1	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	11	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	12	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	13	explain	explain	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	15	success	success	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	18	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	19	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	21	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	22	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	23	2	2	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	24	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	25	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	26	provide	provide	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	27	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	28	understanding	understanding	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	29	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	30	qualitative	qualitative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	31	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	32	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	33	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	34	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	35	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	36	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	37	essential	essential	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	38	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	39	success	success	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	40	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	41	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	42	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	43	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	44	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	45	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	46	goal	goal	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	47	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	48	aiding	aid	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	49	future	future	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	50	research	research	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	51	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	52	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	31	53	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	1	Given	give	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	2	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	4	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	5	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	6	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	7	been	be	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	9	top	top	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	10	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	11	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	12	four	four	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	13	competitive	competitive	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	14	evaluations	evaluation	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	15	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	17	last	last	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	18	few	few	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	19	years	year	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	20	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	21	HOO	HOO	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	22	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	23	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	24	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	26	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	27	believe	believe	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	28	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	29	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	30	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	31	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	32	propose	propose	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	33	will	will	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	34	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	35	useful	useful	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	36	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	37	researchers	researcher	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	38	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	39	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	40	area	area	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	32	41	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	3	next	next	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	4	section	section	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	6	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	7	present	present	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	9	CoNLL-2013	conll-2013	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	10	competition	competition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	33	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	34	1	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	34	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	1	3	3	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	2	gives	give	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	3	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	4	overview	overview	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	7	ap-	ap-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	8	proaches	proache	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	9	adopted	adopt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	10	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	12	top	top	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	13	five	five	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	14	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	35	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	36	1	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	36	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	37	1	4	4	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	37	2	de-	de-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	37	3	scribes	scribe	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	37	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	37	5	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	37	6	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	37	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	38	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	38	2	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	38	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	1	5	5	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	4	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	8	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	9	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	10	presented	present	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	39	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	40	1	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	40	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	41	1	6	6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	41	2	offers	offer	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	41	3	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	41	4	brief	brief	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	41	5	discussion	discussion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	41	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	41	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	41	8	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	41	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	42	1	7	7	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	42	2	concludes	conclude	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	42	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	42	4	paper	paper	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	42	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	1	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	2	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	3	Description	description	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	4	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	5	CoNLL-2013	conll-2013	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	6	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	7	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	8	focuses	focus	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	9	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	10	five	five	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	11	common	common	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	12	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	13	made	make	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	14	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	15	ESL	ESL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	16	writers	writer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	17	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	18	arti-	arti-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	19	cle	cle	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	20	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	21	determiner	determiner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	23	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	24	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	25	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	26	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	28	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	29	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	30	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	31	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	32	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	43	33	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	2	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	3	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	6	shared	shared	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	7	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	8	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	10	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	11	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	12	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	13	Dahlmeier	Dahlmeier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	14	et	et	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	15	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	16	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	18	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	19	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	21	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	22	contains	contain	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	23	essays	essay	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	24	written	write	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	25	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	26	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	27	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	28	English	English	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	29	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	30	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	31	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	32	refer	refer	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	33	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	34	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	35	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	36	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	37	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	38	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	39	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	40	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	41	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	42	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	43	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	44	44	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	2	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	3	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	4	consists	consist	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	6	50	50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	7	essays	essay	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	8	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	9	students	student	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	10	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	12	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	13	linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	14	back-	back-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	15	ground	ground	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	45	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	2	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	5	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	6	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	7	contain	contain	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	8	1.2	1.2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	9	M	m	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	11	29	29	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	12	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	13	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	15	respectively	respectively	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	46	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	1	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	2	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	3	shows	show	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	5	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	7	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	8	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	9	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	12	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	13	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	47	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	1	Determiner	determiner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	2	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	5	most	most	RBS
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	6	com-	com-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	7	mon	mon	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	9	account	account	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	10	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	11	42.1	42.1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	12	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	13	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	14	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	15	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	16	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	17	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	48	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	1	Note	note	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	2	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	4	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	5	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	6	contains	contain	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	7	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	8	much	much	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	9	larger	large	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	10	pro-	pro-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	11	portion	portion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	12	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	13	annotated	annotated	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	14	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	15	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	49	16	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	1	determiner	determiner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	2	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	3	occur	occur	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	4	four	four	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	5	times	time	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	6	more	more	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	7	often	often	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	10	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	11	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	12	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	13	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	15	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	16	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	17	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	18	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	19	2.4	2.4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	20	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	21	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	22	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	23	phrases	phrase	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	24	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	25	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	26	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	27	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	28	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	29	determiner	determiner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	30	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	31	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	32	versus	versus	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	33	10	10	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	34	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	35	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	36	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	37	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	38	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	39	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	50	40	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	2	differences	difference	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	3	might	may	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	4	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	5	attributed	attribute	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	6	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	7	differences	difference	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	9	annotation	annotation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	10	standards	standard	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	12	annotators	annotator	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	14	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	15	writers	writer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	17	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	19	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	20	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	21	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	22	annotated	annotate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	23	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	24	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	25	later	later	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	26	time	time	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	51	27	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	2	shared	shared	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	3	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	4	provided	provide	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	5	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	6	sets	set	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	8	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	9	an-	an-	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	10	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	11	Number	Number	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	12	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	13	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	14	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	15	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	16	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	17	Train	train	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	18	Test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	19	Art	art	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	52	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	1	6658	6658	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	3	2.4	2.4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	4	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	6	690	690	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	8	10.0	10.0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	9	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	10	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	11	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	53	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	1	2404	2404	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	3	2.0	2.0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	4	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	6	311	311	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	8	10.7	10.7	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	9	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	10	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	11	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	12	3779	3779	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	13	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	14	1.6	1.6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	15	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	16	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	17	396	396	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	18	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	19	6.0	6.0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	20	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	21	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	22	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	23	agr	agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	54	24	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	1	1527	1527	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	3	2.0	2.0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	4	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	6	124	124	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	8	5.2	5.2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	9	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	10	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	11	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	12	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	13	1453	1453	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	14	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	15	0.8	0.8	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	16	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	17	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	18	122	122	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	19	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	20	2.5	2.5	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	21	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	22	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	23	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	24	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	25	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	26	Statistics	statistic	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	27	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	28	annotated	annotate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	29	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	30	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	31	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	32	CoNLL-2013	CoNLL-2013	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	33	shared	share	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	34	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	35	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	55	36	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	56	1	Percentage	percentage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	56	2	denotes	denote	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	56	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	56	4	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	56	5	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	56	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	56	7	i.e.	i.e.	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	1	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	2	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	4	erroneous	erroneous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	5	instances	instance	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	6	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	7	respect	respect	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	8	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	10	total	total	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	11	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	12	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	13	relevant	relevant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	14	instances	instance	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	15	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	17	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	57	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	1	notations	notation	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	2	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	4	original	original	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	5	annotated	annotate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	6	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	8	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	9	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	10	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	11	additional	additional	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	12	revisions	revision	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	13	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	14	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	15	includes	include	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	16	alternative	alternative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	17	an-	an-	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	18	notations	notation	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	19	proposed	propose	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	20	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	21	participants	participant	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	58	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	1	Clearly	clearly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	3	having	have	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	4	alternative	alternative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	5	answers	answer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	6	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	8	right	right	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	9	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	10	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	11	there	there	EX
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	12	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	13	typically	typically	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	14	multiple	multiple	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	15	ways	way	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	16	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	17	correct	correct	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	18	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	19	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	59	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	1	How-	How-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	2	ever	ever	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	4	because	because	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	6	alternatives	alternative	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	7	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	8	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	9	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	11	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	12	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	13	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	15	participating	participate	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	16	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	19	revised	revise	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	20	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	21	may	may	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	22	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	23	biased	biased	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	24	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	25	Ng	Ng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	26	et	et	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	27	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	28	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	29	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	30	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	31	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	60	32	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	1	Consequently	consequently	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	3	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	4	report	report	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	5	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	6	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	8	original	original	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	9	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	61	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	1	3	3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	2	Model	Model	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	3	Dimensions	Dimensions	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	4	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	5	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	6	summarizes	summarize	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	7	approaches	approach	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	9	methodologies	methodology	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	12	top	top	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	13	five	five	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	14	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	62	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	2	prevailing	prevailing	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	3	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	4	consists	consist	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	5	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	6	building	build	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	7	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	8	statistical	statistical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	9	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	10	either	either	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	11	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	12	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	13	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	14	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	15	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	16	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	17	much	much	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	18	larger	large	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	19	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	20	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	21	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	22	En-	En-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	23	glish	glish	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	24	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	63	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	2	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	3	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	5	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	6	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	7	make	make	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	8	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	11	Web	web	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	12	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	13	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	14	5-gram	5-gram	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	15	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	16	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	17	henceforth	henceforth	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	18	Web1	web1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	19	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	21	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	22	Brants	Brants	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	23	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	24	Franz	Franz	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	26	2006	2006	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	27	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	28	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	64	29	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	1	NARA	NARA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	2	employs	employ	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	3	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	4	statis-	statis-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	5	tical	tical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	6	machine	machine	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	7	translation	translation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	8	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	9	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	10	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	11	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	12	types	type	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	13	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	14	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	15	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	16	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	17	rule	rule	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	18	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	19	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	20	components	component	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	21	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	22	se-	se-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	23	lected	lecte	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	24	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	65	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	1	Based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	2	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	4	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	8	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	10	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	11	identify	identify	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	13	following	follow	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	15	inter	inter	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	16	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	17	dependent	dependent	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	19	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	20	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	21	will	will	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	22	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	23	examined	examine	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	24	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	25	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	26	work	work	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	27	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	28	1	1	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	66	29	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	1	Learning	learn	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	2	algorithm	algorithm	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	3	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	4	Most	Most	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	7	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	9	includ-	includ-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	10	ing	ing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	11	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	13	built	build	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	14	statistical	statistical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	15	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	67	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	2	show	show	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	3	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	5	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	8	learning	learn	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	9	algorithm	algorithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	10	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	11	very	very	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	12	impor-	impor-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	13	tant	tant	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	14	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	15	affects	affect	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	17	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	18	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	20	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	68	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	69	1	2	2	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	69	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	1	Adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	2	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	3	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	6	Previous	previous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	7	stud-	stud-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	8	ies	ie	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	70	10	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	2	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	4	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	6	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	7	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	8	showed	show	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	9	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	10	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	71	12	i.e.	i.e.	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	1	developing	develop	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	2	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	3	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	4	utilize	utilize	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	5	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	6	about	about	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	7	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	8	patterns	pattern	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	11	non	non	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	12	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	13	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	14	writers	writer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	16	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	17	extremely	extremely	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	18	important	important	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	72	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	2	summarize	summarize	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	3	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	4	techniques	technique	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	5	proposed	propose	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	6	earlier	early	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	8	examine	examine	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	9	their	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	10	impact	impact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	11	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	13	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	14	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	16	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	73	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	74	1	3	3	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	74	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	1	Linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	2	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	3	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	4	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	6	essential	essential	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	7	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	8	use	use	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	9	some	some	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	10	linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	11	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	12	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	13	developing	develop	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	14	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	15	correc-	correc-	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	16	tion	tion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	17	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	19	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	21	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	22	identify	identify	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	23	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	24	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	25	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	26	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	27	420	420	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	28	System	System	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	29	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	30	Approach	Approach	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	31	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	32	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	33	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	34	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	35	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	36	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	37	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	38	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	39	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	40	Art	art	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	75	41	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	1	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	2	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	3	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	4	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	5	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	6	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	8	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	10	shallow	shallow	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	11	parse	parse	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	12	features	feature	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	13	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	76	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	1	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	2	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	3	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	5	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	6	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	8	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	9	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	10	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	11	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	12	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	13	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	14	Agr./Form	Agr./Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	15	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	16	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	17	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	18	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	19	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	20	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	21	NTHU	NTHU	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	22	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	23	Kao	Kao	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	24	et	et	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	25	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	26	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	28	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	29	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	30	All	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	31	Count	Count	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	32	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	33	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	34	backoff	backoff	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	35	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	36	on	on	RP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	37	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	38	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	39	HIT	hit	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	40	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	41	Xiang	Xiang	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	42	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	43	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	44	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	45	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	46	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	47	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	48	Art./Prep./Noun	art./prep./noun	ADD
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	49	ME	ME	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	50	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	51	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	52	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	53	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	54	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	55	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	56	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	57	dependency	dependency	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	58	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	59	Agr./Form	Agr./Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	60	Rule	Rule	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	61	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	62	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	63	NARA	NARA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	64	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	65	Yoshimoto	Yoshimoto	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	66	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	67	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	68	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	69	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	70	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	71	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	72	Art./Prep	Art./Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	77	73	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	1	SMT	SMT	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	2	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	3	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	5	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	6	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	7	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	8	Lang-8	Lang-8	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	9	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	10	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	11	ME	ME	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	12	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	13	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	14	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	15	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	16	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	18	POS	pos	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	19	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	20	dependency	dependency	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	21	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	22	Agr./Form	Agr./Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	23	Treelet	Treelet	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	24	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	25	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	26	Gigaword	Gigaword	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	27	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	28	Penn	Penn	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	29	TreeBank	TreeBank	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	30	corpora	corpora	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	31	UMC	UMC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	32	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	33	Xing	Xing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	34	et	et	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	35	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	36	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	37	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	38	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	39	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	40	Art./Prep	Art./Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	78	41	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	1	Two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	2	LMs	lm	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	3	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	5	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	6	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	7	Web1	web1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	8	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	9	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	10	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	11	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	12	voting	vote	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	13	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	14	Rules	Rules	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	15	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	16	ME	ME	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	17	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	18	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	19	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	20	+	+	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	21	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	22	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	23	on	on	RP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	24	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	25	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	26	Agr./Form	Agr./Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	27	ME	ME	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	28	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	29	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	30	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	31	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	32	agr	agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	33	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	79	34	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	1	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	2	rules	rule	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	3	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	4	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	6	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	7	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	8	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	9	Top	top	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	10	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	11	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	13	CoNLL-2013	CoNLL-2013	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	14	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	15	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	80	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	2	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	3	column	column	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	4	indicates	indicate	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	6	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	7	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	8	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	10	third	third	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	11	column	column	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	12	describes	describe	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	14	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	15	adopted	adopt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	16	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	18	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	81	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	1	ME	ME	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	2	stands	stand	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	3	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	4	Maximum	Maximum	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	5	Entropy	Entropy	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	6	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	7	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	8	stands	stand	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	9	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	10	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	11	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	12	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	13	SMT	SMT	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	14	stands	stand	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	15	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	16	Statistical	Statistical	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	17	Machine	Machine	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	18	Translation	Translation	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	19	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	20	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	21	stands	stand	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	22	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	23	Averaged	Averaged	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	24	Perceptron	Perceptron	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	25	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	26	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	27	stands	stand	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	28	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	29	Naı̈ve	naı̈ve	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	30	Bayes	Bayes	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	82	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	83	1	Classifier	Classifier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	83	2	Art	art	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	83	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	84	1	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	84	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	85	1	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	85	2	Agr	Agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	85	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	1	Form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	2	Train	train	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	3	254	254	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	4	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	5	103	103	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	6	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	7	240	240	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	8	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	9	75	75	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	10	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	11	175	175	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	12	K	K	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	13	Test	Test	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	14	6	6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	15	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	16	2.5	2.5	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	17	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	18	2.6	2.6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	19	K	K	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	20	2.4	2.4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	21	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	22	4.8	4.8	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	23	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	24	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	25	3	3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	26	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	27	Number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	28	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	29	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	30	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	31	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	32	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	33	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	86	34	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	1	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	2	occurs	occur	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	3	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	4	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	5	given	give	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	6	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	8	before	before	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	10	appropri-	appropri-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	11	ate	eat	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	12	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	13	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	14	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	15	employed	employ	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	87	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	2	describe	describe	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	4	evaluate	evaluate	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	6	contribution	contribution	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	8	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	9	elements	element	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	88	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	89	1	4	4	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	89	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	1	Training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	2	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	3	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	4	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	5	discuss	discuss	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	7	advantages	advantage	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	8	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	9	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	10	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	11	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	12	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	13	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	14	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	15	English	english	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	16	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	17	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	19	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	20	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	22	shared	shared	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	23	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	24	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	25	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	26	broader	broad	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	27	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	90	28	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	1	4	4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	2	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	3	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	4	System	System	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	5	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	6	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	7	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	8	consists	consist	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	10	five	five	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	11	machine	machine	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	12	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	13	learning	learn	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	14	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	16	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	17	specializing	specialize	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	18	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	19	correcting	correct	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	20	one	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	21	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	22	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	23	er-	er-	XX
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	24	rors	ror	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	25	described	describe	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	26	above	above	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	91	27	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	2	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	3	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	4	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	5	selected	select	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	6	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	7	input	input	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	8	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	9	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	10	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	11	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	12	called	call	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	13	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	14	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	15	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	16	3	3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	17	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	92	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	3	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	4	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	6	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	7	example	example	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	9	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	10	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	11	determined	determine	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	12	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	13	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	14	forms	form	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	93	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	2	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	3	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	5	determining	determine	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	7	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	8	might	may	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	9	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	10	more	more	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	11	involved	involved	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	94	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	1	All	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	2	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	3	take	take	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	4	as	as	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	5	input	input	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	7	corpus	corpus	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	8	documents	document	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	9	pre	pre	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	10	-	-	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	11	processed	process	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	12	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	13	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	14	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	15	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	17	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	18	speech	speech	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	19	tagger4	tagger4	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	20	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	21	Even-	Even-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	22	Zohar	Zohar	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	23	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	24	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	26	2001	2001	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	27	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	28	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	29	shallow	shallow	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	30	parser5	parser5	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	31	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	32	Pun-	Pun-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	33	yakanok	yakanok	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	34	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	35	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	36	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	37	2001	2001	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	38	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	95	39	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	3	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	4	submis-	submis-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	5	sion	sion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	7	some	some	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	8	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	9	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	10	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	11	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	12	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	13	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	15	oth-	oth-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	16	ers	er	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	17	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	18	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	19	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	96	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	2	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	3	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	5	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	6	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	7	make	make	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	8	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	10	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	11	discriminative	discriminative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	12	algorithm	algorithm	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	14	while	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	15	4http://cogcomp.cs.illinois.edu/page/	4http://cogcomp.cs.illinois.edu/page/	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	16	software	software	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	17	view	view	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	18	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	19	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	20	5http://cogcomp.cs.illinois.edu/page/	5http://cogcomp.cs.illinois.edu/page/	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	21	software	software	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	22	view	view	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	23	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	24	Chunker	Chunker	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	25	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	26	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	27	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	28	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	29	make	make	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	30	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	31	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	32	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	33	Naı̈ve	naı̈ve	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	34	Bayes	Bayes	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	35	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	36	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	37	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	38	algorithm	algorithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	97	39	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	3	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	4	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	5	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	6	option	option	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	7	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	8	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	9	post	post	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	10	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	11	processing	processing	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	12	step	step	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	13	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	14	corrections	correction	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	15	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	16	al-	al-	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	17	ways	way	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	18	result	result	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	19	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	20	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	21	false	false	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	22	positive	positive	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	23	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	24	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	25	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	26	ignored	ignore	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	27	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	28	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	29	option	option	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	30	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	31	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	32	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	33	here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	98	34	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	1	4.1	4.1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	2	Determiner	Determiner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	3	Errors	Errors	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	4	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	5	majority	majority	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	7	determiner	determiner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	8	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	9	involve	involve	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	10	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	12	although	although	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	13	some	some	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	14	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	15	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	16	involve	involve	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	17	pronouns	pronoun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	99	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	100	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	100	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	100	3	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	100	4	addresses	address	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	100	5	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	100	6	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	100	7	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	100	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	1	Can-	Can-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	2	didates	didate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	3	include	include	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	4	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	5	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	6	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	7	a”,“an”,“the”)6	a”,“an”,“the”)6	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	9	omis-	omis-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	10	sions	sion	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	12	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	13	considering	consider	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	14	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	15	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	16	phrase	phrase	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	17	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	18	initial	initial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	19	contexts	contexts	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	20	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	21	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	22	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	23	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	24	likely	likely	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	25	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	26	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	27	omitted	omit	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	101	28	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	2	con-	con-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	3	fusion	fusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	4	set	set	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	5	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	6	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	8	thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	9	{	{	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	10	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	14	∅	∅	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	15	}	}	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	102	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	2	ar-	ar-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	3	ticle	ticle	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	4	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	7	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	8	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	10	one	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	11	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	13	HOO	HOO	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	14	shared	share	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	15	tasks	task	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	16	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	17	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	18	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	19	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	20	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	21	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	22	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	23	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	24	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	25	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	26	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	27	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	28	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	29	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	30	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	31	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	32	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	33	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	34	demonstrated	demonstrate	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	35	superior	superior	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	36	per-	per-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	37	formance	formance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	103	38	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	3	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	4	discriminative	discriminative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	5	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	6	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	7	makes	make	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	8	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	11	Averaged	averaged	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	12	Perceptron	Perceptron	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	13	algorithm	algorithm	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	14	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	15	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	17	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	18	Fre-	Fre-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	19	und	und	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	20	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	21	Schapire	Schapire	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	23	1996	1996	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	24	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	25	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	26	implemented	implement	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	27	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	28	LBJava	LBJava	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	29	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	30	Rizzolo	Rizzolo	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	31	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	32	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	33	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	34	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	35	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	36	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	37	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	38	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	39	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	40	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	41	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	42	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	43	rich	rich	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	44	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	45	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	46	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	47	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	48	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	49	er-	er-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	50	rors	ror	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	104	51	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	105	1	See	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	105	2	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	105	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	106	1	5.2	5.2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	106	2	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	106	3	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	106	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	107	1	5.3	5.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	107	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	1	4.2	4.2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	2	Preposition	Preposition	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	3	Errors	Errors	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	4	Similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	5	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	6	determiners	determiner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	8	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	9	distinguish	distinguish	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	10	three	three	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	11	types	type	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	12	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	13	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	14	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	15	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	16	choosing	choose	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	17	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	18	incorrect	incorrect	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	19	prepo-	prepo-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	20	sition	sition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	21	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	22	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	23	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	24	superfluous	superfluous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	25	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	26	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	27	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	28	omitting	omit	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	29	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	30	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	108	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	2	contrast	contrast	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	3	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	4	determiners	determiner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	6	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	7	learn-	learn-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	8	ers	er	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	10	many	many	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	11	first	first	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	12	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	13	backgrounds	background	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	15	most	most	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	18	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	19	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	20	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	21	replacements	replacement	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	23	i.e.	i.e.	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	24	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	25	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	26	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	27	6The	6the	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	28	variants	variant	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	29	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	30	a	a	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	31	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	32	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	33	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	34	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	35	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	36	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	37	collapsed	collapse	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	38	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	39	one	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	40	class	class	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	109	41	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	1	421	421	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	2	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	3	Hence	hence	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	6	environmental	environmental	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	7	factors	factor	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	8	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	9	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	10	contributes/	contributes/	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	11	contribute	contribute	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	12	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	13	various	various	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	14	difficulties	difficulty	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	16	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	17	giving	give	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	18	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	19	given	give	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	20	prob-	prob-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	21	lems	lem	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	22	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	23	nuclear	nuclear	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	24	technology	technology	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	26	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	27	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	28	Confusion	Confusion	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	29	set	set	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	30	Agr	Agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	110	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	1	{	{	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	2	INF	INF	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	3	=	=	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	4	contribute	contribute	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	6	S	s	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	7	=	=	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	8	contributes	contribute	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	9	}	}	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	10	Form	Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	11	{	{	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	12	INF	INF	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	13	=	=	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	14	give	give	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	16	ED	ED	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	17	=	=	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	18	given	give	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	20	ING	ING	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	21	=	=	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	22	giving	giving	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	24	S	S	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	25	=	=	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	26	gives	give	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	27	}	}	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	28	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	29	4	4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	30	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	31	Confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	32	sets	set	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	33	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	34	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	35	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	36	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	111	37	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	2	irreg-	irreg-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	3	ular	ular	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	4	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	7	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	8	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	9	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	11	confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	12	set	set	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	13	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	14	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	15	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	16	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	18	past	past	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	19	participle	participle	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	112	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	1	author	author	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	2	correctly	correctly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	3	recognized	recognize	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	5	need	need	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	6	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	7	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	8	prepo-	prepo-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	9	sition	sition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	11	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	12	chose	choose	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	14	wrong	wrong	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	15	one	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	16	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	17	Leacock	Leacock	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	18	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	19	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	20	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	21	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	22	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	23	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	113	24	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	1	However	however	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	3	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	5	depend	depend	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	6	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	8	first	first	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	9	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	10	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	11	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	12	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	14	spurious	spurious	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	15	prepositions	preposition	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	16	occur	occur	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	17	more	more	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	18	frequently	frequently	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	19	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	20	29	29	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	21	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	22	versus	versus	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	23	18	18	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	24	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	25	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	26	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	27	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	28	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	29	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	30	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	31	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	32	corpora	corpora	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	33	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	34	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	35	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	36	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	37	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	38	2010a	2010a	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	39	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	40	Yannakoudakis	Yannakoudakis	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	41	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	42	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	43	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	44	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	45	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	46	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	114	47	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	3	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	4	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	6	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	7	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	8	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	9	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	10	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	11	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	12	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	13	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	14	uses	use	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	15	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	16	n	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	17	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	18	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	19	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	20	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	22	4-word	4-word	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	23	window	window	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	24	around	around	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	25	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	26	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	115	27	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	2	4-word	4-word	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	3	window	window	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	4	refers	refer	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	5	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	7	four	four	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	8	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	9	before	before	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	12	four	four	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	13	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	14	after	after	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	16	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	116	18	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	1	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	2	problem	problem	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	3	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	5	search	search	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	7	alternative	alternative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	8	resources	resource	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	9	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	11	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	12	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	14	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	15	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	17	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	117	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	1	Features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	2	consist	consist	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	4	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	5	n	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	6	-	-	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	7	grams	gram	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	8	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	9	various	various	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	10	lengths	length	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	11	spanning	span	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	13	target	target	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	14	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	118	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	2	example	example	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	4	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	6	search	search	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	8	”	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	9	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	10	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	11	3-gram	3-gram	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	12	feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	119	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	2	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	4	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	5	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	6	likely	likely	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	7	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	8	confu-	confu-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	9	sions	sion	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	10	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	12	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	13	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	14	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	15	see	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	16	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	120	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	121	1	5.2	5.2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	121	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	121	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	3	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	4	targets	target	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	5	replacement	replacement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	6	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	9	12	12	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	10	most	most	RBS
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	11	common	common	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	12	English	english	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	13	prepositions	preposition	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	122	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	1	Here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	2	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	3	aug-	aug-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	4	ment	ment	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	5	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	6	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	7	identify	identify	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	8	spurious	spurious	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	9	prepositions	preposition	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	123	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	2	con-	con-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	3	fusion	fusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	4	set	set	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	5	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	6	prepositions	preposition	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	8	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	9	follows	follow	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	10	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	11	{	{	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	12	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	14	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	16	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	18	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	20	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	21	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	22	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	24	about	about	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	26	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	28	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	29	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	30	by	by	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	31	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	32	into	into	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	33	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	34	during	during	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	35	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	36	∅	∅	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	37	}	}	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	124	38	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	1	4.3	4.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	2	Agreement	Agreement	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	4	Form	Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	5	Errors	Errors	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	6	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	8	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	9	implements	implement	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	10	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	11	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	12	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	13	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	14	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	15	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	16	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	17	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	18	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	19	consist	consist	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	20	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	22	following	follow	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	23	components	component	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	24	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	25	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	26	1	1	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	27	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	28	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	29	identification	identification	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	30	;	;	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	31	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	32	2	2	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	33	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	34	deter-	deter-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	35	mining	mine	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	36	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	37	relevant	relevant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	38	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	39	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	40	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	41	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	42	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	43	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	44	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	45	finiteness	finiteness	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	46	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	47	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	48	3	3	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	49	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	50	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	51	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	52	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	53	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	54	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	55	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	125	56	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	2	confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	3	set	set	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	4	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	5	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	6	depends	depend	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	9	target	target	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	10	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	11	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	12	includes	include	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	13	its	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	14	morphological	morphological	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	15	vari-	vari-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	16	ants	ant	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	17	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	18	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	19	4	4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	20	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	126	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	2	irregular	irregular	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	3	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	6	past	past	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	7	partici-	partici-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	8	ple	ple	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	9	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	10	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	11	included	include	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	13	while	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	15	past	past	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	16	tense	tense	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	17	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	18	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	19	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	20	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	127	21	i.e.	i.e.	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	1	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	2	given	give	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	3	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	4	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	5	included	include	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	6	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	7	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	8	gave	give	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	9	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	10	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	11	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	12	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	14	since	since	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	15	tense	tense	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	16	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	17	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	18	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	19	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	20	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	22	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	128	23	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	1	To	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	2	generate	generate	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	3	morphological	morphological	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	4	variants	variant	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	7	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	8	makes	make	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	9	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	11	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	12	morphological	morphological	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	13	analyzer	analyzer	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	14	verbMorph	verbmorph	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	15	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	16	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	17	assumes	assume	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	18	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	19	1	1	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	20	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	21	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	22	list	list	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	23	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	24	valid	valid	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	25	verb	verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	26	lemmas	lemmas	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	27	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	28	compiled	compile	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	29	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	30	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	31	POS-	POS-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	32	Dimension	Dimension	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	33	Systems	Systems	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	34	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	35	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	36	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	37	comparison	comparison	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	38	Learn	Learn	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	129	39	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	130	1	alg	alg	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	130	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	131	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	131	2	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	131	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	132	1	5.1	5.1	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	132	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	132	3	NTHU	NTHU	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	132	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	132	5	UMC	UMC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	132	6	Adaptation	Adaptation	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	132	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	132	8	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	132	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	133	1	5.2	5.2	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	133	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	133	3	Error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	133	4	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	133	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	133	6	HIT	HIT	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	133	7	Ling	Ling	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	133	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	134	1	knowledge	knowledge	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	134	2	Cand	Cand	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	134	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	135	1	identification	identification	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	135	2	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	135	3	NTHU	NTHU	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	135	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	135	5	HIT	HIT	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	135	6	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	135	7	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	135	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	136	1	5.3	5.3	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	136	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	136	3	Verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	136	4	finiteness	finiteness	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	136	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	136	6	NTHU	NTHU	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	136	7	Train	Train	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	136	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	137	1	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	137	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	137	3	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	137	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	1	5.4	5.4	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	3	HIT	hit	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	5	NARA	NARA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	6	Table	Table	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	7	5	5	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	8	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	9	System	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	10	comparisons	comparison	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	138	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	1	Column	column	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	2	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	3	indicates	indicate	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	5	di-	di-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	6	mension	mension	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	9	column	column	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	10	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	11	lists	list	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	12	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	13	whose	whose	WP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	14	approaches	approach	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	15	provide	provide	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	16	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	17	relevant	relevant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	18	point	point	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	19	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	20	comparison	comparison	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	139	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	1	tagged	tag	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	2	version	version	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	5	NYT	NYT	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	6	section	section	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	9	Gigaword	Gigaword	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	10	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	11	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	12	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	13	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	14	2	2	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	15	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	16	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	17	list	list	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	18	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	19	irregular	irregular	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	20	English	English	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	21	verbs.7	verbs.7	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	22	Candidate	Candidate	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	23	Identification	Identification	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	24	stage	stage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	25	selects	select	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	26	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	27	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	28	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	29	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	30	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	31	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	32	presented	present	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	33	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	34	input	input	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	35	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	36	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	37	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	140	38	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	1	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	3	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	4	crucial	crucial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	5	step	step	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	6	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	7	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	8	missed	miss	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	9	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	10	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	11	stage	stage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	12	will	will	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	13	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	14	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	15	detected	detect	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	16	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	18	later	later	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	19	stages	stage	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	141	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	142	1	See	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	142	2	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	142	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	143	1	5.3	5.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	143	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	1	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	2	Finiteness	Finiteness	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	4	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	5	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	8	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	9	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	10	sep-	sep-	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	11	arately	arately	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	12	process	process	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	13	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	14	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	15	fulfill	fulfill	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	16	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	17	grammati-	grammati-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	18	cal	cal	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	19	functions	function	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	20	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	21	thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	22	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	23	marked	mark	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	24	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	25	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	26	gram-	gram-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	27	matical	matical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	28	properties	property	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	144	29	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	145	1	See	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	145	2	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	145	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	146	1	5.3	5.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	146	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	1	Correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	2	Modules	Modules	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	3	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	4	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	5	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	6	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	7	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	8	bi-	bi-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	9	nary	nary	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	10	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	147	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	148	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	148	2	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	148	3	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	148	4	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	148	5	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	148	6	4-class	4-class	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	148	7	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	148	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	1	Both	both	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	2	classifiers	classifier	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	4	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	5	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	7	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	8	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	9	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	149	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	1	4.4	4.4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	2	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	3	Errors	Errors	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	4	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	5	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	6	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	7	involve	involve	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	8	confusing	confuse	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	9	singular	singular	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	11	plural	plural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	12	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	13	forms	form	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	14	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	150	15	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	1	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	2	phone	phone	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	3	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	4	instead	instead	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	6	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	7	phones	phone	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	8	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	9	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	10	Fig	Fig	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	151	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	1	1	1	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	4	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	6	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	7	most	most	RBS
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	8	common	common	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	9	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	10	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	11	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	13	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	14	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	15	after	after	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	16	determiner	determiner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	17	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	18	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	19	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	20	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	21	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	152	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	3	noun	noun	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	4	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	6	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	9	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	10	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	11	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	12	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	13	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	153	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	1	Similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	2	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	3	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	5	candi-	candi-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	6	date	date	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	7	identification	identification	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	8	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	9	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	10	important	important	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	11	step	step	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	12	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	14	noun	noun	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	15	classifier	classifier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	154	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	155	1	See	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	155	2	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	155	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	156	1	5.3	5.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	156	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	1	5	5	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	2	System	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	3	Analysis	Analysis	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	4	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	5	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	6	section	section	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	8	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	9	evaluate	evaluate	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	11	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	12	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	13	along	along	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	15	four	four	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	16	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	17	identified	identify	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	18	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	19	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	157	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	1	3	3	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	3	compare	compare	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	4	its	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	5	components	component	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	6	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	7	alternative	alternative	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	8	configurations	configuration	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	9	imple-	imple-	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	10	mented	mente	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	11	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	12	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	13	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	15	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	16	present	present	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	17	additional	additional	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	18	exper-	exper-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	19	iments	iment	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	20	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	21	further	further	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	22	analyze	analyze	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	23	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	24	dimension	dimension	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	158	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	1	While	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	2	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	3	direct	direct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	4	comparison	comparison	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	5	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	6	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	7	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	8	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	9	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	10	always	always	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	11	possible	possible	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	12	due	due	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	13	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	14	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	15	differences	difference	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	16	between	between	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	18	sys-	sys-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	19	tems	tem	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	21	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	22	believe	believe	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	23	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	24	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	25	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	26	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	27	still	still	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	28	useful	useful	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	159	29	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	160	1	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	160	2	5	5	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	160	3	lists	list	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	160	4	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	160	5	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	160	6	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	160	7	comparion	comparion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	160	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	3	im-	im-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	4	portant	portant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	5	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	6	note	note	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	7	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	9	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	10	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	11	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	12	indepen-	indepen-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	13	dent	dent	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	161	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	2	instance	instance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	4	there	there	EX
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	6	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	7	correlation	correlation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	8	between	between	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	9	algorithm	algorithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	10	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	11	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	12	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	13	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	162	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	1	7The	7The	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	2	tool	tool	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	4	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	5	detail	detail	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	6	about	about	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	7	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	8	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	9	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	10	found	find	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	11	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	12	http://cogcomp.cs.illinois.edu/page/publication	http://cogcomp.cs.illinois.edu/page/publication	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	13	view/743	view/743	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	14	422	422	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	15	Results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	16	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	17	reported	report	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	18	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	20	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	21	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	22	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	23	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	24	com-	com-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	25	puted	put	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	26	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	28	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	29	scorer	scorer	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	30	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	31	Dahlmeier	Dahlmeier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	32	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	33	Ng	Ng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	34	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	35	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	36	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	163	37	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	1	Error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	3	specific	specific	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	4	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	5	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	6	generated	generate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	7	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	8	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	10	output	output	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	12	individual	individual	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	13	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	164	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	1	Note	note	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	2	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	3	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	4	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	5	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	6	directly	directly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	7	comparable	comparable	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	8	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	9	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	10	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	11	specific	specific	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	12	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	13	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	15	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	16	overview	overview	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	17	paper	paper	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	18	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	20	latter	latter	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	21	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	22	approx-	approx-	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	23	imate	imate	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	24	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	25	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	26	organizers	organizer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	27	did	do	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	28	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	29	have	have	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	31	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	32	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	33	information	information	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	34	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	35	corrections	correction	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	36	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	37	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	38	output	output	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	165	39	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	2	com-	com-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	3	plete	plete	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	4	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	5	includes	include	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	7	union	union	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	8	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	9	corrections	correction	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	10	made	make	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	11	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	12	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	13	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	14	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	15	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	17	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	19	corrections	correction	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	20	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	21	applied	apply	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	22	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	23	order	order	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	166	24	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	1	Ordering	order	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	2	overlapping	overlapping	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	3	candidates8	candidates8	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	4	might	may	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	5	potentially	potentially	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	6	affect	affect	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	8	final	final	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	9	output	output	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	11	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	12	mod-	mod-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	13	ules	ule	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	14	correctly	correctly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	15	identify	identify	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	16	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	17	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	18	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	19	propose	propose	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	20	differ-	differ-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	21	ent	ent	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	22	corrections	correction	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	24	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	25	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	26	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	27	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	28	happen	happen	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	29	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	30	practice	practice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	167	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	1	Modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	2	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	4	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	8	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	9	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	10	marked	mark	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	11	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	12	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	13	asterisk	asterisk	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	14	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	15	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	16	tables	table	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	168	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	1	To	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	2	demonstrate	demonstrate	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	3	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	4	our	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	5	findings	finding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	6	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	7	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	8	spe-	spe-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	9	cific	cific	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	10	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	11	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	13	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	14	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	15	show	show	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	16	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	17	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	19	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	20	dataset	dataset	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	169	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	3	produced	produce	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	4	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	5	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	6	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	7	seventeen	seventeen	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	8	first	first	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	9	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	10	backgrounds	background	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	11	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	12	contains	contain	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	13	500,000	500,000	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	14	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	15	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	17	Cambridge	Cambridge	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	18	Learner	Learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	19	Corpus	Corpus	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	20	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	21	CLC	CLC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	22	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	23	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	24	Yannakoudakis	Yannakoudakis	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	25	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	26	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	27	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	28	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	29	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	30	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	170	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	2	split	split	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	4	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	5	into	into	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	6	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	7	equal	equal	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	8	parts	part	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	9	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	10	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	11	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	12	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	171	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	2	statis-	statis-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	3	tics	tic	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	4	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	5	shown	show	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	6	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	7	Appendix	Appendix	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	8	Tables	Tables	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	9	A.16	a.16	ADD
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	11	A.17	a.17	ADD
work_pdhkjvfkbnb4ngi7bpbilobcq4	172	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	173	1	5.1	5.1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	173	2	Dim	Dim	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	173	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	1	1	1	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	2	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	3	Learning	Learning	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	4	Algorithm	Algorithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	5	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	6	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	7	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	8	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	9	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	11	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	174	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	1	3	3	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	3	discuss	discuss	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	5	re-	re-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	6	lations	lation	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	7	between	between	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	9	amount	amount	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	11	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	12	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	14	learn-	learn-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	15	ing	ing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	16	algorithms	algorithms	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	18	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	20	resulting	result	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	21	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	175	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	1	They	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	2	show	show	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	3	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	5	training	train	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	6	sets	set	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	8	similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	9	sizes	size	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	11	discrimi-	discrimi-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	12	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	13	classifiers	classifiers	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	14	outperform	outperform	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	15	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	16	machine	machine	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	17	learning	learn	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	18	methods	method	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	19	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	20	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	21	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	176	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	1	Following	follow	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	2	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	3	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	6	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	7	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	8	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	9	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	10	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	11	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	12	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	14	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	15	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	16	uses	use	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	18	discriminative	discriminative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	19	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	20	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	177	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	1	Most	Most	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	2	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	4	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	5	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	6	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	7	train	train	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	8	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	10	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	11	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	12	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	13	use	use	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	14	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	15	discriminative	discriminative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	16	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	178	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	1	However	however	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	3	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	4	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	5	very	very	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	6	large	large	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	7	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	8	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	9	set	set	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	10	such	such	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	11	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	13	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	14	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	15	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	16	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	17	available	available	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	19	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	20	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	21	often	often	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	22	ad-	ad-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	23	vantageous	vantageous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	24	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	25	use	use	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	26	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	179	27	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	2	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	3	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	4	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	6	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	7	collec-	collec-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	8	tion	tion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	10	n	n	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	11	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	12	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	13	counts	count	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	14	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	15	length	length	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	16	one	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	17	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	18	five	five	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	19	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	20	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	21	cor-	cor-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	22	pus	pus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	23	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	24	1012	1012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	25	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	180	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	1	Since	since	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	3	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	4	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	5	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	6	come	come	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	7	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	8	complete	complete	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	9	sentences	sentence	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	11	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	12	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	13	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	14	straightforward	straightforward	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	15	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	16	make	make	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	17	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	18	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	19	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	20	discriminative	discriminative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	21	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	22	because	because	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	23	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	24	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	25	limited	limited	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	26	window	window	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	27	provided	provide	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	28	around	around	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	29	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	30	example	example	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	31	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	32	training	train	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	33	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	34	discriminative	discriminative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	35	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	36	would	would	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	37	limit	limit	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	38	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	39	sur-	sur-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	40	8Overlapping	8overlapping	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	41	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	42	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	43	included	include	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	44	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	45	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	46	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	47	one	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	48	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	49	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	50	if	if	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	51	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	52	work	work	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	53	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	54	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	55	tagged	tag	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	56	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	57	NN	NN	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	58	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	59	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	60	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	61	included	include	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	62	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	63	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	64	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	65	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	66	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	67	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	68	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	69	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	70	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	71	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	72	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	73	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	74	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	75	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	76	valid	valid	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	77	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	78	lemma	lemma	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	79	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	181	80	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	182	1	rounding	round	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	182	2	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	182	3	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	182	4	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	182	5	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	182	6	2-word	2-word	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	182	7	window	window	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	182	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	1	Be-	be-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	2	cause	cause	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	3	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	4	wish	wish	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	5	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	6	make	make	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	7	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	8	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	10	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	11	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	12	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	13	extend	extend	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	14	beyond	beyond	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	16	2-word	2-word	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	17	window	window	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	19	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	20	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	21	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	22	possible	possible	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	23	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	24	use	use	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	25	count	count	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	26	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	27	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	28	methods	method	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	29	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	30	such	such	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	31	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	32	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	33	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	34	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	183	35	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	1	Several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	2	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	3	make	make	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	4	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	7	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	8	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	9	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	10	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	11	UMC	UMC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	12	uses	use	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	13	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	14	count	count	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	15	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	16	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	17	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	18	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	19	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	21	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	23	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	24	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	25	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	26	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	27	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	28	NTHU	NTHU	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	29	addresses	address	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	30	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	31	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	32	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	33	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	34	count	count	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	35	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	36	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	37	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	38	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	39	backoff	backoff	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	40	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	41	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	42	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	43	es-	es-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	44	sentially	sentially	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	45	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	46	variation	variation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	47	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	48	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	49	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	50	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	51	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	52	back-	back-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	53	off	off	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	184	54	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	3	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	4	employs	employ	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	6	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	7	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	8	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	9	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	10	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	11	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	13	except	except	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	14	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	16	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	17	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	185	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	1	Training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	2	Naı̈ve	naı̈ve	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	3	Bayes	Bayes	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	4	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	5	Deletions	Deletions	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	6	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	7	Inser-	Inser-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	8	tions	tion	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	9	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	10	reason	reason	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	11	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	12	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	13	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	15	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	16	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	17	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	18	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	19	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	20	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	21	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	22	that	that	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	23	training	train	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	24	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	25	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	26	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	27	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	28	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	29	deletions	deletion	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	30	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	31	insertions	insertion	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	32	presents	present	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	33	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	34	problem	problem	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	35	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	36	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	37	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	38	majority	majority	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	39	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	40	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	41	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	42	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	43	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	44	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	45	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	186	46	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	1	Recall	recall	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	2	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	3	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	4	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	5	contains	contain	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	6	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	7	n	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	8	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	9	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	10	counts	count	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	12	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	13	makes	make	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	14	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	15	difficult	difficult	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	16	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	17	estimate	estimate	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	19	prior	prior	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	20	count	count	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	21	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	22	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	23	∅	∅	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	24	candi-	candi-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	25	date	date	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	187	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	2	With	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	3	access	access	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	4	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	5	complete	complete	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	6	sentences	sentence	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	9	prior	prior	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	11	∅	∅	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	12	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	13	estimated	estimate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	14	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	15	counting	count	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	17	total	total	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	18	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	19	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	20	∅	∅	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	21	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	22	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	23	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	24	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	25	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	26	case	case	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	27	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	28	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	29	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	31	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	32	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	33	NPs	NPs	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	34	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	35	∅	∅	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	36	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	37	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	38	computed	compute	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	39	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	188	40	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	2	solve	solve	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	4	prob-	prob-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	5	lem	lem	XX
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	6	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	7	treating	treat	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	9	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	12	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	13	following	follow	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	14	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	15	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	16	one	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	17	target	target	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	189	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	2	instance	instance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	4	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	5	estimate	estimate	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	6	prior	prior	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	7	counts	count	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	8	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	10	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	11	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	12	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	13	front	front	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	14	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	16	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	17	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	18	cam-	cam-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	19	era	era	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	20	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	21	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	22	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	23	including	include	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	24	camera	camera	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	25	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	26	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	27	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	28	obtain	obtain	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	29	counts	count	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	30	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	31	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	32	camera	camera	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	33	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	34	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	35	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	36	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	37	camera	camera	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	38	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	39	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	40	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	41	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	42	camera	camera	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	43	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	190	44	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	3	case	case	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	6	∅	∅	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	7	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	10	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	11	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	12	camera	camera	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	13	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	14	acts	act	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	15	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	17	tar-	tar-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	18	get	get	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	191	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	1	Thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	4	confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	5	set	set	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	6	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	8	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	9	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	10	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	11	modified	modify	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	12	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	13	follows	follow	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	14	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	15	instead	instead	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	18	three	three	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	19	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	20	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	21	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	22	shown	show	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	23	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	24	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	192	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	1	4.1	4.1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	4	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	5	member	member	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	8	confu-	confu-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	9	sion	sion	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	10	set	set	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	11	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	12	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	13	concatenation	concatenation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	14	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	16	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	17	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	19	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	20	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	21	follows	follow	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	22	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	193	24	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	1	{	{	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	2	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	3	camera	camera	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	6	camera	camera	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	8	cam-	cam-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	9	era	era	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	10	}	}	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	194	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	2	counts	count	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	3	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	4	contextual	contextual	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	5	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	6	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	7	obtained	obtain	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	8	similarly	similarly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	195	10	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	1	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	2	feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	3	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	4	includes	include	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	5	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	6	preceding	precede	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	7	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	8	would	would	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	9	correspond	correspond	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	10	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	12	count	count	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	13	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	14	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	15	including	include	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	16	x	x	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	17	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	19	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	20	x	x	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	21	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	22	take	take	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	23	any	any	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	24	value	value	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	25	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	26	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	27	confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	28	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	196	29	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	2	above	above	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	3	solution	solution	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	4	allows	allow	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	5	us	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	6	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	7	train	train	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	8	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	9	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	10	ar-	ar-	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	11	ticle	ticle	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	12	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	13	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	14	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	15	extend	extend	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	17	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	18	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	19	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	20	handle	handle	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	21	extraneous	extraneous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	22	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	23	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	24	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	25	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	26	6	6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	27	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	197	28	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	1	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	2	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	3	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	4	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	5	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	6	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	7	study	study	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	8	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	9	algo-	algo-	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	10	rithms	rithms	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	11	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	12	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	14	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	15	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	16	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	17	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	18	observe	observe	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	19	that	that	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	21	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	22	evaluated	evaluate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	23	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	24	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	25	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	26	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	27	win-	win-	WP
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	28	dow	dow	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	29	size	size	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	30	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	31	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	32	performs	perform	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	33	better	well	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	34	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	35	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	36	count	count	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	37	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	38	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	39	methods	method	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	198	40	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	2	order	order	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	3	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	4	show	show	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	6	impact	impact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	9	algo-	algo-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	10	rithm	rithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	11	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	13	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	14	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	15	6	6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	17	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	18	compare	compare	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	19	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	20	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	21	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	22	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	199	23	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	1	Both	both	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	2	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	3	use	use	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	4	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	5	n	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	6	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	7	grams	gram	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	8	spanning	span	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	10	target	target	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	11	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	12	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	14	4-word	4-word	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	15	window	window	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	200	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	2	train	train	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	3	LMs	lm	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	4	423	423	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	5	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	6	Model	Model	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	7	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	8	CoNLL	CoNLL	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	9	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	10	Art	Art	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	201	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	202	1	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	202	2	21.11	21.11	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	202	3	24.15	24.15	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	202	4	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	202	5	32.45	32.45	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	202	6	30.78	30.78	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	202	7	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	202	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	1	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	2	12.09	12.09	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	3	30.01	30.01	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	4	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	5	14.04	14.04	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	6	29.40	29.40	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	7	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	8	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	9	40.72	40.72	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	10	32.41	32.41	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	11	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	12	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	13	42.60	42.60	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	14	34.40	34.40	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	15	Agr	Agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	203	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	1	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	2	20.65	20.65	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	3	33.53	33.53	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	4	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	5	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	6	26.46	26.46	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	7	36.42	36.42	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	8	Form	Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	9	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	10	13.40	13.40	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	11	08.46	08.46	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	12	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	13	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	14	14.50	14.50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	15	12.16	12.16	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	16	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	17	6	6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	18	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	19	Comparison	comparison	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	20	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	21	learning	learning	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	22	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	204	23	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	205	1	Web1	web1	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	205	2	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	205	3	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	205	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	1	Modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	2	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	4	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	8	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	9	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	10	marked	mark	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	11	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	12	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	13	asterisk	asterisk	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	206	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	1	Source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	2	Candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	3	ED	ED	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	4	INF	INF	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	5	ING	ING	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	6	S	S	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	7	ED	ED	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	8	0.99675	0.99675	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	9	0.00192	0.00192	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	10	0.00103	0.00103	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	11	0.00030	0.00030	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	12	INF	INF	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	13	0.00177	0.00177	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	14	0.99630	0.99630	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	15	0.00168	0.00168	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	16	0.00025	0.00025	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	17	ING	ING	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	18	0.00124	0.00124	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	19	0.00447	0.00447	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	20	0.99407	0.99407	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	21	0.00022	0.00022	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	22	S	s	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	23	0.00054	0.00054	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	24	0.00544	0.00544	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	25	0.00132	0.00132	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	26	0.99269	0.99269	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	27	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	28	7	7	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	29	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	30	Priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	31	confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	32	matrix	matrix	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	33	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	34	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	35	adapting	adapt	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	36	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	207	37	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	1	Each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	2	entry	entry	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	3	shows	show	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	4	Prob(candidate|source	Prob(candidate|source	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	7	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	8	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	9	corre-	corre-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	10	sponds	spond	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	11	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	13	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	14	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	15	chosen	choose	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	16	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	18	author	author	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	208	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	1	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	2	SRILM	SRILM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	3	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	4	Stolcke	Stolcke	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	6	2002	2002	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	7	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	8	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	9	Jelinek	Jelinek	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	10	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	11	Mercer	Mercer	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	12	linear	linear	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	13	interpolation	interpolation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	14	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	15	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	16	smoothing	smoothing	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	17	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	18	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	19	Chen	Chen	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	20	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	21	Goodman	Goodman	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	23	1996	1996	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	24	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	209	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	1	On	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	3	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	4	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	5	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	7	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	8	outperforms	outperform	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	9	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	10	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	11	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	12	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	13	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	14	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	16	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	17	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	19	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	20	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	21	superior	superior	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	22	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	23	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	24	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	26	except	except	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	27	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	28	er-	er-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	29	rors	ror	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	30	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	31	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	32	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	33	outperforms	outperform	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	34	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	35	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	36	very	very	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	37	slightly	slightly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	210	38	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	2	attribute	attribute	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	4	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	6	fact	fact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	7	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	9	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	10	prob-	prob-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	11	lem	lem	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	12	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	13	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	14	labels	label	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	15	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	16	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	17	there	there	EX
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	18	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	19	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	20	big	big	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	21	confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	22	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	24	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	25	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	26	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	27	default	default	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	28	smooth	smooth	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	29	weights	weight	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	30	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	31	so	so	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	32	there	there	EX
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	33	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	34	no	no	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	35	advantage	advantage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	36	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	37	running	run	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	38	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	211	39	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	2	found	find	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	3	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	4	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	5	fewer	few	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	6	classes	class	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	8	6	6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	9	rather	rather	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	10	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	11	12	12	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	12	prepositions	preposition	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	13	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	15	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	16	outperforms	outperform	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	17	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	212	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	3	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	4	possible	possible	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	5	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	6	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	7	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	8	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	9	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	10	lot	lot	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	12	labels	label	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	15	theoretical	theoretical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	16	difference	difference	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	17	be-	be-	XX
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	18	tween	tween	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	20	algorithms	algorithm	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	21	disappears	disappear	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	213	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	1	Note	note	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	2	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	3	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	4	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	5	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	6	improved	improve	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	7	via	via	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	8	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	9	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	10	next	next	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	11	section	section	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	12	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	13	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	14	then	then	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	15	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	16	outperforms	outperform	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	18	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	19	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	20	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	21	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	22	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	214	23	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	215	1	5.2	5.2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	215	2	Dim	Dim	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	215	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	1	2	2	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	2	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	3	Adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	4	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	5	Learner	Learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	6	Errors	Errors	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	7	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	9	previous	previous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	10	section	section	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	13	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	14	were	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	15	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	16	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	17	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	18	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	216	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	1	These	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	2	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	3	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	4	no	no	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	5	notion	notion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	8	er-	er-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	9	ror	ror	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	10	patterns	pattern	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	13	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	217	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	1	Here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	2	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	3	discuss	discuss	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	4	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	5	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	6	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	7	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	8	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	218	10	i.e.	i.e.	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	1	developing	develop	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	2	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	3	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	4	utilize	utilize	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	6	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	7	about	about	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	9	types	type	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	11	mis-	mis-	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	12	takes	take	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	13	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	14	make	make	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	219	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	1	Adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	3	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	6	fact	fact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	7	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	8	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	9	make	make	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	10	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	11	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	12	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	13	systematic	systematic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	14	manner	manner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	220	16	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	1	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	2	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	3	influenced	influence	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	4	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	6	writer	writer	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	7	’s	’s	POS
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	8	first	first	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	9	lan-	lan-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	10	guage	guage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	11	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	12	Gass	Gass	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	13	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	14	Selinker	Selinker	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	16	1992	1992	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	17	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	18	Ionin	Ionin	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	19	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	20	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	21	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	23	2008	2008	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	24	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	221	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	1	There	there	EX
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	2	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	3	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	4	ways	way	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	5	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	6	adapt	adapt	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	7	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	8	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	9	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	10	de-	de-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	11	pend	pend	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	12	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	14	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	15	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	16	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	17	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	18	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	19	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	20	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	21	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	22	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	23	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	24	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	25	algorithm	algorithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	26	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	222	27	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	2	key	key	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	3	application	application	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	5	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	6	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	7	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	8	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	9	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	10	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	11	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	12	English	english	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	13	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	15	because	because	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	17	learned	learn	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	18	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	19	do	do	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	20	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	21	know	know	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	22	any-	any-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	23	thing	thing	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	24	about	about	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	25	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	26	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	27	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	28	make	make	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	223	29	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	1	With	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	2	adapta-	adapta-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	3	tion	tion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	5	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	6	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	8	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	9	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	10	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	11	use	use	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	13	au-	au-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	14	thor	thor	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	15	’s	’s	POS
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	16	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	17	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	19	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	20	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	21	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	22	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	23	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	24	feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	25	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	26	thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	27	propose	propose	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	28	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	29	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	30	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	31	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	32	what	what	WP
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	33	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	34	author	author	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	35	orig-	orig-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	36	inally	inally	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	37	wrote	write	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	224	38	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	1	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	3	crucial	crucial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	5	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	7	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	8	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	9	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	10	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	11	important	important	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	12	piece	piece	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	13	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	14	information	information	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	15	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	16	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	17	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	18	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	20	2010b	2010b	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	21	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	225	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	1	Below	below	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	3	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	4	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	5	techniques	technique	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	6	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	7	summarized	summarize	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	9	evaluated	evaluate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	226	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	3	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	4	makes	make	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	5	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	7	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	10	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	11	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	12	via	via	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	14	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	15	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	16	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	17	adapts	adapt	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	18	its	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	19	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	20	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	21	clas-	clas-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	22	sifier	sifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	23	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	24	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	25	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	26	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	27	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	28	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	29	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	30	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	227	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	1	Adapting	adapt	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	2	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	3	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	4	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	5	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	6	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	7	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	9	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	11	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	13	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	228	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	1	4	4	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	4	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	5	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	6	technique	technique	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	7	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	8	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	9	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	10	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	11	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	12	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	13	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	14	English	english	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	15	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	16	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	17	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	18	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	19	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	20	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	21	changing	change	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	22	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	23	distribution	distribution	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	24	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	25	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	26	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	28	cor-	cor-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	29	rection	rection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	30	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	229	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	1	Candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	2	prior	prior	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	4	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	5	special	special	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	6	pa-	pa-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	7	rameter	rameter	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	9	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	10	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	11	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	12	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	13	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	14	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	15	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	16	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	17	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	19	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	20	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	21	correspond	correspond	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	22	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	23	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	24	relative	relative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	25	frequen-	frequen-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	26	cies	cie	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	27	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	28	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	29	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	30	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	31	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	32	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	33	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	34	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	35	do	do	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	36	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	37	provide	provide	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	38	any	any	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	39	information	information	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	40	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	41	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	42	real	real	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	43	distribution	distribution	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	44	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	45	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	46	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	47	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	48	dependence	dependence	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	49	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	50	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	51	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	52	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	53	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	54	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	55	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	56	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	57	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	58	author	author	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	230	59	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	3	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	4	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	6	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	7	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	8	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	9	changed	change	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	10	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	11	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	12	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	13	confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	14	matrix	matrix	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	15	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	16	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	17	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	18	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	19	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	20	specifies	specify	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	21	how	how	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	22	likely	likely	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	23	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	24	confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	25	pair	pair	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	26	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	231	27	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	1	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	2	7	7	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	3	shows	show	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	5	confusion	confusion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	6	matrix	matrix	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	7	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	8	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	9	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	10	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	12	computed	compute	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	13	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	15	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	16	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	232	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	1	Adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	2	pri-	pri-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	3	ors	or	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	4	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	5	dependent	dependent	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	6	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	8	author	author	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	9	’s	’s	POS
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	10	original	original	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	11	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	12	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	13	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	14	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	15	let	let	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	16	s	s	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	17	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	18	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	19	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	20	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	22	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	23	appearing	appear	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	24	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	25	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	26	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	27	text	text	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	28	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	29	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	30	c	c	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	31	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	32	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	33	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	233	34	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	1	Then	then	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	3	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	4	prior	prior	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	6	c	c	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	7	given	give	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	8	s	s	POS
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	9	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	10	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	11	prior(c|s	prior(c|s	ADD
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	12	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	13	=	=	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	14	C(s	C(s	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	16	c	c	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	17	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	18	C(s	C(s	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	19	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	20	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	21	C(s	C(s	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	22	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	23	denotes	denote	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	24	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	25	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	26	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	27	times	time	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	28	s	s	POS
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	29	appeared	appear	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	30	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	31	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	32	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	33	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	34	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	35	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	36	C(s	C(s	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	37	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	38	c	c	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	39	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	40	denotes	denote	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	41	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	42	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	43	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	44	times	time	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	45	c	c	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	46	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	47	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	48	correct	correct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	49	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	50	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	51	s	s	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	52	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	53	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	54	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	55	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	56	writer	writer	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	234	57	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	2	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	3	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	4	differ	differ	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	5	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	7	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	8	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	10	probability	probability	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	12	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	13	INF	INF	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	14	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	16	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	17	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	18	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	19	S	S	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	21	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	22	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	23	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	24	twice	twice	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	25	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	26	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	28	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	29	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	30	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	31	424	424	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	32	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	33	Model	Model	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	34	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	35	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	36	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	37	Train	Train	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	38	Test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	39	Art	art	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	235	40	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	1	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	2	18.28	18.28	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	3	32.45	32.45	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	4	30.78	30.78	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	5	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	6	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	7	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	8	19.18	19.18	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	9	34.49	34.49	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	10	31.76	31.76	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	11	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	236	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	1	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	2	09.03	09.03	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	3	14.04	14.04	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	4	29.40	29.40	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	5	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	6	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	7	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	8	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	9	10.94	10.94	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	10	12.14	12.14	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	11	32.22	32.22	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	12	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	13	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	14	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	15	23.06	23.06	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	16	42.60	42.60	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	17	34.40	34.40	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	18	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	19	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	20	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	21	22.89	22.89	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	22	42.31	42.31	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	23	32.38	32.38	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	24	Agr	Agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	237	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	1	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	2	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	3	16.72	16.72	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	4	26.46	26.46	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	5	36.42	36.42	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	6	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	7	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	8	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	9	17.62	17.62	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	10	23.46	23.46	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	11	38.57	38.57	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	12	Form	Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	13	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	14	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	15	11.93	11.93	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	16	14.50	14.50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	17	12.16	12.16	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	18	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	19	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	20	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	21	14.63	14.63	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	22	18.35	18.35	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	23	16.67	16.67	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	24	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	25	8	8	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	26	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	27	Adapting	adapt	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	28	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	29	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	31	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	32	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	238	33	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	1	All	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	2	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	4	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	5	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	7	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	8	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	9	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	239	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	1	Modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	2	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	4	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	8	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	9	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	10	marked	mark	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	11	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	12	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	13	asterisk	asterisk	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	240	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	1	ED	ED	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	2	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	4	probability	probability	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	5	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	6	S	S	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	9	correct	correct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	10	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	11	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	12	very	very	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	13	high	high	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	15	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	16	reflects	reflect	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	18	low	low	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	19	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	20	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	241	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	1	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	2	8	8	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	3	compares	compare	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	4	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	5	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	6	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	7	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	8	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	9	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	242	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	1	Because	because	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	2	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	4	dichotomy	dichotomy	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	5	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	7	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	8	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	9	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	10	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	11	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	12	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	13	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	14	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	16	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	17	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	18	show	show	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	19	exper-	exper-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	20	iments	iment	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	21	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	22	5-fold	5-fold	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	23	cross	cross	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	24	-	-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	25	validation	validation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	26	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	28	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	29	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	243	30	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	1	Adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	2	always	always	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	3	helps	help	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	6	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	7	train-	train-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	8	ing	ing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	9	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	12	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	13	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	14	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	15	except	except	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	16	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	17	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	18	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	20	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	21	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	22	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	23	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	24	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	25	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	26	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	27	helps	help	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	28	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	29	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	30	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	31	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	32	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	33	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	244	34	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	1	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	3	due	due	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	4	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	5	discrepancies	discrepancy	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	6	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	8	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	9	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	11	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	12	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	13	exploits	exploit	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	15	property	property	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	16	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	17	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	18	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	19	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	20	systematic	systematic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	245	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	1	Indeed	indeed	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	3	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	4	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	5	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	6	estimated	estimate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	9	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	10	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	11	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	12	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	13	5-fold	5-fold	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	14	cross	cross	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	15	-	-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	16	validation	validation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	17	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	20	perfor-	perfor-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	21	mance	mance	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	22	improves	improve	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	246	24	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	1	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	2	preposition	preposition	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	3	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	4	attains	attain	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	5	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	6	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	8	18.05	18.05	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	9	instead	instead	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	11	12.14	12.14	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	247	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	1	Concerning	concern	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	2	lack	lack	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	4	improvement	improvement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	5	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	6	noun	noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	7	num-	num-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	8	ber	ber	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	9	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	11	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	12	hypothesize	hypothesize	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	13	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	14	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	15	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	16	differ	differ	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	17	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	19	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	20	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	21	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	22	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	23	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	24	appropriate	appropriate	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	25	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	26	strongly	strongly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	27	depends	depend	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	28	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	29	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	30	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	31	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	32	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	33	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	34	noun	noun	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	35	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	36	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	37	would	would	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	38	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	39	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	40	turn	turn	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	41	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	42	suggest	suggest	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	43	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	44	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	45	dependency	dependency	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	46	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	47	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	48	label	label	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	49	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	50	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	51	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	52	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	53	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	54	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	55	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	56	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	57	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	58	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	59	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	60	trying	try	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	61	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	62	discover	discover	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	63	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	64	weak	weak	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	248	65	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	1	In-	in-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	2	deed	deed	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	5	prior	prior	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	6	distribution	distribution	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	8	{	{	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	9	singular	singular	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	11	plural	plural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	12	}	}	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	13	la-	la-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	14	bel	bel	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	15	space	space	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	16	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	17	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	18	change	change	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	19	much	much	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	20	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	22	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	23	feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	24	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	25	taken	take	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	26	into	into	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	27	account	account	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	249	28	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	2	unadapted	unadapted	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	3	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	4	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	5	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	6	singular	singular	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	7	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	9	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	10	plural	plural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	11	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	12	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	13	0.75	0.75	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	14	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	15	0.25	0.25	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	17	respec-	respec-	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	18	tively	tively	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	250	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	1	Similarly	similarly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	4	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	5	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	6	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	7	singular|plural	singular|plural	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	8	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	9	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	10	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	11	plural|singular	plural|singular	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	12	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	13	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	14	0.034	0.034	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	15	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	16	0.016	0.016	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	18	respec-	respec-	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	19	tively	tively	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	251	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	2	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	3	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	6	unadapted	unadapted	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	7	prior	prior	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	8	probabil-	probabil-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	9	ity	ity	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	10	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	11	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	12	plural	plural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	13	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	14	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	15	three	three	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	16	times	time	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	17	lower	low	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	18	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	19	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	20	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	21	singu-	singu-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	22	lar	lar	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	23	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	24	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	25	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	26	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	27	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	28	change	change	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	29	much	much	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	30	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	31	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	252	32	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	253	1	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	253	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	253	3	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	253	4	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	253	5	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	253	6	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	253	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	2	instance	instance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	4	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	5	case	case	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	7	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	8	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	11	unadapted	unadapted	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	12	prior	prior	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	13	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	14	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	15	plu-	plu-	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	16	ral	ral	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	17	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	18	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	19	0.617	0.617	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	21	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	22	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	23	three	three	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	24	times	time	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	25	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	26	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	27	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	28	sin-	sin-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	29	gular	gular	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	30	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	31	prior	prior	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	32	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	33	0.20	0.20	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	254	34	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	1	With	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	2	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	4	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	5	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	6	become	become	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	7	almost	almost	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	9	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	10	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	11	0.016	0.016	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	12	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	13	0.012	0.012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	14	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	255	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	1	Adapting	adapt	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	2	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	3	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	4	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	6	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	7	discriminative	discriminative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	8	learning	learning	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	9	algorithm	algorithm	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	11	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	12	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	13	use	use	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	14	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	15	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	17	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	18	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	19	can-	can-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	20	didates	didate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	256	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	2	order	order	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	3	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	4	reflect	reflect	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	5	our	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	6	estimate	estimate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	9	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	10	distribution	distribution	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	13	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	14	algorithm	algorithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	15	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	16	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	17	differently	differently	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	19	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	20	introducing	introduce	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	21	into	into	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	22	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	23	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	24	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	25	artificial	artificial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	26	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	28	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	29	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	30	rate	rate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	31	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	32	reflects	reflect	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	33	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	34	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	35	made	make	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	36	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	37	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	38	ESL	ESL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	39	writers	writer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	40	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	41	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	42	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	43	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	44	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	45	2010b	2010b	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	46	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	257	47	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	2	idea	idea	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	4	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	5	simulate	simulate	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	6	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	7	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	9	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	11	through	through	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	12	arti-	arti-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	13	ficial	ficial	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	14	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	15	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	16	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	17	produced	produce	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	18	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	19	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	20	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	21	confu-	confu-	XX
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	22	sion	sion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	23	matrix).9	matrix).9	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	24	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	25	original	original	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	26	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	27	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	28	proposed	propose	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	29	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	30	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	31	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	32	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	33	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	34	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	258	35	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	1	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	2	technique	technique	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	3	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	4	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	5	further	further	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	6	enhanced	enhance	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	7	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	9	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	10	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	11	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	12	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	13	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	14	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	15	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	16	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	18	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	20	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	259	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	1	6	6	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	3	applied	apply	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	4	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	5	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	6	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	8	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	9	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	10	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	11	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	260	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	3	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	4	uses	use	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	5	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	6	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	7	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	8	its	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	9	ar-	ar-	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	10	ticle	ticle	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	11	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	261	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	1	Because	because	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	2	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	3	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	4	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	5	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	6	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	7	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	8	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	11	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	12	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	13	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	14	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	15	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	16	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	17	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	18	fea-	fea-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	19	ture	ture	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	262	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	1	However	however	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	3	since	since	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	4	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	5	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	6	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	7	sparse	sparse	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	10	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	11	feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	12	encourages	encourage	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	14	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	15	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	16	abstain	abstain	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	17	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	18	flagging	flag	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	19	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	20	mistake	mistake	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	21	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	22	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	23	results	result	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	24	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	25	low	low	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	26	recall	recall	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	263	27	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	2	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	3	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	4	technique	technique	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	5	addresses	address	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	6	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	7	problem	problem	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	8	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	9	boosting	boost	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	11	proportion	proportion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	12	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	13	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	14	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	16	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	17	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	264	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	2	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	4	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	5	generating	generate	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	6	additional	additional	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	7	artificial	artificial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	8	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	9	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	11	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	12	distribution	distribution	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	13	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	15	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	16	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	265	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	1	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	2	9	9	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	3	shows	show	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	5	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	7	adapting	adapt	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	9	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	10	clas-	clas-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	11	sifier	sifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	12	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	13	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	14	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	266	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	2	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	3	omit	omit	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	4	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	5	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	7	since	since	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	9	noun	noun	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	10	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	11	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	12	performs	perform	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	13	better	well	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	14	without	without	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	16	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	17	feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	19	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	20	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	21	similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	22	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	23	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	24	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	25	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	26	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	28	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	29	discussed	discuss	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	30	above	above	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	267	32	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	2	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	3	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	4	improves	improve	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	5	recall	recall	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	6	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	8	consequently	consequently	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	10	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	268	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	2	should	should	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	3	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	4	noted	note	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	5	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	6	although	although	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	7	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	8	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	9	decreases	decrease	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	10	preci-	preci-	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	11	sion	sion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	12	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	13	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	14	still	still	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	15	helpful	helpful	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	269	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	2	fact	fact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	4	because	because	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	7	low	low	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	8	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	9	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	11	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	12	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	14	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	15	dataset	dataset	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	16	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	17	natural	natural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	18	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	19	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	20	very	very	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	21	poor	poor	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	23	often	often	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	24	resulting	result	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	25	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	26	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	27	be-	be-	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	28	ing	e	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	29	equal	equal	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	30	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	31	0	0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	32	due	due	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	33	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	34	no	no	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	35	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	36	being	be	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	37	detected	detect	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	270	38	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	1	Inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	2	vs.	vs.	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	3	Sampling	Sampling	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	4	To	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	5	demonstrate	demonstrate	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	7	impact	impact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	8	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	9	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	10	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	12	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	13	compare	compare	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	14	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	15	against	against	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	16	sampling	sampling	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	18	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	19	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	20	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	21	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	22	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	23	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	24	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	271	25	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	1	HIT	hit	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	2	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	3	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	4	improves	improve	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	5	recall	recall	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	6	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	7	removing	remove	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	8	correct	correct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	9	examples	example	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	10	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	11	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	272	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	2	HIT	hit	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	3	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	4	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	6	similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	7	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	9	9The	9the	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	10	idea	idea	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	12	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	13	artificial	artificial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	14	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	15	goes	go	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	16	back	back	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	17	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	18	Izumi	Izumi	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	19	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	20	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	273	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	2	2003	2003	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	3	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	4	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	5	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	6	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	7	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	9	Foster	Foster	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	11	Andersen	Andersen	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	12	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	13	2009	2009	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	14	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	274	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	2	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	3	discussed	discuss	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	4	here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	5	refers	refer	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	6	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	8	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	9	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	10	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	11	Ro-	Ro-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	12	zovskaya	zovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	13	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	14	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	15	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	16	2010b	2010b	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	17	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	18	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	19	generates	generate	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	20	artificial	artificial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	21	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	22	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	23	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	24	distribution	distribution	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	25	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	26	naturally	naturally	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	27	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	28	occurring	occur	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	29	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	275	30	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	276	1	425	425	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	276	2	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	276	3	Model	Model	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	276	4	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	276	5	CoNLL	CoNLL	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	276	6	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	276	7	Art	Art	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	276	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	1	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	3	natural	natural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	6	07.06	07.06	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	7	27.65	27.65	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	8	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	9	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	10	infl	infl	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	277	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	278	1	const	const	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	278	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	279	1	0.9	0.9	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	279	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	279	3	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	279	4	24.61	24.61	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	279	5	30.96	30.96	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	279	6	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	279	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	1	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	3	natural	natural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	6	0.0	0.0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	7	14.69	14.69	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	8	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	9	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	10	infl	infl	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	280	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	281	1	const	const	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	281	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	282	1	0.7	0.7	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	282	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	282	3	07.37	07.37	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	282	4	34.77	34.77	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	282	5	Agr	Agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	282	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	1	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	3	natural	natural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	6	0.0	0.0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	7	08.05	08.05	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	8	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	9	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	10	infl	infl	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	283	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	284	1	const	const	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	284	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	1	0.8	0.8	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	3	17.06	17.06	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	4	31.03	31.03	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	5	Form	Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	6	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	8	natural	natural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	9	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	10	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	11	0.0	0.0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	12	01.56	01.56	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	13	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	14	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	15	infl	infl	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	285	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	286	1	const	const	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	286	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	1	0.9	0.9	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	3	10.53	10.53	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	4	09.43	09.43	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	5	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	6	9	9	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	7	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	8	Adapting	adapt	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	9	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	10	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	11	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	12	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	287	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	1	Models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	2	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	3	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	5	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	6	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	7	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	8	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	9	n	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	10	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	11	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	12	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	13	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	15	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	16	feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	288	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	1	Inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	2	constant	constant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	3	shows	show	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	4	how	how	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	5	many	many	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	6	correct	correct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	7	instances	instance	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	8	remain	remain	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	9	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	289	10	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	1	0.9	0.9	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	2	indicates	indicate	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	3	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	4	90	90	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	5	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	7	correct	correct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	8	examples	example	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	9	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	10	un-	un-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	11	changed	change	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	13	while	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	14	10	10	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	15	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	16	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	17	converted	convert	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	18	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	19	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	290	21	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	1	Modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	2	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	4	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	8	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	9	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	10	marked	mark	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	11	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	12	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	13	asterisk	asterisk	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	291	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	292	1	Infl	infl	UH
work_pdhkjvfkbnb4ngi7bpbilobcq4	292	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	1	constant	constant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	2	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	3	Sampling	Sampling	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	4	Inflation	Inflation	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	5	0.90	0.90	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	6	23.22	23.22	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	7	24.61	24.61	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	8	0.85	0.85	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	9	27.75	27.75	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	10	29.29	29.29	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	11	0.80	0.80	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	12	30.04	30.04	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	13	33.47	33.47	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	14	0.70	0.70	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	15	33.02	33.02	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	16	35.52	35.52	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	17	0.60	0.60	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	18	32.78	32.78	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	19	35.03	35.03	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	20	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	21	10	10	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	22	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	23	Comparison	comparison	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	24	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	25	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	26	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	27	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	28	sampling	sample	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	29	meth-	meth-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	30	ods	od	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	31	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	32	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	33	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	34	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	35	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	36	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	293	37	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	2	proportion	proportion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	5	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	6	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	7	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	8	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	9	row	row	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	10	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	11	identical	identical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	294	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	295	1	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	295	2	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	295	3	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	295	4	scored	score	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	295	5	three	three	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	295	6	points	point	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	295	7	below	below	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	295	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	1	Ta-	Ta-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	2	ble	ble	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	3	10	10	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	4	shows	show	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	5	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	6	sampling	sampling	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	7	falls	fall	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	8	behind	behind	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	10	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	11	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	13	since	since	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	14	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	15	considerably	considerably	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	16	reduces	reduce	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	18	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	19	size	size	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	20	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	21	achieve	achieve	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	22	similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	23	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	24	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	296	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	2	proportion	proportion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	5	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	6	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	7	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	8	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	9	row	row	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	10	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	11	identical	identical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	12	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	13	sampling	sample	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	14	achieves	achieve	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	16	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	17	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	18	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	19	removing	remove	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	20	correct	correct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	21	exam-	exam-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	22	ples	ple	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	24	whereas	whereas	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	25	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	26	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	27	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	28	converts	convert	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	29	some	some	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	30	positive	positive	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	31	examples	example	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	32	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	33	artificial	artificial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	34	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	297	35	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	1	Inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	2	constant	constant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	3	shows	show	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	4	how	how	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	5	many	many	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	6	correct	correct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	7	instances	instance	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	8	remain	remain	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	9	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	10	smaller	small	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	11	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	12	values	value	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	13	correspond	correspond	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	14	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	15	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	16	erro-	erro-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	17	neous	neous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	18	instances	instance	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	19	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	20	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	21	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	22	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	23	sampling	sampling	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	24	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	26	correspondingly	correspondingly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	28	removes	remove	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	29	more	more	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	30	positive	positive	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	31	examples	example	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	298	32	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	1	To	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	2	summarize	summarize	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	4	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	5	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	6	demonstrated	demonstrate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	8	impact	impact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	10	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	11	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	12	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	13	comparing	compare	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	14	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	15	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	16	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	17	similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	18	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	19	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	20	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	21	another	another	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	22	team	team	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	23	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	24	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	25	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	26	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	27	shown	show	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	28	that	that	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	29	fur-	fur-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	30	ther	ther	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	31	improvements	improvement	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	32	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	33	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	34	obtained	obtain	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	35	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	36	adapting	adapt	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	37	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	38	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	39	learner	learner	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	40	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	41	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	42	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	43	priors	prior	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	44	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	45	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	46	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	47	train-	train-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	48	ing	ing	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	49	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	50	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	51	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	52	exhibit	exhibit	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	53	similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	54	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	55	patterns	pattern	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	299	56	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	300	1	5.3	5.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	300	2	Dim	Dim	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	300	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	1	3	3	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	2	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	3	Linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	4	Knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	5	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	6	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	8	linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	9	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	10	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	11	important	important	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	12	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	13	sev-	sev-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	14	eral	eral	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	15	components	component	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	18	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	19	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	20	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	21	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	22	fea-	fea-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	23	ture	ture	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	24	engineering	engineering	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	26	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	27	identification	identification	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	28	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	29	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	30	spe-	spe-	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	31	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	32	Features	Features	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	33	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	34	CoNLL	CoNLL	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	35	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	36	Art	Art	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	301	37	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	1	n	n	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	3	gram	gram	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	4	24.61	24.61	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	5	30.96	30.96	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	6	n	n	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	7	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	8	gram+POS+chunk	gram+pos+chunk	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	9	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	10	33.50	33.50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	11	35.66	35.66	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	12	Agr	Agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	302	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	1	n	n	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	3	gram	gram	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	4	17.06	17.06	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	5	31.03	31.03	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	6	n	n	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	7	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	8	gram+POS	gram+POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	9	24.14	24.14	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	10	35.29	35.29	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	11	n	n	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	12	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	13	gram+POS+syntax	gram+POS+syntax	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	14	27.93	27.93	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	15	41.23	41.23	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	16	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	17	11	11	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	18	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	19	Feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	20	evaluation	evaluation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	303	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	1	Models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	2	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	3	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	5	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	6	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	8	use	use	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	10	source	source	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	11	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	12	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	13	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	14	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	304	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	1	Modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	2	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	4	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	8	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	9	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	10	marked	mark	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	11	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	12	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	13	asterisk	asterisk	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	305	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	306	1	cial	cial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	306	2	techniques	technique	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	306	3	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	306	4	correcting	correct	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	306	5	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	306	6	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	306	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	1	Features	feature	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	2	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	4	known	know	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	5	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	6	many	many	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	7	NLP	NLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	8	tasks	task	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	9	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	10	feature	feature	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	11	engineering	engineering	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	12	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	13	important	important	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	15	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	16	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	17	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	19	case	case	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	20	here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	307	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	1	Note	note	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	2	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	4	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	5	relevant	relevant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	6	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	7	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	8	training	train	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	9	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	10	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	11	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	13	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	14	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	15	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	16	on	on	RP
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	17	Web1	web1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	18	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	19	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	20	make	make	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	21	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	22	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	23	n	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	24	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	25	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	26	features	feature	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	27	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	28	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	29	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	31	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	32	cor-	cor-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	33	pus	pus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	34	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	35	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	36	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	37	layers	layer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	38	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	39	linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	40	annotation.10	annotation.10	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	41	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	42	found	find	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	43	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	44	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	45	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	46	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	47	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	48	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	49	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	50	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	51	deeper	deep	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	52	linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	53	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	54	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	55	especially	especially	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	56	beneficial	beneficial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	308	57	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	2	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	3	features	feature	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	4	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	6	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	7	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	9	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	10	addi-	addi-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	11	tion	tion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	12	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	14	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	15	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	18	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	20	encode	encode	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	21	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	22	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	23	shallow	shallow	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	24	parse	parse	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	25	properties	property	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	309	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	1	These	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	2	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	4	pre-	pre-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	5	sented	sente	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	6	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	7	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	8	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	9	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	310	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	2	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	4	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	5	3	3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	6	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	8	Ap-	Ap-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	9	pendix	pendix	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	10	Table	Table	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	11	A.19	A.19	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	311	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	3	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	4	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	6	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	8	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	9	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	10	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	11	further	further	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	12	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	13	reveals	reveal	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	14	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	15	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	16	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	17	better	well	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	18	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	19	train	train	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	20	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	21	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	22	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	23	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	24	rich	rich	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	25	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	312	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	2	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	3	n	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	4	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	5	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	6	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	7	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	8	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	9	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	10	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	12	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	13	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	14	those	those	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	15	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	17	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	18	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	313	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	1	Syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	2	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	3	encode	encode	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	4	properties	property	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	7	subject	subject	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	8	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	10	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	11	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	12	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	13	presented	present	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	14	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	15	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	16	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	17	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	314	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	2	2014b	2014b	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	4	Table	Table	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	5	7	7	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	6	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	8	Appendix	Appendix	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	9	Table	Table	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	10	A.18	A.18	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	11	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	12	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	13	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	14	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	15	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	17	syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	18	parser	parser	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	19	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	20	Klein	Klein	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	21	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	22	Manning	Manning	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	24	2003	2003	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	25	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	26	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	28	dependency	dependency	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	29	converter	converter	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	30	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	31	Marneffe	Marneffe	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	32	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	33	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	34	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	35	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	36	2006	2006	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	37	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	315	38	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	1	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	2	11	11	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	3	shows	show	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	4	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	5	adding	add	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	6	rich	rich	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	7	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	8	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	9	help-	help-	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	10	ful	ful	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	316	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	1	Notably	notably	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	3	adding	add	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	4	deeper	deep	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	5	syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	6	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	7	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	9	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	10	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	11	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	12	useful	useful	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	14	although	although	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	15	parse	parse	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	16	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	17	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	18	likely	likely	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	19	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	20	contain	contain	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	21	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	22	noise.11	noise.11	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	23	Foster	Foster	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	24	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	25	2007	2007	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	26	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	27	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	28	Lee	Lee	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	29	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	30	Seneff	Seneff	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	31	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	32	2008	2008	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	33	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	34	observe	observe	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	35	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	36	degrade	degrade	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	37	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	38	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	39	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	40	syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	41	parsers	parser	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	42	due	due	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	43	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	44	grammat-	grammat-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	45	ical	ical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	46	noise	noise	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	47	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	48	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	49	includes	include	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	50	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	51	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	317	52	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	2	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	4	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	5	chose	choose	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	6	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	7	add	add	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	8	syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	9	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	10	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	11	shallow	shallow	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	12	parse	parse	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	13	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	14	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	15	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	16	likely	likely	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	17	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	18	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	19	sufficient	sufficient	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	20	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	21	arti-	arti-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	22	cles	cles	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	23	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	24	more	more	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	25	accurate	accurate	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	26	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	27	full	full	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	28	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	29	parse	parse	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	30	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	318	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	1	Candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	2	Identification	identification	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	3	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	5	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	6	open	open	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	7	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	8	class	class	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	9	10Feature	10Feature	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	10	engineering	engineering	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	11	will	will	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	12	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	13	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	14	relevant	relevant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	15	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	16	training	train	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	17	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	18	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	19	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	20	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	21	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	22	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	23	linguistic	linguistic	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	24	annotation	annotation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	319	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	1	11Parse	11parse	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	2	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	3	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	4	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	5	been	be	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	6	found	find	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	7	useful	useful	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	9	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	10	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	11	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	12	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	13	Tetreault	Tetreault	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	14	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	15	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	16	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	18	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	19	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	320	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	1	426	426	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	2	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	4	rarely	rarely	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	5	discussed	discuss	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	6	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	8	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	9	crucial	crucial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	10	step	step	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	11	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	12	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	13	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	14	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	15	possible	possible	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	16	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	17	identify	identify	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	19	relevant	relevant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	20	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	21	us-	us-	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	22	ing	e	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	23	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	24	closed	close	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	25	list	list	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	26	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	27	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	28	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	29	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	31	procedure	procedure	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	32	needs	need	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	33	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	34	rely	rely	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	35	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	36	pre	pre	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	37	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	38	processing	processing	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	39	tools	tool	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	40	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	41	whose	whose	WP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	42	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	43	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	44	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	45	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	46	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	47	suboptimal.12	suboptimal.12	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	48	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	49	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	50	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	321	51	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	322	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	322	2	2014b	2014b	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	322	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	322	4	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	322	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	1	5.1	5.1	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	3	describe	describe	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	4	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	5	evaluate	evaluate	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	6	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	7	can-	can-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	8	didate	didate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	9	selection	selection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	10	methods	method	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	11	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	12	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	323	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	3	sys-	sys-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	4	tem	tem	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	5	implements	implement	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	6	their	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	7	best	good	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	8	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	9	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	10	addresses	address	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	11	pre	pre	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	12	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	13	processing	processing	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	14	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	16	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	17	selecting	select	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	18	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	19	tagged	tag	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	20	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	21	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	22	as	as	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	23	well	well	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	24	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	25	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	26	tagged	tag	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	27	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	28	NN	NN	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	29	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	30	whose	whose	WP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	31	lemma	lemma	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	32	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	33	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	34	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	35	list	list	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	36	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	37	valid	valid	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	38	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	39	lemmas	lemma	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	40	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	41	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	324	42	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	325	1	4.3	4.3	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	325	2	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	325	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	1	Following	follow	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	2	descriptions	description	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	3	provided	provide	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	4	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	5	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	6	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	8	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	9	evaluate	evaluate	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	10	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	11	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	12	selection	selection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	13	methods	method	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	14	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	15	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	326	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	2	first	first	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	3	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	4	includes	include	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	5	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	6	tagged	tag	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	7	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	8	NN	NN	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	9	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	10	NNS	NNS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	11	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	12	head	head	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	13	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	14	NP	NP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	327	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	1	NTHU	NTHU	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	2	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	3	HIT	hit	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	4	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	5	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	6	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	7	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	8	NTHU	NTHU	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	9	obtained	obtain	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	11	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	12	best	good	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	13	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	14	score	score	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	16	after	after	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	18	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	19	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	20	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	21	its	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	22	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	23	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	24	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	25	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	26	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	328	27	Web1T.	Web1T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	2	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	3	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	4	includes	include	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	5	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	6	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	7	tagged	tag	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	8	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	9	NN	NN	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	11	NNS	NNS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	12	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	13	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	14	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	15	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	16	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	17	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	18	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	329	20	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	1	SZEG	SZEG	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	3	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	4	Berend	Berend	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	5	et	et	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	6	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	7	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	9	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	10	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	330	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	2	above	above	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	3	procedures	procedure	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	4	suffer	suffer	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	5	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	6	pre	pre	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	7	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	8	processing	processing	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	9	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	331	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	3	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	4	addresses	address	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	5	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	6	problem	problem	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	7	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	8	adding	add	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	9	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	10	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	11	end	end	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	12	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	13	common	common	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	14	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	15	suffixes	suffix	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	332	17	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	1	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	2	ment	ment	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	3	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	5	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	6	ments	ment	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	7	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	9	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	10	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	11	ist	ist	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	12	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	333	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	2	percentage	percentage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	4	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	5	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	6	selected	select	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	7	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	8	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	9	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	10	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	11	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	12	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	14	impact	impact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	15	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	16	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	17	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	18	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	20	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	21	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	22	shown	show	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	23	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	24	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	25	12	12	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	334	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	3	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	4	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	6	best	good	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	7	result	result	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	8	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	9	both	both	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	10	datasets	dataset	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	11	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	12	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	13	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	15	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	16	improves	improve	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	17	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	18	score	score	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	19	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	20	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	21	points	point	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	22	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	23	recovers	recover	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	24	43	43	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	25	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	26	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	28	candi-	candi-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	29	dates	date	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	30	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	31	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	32	missed	miss	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	33	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	34	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	35	first	first	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	36	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	335	37	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	1	On	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	2	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	5	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	6	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	8	able	able	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	9	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	10	recover	recover	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	11	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	12	erroneous	erroneous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	13	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	15	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	16	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	17	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	18	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	19	perform	perform	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	20	as	as	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	21	well	well	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	22	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	23	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	24	last	last	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	25	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	26	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	27	possibly	possibly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	28	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	29	due	due	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	30	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	31	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	32	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	33	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	34	noisy	noisy	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	35	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	36	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	37	generates	generate	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	336	38	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	1	To	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	2	conclude	conclude	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	4	pre	pre	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	5	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	6	processing	processing	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	7	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	8	should	should	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	9	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	10	taken	take	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	11	into	into	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	12	consideration	consideration	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	14	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	15	correcting	correct	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	16	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	18	especially	especially	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	19	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	20	open	open	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	21	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	22	class	class	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	23	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	337	24	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	1	Using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	2	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	3	Finiteness	Finiteness	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	4	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	5	Correct	correct	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	6	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	7	Errors	Errors	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	8	As	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	9	shown	show	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	10	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	11	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	12	4	4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	15	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	16	realizations	realization	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	17	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	18	cor-	cor-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	19	respond	respond	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	20	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	22	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	23	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	24	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	25	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	26	subset	subset	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	27	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	28	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	29	possible	possible	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	30	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	31	realizations	realization	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	32	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	33	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	34	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	35	classi-	classi-	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	36	fier	fier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	338	37	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	1	One	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	2	natural	natural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	3	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	5	thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	8	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	9	train	train	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	10	one	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	11	clas-	clas-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	12	sifier	sifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	13	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	14	predict	predict	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	16	correct	correct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	17	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	18	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	19	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	20	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	21	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	339	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	1	However	however	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	4	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	5	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	6	realization	realization	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	7	may	may	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	8	corre-	corre-	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	9	spond	spond	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	10	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	11	multiple	multiple	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	12	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	13	properties	property	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	340	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	1	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	2	ob-	ob-	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	3	12Candidate	12Candidate	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	4	selection	selection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	6	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	7	difficult	difficult	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	8	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	9	closed	closed	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	10	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	11	class	class	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	12	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	13	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	15	case	case	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	17	omissions	omission	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	341	19	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	1	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	3	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	4	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	5	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	6	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	7	been	be	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	8	studied	study	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	9	rather	rather	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	10	extensively	extensively	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	342	12	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	2	Han	Han	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	3	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	4	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	5	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	7	2006	2006	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	8	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	11	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	12	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	13	no	no	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	14	room	room	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	15	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	16	elaborate	elaborate	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	17	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	18	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	19	here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	343	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	344	1	Candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	344	2	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	344	3	recall	recall	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	344	4	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	344	5	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	344	6	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	344	7	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	344	8	ident	ident	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	344	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	1	method	method	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	2	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	3	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	5	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	6	NP	NP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	7	heads	head	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	8	87.72	87.72	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	9	92.32	92.32	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	10	40.47	40.47	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	11	34.16	34.16	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	12	All	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	13	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	14	89.50	89.50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	15	95.29	95.29	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	16	41.08	41.08	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	17	33.16	33.16	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	18	Nouns+heuristics	Nouns+heuristics	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	19	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	20	92.84	92.84	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	21	94.86	94.86	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	22	42.60	42.60	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	23	34.40	34.40	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	24	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	25	12	12	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	26	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	27	Nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	28	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	29	effect	effect	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	30	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	31	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	32	identification	identification	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	33	methods	method	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	34	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	35	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	36	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	37	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	345	38	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	346	1	Models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	346	2	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	346	3	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	346	4	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	346	5	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	346	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	1	Error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	2	recall	recall	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	3	denotes	denote	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	5	percentage	percentage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	7	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	8	containing	contain	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	9	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	10	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	11	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	12	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	13	selected	select	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	14	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	15	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	347	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	1	Modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	2	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	4	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	8	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	9	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	10	marked	mark	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	11	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	12	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	13	asterisk	asterisk	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	348	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	1	Training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	2	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	3	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	5	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	6	One	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	7	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	8	16.43	16.43	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	9	21.14	21.14	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	10	Finiteness	Finiteness	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	11	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	12	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	13	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	14	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	15	I	i	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	16	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	17	18.59	18.59	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	18	27.72	27.72	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	19	Finiteness	Finiteness	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	20	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	21	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	22	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	23	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	24	II	II	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	25	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	26	21.08	21.08	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	27	29.98	29.98	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	28	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	29	13	13	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	30	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	31	Improvement	improvement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	32	due	due	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	33	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	34	separate	separate	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	35	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	36	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	37	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	38	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	349	39	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	350	1	Models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	350	2	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	350	3	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	350	4	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	350	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	350	6	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	350	7	algorithm	algorithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	350	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	1	servation	servation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	2	motivates	motivate	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	4	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	5	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	6	corrects	correct	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	7	agree-	agree-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	8	ment	ment	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	9	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	10	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	11	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	12	separately	separately	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	13	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	14	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	15	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	16	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	17	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	19	2014b	2014b	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	20	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	351	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	2	uses	use	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	4	linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	5	notion	notion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	7	verb	verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	8	finite-	finite-	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	9	ness	ness	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	10	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	11	Radford	Radford	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	13	1988	1988	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	14	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	15	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	16	distinguishes	distinguish	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	17	between	between	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	18	fi-	fi-	XX
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	19	nite	nite	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	20	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	21	non	non	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	22	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	23	finite	finite	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	24	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	26	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	27	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	28	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	29	fulfill	fulfill	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	30	differ-	differ-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	31	ent	ent	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	32	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	33	functions	function	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	34	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	35	thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	36	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	37	marked	mark	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	38	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	39	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	40	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	41	properties	property	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	352	42	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	1	Verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	2	finiteness	finiteness	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	4	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	5	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	6	direct	direct	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	7	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	8	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	9	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	11	appropriate	appropriate	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	12	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	353	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	2	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	3	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	5	agree-	agree-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	6	ment	ment	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	7	module	module	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	8	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	9	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	10	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	11	take	take	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	12	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	13	markers	marker	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	14	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	16	finite	finite	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	17	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	18	forms	form	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	19	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	20	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	21	be	be	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	22	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	23	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	24	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	25	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	26	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	27	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	28	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	29	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	30	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	31	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	32	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	33	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	34	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	35	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	36	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	37	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	38	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	39	were	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	40	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	41	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	42	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	43	auxiliaries	auxiliary	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	44	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	45	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	46	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	47	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	48	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	49	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	50	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	51	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	52	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	53	finite	finite	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	54	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	55	tagged	tag	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	56	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	57	VB	VB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	58	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	59	VBZ	VBZ	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	60	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	61	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	62	ex-	ex-	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	63	plicit	plicit	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	64	subjects	subject	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	65	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	66	identified	identify	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	67	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	68	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	69	parser	parser	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	70	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	354	71	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	2	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	3	candidates	candidate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	4	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	5	non	non	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	6	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	7	finite	finite	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	8	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	9	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	10	some	some	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	13	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	14	whose	whose	WP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	15	finiteness	finiteness	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	16	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	17	ambiguous	ambiguous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	355	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	1	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	2	13	13	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	3	compares	compare	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	5	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	6	approaches	approach	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	7	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	8	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	9	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	10	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	11	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	12	handled	handle	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	13	together	together	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	14	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	15	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	16	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	17	verbs	verb	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	18	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	19	pro-	pro-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	20	cessed	cesse	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	21	separately	separately	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	356	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	1	All	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	2	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	4	classifiers	classifier	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	5	use	use	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	6	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	7	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	9	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	10	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	13	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	14	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	16	4-word	4-word	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	17	window	window	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	18	around	around	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	20	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	357	21	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	1	Several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	2	subsets	subset	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	4	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	5	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	6	were	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	7	tried	try	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	8	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	10	single	single	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	11	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	12	uses	use	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	14	best	good	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	15	combination	combination	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	17	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	18	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	20	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	21	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	22	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	23	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	24	fea-	fea-	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	25	tures	ture	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	26	shown	show	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	27	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	28	Appendix	Appendix	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	29	Table	Table	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	30	A.19	A.19	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	358	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	1	Finiteness-	Finiteness-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	2	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	3	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	4	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	5	I	i	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	6	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	7	uses	use	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	9	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	10	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	11	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	12	agree-	agree-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	13	ment	ment	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	14	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	15	form	form	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	16	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	18	single	single	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	19	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	359	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	1	When	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	2	training	train	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	3	separately	separately	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	5	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	6	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	7	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	8	explore	explore	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	9	whether	whether	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	10	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	11	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	12	benefit	benefit	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	13	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	14	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	15	fea-	fea-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	16	tures	ture	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	17	;	;	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	18	finiteness	finiteness	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	19	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	20	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	21	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	22	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	23	II	II	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	24	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	25	optimizes	optimize	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	26	fea-	fea-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	27	tures	ture	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	28	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	29	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	30	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	360	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	2	differences	difference	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	3	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	5	fea-	fea-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	6	ture	ture	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	7	sets	set	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	8	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	9	minor	minor	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	11	consist	consist	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	12	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	13	removing	remove	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	14	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	15	427	427	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	16	unigram	unigram	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	17	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	18	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	19	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	20	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	21	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	22	tokens	token	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	23	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	24	do	do	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	25	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	26	appear	appear	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	27	immediately	immediately	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	28	next	next	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	29	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	31	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	361	32	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	1	Recall	recall	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	2	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	4	discussion	discussion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	5	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	6	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	7	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	9	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	10	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	11	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	12	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	13	further	further	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	14	improved	improve	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	15	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	16	adding	add	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	17	syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	18	knowl-	knowl-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	19	edge	edge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	362	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	3	next	next	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	4	section	section	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	6	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	8	shown	show	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	9	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	10	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	11	even	even	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	12	better	well	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	13	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	14	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	15	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	16	train	train	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	17	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	18	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	19	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	20	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	21	agree-	agree-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	22	ment	ment	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	23	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	24	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	25	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	26	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	27	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	28	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	29	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	30	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	363	31	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	2	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	3	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	4	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	5	13	13	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	6	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	7	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	8	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	9	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	10	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	11	sim-	sim-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	12	ilar	ilar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	13	improvements	improvement	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	14	due	due	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	15	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	16	separate	separate	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	17	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	18	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	19	ob-	ob-	XX
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	20	served	serve	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	21	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	22	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	23	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	24	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	25	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	364	26	Web1T.	Web1T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	1	Note	note	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	2	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	4	NTHU	NTHU	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	5	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	6	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	7	corrects	correct	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	8	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	9	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	10	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	11	us-	us-	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	12	ing	e	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	13	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	14	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	15	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	16	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	17	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	18	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	19	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	20	handles	handle	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	21	all	all	PDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	22	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	23	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	24	together	together	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	25	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	26	its	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	27	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	28	module	module	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	29	scored	score	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	30	8	8	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	31	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	32	points	point	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	33	below	below	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	34	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	35	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	36	one	one	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	365	37	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	1	While	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	2	there	there	EX
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	4	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	5	differ-	differ-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	6	ences	ence	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	7	between	between	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	9	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	10	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	13	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	14	suggest	suggest	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	15	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	16	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	17	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	19	improvement	improvement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	20	within	within	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	22	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	23	sys-	sys-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	24	tem	tem	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	25	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	26	indeed	indeed	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	27	due	due	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	28	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	29	handling	handle	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	31	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	32	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	33	sepa-	sepa-	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	34	rately	rately	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	366	35	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	367	1	5.4	5.4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	367	2	Dim	Dim	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	367	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	1	4	4	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	2	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	3	Training	Training	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	4	Data	Data	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	5	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	6	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	7	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	8	large	large	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	9	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	10	produced	produce	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	11	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	12	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	13	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	15	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	16	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	17	background	background	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	18	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	20	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	21	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	368	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	1	Because	because	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	2	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	3	its	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	4	large	large	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	5	size	size	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	7	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	8	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	9	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	10	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	11	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	12	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	13	natural	natural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	14	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	369	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	370	1	Indeed	indeed	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	370	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	370	3	many	many	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	370	4	teams	team	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	370	5	follow	follow	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	370	6	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	370	7	approach	approach	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	370	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	1	On	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	3	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	4	hand	hand	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	6	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	7	important	important	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	8	issue	issue	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	9	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	11	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	12	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	13	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	15	difference	difference	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	16	between	between	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	18	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	19	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	20	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	21	sets	set	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	23	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	24	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	25	impact	impact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	26	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	28	selection	selection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	29	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	31	train-	train-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	32	ing	ing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	33	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	34	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	35	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	36	large	large	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	37	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	38	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	39	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	40	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	41	coverage	coverage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	42	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	43	allows	allow	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	44	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	45	better	well	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	46	generalization	generalization	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	371	47	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	2	show	show	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	3	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	4	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	5	some	some	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	6	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	7	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	8	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	9	especially	especially	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	10	advantageous	advantageous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	11	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	12	train	train	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	13	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	14	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	15	larger	large	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	16	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	17	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	18	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	19	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	372	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	2	should	should	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	3	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	4	noted	note	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	5	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	6	while	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	7	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	8	refer	refer	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	9	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	11	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	12	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	13	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	14	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	15	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	16	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	17	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	19	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	20	certainly	certainly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	21	contains	contain	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	22	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	23	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	24	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	25	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	26	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	27	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	28	assume	assume	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	29	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	31	noise	noise	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	32	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	33	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	34	neglected	neglect	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	373	35	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	1	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	2	14	14	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	3	compares	compare	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	4	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	5	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	6	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	7	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	9	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	10	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	11	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	12	their	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	13	best	good	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	14	configurations	configuration	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	15	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	16	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	18	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	19	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	374	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	1	Overall	overall	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	3	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	4	find	find	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	5	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	6	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	7	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	8	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	9	clearly	clearly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	10	preferable	preferable	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	11	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	12	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	13	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	375	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	2	attribute	attribute	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	4	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	6	observation	observation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	7	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	8	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	9	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	10	usage	usage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	11	strongly	strongly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	12	depends	depend	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	13	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	15	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	16	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	17	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	19	noun	noun	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	21	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	22	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	23	just	just	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	24	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	25	contextual	contextual	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	26	cues	cue	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	27	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	28	syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	29	structure	structure	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	376	30	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	2	example	example	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	4	certain	certain	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	5	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	6	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	7	English	English	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	8	tend	tend	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	9	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	10	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	11	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	12	exclusively	exclusively	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	13	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	14	singular	singular	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	15	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	16	plural	plural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	17	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	377	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	1	Thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	3	con-	con-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	4	siderably	siderably	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	5	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	6	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	7	compared	compare	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	8	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	9	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	10	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	11	types	type	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	12	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	13	required	require	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	14	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	15	learn	learn	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	16	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	17	parameters	parameter	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	378	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	1	On	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	2	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	4	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	5	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	7	native	native	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	8	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	9	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	10	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	11	perform	perform	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	12	slightly	slightly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	13	better	well	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	14	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	15	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	17	while	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	18	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	19	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	20	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	21	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	22	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	23	better	well	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	24	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	25	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	379	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	380	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	380	2	con-	con-	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	380	3	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	380	4	Train	Train	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	380	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	381	1	Learning	Learning	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	381	2	Features	Features	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	381	3	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	381	4	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	381	5	algorithm	algorithm	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	381	6	CoNLL	CoNLL	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	381	7	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	381	8	Art	Art	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	381	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	382	1	Native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	382	2	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	382	3	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	382	4	adapt	adapt	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	382	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	1	n	n	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	3	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	4	34.49	34.49	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	5	31.76	31.76	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	6	Learner	Learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	7	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	8	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	9	infl	infl	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	383	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	384	1	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	384	2	+	+	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	384	3	POS+chunk	pos+chunk	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	384	4	33.50	33.50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	384	5	35.66	35.66	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	384	6	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	384	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	385	1	Native	Native	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	385	2	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	385	3	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	385	4	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	385	5	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	385	6	adapt	adapt	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	385	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	1	n	n	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	3	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	4	12.09	12.09	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	5	32.22	32.22	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	6	Learner	Learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	7	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	8	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	9	infl	infl	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	386	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	1	n	n	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	3	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	4	10.26	10.26	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	5	33.93	33.93	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	6	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	7	Native	Native	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	8	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	9	*	*	NFP
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	10	n	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	11	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	12	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	13	42.60	42.60	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	14	32.38	32.38	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	15	Learner	Learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	16	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	17	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	18	infl	infl	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	387	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	388	1	+	+	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	388	2	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	388	3	19.22	19.22	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	388	4	17.28	17.28	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	388	5	Agr	agr	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	388	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	389	1	Native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	389	2	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	389	3	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	389	4	adapt	adapt	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	389	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	1	n	n	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	3	gram	gram	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	4	23.46	23.46	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	5	38.57	38.57	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	6	Learner	Learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	7	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	8	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	9	infl	infl	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	390	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	1	+	+	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	2	POS+syntax	pos+syntax	XX
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	3	27.93	27.93	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	4	41.23	41.23	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	5	Form	Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	6	Native	Native	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	7	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	8	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	9	adapt	adapt	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	391	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	1	n	n	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	3	gram	gram	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	4	18.35	18.35	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	5	16.67	16.67	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	6	Learner	Learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	7	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	8	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	9	infl	infl	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	392	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	1	+	+	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	2	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	3	12.32	12.32	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	4	12.02	12.02	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	5	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	6	14	14	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	7	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	8	Choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	10	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	11	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	12	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	13	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	14	vs.	vs.	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	15	native	native	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	16	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	17	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	18	T	T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	19	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	393	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	2	prepositions	preposition	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	4	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	5	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	6	chosen	choose	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	7	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	8	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	11	NB-	NB-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	12	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	13	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	14	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	394	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	1	Modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	2	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	4	part	part	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	7	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	8	submis-	submis-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	9	sion	sion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	10	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	11	marked	mark	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	12	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	13	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	14	asterisk	asterisk	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	395	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	1	jecture	jecture	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	2	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	4	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	5	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	6	set	set	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	8	more	more	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	9	similar	similar	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	10	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	12	respective	respective	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	13	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	14	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	15	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	16	thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	17	provides	provide	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	18	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	19	advan-	advan-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	20	tage	tage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	21	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	22	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	23	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	24	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	25	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	396	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	1	On	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	2	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	3	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	6	native	native	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	7	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	8	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	9	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	10	perform	perform	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	11	better	well	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	12	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	13	those	those	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	14	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	15	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	16	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	17	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	19	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	20	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	21	same	same	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	22	n	n	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	23	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	24	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	25	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	26	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	27	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	397	28	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	1	However	however	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	3	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	4	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	5	add	add	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	6	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	8	syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	9	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	11	train-	train-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	12	ing	ing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	13	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	14	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	15	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	16	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	17	advantageous	advantageous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	398	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	1	Finally	finally	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	3	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	4	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	5	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	6	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	8	there	there	EX
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	9	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	10	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	11	advantage	advantage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	12	when	when	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	13	training	train	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	14	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	15	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	16	lot	lot	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	17	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	18	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	19	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	21	although	although	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	22	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	23	difference	difference	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	24	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	25	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	26	as	as	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	27	substantial	substantial	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	28	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	29	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	30	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	31	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	399	32	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	1	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	2	suggests	suggest	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	3	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	4	unlike	unlike	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	5	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	6	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	7	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	8	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	9	better	well	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	10	addressed	address	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	11	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	12	syntax	syntax	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	14	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	15	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	16	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	17	similarly	similarly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	18	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	19	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	20	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	21	benefit	benefit	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	22	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	23	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	24	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	25	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	26	lot	lot	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	27	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	28	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	29	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	30	n	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	31	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	32	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	33	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	400	34	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	1	To	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	2	summarize	summarize	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	4	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	7	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	8	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	9	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	10	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	11	important	important	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	12	consideration	consideration	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	13	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	14	building	build	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	15	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	16	robust	robust	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	17	sys-	sys-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	18	tem	tem	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	401	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	1	Researchers	researcher	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	2	compared	compare	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	3	native-	native-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	4	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	5	learner-	learner-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	6	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	7	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	8	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	9	prepositions	preposition	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	10	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	11	Han	Han	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	12	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	13	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	14	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	16	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	17	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	18	Cahill	Cahill	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	19	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	20	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	21	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	23	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	24	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	26	while	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	28	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	29	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	30	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	31	work	work	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	32	addresses	address	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	33	five	five	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	34	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	35	types	type	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	36	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	37	showing	show	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	38	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	39	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	40	be-	be-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	41	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	42	differently	differently	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	43	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	44	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	45	evaluates	evaluate	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	46	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	47	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	48	corpora.13	corpora.13	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	49	6	6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	50	Discussion	Discussion	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	51	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	52	Table	Table	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	53	15	15	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	54	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	55	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	56	show	show	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	57	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	58	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	59	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	60	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	61	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	62	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	63	where	where	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	64	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	65	best	good	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	66	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	67	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	68	selected	select	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	69	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	70	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	71	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	72	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	73	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	74	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	75	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	76	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	402	77	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	2	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	3	show	show	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	5	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	6	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	8	without	without	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	9	post	post	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	10	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	11	processing	processing	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	12	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	403	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	2	fol-	fol-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	3	lowing	low	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	4	changes	change	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	5	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	6	made	make	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	7	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	8	respect	respect	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	9	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	11	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	12	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	13	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	15	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	16	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	17	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	18	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	19	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	20	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	21	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	22	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	23	enhanced	enhance	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	24	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	25	handle	handle	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	26	spurious	spurious	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	27	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	28	er-	er-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	29	rors	ror	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	30	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	31	thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	32	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	33	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	34	result	result	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	35	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	36	7.10	7.10	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	37	shown	show	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	38	here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	39	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	40	13For	13for	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	41	studies	study	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	42	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	43	directly	directly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	44	combine	combine	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	45	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	46	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	47	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	48	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	49	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	50	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	51	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	52	see	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	53	Gamon	Gamon	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	54	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	55	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	56	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	57	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	58	Dahlmeier	Dahlmeier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	59	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	60	Ng	Ng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	61	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	62	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	63	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	404	64	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	1	428	428	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	2	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	3	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	4	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	5	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	6	work	work	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	7	Model	Model	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	8	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	9	Model	Model	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	10	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	11	Art	Art	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	405	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	406	1	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	406	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	406	3	infl	infl	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	406	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	407	1	33.50	33.50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	407	2	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	407	3	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	407	4	infl	infl	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	407	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	408	1	33.50	33.50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	408	2	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	408	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	409	1	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	409	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	409	3	adapt	adapt	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	409	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	1	07.10	07.10	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	2	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	3	12.09	12.09	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	4	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	5	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	6	42.60	42.60	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	7	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	8	42.60	42.60	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	9	Agr	Agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	410	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	411	1	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	411	2	26.14	26.14	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	411	3	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	411	4	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	411	5	infl	infl	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	411	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	412	1	27.93	27.93	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	412	2	Form	Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	412	3	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	412	4	14.50	14.50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	412	5	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	412	6	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	412	7	adapt	adapt	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	412	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	1	18.35	18.35	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	2	All	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	3	31.43	31.43	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	4	31.75	31.75	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	5	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	6	15	15	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	7	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	8	Results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	9	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	10	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	13	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	14	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	15	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	16	with-	with-	EX
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	17	out	out	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	18	post	post	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	19	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	20	processing	processing	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	21	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	22	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	23	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	24	work	work	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	413	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	1	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	2	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	3	LM	LM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	4	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	5	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	6	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	8	Web1	Web1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	9	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	10	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	11	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	12	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	13	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	14	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	15	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	16	NUCLE	NUCLE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	414	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	1	Modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	2	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	3	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	5	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	6	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	7	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	9	bold	bold	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	415	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	1	different	different	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	2	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	4	12.14	12.14	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	5	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	6	Table	Table	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	7	8)	8)	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	8	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	10	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	11	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	12	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	13	trained	train	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	14	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	16	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	17	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	18	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	19	AP	AP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	20	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	21	rich	rich	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	22	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	23	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	24	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	25	inflation	inflation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	26	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	28	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	29	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	30	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	31	adapted	adapt	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	32	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	33	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	34	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	35	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	36	whereas	whereas	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	37	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	38	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	39	submission	submission	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	40	trains	train	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	41	NB	NB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	42	without	without	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	43	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	416	44	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	2	key	key	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	3	improvements	improvement	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	4	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	5	observed	observe	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	6	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	7	respect	respect	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	8	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	9	least	least	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	10	fre-	fre-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	11	quent	quent	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	12	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	14	so	so	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	16	overall	overall	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	17	improvement	improvement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	18	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	19	small	small	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	417	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	1	Importantly	importantly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	4	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	5	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	6	already	already	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	7	takes	take	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	8	into	into	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	9	account	account	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	11	four	four	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	12	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	13	analyzed	analyze	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	14	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	15	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	16	paper	paper	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	418	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	419	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	419	2	CoNLL-2013	CoNLL-2013	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	419	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	419	4	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	419	5	were	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	419	6	compared	compare	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	419	7	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	419	8	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	419	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	1	Practical	practical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	2	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	4	however	however	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	6	should	should	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	7	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	8	tuned	tune	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	9	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	10	good	good	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	11	precision	precision	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	12	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	13	guarantee	guarantee	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	14	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	16	overall	overall	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	17	qual-	qual-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	18	ity	ity	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	19	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	20	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	21	text	text	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	22	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	23	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	24	go	go	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	25	down	down	RP
work_pdhkjvfkbnb4ngi7bpbilobcq4	420	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	1	Clearly	clearly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	3	optimiz-	optimiz-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	4	ing	ing	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	5	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	6	F1	F1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	7	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	8	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	9	ensure	ensure	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	10	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	12	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	13	improves	improve	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	15	quality	quality	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	18	text	text	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	19	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	20	see	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	21	Appendix	Appendix	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	22	B	B	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	23	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	421	24	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	1	A	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	2	differ-	differ-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	3	ent	ent	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	4	evaluation	evaluation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	5	metric	metric	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	6	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	9	accuracy	accuracy	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	12	data	data	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	13	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	14	proposed	propose	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	15	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	16	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	17	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	18	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	19	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	20	2010b	2010b	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	21	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	422	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	2	further	further	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	3	discussion	discussion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	5	evaluation	evaluation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	6	metrics	metric	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	8	see	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	9	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	10	Wagner	Wagner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	11	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	12	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	13	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	14	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	15	Chodorow	Chodorow	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	16	et	et	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	17	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	423	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	424	1	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	424	2	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	424	3	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	424	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	3	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	4	worth	worth	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	5	noting	note	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	6	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	8	obtained	obtain	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	9	results	result	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	10	underestimate	underestimate	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	12	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	13	because	because	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	15	agree-	agree-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	16	ment	ment	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	17	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	18	what	what	WP
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	19	constitutes	constitute	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	20	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	21	mistake	mistake	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	22	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	23	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	24	quite	quite	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	25	low	low	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	26	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	27	Madnani	Madnani	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	28	et	et	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	29	al	al	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	30	.	.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	31	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	32	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	33	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	34	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	35	so	so	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	36	providing	provide	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	37	alternative	alternative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	38	cor-	cor-	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	39	rections	rection	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	40	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	41	important	important	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	425	42	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	426	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	426	2	revised	revised	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	426	3	annotations	annotation	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	426	4	ad-	ad-	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	426	5	dress	dress	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	426	6	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	426	7	problem	problem	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	426	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	2	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	3	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	4	improves	improve	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	5	its	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	6	F1	f1	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	7	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	8	31.20	31.20	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	9	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	10	42.14	42.14	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	11	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	12	revised	revise	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	13	annotations	annotation	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	427	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	1	However	however	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	3	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	4	numbers	number	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	5	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	6	still	still	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	7	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	8	underestimation	underestimation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	9	because	because	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	11	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	12	typically	typically	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	13	eliminates	eliminate	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	14	precision	precision	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	15	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	16	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	17	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	18	recall	recall	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	19	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	428	20	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	1	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	3	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	4	specific	specific	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	5	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	6	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	7	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	8	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	9	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	10	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	13	false	false	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	14	positives	positive	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	15	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	16	CLC	CLC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	17	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	18	includes	include	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	20	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	21	showed	show	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	22	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	23	increase	increase	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	24	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	25	precision	precision	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	26	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	27	33	33	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	28	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	29	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	30	85	85	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	31	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	32	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	33	33	33	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	34	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	35	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	36	75	75	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	37	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	38	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	39	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	40	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	41	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	42	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	43	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	44	Gamon	Gamon	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	45	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	46	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	47	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	429	48	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	1	An	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	2	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	3	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	6	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	7	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	8	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	9	al-	al-	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	10	lows	low	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	11	us	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	12	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	13	determine	determine	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	14	prominent	prominent	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	15	groups	group	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	17	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	18	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	19	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	20	identify	identify	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	21	areas	area	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	22	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	23	potential	potential	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	24	improvement	improvement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	26	which	which	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	27	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	28	outline	outline	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	29	below	below	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	430	30	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	1	Cascading	Cascading	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	2	NLP	NLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	3	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	4	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	5	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	7	example	example	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	8	below	below	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	11	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	12	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	13	incorrectly	incorrectly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	14	changes	change	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	15	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	16	need	need	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	17	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	18	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	19	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	20	needs	need	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	21	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	22	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	23	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	24	considers	consider	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	25	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	26	victim	victim	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	27	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	28	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	29	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	31	subject	subject	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	32	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	33	that	that	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	34	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	35	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	36	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	37	Also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	38	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	39	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	40	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	41	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	42	kid-	kid-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	43	nappers	napper	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	44	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	45	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	46	victim	victim	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	47	needs	need	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	48	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	49	be	be	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	50	tracked	track	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	51	down	down	RP
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	52	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	53	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	54	also	also	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	55	jailbreakers	jailbreaker	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	56	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	57	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	58	Errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	59	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	60	interacting	interact	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	61	linguis-	linguis-	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	62	tic	tic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	63	structures	structure	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	64	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	65	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	66	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	67	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	68	considers	consider	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	69	every	every	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	70	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	71	independently	independently	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	72	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	73	thus	thus	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	74	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	75	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	76	handle	handle	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	77	interact-	interact-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	78	ing	ing	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	79	phenomena	phenomenon	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	431	80	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	3	example	example	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	4	below	below	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	7	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	10	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	11	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	12	classifiers	classifier	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	13	propose	propose	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	14	corrections	correction	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	15	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	16	result	result	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	17	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	18	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	19	ungrammatical	ungrammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	20	structure	structure	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	21	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	22	such	such	PDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	23	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	24	sit-	sit-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	25	uations	uation	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	26	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	27	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	28	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	29	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	30	such	such	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	31	situation	situation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	32	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	33	individuals	individual	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	34	will	will	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	35	lose	lose	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	36	their	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	37	basic	basic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	38	privacy	privacy	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	39	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	40	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	41	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	42	problem	problem	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	43	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	44	addressed	address	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	45	via	via	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	46	global	global	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	47	models	model	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	48	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	49	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	50	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	51	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	52	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	53	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	54	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	55	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	56	re-	re-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	57	sults	sult	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	58	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	59	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	60	improvement	improvement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	61	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	62	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	63	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	64	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	432	65	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	1	Er-	Er-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	2	rors	ror	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	3	due	due	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	4	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	5	limited	limit	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	6	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	7	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	8	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	9	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	10	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	11	does	do	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	12	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	13	consider	consider	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	14	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	15	beyond	beyond	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	16	sentence	sentence	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	17	level	level	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	433	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	2	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	3	example	example	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	4	below	below	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	7	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	8	incorrectly	incorrectly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	9	proposes	propose	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	10	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	11	delete	delete	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	12	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	14	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	15	but	but	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	16	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	17	wider	wide	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	18	context	context	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	19	indicates	indicate	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	20	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	22	definite	definite	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	23	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	24	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	25	more	more	RBR
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	26	appropriate	appropriate	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	27	here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	28	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	29	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	30	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	31	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	32	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	33	admit	admit	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	34	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	35	how	how	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	36	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	37	prevent	prevent	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	38	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	39	abuse	abuse	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	40	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	41	how	how	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	42	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	43	use	use	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	44	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	45	reasonably	reasonably	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	46	depend	depend	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	47	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	48	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	49	sound	sound	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	50	legal	legal	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	51	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	52	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	53	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	54	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	55	means	mean	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	56	surveillance	surveillance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	57	has	have	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	58	its	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	59	own	own	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	60	restriction	restriction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	61	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	62	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	63	7	7	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	64	Conclusion	Conclusion	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	65	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	66	identified	identify	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	67	key	key	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	68	design	design	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	69	principles	principle	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	70	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	71	developing	develop	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	72	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	73	state	state	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	74	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	75	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	76	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	77	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	78	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	79	art	art	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	80	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	81	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	82	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	434	83	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	2	did	do	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	4	through	through	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	5	analysis	analysis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	8	top	top	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	9	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	10	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	12	CoNLL-	conll-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	13	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	14	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	15	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	16	along	along	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	17	several	several	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	18	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	435	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	2	key	key	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	3	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	4	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	5	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	6	identified	identify	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	8	analyzed	analyze	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	9	con-	con-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	10	cern	cern	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	12	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	13	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	14	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	15	learning	learn	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	16	algorithm	algorithm	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	18	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	19	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	20	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	21	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	23	linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	24	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	26	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	28	choice	choice	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	29	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	30	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	31	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	32	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	436	33	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	2	showed	show	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	3	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	5	de-	de-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	6	cisions	cision	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	7	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	8	each	each	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	9	case	case	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	10	depend	depend	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	11	both	both	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	12	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	14	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	15	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	16	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	17	mistake	mistake	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	18	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	20	specific	specific	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	21	setting	setting	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	22	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	437	23	e.g.	e.g.	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	438	1	how	how	WRB
work_pdhkjvfkbnb4ngi7bpbilobcq4	438	2	much	much	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	438	3	an-	an-	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	438	4	notated	notate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	438	5	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	438	6	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	438	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	438	8	available	available	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	438	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	1	Furthermore	furthermore	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	3	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	4	provided	provide	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	5	points	point	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	7	comparison	comparison	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	8	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	9	other	other	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	10	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	11	along	along	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	12	these	these	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	13	four	four	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	14	dimensions	dimension	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	439	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	1	Acknowledgments	acknowledgment	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	2	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	3	thank	thank	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	4	Peter	Peter	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	5	Chew	Chew	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	6	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	8	anonymous	anonymous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	9	reviewers	reviewer	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	10	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	12	feedback	feedback	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	440	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	1	Most	Most	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	2	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	3	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	4	work	work	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	5	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	6	done	do	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	7	while	while	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	8	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	9	first	first	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	10	author	author	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	11	was	be	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	12	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	14	University	University	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	15	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	16	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	441	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	1	This	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	2	material	material	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	4	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	5	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	6	re-	re-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	7	search	search	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	8	sponsored	sponsor	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	9	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	10	DARPA	DARPA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	11	under	under	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	12	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	13	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	14	FA8750-	FA8750-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	15	13	13	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	16	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	17	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	18	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	19	0008	0008	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	20	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	21	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	22	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	23	Army	Army	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	24	Research	Research	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	25	Laboratory	Laboratory	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	26	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	27	ARL	ARL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	28	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	29	under	under	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	30	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	31	W911NF-09	W911NF-09	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	32	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	33	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	34	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	35	0053	0053	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	442	36	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	1	Any	any	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	2	opinions	opinion	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	4	findings	finding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	6	con-	con-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	7	clusions	clusion	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	8	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	9	recommendations	recommendation	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	10	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	11	those	those	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	12	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	14	authors	author	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	15	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	16	do	do	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	17	not	not	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	18	necessarily	necessarily	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	19	reflect	reflect	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	20	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	21	view	view	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	22	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	23	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	24	agencies	agency	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	443	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	1	429	429	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	2	References	References	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	3	G.	G.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	4	Berend	Berend	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	5	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	6	V.	V.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	7	Vincze	Vincze	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	9	S.	S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	10	Zarrieß	Zarrieß	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	12	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	13	R.	R.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	14	Farkas	Farkas	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	444	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	445	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	445	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	1	Lfg	lfg	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	3	based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	4	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	5	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	6	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	7	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	9	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	10	gram-	gram-	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	11	matical	matical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	12	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	446	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	447	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	447	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	447	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	447	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	447	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	447	6	Shared	Shared	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	447	7	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	447	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	448	1	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	448	2	Brants	Brants	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	448	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	448	4	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	448	5	Franz	Franz	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	448	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	449	1	2006	2006	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	449	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	450	1	Web	web	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	450	2	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	450	3	T	t	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	450	4	5-gram	5-gram	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	450	5	Version	version	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	450	6	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	450	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	451	1	Linguistic	Linguistic	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	451	2	Data	Data	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	451	3	Consortium	Consortium	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	451	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	2	Cahill	Cahill	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	4	N.	N.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	5	Madnani	Madnani	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	7	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	8	Tetreault	Tetreault	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	11	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	12	Napolitano	Napolitano	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	452	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	453	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	453	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	1	Robust	robust	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	2	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	3	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	4	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	5	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	6	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	7	using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	8	wikipedia	wikipedia	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	9	revisions	revision	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	454	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	455	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	455	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	455	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	455	4	NAACL	NAACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	455	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	456	1	S.	S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	456	2	Chen	Chen	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	456	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	456	4	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	456	5	Goodman	Goodman	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	456	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	457	1	1996	1996	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	457	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	1	An	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	2	empirical	empirical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	3	study	study	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	5	smoothing	smooth	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	6	techniques	technique	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	7	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	8	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	9	modeling	modeling	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	458	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	459	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	459	2	Pro-	Pro-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	459	3	ceedings	ceeding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	459	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	459	5	ACL	ACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	459	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	1	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	2	Chodorow	Chodorow	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	4	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	5	Dickinson	Dickinson	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	7	R.	R.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	8	Israel	Israel	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	11	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	12	Tetreault	Tetreault	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	460	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	461	1	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	461	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	462	1	Problems	problem	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	462	2	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	462	3	evaluating	evaluate	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	462	4	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	462	5	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	462	6	de-	de-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	462	7	tection	tection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	462	8	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	462	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	463	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	463	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	463	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	463	4	COLING	COLING	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	463	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	464	1	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	464	2	Dahlmeier	Dahlmeier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	464	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	464	4	H.	H.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	464	5	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	464	6	Ng	Ng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	464	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	465	1	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	465	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	466	1	Grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	466	2	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	466	3	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	466	4	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	466	5	alternating	alternate	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	466	6	structure	structure	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	466	7	optimization	optimization	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	466	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	467	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	467	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	467	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	467	4	ACL	ACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	467	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	468	1	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	468	2	Dahlmeier	Dahlmeier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	468	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	468	4	H.T	H.T	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	468	5	Ng	ng	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	468	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	469	1	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	469	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	1	A	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	2	beam	beam	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	3	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	4	search	search	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	5	decoder	decoder	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	6	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	7	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	8	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	9	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	470	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	471	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	471	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	471	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	471	4	EMNLP	EMNLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	471	5	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	471	6	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	471	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	472	1	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	472	2	Dahlmeier	Dahlmeier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	472	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	472	4	H.T.	H.T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	473	1	Ng	ng	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	473	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	473	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	473	4	S.M.	S.M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	474	1	Wu	Wu	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	474	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	475	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	475	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	1	Build-	build-	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	2	ing	e	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	3	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	4	large	large	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	5	annotated	annotate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	6	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	8	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	9	English	English	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	10	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	11	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	12	NUS	NUS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	13	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	14	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	15	learner	learner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	16	English	English	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	476	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	2	Proceedings	proceeding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	5	NAACL	NAACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	6	Workshop	Workshop	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	8	Innovative	innovative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	9	Use	Use	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	11	NLP	NLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	12	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	13	Build-	Build-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	14	ing	ing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	15	Educational	Educational	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	16	Applications	Applications	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	477	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	478	1	R.	R.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	478	2	Dale	Dale	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	478	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	478	4	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	478	5	Kilgarriff	Kilgarriff	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	478	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	479	1	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	479	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	1	Helping	help	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	2	Our	-PRON-	PRP$
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	3	Own	own	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	4	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	5	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	6	HOO	HOO	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	7	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	8	pilot	pilot	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	9	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	10	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	480	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	2	Proceedings	proceeding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	5	13th	13th	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	6	European	European	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	7	Workshop	Workshop	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	8	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	9	Natural	Natural	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	10	Language	Language	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	11	Gen-	Gen-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	12	eration	eration	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	481	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	1	R.	R.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	2	Dale	Dale	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	4	I.	I.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	5	Anisimoff	Anisimoff	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	8	G.	G.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	9	Narroway	Narroway	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	482	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	483	1	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	483	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	1	A	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	2	re-	re-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	3	port	port	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	6	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	8	determiner	determiner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	9	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	10	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	11	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	12	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	484	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	2	Proceedings	proceeding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	5	NAACL	NAACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	6	Workshop	Workshop	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	8	Innovative	innovative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	9	Use	Use	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	11	NLP	NLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	12	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	13	Building	Building	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	14	Educational	Educational	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	15	Applications	Applications	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	485	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	486	1	Y.	Y.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	486	2	Even	Even	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	486	3	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	486	4	Zohar	Zohar	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	486	5	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	486	6	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	486	7	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	486	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	487	1	2001	2001	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	487	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	488	1	A	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	488	2	sequential	sequential	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	488	3	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	488	4	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	488	5	multi	multi	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	488	6	class	class	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	488	7	classification	classification	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	488	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	489	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	489	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	489	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	489	4	EMNLP	EMNLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	489	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	490	1	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	491	1	Foster	Foster	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	491	2	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	491	3	Ø.	Ø.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	491	4	Andersen	Andersen	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	491	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	492	1	2009	2009	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	492	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	1	Generrate	generrate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	2	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	3	Generating	generate	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	5	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	6	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	7	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	8	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	9	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	10	detection	detection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	493	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	2	Pro-	Pro-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	3	ceedings	ceeding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	6	NAACL	NAACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	7	Workshop	Workshop	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	8	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	9	Innovative	innovative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	10	Use	Use	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	12	NLP	NLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	13	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	14	Building	Building	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	15	Educational	Educational	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	16	Applications	Applications	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	494	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	495	1	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	496	1	Foster	foster	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	496	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	497	1	2007	2007	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	497	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	1	Treebanks	treebank	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	2	gone	go	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	3	bad	bad	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	4	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	5	Generating	generate	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	6	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	7	tree-	tree-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	8	bank	bank	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	9	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	10	ungrammatical	ungrammatical	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	11	english	english	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	498	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	2	Proceedings	proceeding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	5	IJCAI	IJCAI	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	6	Workshop	Workshop	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	8	Analytics	Analytics	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	9	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	10	Noisy	Noisy	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	11	Unstructures	Unstructures	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	12	Data	Data	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	499	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	500	1	Y.	Y.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	500	2	Freund	Freund	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	500	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	500	4	R.	R.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	500	5	E.	E.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	500	6	Schapire	Schapire	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	500	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	501	1	1996	1996	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	501	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	502	1	Experiments	experiment	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	502	2	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	502	3	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	502	4	new	new	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	502	5	boosting	boost	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	502	6	algorithm	algorithm	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	502	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	5	13th	13th	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	6	International	International	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	7	Conference	Conference	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	8	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	9	Machine	Machine	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	10	Learning	Learning	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	503	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	504	1	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	504	2	Gamon	Gamon	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	504	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	505	1	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	505	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	1	Using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	2	mostly	mostly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	3	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	4	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	5	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	6	correct	correct	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	7	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	9	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	10	’	’	POS
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	11	writing	writing	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	506	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	507	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	507	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	507	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	507	4	NAACL	NAACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	507	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	508	1	S.	S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	508	2	Gass	Gass	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	508	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	508	4	L.	L.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	508	5	Selinker	Selinker	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	508	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	509	1	1992	1992	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	509	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	510	1	Language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	510	2	transfer	transfer	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	510	3	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	510	4	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	510	5	learning	learning	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	510	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	511	1	John	John	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	511	2	Benjamins	Benjamins	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	511	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	1	N.	N.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	2	Han	Han	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	4	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	5	Chodorow	Chodorow	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	8	C.	C.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	9	Leacock	Leacock	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	512	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	513	1	2006	2006	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	513	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	1	Detecting	detect	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	2	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	3	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	4	English	english	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	5	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	6	usage	usage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	7	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	8	non	non	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	9	-	-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	10	native	native	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	11	speakers	speaker	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	514	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	1	Journal	Journal	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	2	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	3	Natural	Natural	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	4	Language	Language	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	5	Engineering	Engineering	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	7	12(2):115	12(2):115	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	8	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	9	129	129	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	515	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	1	N.	N.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	2	Han	Han	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	4	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	5	Tetreault	Tetreault	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	7	S.	S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	8	Lee	Lee	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	11	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	12	Ha	Ha	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	516	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	517	1	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	517	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	1	Us-	us-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	2	ing	e	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	3	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	4	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	5	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	6	annotated	annotate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	7	learner	learner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	8	corpus	corpus	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	9	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	10	develop	develop	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	11	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	12	ESL	ESL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	13	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	14	EFL	EFL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	15	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	16	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	17	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	518	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	519	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	519	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	519	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	519	4	LREC	LREC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	519	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	520	1	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	520	2	Ionin	Ionin	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	520	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	520	4	M.L.	M.L.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	521	1	Zubizarreta	Zubizarreta	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	521	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	521	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	521	4	S.	S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	521	5	Bautista	Bautista	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	521	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	522	1	2008	2008	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	522	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	1	Sources	source	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	2	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	3	linguistic	linguistic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	4	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	5	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	7	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	8	lan-	lan-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	9	guage	guage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	10	acquisition	acquisition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	12	English	english	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	13	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	523	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	524	1	Lingua	Lingua	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	524	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	524	3	118:554–576	118:554–576	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	524	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	1	E.	E.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	2	Izumi	Izumi	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	4	K.	K.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	5	Uchimoto	Uchimoto	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	7	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	8	Saiga	Saiga	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	10	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	11	Supnithi	Supnithi	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	13	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	14	H.	H.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	15	Isa-	Isa-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	16	hara	hara	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	525	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	526	1	2003	2003	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	526	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	1	Automatic	automatic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	2	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	3	detection	detection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	4	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	6	Japanese	japanese	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	7	learners	learner	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	8	’	’	POS
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	9	English	english	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	10	spoken	speak	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	11	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	527	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	528	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	528	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	528	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	528	4	ACL	ACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	528	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	1	T.-H.	t.-h.	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	2	Kao	Kao	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	4	Y.-W.	Y.-W.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	5	Chang	Chang	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	7	H.-W.	H.-W.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	8	Chiu	Chiu	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	529	10	T-.H.	t-.h.	ADD
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	1	Yen	Yen	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	3	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	4	Bois-	Bois-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	5	son	son	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	7	J.-C.	J.-C.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	8	Wu	Wu	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	530	11	J.S.	J.S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	531	1	Chang	Chang	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	531	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	532	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	532	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	1	CoNLL-2013	CoNLL-2013	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	2	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	3	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	4	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	5	Grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	6	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	7	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	8	NTHU	NTHU	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	9	sys-	sys-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	10	tem	tem	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	11	description	description	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	533	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	534	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	534	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	534	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	534	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	534	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	534	6	Shared	Shared	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	534	7	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	534	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	535	1	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	535	2	Klein	Klein	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	535	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	535	4	C.	C.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	535	5	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	535	6	Manning	Manning	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	535	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	536	1	2003	2003	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	536	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	1	Fast	fast	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	2	exact	exact	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	3	inference	inference	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	4	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	5	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	6	factored	factored	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	7	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	8	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	9	natural	natural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	10	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	11	parsing	parsing	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	537	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	538	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	538	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	538	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	538	4	NIPS	NIPS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	538	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	1	C.	C.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	2	Leacock	Leacock	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	4	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	5	Chodorow	Chodorow	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	7	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	8	Gamon	Gamon	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	11	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	12	Tetreault	Tetreault	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	539	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	540	1	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	540	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	541	1	Automated	Automated	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	541	2	Grammatical	Grammatical	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	541	3	Error	Error	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	541	4	Detection	Detection	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	541	5	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	541	6	Language	Language	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	541	7	Learners	Learners	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	541	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	542	1	Morgan	Morgan	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	542	2	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	542	3	Claypool	Claypool	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	542	4	Publish-	Publish-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	542	5	ers	er	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	542	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	543	1	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	543	2	Lee	Lee	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	543	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	543	4	S.	S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	543	5	Seneff	Seneff	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	543	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	544	1	2008	2008	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	544	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	545	1	Correcting	correct	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	545	2	misuse	misuse	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	545	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	545	4	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	545	5	forms	form	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	545	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	546	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	546	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	546	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	546	4	ACL	ACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	546	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	1	N.	N.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	2	Madnani	Madnani	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	4	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	5	Chodorow	Chodorow	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	7	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	8	Tetreault	Tetreault	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	11	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	12	Ro-	ro-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	13	zovskaya	zovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	547	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	548	1	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	548	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	1	They	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	2	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	3	help	help	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	4	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	5	Using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	6	crowdsourcing	crowdsource	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	7	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	8	improve	improve	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	10	evaluation	evaluation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	11	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	12	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	13	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	14	detec-	detec-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	15	tion	tion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	16	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	549	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	550	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	550	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	550	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	550	4	ACL	ACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	550	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	551	1	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	551	2	Marneffe	Marneffe	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	551	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	551	4	B.	B.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	551	5	MacCartney	MacCartney	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	551	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	551	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	551	8	Ch	Ch	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	551	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	552	1	Manning	manning	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	552	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	553	1	2006	2006	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	553	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	554	1	Generating	generate	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	554	2	typed	type	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	554	3	dependency	dependency	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	554	4	parses	parse	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	554	5	from	from	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	554	6	phrase	phrase	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	554	7	structure	structure	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	554	8	parses	parse	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	554	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	555	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	555	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	555	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	555	4	LREC	LREC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	555	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	1	H.	H.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	2	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	3	Ng	Ng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	5	S.	S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	6	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	7	Wu	Wu	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	9	Y.	Y.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	10	Wu	Wu	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	12	Ch	Ch	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	556	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	557	1	Hadiwinoto	hadiwinoto	UH
work_pdhkjvfkbnb4ngi7bpbilobcq4	557	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	557	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	557	4	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	557	5	Tetreault	Tetreault	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	557	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	558	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	558	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	559	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	559	2	CoNLL-2013	conll-2013	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	559	3	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	559	4	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	559	5	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	559	6	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	559	7	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	559	8	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	559	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	560	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	560	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	560	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	560	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	560	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	560	6	Shared	Shared	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	560	7	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	560	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	1	H.	H.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	2	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	3	Ng	Ng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	5	S.	S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	6	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	7	Wu	Wu	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	9	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	10	Briscoe	Briscoe	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	12	C.	C.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	13	Hadiwinoto	Hadiwinoto	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	14	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	15	R.	R.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	16	H.	H.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	17	Su-	Su-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	18	santo	santo	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	20	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	21	C.	C.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	22	Bryant	Bryant	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	561	23	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	562	1	2014	2014	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	562	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	563	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	563	2	CoNLL-2014	CoNLL-2014	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	563	3	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	563	4	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	563	5	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	563	6	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	563	7	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	563	8	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	563	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	564	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	564	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	564	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	564	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	564	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	564	6	Shared	Shared	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	564	7	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	564	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	565	1	V.	V.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	565	2	Punyakanok	Punyakanok	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	565	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	565	4	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	565	5	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	565	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	566	1	2001	2001	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	566	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	567	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	567	2	use	use	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	567	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	567	4	classifiers	classifier	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	567	5	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	567	6	sequential	sequential	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	567	7	inference	inference	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	567	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	568	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	568	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	568	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	568	4	NIPS	NIPS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	568	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	569	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	569	2	Radford	Radford	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	569	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	570	1	1988	1988	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	570	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	571	1	Transformational	Transformational	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	571	2	Grammar	Grammar	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	571	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	572	1	Cam-	Cam-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	572	2	bridge	bridge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	572	3	University	University	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	572	4	Press	Press	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	572	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	573	1	N.	N.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	573	2	Rizzolo	Rizzolo	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	573	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	573	4	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	573	5	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	573	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	574	1	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	574	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	1	Learning	learn	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	2	Based	base	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	3	Java	Java	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	4	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	5	Rapid	Rapid	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	6	Development	Development	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	8	NLP	NLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	9	Systems	Systems	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	575	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	576	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	576	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	576	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	576	4	LREC	LREC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	576	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	577	1	430	430	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	577	2	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	577	3	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	577	4	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	577	5	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	577	6	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	577	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	578	1	2010a	2010a	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	578	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	579	1	Annotating	annotate	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	579	2	ESL	ESL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	579	3	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	579	4	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	579	5	Challenges	challenge	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	579	6	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	579	7	rewards	reward	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	579	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	2	Proceedings	proceeding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	5	NAACL	NAACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	6	Workshop	Workshop	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	8	Innovative	innovative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	9	Use	Use	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	11	NLP	NLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	12	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	13	Build-	Build-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	14	ing	ing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	15	Educational	Educational	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	16	Applications	Applications	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	580	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	581	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	581	2	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	581	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	581	4	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	581	5	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	581	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	582	1	2010b	2010b	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	582	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	1	Training	train	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	2	paradigms	paradigms	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	3	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	4	correcting	correct	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	5	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	6	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	7	grammar	grammar	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	8	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	9	usage	usage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	583	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	584	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	584	2	Pro-	Pro-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	584	3	ceedings	ceeding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	584	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	584	5	NAACL	naacl	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	584	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	585	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	585	2	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	585	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	585	4	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	585	5	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	585	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	586	1	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	586	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	1	Algorithm	Algorithm	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	2	selec-	selec-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	3	tion	tion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	4	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	5	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	6	adaptation	adaptation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	7	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	8	ESL	ESL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	9	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	10	tasks	task	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	587	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	588	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	588	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	588	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	588	4	ACL	ACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	588	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	589	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	589	2	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	589	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	589	4	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	589	5	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	589	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	590	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	590	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	1	Joint	joint	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	2	learning	learning	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	4	in-	in-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	5	ference	ference	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	6	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	7	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	8	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	9	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	591	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	592	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	592	2	Proceed-	Proceed-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	592	3	ings	ing	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	592	4	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	592	5	EMNLP	EMNLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	592	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	2	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	4	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	5	Sammons	Sammons	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	7	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	8	Gioja	Gioja	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	11	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	12	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	593	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	594	1	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	594	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	1	University	University	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	2	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	3	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	4	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	5	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	6	HOO	HOO	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	7	text	text	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	8	cor-	cor-	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	9	rection	rection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	10	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	11	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	595	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	5	European	European	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	6	Workshop	Workshop	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	8	Natural	Natural	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	9	Language	Language	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	10	Generation	Generation	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	11	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	12	ENLG	ENLG	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	13	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	596	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	2	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	4	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	5	Sammons	Sammons	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	8	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	9	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	597	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	598	1	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	598	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	2	UI	UI	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	3	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	4	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	6	HOO	HOO	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	7	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	8	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	9	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	10	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	11	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	12	cor-	cor-	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	13	rection	rection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	599	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	2	Proceedings	proceeding	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	5	NAACL	NAACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	6	Workshop	Workshop	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	8	Innovative	innovative	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	9	Use	Use	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	11	NLP	NLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	12	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	13	Building	Building	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	14	Educational	Educational	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	15	Ap-	Ap-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	16	plications	plication	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	600	17	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	2	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	4	K.-W.	K.-W.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	5	Chang	Chang	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	7	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	8	Sammons	Sammons	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	10	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	11	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	12	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	601	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	602	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	602	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	2	University	University	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	4	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	5	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	6	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	7	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	8	CoNLL-2013	CoNLL-2013	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	9	shared	share	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	10	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	603	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	604	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	604	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	604	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	604	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	604	5	Shared	Shared	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	604	6	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	604	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	2	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	4	K.-W.	K.-W.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	5	Chang	Chang	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	7	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	8	Sammons	Sammons	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	10	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	11	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	13	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	14	N.	N.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	15	Habash	Habash	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	605	16	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	606	1	2014a	2014a	ADD
work_pdhkjvfkbnb4ngi7bpbilobcq4	606	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	2	University	University	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	4	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	5	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	6	Columbia	Columbia	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	7	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	8	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	10	CoNLL-2014	CoNLL-2014	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	11	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	12	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	607	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	608	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	608	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	608	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	608	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	608	5	Shared	Shared	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	608	6	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	608	7	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	2	Rozovskaya	Rozovskaya	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	4	D.	D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	5	Roth	Roth	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	8	V.	V.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	9	Srikumar	Srikumar	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	609	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	610	1	2014b	2014b	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	610	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	611	1	Cor-	Cor-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	611	2	recting	recte	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	611	3	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	611	4	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	611	5	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	611	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	612	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	612	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	612	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	612	4	EACL	EACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	612	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	613	1	A.	A.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	613	2	Stolcke	Stolcke	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	613	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	614	1	2002	2002	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	614	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	615	1	Srilm	srilm	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	615	2	-	-	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	615	3	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	615	4	extensible	extensible	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	615	5	language	language	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	615	6	model-	model-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	615	7	ing	ing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	615	8	toolkit	toolkit	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	615	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	4	International	International	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	5	Confer-	Confer-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	6	ence	ence	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	7	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	8	Spoken	Spoken	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	9	Language	Language	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	10	Processing	Processing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	616	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	617	1	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	617	2	Tetreault	Tetreault	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	617	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	617	4	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	618	1	Foster	foster	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	618	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	618	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	618	4	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	618	5	Chodorow	Chodorow	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	618	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	619	1	2010	2010	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	619	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	1	Using	use	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	2	parse	parse	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	3	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	4	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	5	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	6	selection	selection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	8	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	9	de-	de-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	10	tection	tection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	620	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	621	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	621	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	621	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	621	4	ACL	ACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	621	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	622	1	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	622	2	Wagner	Wagner	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	622	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	623	1	2012	2012	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	623	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	1	Detecting	Detecting	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	2	Grammatical	Grammatical	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	3	Errors	Errors	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	4	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	5	Treebank	Treebank	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	6	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	7	Induced	induce	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	8	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	9	Probabilistic	Probabilistic	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	10	Parsers	Parsers	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	624	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	625	1	Ph.D.	Ph.D.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	625	2	the-	the-	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	625	3	sis	sis	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	625	4	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	1	Y.	Y.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	2	Xiang	Xiang	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	4	B.	B.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	5	Yuan	Yuan	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	7	Y.	Y.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	8	Zhang	Zhang	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	10	X.	X.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	11	Wang	Wang	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	13	W.	W.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	14	Zheng	Zheng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	16	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	17	C.	C.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	18	Wei	Wei	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	626	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	627	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	627	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	628	1	A	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	628	2	hybrid	hybrid	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	628	3	model	model	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	628	4	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	628	5	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	628	6	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	628	7	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	628	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	629	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	629	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	629	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	629	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	629	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	629	6	Shared	Shared	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	629	7	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	629	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	630	1	J.	J.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	630	2	Xing	Xing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	630	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	630	4	L.	L.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	630	5	Wang	Wang	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	630	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	630	7	D.F.	D.F.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	631	1	Wong	Wong	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	631	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	631	3	L.S.	L.S.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	632	1	Chao	Chao	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	632	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	632	3	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	632	4	X.	X.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	632	5	Zeng	Zeng	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	632	6	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	633	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	633	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	1	UM	UM	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	2	-	-	HYPH
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	3	Checker	checker	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	4	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	5	A	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	6	hybrid	hybrid	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	7	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	8	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	9	English	english	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	10	grammatical	grammatical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	11	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	12	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	634	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	635	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	635	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	635	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	635	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	635	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	635	6	Shared	Shared	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	635	7	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	635	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	1	H.	H.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	2	Yannakoudakis	Yannakoudakis	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	4	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	5	Briscoe	Briscoe	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	8	B.	B.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	9	Medlock	Medlock	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	636	10	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	637	1	2011	2011	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	637	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	1	A	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	2	new	new	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	3	dataset	dataset	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	4	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	5	method	method	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	6	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	7	automatically	automatically	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	8	grading	grade	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	9	ESOL	ESOL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	10	texts	text	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	638	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	639	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	639	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	639	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	639	4	ACL	ACL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	639	5	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	1	I.	I.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	2	Yoshimoto	Yoshimoto	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	4	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	5	Kose	Kose	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	7	K.	K.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	8	Mitsuzawa	Mitsuzawa	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	9	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	10	K.	K.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	11	Sakaguchi	Sakaguchi	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	13	T.	T.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	14	Mizumoto	Mizumoto	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	16	Y.	Y.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	17	Hayashibe	Hayashibe	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	18	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	19	M.	M.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	20	Komachi	Komachi	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	21	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	22	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	23	Y.	Y.	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	24	Mat-	Mat-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	25	sumoto	sumoto	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	640	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	641	1	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	641	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	1	NAIST	NAIST	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	2	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	3	2013	2013	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	5	grammat-	grammat-	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	6	ical	ical	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	7	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	8	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	9	shared	share	VBD
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	10	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	642	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	643	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	643	2	Proceedings	Proceedings	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	643	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	643	4	CoNLL	CoNLL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	643	5	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	643	6	Shared	Shared	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	643	7	Task	Task	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	643	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	1	431	431	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	2	Appendix	Appendix	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	3	A	A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	4	Features	Features	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	5	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	6	Additional	Additional	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	7	Information	Information	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	8	about	about	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	10	Data	Data	NNPS
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	11	Classifier	Classifier	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	12	Art	Art	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	644	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	645	1	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	645	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	646	1	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	646	2	Agr	Agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	646	3	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	1	Form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	2	Train	train	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	3	43	43	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	4	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	5	20	20	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	6	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	7	39	39	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	8	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	9	22	22	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	10	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	11	37	37	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	12	K	K	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	13	Test	Test	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	14	43	43	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	15	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	16	20	20	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	17	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	18	39	39	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	19	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	20	22	22	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	21	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	22	37	37	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	23	K	k	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	24	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	25	A.16	a.16	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	26	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	27	Number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	28	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	29	candidate	candidate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	30	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	31	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	32	classifier	classifier	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	33	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	34	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	35	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	36	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	37	test	test	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	38	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	39	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	40	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	41	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	647	42	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	1	Error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	2	Number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	4	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	5	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	6	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	7	rate	rate	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	8	Train	train	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	9	Test	Test	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	10	Art	art	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	648	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	1	2336	2336	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	3	5.4	5.4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	4	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	6	2290	2290	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	8	5.3	5.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	9	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	10	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	11	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	649	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	1	1263	1263	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	3	6.4	6.4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	4	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	6	1205	1205	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	8	6.1	6.1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	9	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	10	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	11	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	12	858	858	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	13	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	14	2.2	2.2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	15	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	16	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	17	805	805	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	18	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	19	2.0	2.0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	20	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	21	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	22	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	23	agr	agr	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	650	24	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	1	319	319	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	2	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	3	1.5	1.5	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	4	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	6	330	330	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	8	1.4	1.4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	9	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	10	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	11	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	12	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	13	104	104	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	14	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	15	0.3	0.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	16	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	17	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	18	127	127	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	19	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	20	0.3	0.3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	21	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	22	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	23	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	24	A.17	A.17	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	25	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	26	Statistics	statistic	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	27	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	28	annotated	annotate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	29	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	30	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	31	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	32	FCE	FCE	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	33	cor-	cor-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	34	pus	pus	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	651	35	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	652	1	Percentage	percentage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	652	2	denotes	denote	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	652	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	652	4	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	652	5	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	652	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	652	7	i.e.	i.e.	FW
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	1	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	2	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	4	er-	er-	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	5	roneous	roneous	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	6	instances	instance	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	7	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	8	respect	respect	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	9	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	10	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	11	total	total	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	12	number	number	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	13	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	14	relevant	relevant	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	15	instances	instance	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	16	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	18	data	datum	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	653	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	1	Features	Features	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	2	Description	Description	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	3	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	4	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	5	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	6	subjHead	subjhead	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	7	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	8	subjPOS	subjpos	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	10	surface	surface	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	11	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	12	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	14	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	15	tag	tag	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	16	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	17	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	18	subject	subject	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	19	head	head	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	20	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	21	2	2	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	22	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	23	subjDet	subjDet	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	24	determiner	determiner	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	25	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	26	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	27	subject	subject	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	28	NP	NP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	29	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	30	3	3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	31	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	32	subjDistance	subjdistance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	33	distance	distance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	34	between	between	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	35	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	36	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	37	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	38	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	39	subject	subject	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	40	head	head	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	41	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	42	4	4	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	43	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	44	subjNumber	subjnumber	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	45	Sing	sing	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	46	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	47	singular	singular	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	48	pro-	pro-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	49	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	50	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	51	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	52	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	53	Pl	Pl	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	54	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	55	plural	plural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	56	pronouns	pronoun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	57	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	58	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	59	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	60	5	5	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	61	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	62	subjPerson	subjperson	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	63	3rdSing	3rdsing	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	64	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	65	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	66	she	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	67	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	68	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	69	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	70	he	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	71	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	72	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	73	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	74	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	75	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	76	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	77	singular	singular	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	78	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	79	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	80	Not3rdSing	Not3rdSing	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	81	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	82	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	83	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	84	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	85	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	86	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	87	you	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	88	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	89	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	90	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	91	they	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	92	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	93	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	94	plural	plural	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	95	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	96	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	97	1stSing	1stsing	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	98	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	99	“	"	``
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	100	I	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	101	”	"	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	102	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	103	6	6	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	104	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	105	conjunctions	conjunction	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	106	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	107	1)&(3	1)&(3	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	108	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	109	;	;	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	110	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	111	4)&(5	4)&(5	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	112	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	113	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	114	A.18	A.18	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	115	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	116	Verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	117	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	118	features	feature	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	119	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	120	use	use	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	121	syntactic	syntactic	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	122	knowledge	knowledge	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	654	123	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	1	Appendix	Appendix	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	2	B	B	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	3	Evaluation	Evaluation	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	4	Metrics	Metrics	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	5	Here	here	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	6	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	7	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	8	discuss	discuss	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	10	CoNLL-2013	conll-2013	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	11	shared	share	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	12	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	13	eval-	eval-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	14	uation	uation	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	15	metric	metric	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	16	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	17	provide	provide	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	18	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	19	little	little	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	20	bit	bit	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	21	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	22	detail	detail	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	23	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	24	0	0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	25	10	10	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	26	20	20	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	27	30	30	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	28	40	40	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	29	50	50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	30	60	60	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	31	70	70	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	32	80	80	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	33	0	0	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	34	10	10	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	35	20	20	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	36	30	30	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	37	40	40	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	38	50	50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	39	P	p	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	40	R	r	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	41	E	e	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	42	C	c	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	43	IS	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	44	IO	IO	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	45	N	n	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	46	RECALL	RECALL	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	47	Article	Article	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	48	Prep	Prep	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	49	Noun	Noun	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	50	Agreement	Agreement	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	51	Form	Form	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	52	Figure	Figure	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	53	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	54	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	55	Precision	Precision	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	56	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	57	Recall	Recall	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	58	curves	curve	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	59	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	60	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	61	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	655	62	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	1	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	2	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	3	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	5	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	6	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	7	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	8	this	this	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	9	con-	con-	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	10	text	text	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	656	11	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	657	1	As	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	657	2	shown	show	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	657	3	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	657	4	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	657	5	1	1	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	657	6	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	657	7	Sec	Sec	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	657	8	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	1	2	2	LS
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	3	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	4	90	90	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	5	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	6	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	7	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	8	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	9	about	about	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	10	98	98	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	11	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	12	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	13	training	training	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	14	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	15	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	16	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	17	correctly	correctly	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	658	18	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	2	low	low	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	3	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	4	rates	rate	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	5	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	7	key	key	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	8	reason	reason	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	10	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	11	cor-	cor-	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	12	rection	rection	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	13	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	14	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	15	so	so	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	16	difficult	difficult	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	17	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	18	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	19	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	20	quite	quite	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	21	challenging	challenging	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	22	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	23	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	24	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	25	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	26	improve	improve	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	27	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	28	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	29	writer	writer	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	30	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	31	already	already	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	32	per-	per-	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	33	forms	form	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	34	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	35	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	36	level	level	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	37	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	38	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	39	90	90	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	40	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	659	41	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	1	Indeed	indeed	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	3	very	very	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	4	few	few	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	5	NLP	NLP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	6	tasks	task	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	7	already	already	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	8	have	have	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	9	systems	system	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	10	that	that	WDT
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	11	perform	perform	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	12	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	13	that	that	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	14	level	level	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	660	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	2	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	3	sparsity	sparsity	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	4	makes	make	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	5	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	6	very	very	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	7	challenging	challenging	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	8	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	9	identify	identify	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	10	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	11	accurately	accurately	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	661	12	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	1	In	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	2	fact	fact	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	5	highest	high	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	6	precision	precision	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	7	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	8	46.45	46.45	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	9	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	11	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	12	calculated	calculate	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	13	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	14	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	15	shared	shared	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	16	task	task	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	17	evaluation	evaluation	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	18	metric	metric	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	20	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	21	achieved	achieve	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	22	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	23	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	24	Illinois	Illinois	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	25	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	662	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	1	However	however	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	2	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	3	once	once	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	4	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	5	precision	precision	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	6	drops	drop	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	7	below	below	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	8	50	50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	9	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	10	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	12	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	13	introduces	introduce	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	14	more	more	JJR
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	15	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	16	than	than	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	17	it	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	18	identifies	identify	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	663	19	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	1	We	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	2	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	3	look	look	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	4	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	5	individual	individual	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	6	modules	module	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	7	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	8	see	see	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	9	whether	whether	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	10	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	11	any	any	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	12	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	13	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	14	mistake	mistake	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	16	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	17	improves	improve	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	18	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	19	quality	quality	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	20	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	22	text	text	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	664	23	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	665	1	Fig	fig	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	665	2	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	1	2	2	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	2	shows	show	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	3	Preci-	Preci-	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	4	sion	sion	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	5	/	/	SYM
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	6	Recall	Recall	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	7	curves	curve	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	8	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	10	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	11	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	12	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	13	15	15	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	666	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	1	It	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	2	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	3	interesting	interesting	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	4	to	to	TO
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	5	note	note	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	6	that	that	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	7	performance	performance	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	8	varies	vary	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	9	widely	widely	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	10	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	11	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	12	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	667	13	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	1	The	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	2	easiest	easy	JJS
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	4	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	5	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	6	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	7	usage	usage	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	8	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	9	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	10	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	11	nouns	noun	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	12	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	13	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	14	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	15	do	do	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	16	pretty	pretty	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	17	well	well	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	18	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	19	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	20	recall	recall	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	21	point	point	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	22	20	20	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	23	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	24	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	25	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	26	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	27	corresponding	corresponding	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	28	precision	precision	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	29	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	30	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	31	60	60	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	32	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	33	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	34	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	35	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	36	articles	article	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	37	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	38	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	39	precision	precision	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	40	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	41	around	around	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	42	50	50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	43	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	44	at	at	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	45	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	46	recall	recall	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	47	value	value	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	48	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	49	20	20	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	50	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	668	51	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	1	For	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	2	agreement	agreement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	3	errors	error	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	4	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	5	we	-PRON-	PRP
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	6	can	can	MD
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	7	get	get	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	8	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	9	precision	precision	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	10	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	11	55	55	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	12	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	13	with	with	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	14	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	15	very	very	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	16	high	high	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	17	threshold	threshold	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	18	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	19	identifying	identify	VBG
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	20	only	only	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	21	5	5	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	22	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	23	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	24	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	25	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	669	26	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	1	Fi-	fi-	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	2	nally	nally	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	3	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	4	on	on	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	5	two	two	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	6	mistakes	mistake	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	7	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	8	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	9	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	10	verb	verb	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	11	form	form	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	12	–	–	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	13	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	14	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	15	never	never	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	16	achieves	achieve	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	17	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	18	precision	precision	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	19	over	over	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	20	50	50	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	21	%	%	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	670	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	1	432	432	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	2	Feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	3	type	type	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	4	Feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	5	group	group	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	6	Features	Features	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	7	Word	Word	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	8	n-	n-	UH
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	9	gram	gram	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	10	wB	wB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	11	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	12	w2B	w2B	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	14	w3B	w3B	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	15	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	16	wA	wA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	17	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	18	w2A	w2A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	19	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	20	w3A	w3A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	21	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	22	wBwA	wBwA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	23	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	24	w2BwB	w2bwb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	25	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	26	wAw2A	wAw2A	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	27	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	28	w3Bw2BwB	w3Bw2BwB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	29	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	30	w2BwBwA	w2bwbwa	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	31	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	32	wBwAw2A	wBwAw2A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	33	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	34	wAw2Aw3A	wAw2Aw3A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	35	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	36	w4Bw3Bw2BwB	w4bw3bw2bwb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	37	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	38	w3Bw2BwBwA	w3bw2bwbwa	ADD
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	39	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	40	w2BwBwAw2A	w2bwbwaw2a	UH
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	41	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	42	wBwAw2Aw3A	wBwAw2Aw3A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	43	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	44	wAw2Aw3w4A	wAw2Aw3w4A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	45	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	46	pB	pB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	47	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	48	p2B	p2B	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	49	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	50	p3B	p3B	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	51	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	52	pA	pA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	53	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	54	p2A	p2A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	55	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	56	p3A	p3A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	57	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	58	pBpA	pBpA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	59	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	60	p2BpB	p2bpb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	61	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	62	pAp2A	pAp2A	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	63	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	64	pBwB	pBwB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	65	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	66	pAwA	pAwA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	67	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	68	p2Bw2B	p2Bw2B	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	69	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	70	p2Aw2A	p2Aw2A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	71	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	72	p2BpBpA	p2BpBpA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	73	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	74	pBpAp2A	pBpAp2A	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	75	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	76	pAp2Ap3A	pAp2Ap3A	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	77	Chunk	Chunk	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	78	NP1	NP1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	79	headWord	headWord	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	80	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	81	npWords	npWords	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	82	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	83	NC	NC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	84	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	85	adj&headWord	adj&headWord	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	86	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	87	adjTag&headWord	adjTag&headWord	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	88	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	89	adj&NC	adj&NC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	90	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	91	adjTag&NC	adjtag&nc	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	92	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	93	npTags&headWord	npTags&headWord	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	94	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	95	npTags&NC	npTags&NC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	96	NP2	np2	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	97	headWord&headPOS	headword&headpos	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	98	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	99	headNumber	headnumber	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	100	wordsAfterNP	wordsAfterNP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	101	headWord&wordAfterNP	headWord&wordAfterNP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	102	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	103	npWords&wordAfterNP	npWords&wordAfterNP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	104	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	105	headWord&2wordsAfterNP	headWord&2wordsAfterNP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	106	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	107	npWords&2wordsAfterNP	npWords&2wordsAfterNP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	108	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	109	headWord&3wordsAfterNP	headWord&3wordsAfterNP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	110	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	111	npWords&3wordsAfterNP	npWords&3wordsAfterNP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	112	wordBeforeNP	wordBeforeNP	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	113	wB&fi	wB&fi	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	114	∀i	∀i	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	115	∈	∈	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	116	NP1	NP1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	117	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	118	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	119	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	120	verb&fi	verb&fi	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	121	∀i	∀i	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	122	∈	∈	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	123	NP1	NP1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	124	Preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	125	prep&fi	prep&fi	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	126	∀i	∀i	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	127	∈	∈	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	128	NP1	NP1	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	129	Table	table	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	130	A.19	A.19	''
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	131	:	:	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	132	Features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	133	used	use	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	134	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	135	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	136	article	article	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	137	error	error	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	138	correction	correction	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	139	system	system	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	671	140	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	1	wB	wB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	2	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	3	wA	wA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	4	denote	denote	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	5	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	6	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	7	immediately	immediately	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	8	before	before	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	9	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	10	after	after	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	12	target	target	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	13	,	,	,
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	14	respectively	respectively	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	15	;	;	:
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	16	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	17	pB	pB	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	18	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	19	pA	pA	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	20	denote	denote	VB
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	21	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	22	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	23	tag	tag	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	24	before	before	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	25	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	26	after	after	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	27	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	28	target	target	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	672	29	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	673	1	headWord	headWord	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	673	2	denotes	denote	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	673	3	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	673	4	head	head	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	673	5	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	673	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	673	7	NP	NP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	673	8	complement	complement	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	673	9	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	1	NC	NC	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	2	stands	stand	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	3	for	for	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	4	noun	noun	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	5	compound	compound	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	6	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	7	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	8	active	active	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	9	if	if	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	10	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	11	to	to	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	12	last	last	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	13	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	14	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	15	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	16	NP	NP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	17	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	18	tagged	tag	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	19	as	as	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	20	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	21	noun	noun	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	674	22	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	1	Verb	Verb	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	2	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	4	active	active	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	5	if	if	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	7	NP	NP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	8	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	9	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	10	direct	direct	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	11	object	object	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	12	of	of	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	13	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	14	verb	verb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	675	15	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	1	Preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	2	features	feature	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	3	are	be	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	4	active	active	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	5	if	if	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	7	NP	NP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	8	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	9	immediately	immediately	RB
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	10	preceded	precede	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	11	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	12	a	a	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	13	preposition	preposition	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	676	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	1	Adj	adj	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	2	feature	feature	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	3	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	4	active	active	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	5	if	if	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	6	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	7	first	first	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	8	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	9	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	10	or	or	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	11	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	12	second	second	JJ
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	13	word	word	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	14	preceded	precede	VBN
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	15	by	by	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	16	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	17	adverb	adverb	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	18	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	19	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	20	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	21	NP	NP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	22	is	be	VBZ
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	23	an	an	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	24	adjective	adjective	NN
work_pdhkjvfkbnb4ngi7bpbilobcq4	677	25	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	1	NpWords	NpWords	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	2	and	and	CC
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	3	npTags	npTags	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	4	denote	denote	VBP
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	5	all	all	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	6	words	word	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	7	(	(	-LRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	8	POS	POS	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	9	tags	tag	NNS
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	10	)	)	-RRB-
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	11	in	in	IN
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	12	the	the	DT
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	13	NP	NP	NNP
work_pdhkjvfkbnb4ngi7bpbilobcq4	678	14	.	.	.
work_pdhkjvfkbnb4ngi7bpbilobcq4	679	1	433	433	CD
work_pdhkjvfkbnb4ngi7bpbilobcq4	679	2	434	434	CD