id	sid	tid	token	lemma	pos
work_nu5xpentbrcxdlcdt3lmatbur4	1	1	Transactions	transaction	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	1	2	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	1	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	1	4	Association	Association	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	1	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	1	6	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	1	7	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	1	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	1	9	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	1	10	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	1	11	2013	2013	CD
work_nu5xpentbrcxdlcdt3lmatbur4	1	12	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	1	13	231–242	231–242	CD
work_nu5xpentbrcxdlcdt3lmatbur4	1	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	2	1	Action	Action	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	2	2	Editor	Editor	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	2	3	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	2	4	Noah	Noah	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	2	5	Smith	Smith	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	2	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	3	1	Submitted	submit	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	3	2	11/2012	11/2012	CD
work_nu5xpentbrcxdlcdt3lmatbur4	3	3	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	3	4	Revised	revise	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	3	5	2/2013	2/2013	CD
work_nu5xpentbrcxdlcdt3lmatbur4	3	6	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	3	7	Published	publish	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	3	8	5/2013	5/2013	CD
work_nu5xpentbrcxdlcdt3lmatbur4	3	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	4	1	c	c	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	4	2	©	©	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	4	3	2013	2013	CD
work_nu5xpentbrcxdlcdt3lmatbur4	4	4	Association	Association	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	4	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	4	6	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	4	7	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	4	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	5	1	Modeling	model	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	5	2	Semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	5	3	Relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	5	4	Expressed	express	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	5	5	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	5	6	Prepositions	Prepositions	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	5	7	Vivek	Vivek	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	8	Srikumar	Srikumar	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	5	10	Dan	Dan	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	11	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	12	University	University	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	5	14	Illinois	Illinois	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	5	16	Urbana	Urbana	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	17	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	5	18	Champaign	Champaign	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	19	Urbana	Urbana	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	5	21	IL	IL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	5	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	6	1	61801	61801	CD
work_nu5xpentbrcxdlcdt3lmatbur4	6	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	7	1	{	{	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	7	2	vsrikum2	vsrikum2	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	7	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	7	4	danr}@illinois.edu	danr}@illinois.edu	.
work_nu5xpentbrcxdlcdt3lmatbur4	7	5	Abstract	Abstract	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	7	6	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	7	7	paper	paper	NN
work_nu5xpentbrcxdlcdt3lmatbur4	7	8	introduces	introduce	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	7	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	7	10	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	7	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	7	12	predict-	predict-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	7	13	ing	e	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	7	14	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	7	15	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	7	16	expressed	express	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	7	17	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	7	18	preposi-	preposi-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	7	19	tions	tion	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	7	20	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	7	21	develops	develop	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	7	22	statistical	statistical	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	7	23	learning	learning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	7	24	models	model	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	7	25	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	7	26	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	7	27	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	7	28	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	7	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	7	30	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	7	31	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	7	32	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	7	33	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	7	34	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	7	35	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	7	36	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	7	37	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	7	38	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	7	39	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	8	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	8	2	define	define	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	8	3	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	8	4	inventory	inventory	NN
work_nu5xpentbrcxdlcdt3lmatbur4	8	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	8	6	32	32	CD
work_nu5xpentbrcxdlcdt3lmatbur4	8	7	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	8	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	8	9	build-	build-	VB
work_nu5xpentbrcxdlcdt3lmatbur4	8	10	ing	ing	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	8	11	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	8	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	8	13	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	8	14	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	8	15	disambiguation	disambiguation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	8	16	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	8	17	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	8	18	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	8	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	8	20	collapsing	collapse	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	8	21	related	related	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	8	22	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	8	23	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	8	24	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	8	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	9	1	Given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	9	2	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	9	3	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	9	4	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	9	5	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	9	6	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	9	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	9	8	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	9	9	computational	computational	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	9	10	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	9	11	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	9	12	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	9	13	model	model	VB
work_nu5xpentbrcxdlcdt3lmatbur4	9	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	9	15	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	9	16	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	9	17	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	9	18	its	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	9	19	argu-	argu-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	9	20	ments	ment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	9	21	along	along	IN
work_nu5xpentbrcxdlcdt3lmatbur4	9	22	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	9	23	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	9	24	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	9	25	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	9	26	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	9	27	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	9	28	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	9	29	way	way	NN
work_nu5xpentbrcxdlcdt3lmatbur4	9	30	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	9	31	support	support	VB
work_nu5xpentbrcxdlcdt3lmatbur4	9	32	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	9	33	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	9	34	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	9	35	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	10	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	10	2	an-	an-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	10	3	notated	notate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	10	4	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	10	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	10	6	however	however	RB
work_nu5xpentbrcxdlcdt3lmatbur4	10	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	10	8	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	10	9	provides	provide	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	10	10	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	10	11	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	10	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	10	13	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	10	14	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	10	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	10	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	10	17	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	10	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	10	19	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	10	20	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	10	21	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	10	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	11	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	11	2	address	address	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	11	3	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	11	4	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	11	5	presenting	present	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	11	6	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	11	7	mod-	mod-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	11	8	els	el	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	11	9	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	11	10	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	11	11	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	11	12	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	11	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	12	1	Our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	12	2	gen-	gen-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	12	3	eralization	eralization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	12	4	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	12	5	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	12	6	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	12	7	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	12	8	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	12	9	close	close	RB
work_nu5xpentbrcxdlcdt3lmatbur4	12	10	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	12	11	90	90	CD
work_nu5xpentbrcxdlcdt3lmatbur4	12	12	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	12	13	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	12	14	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	12	15	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	12	16	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	12	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	13	1	Further	further	RB
work_nu5xpentbrcxdlcdt3lmatbur4	13	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	13	3	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	13	4	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	13	5	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	13	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	13	7	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	13	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	13	9	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	13	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	13	11	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	13	12	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	13	13	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	13	14	along	along	IN
work_nu5xpentbrcxdlcdt3lmatbur4	13	15	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	13	16	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	13	17	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	13	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	13	19	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	13	20	show	show	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	13	21	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	13	22	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	13	23	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	13	24	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	13	25	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	13	26	improve	improve	VB
work_nu5xpentbrcxdlcdt3lmatbur4	13	27	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	13	28	re-	re-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	13	29	lation	lation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	13	30	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	13	31	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	13	32	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	13	33	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	13	34	significantly	significantly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	13	35	improve	improve	VB
work_nu5xpentbrcxdlcdt3lmatbur4	13	36	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	13	37	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	13	38	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	13	39	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	14	1	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	14	2	Introduction	introduction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	14	3	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	14	4	paper	paper	NN
work_nu5xpentbrcxdlcdt3lmatbur4	14	5	addresses	address	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	14	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	14	7	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	14	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	14	9	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	14	10	se-	se-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	14	11	mantic	mantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	14	12	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	14	13	conveyed	convey	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	14	14	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	14	15	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	14	16	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	14	17	text	text	NN
work_nu5xpentbrcxdlcdt3lmatbur4	14	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	15	1	Prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	15	2	express	express	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	15	3	many	many	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	15	4	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	15	5	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	15	6	be-	be-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	15	7	tween	tween	NN
work_nu5xpentbrcxdlcdt3lmatbur4	15	8	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	15	9	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	15	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	15	11	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	15	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	16	1	Predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	16	2	these	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	16	3	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	16	4	help	help	VB
work_nu5xpentbrcxdlcdt3lmatbur4	16	5	advancing	advance	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	16	6	text	text	NN
work_nu5xpentbrcxdlcdt3lmatbur4	16	7	understanding	understanding	NN
work_nu5xpentbrcxdlcdt3lmatbur4	16	8	tasks	task	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	16	9	like	like	IN
work_nu5xpentbrcxdlcdt3lmatbur4	16	10	question	question	NN
work_nu5xpentbrcxdlcdt3lmatbur4	16	11	answering	answering	NN
work_nu5xpentbrcxdlcdt3lmatbur4	16	12	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	16	13	textual	textual	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	16	14	entailment	entailment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	16	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	17	1	Consider	consider	VB
work_nu5xpentbrcxdlcdt3lmatbur4	17	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	17	3	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	17	4	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	17	5	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	17	6	1	1	LS
work_nu5xpentbrcxdlcdt3lmatbur4	17	7	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	17	8	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	17	9	book	book	NN
work_nu5xpentbrcxdlcdt3lmatbur4	17	10	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	17	11	Prof.	Prof.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	17	12	Alexander	Alexander	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	17	13	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	17	14	primary	primary	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	17	15	school	school	NN
work_nu5xpentbrcxdlcdt3lmatbur4	17	16	methods	method	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	17	17	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	17	18	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	17	19	valuable	valuable	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	17	20	teaching	teaching	NN
work_nu5xpentbrcxdlcdt3lmatbur4	17	21	resource	resource	NN
work_nu5xpentbrcxdlcdt3lmatbur4	17	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	18	1	Here	here	RB
work_nu5xpentbrcxdlcdt3lmatbur4	18	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	18	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	18	4	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	18	5	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	18	6	indicates	indicate	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	18	7	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	18	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	18	9	book	book	NN
work_nu5xpentbrcxdlcdt3lmatbur4	18	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	18	11	primary	primary	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	18	12	school	school	NN
work_nu5xpentbrcxdlcdt3lmatbur4	18	13	methods	method	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	18	14	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	18	15	connected	connect	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	18	16	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	18	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	18	18	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	18	19	Topic	Topic	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	18	20	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	18	21	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	18	22	indicates	indicate	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	18	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	18	24	Creator-	Creator-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	18	25	Creation	Creation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	18	26	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	18	27	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	18	28	Prof.	Prof.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	18	29	Alexander	Alexander	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	18	30	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	18	31	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	18	32	book	book	NN
work_nu5xpentbrcxdlcdt3lmatbur4	18	33	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	19	1	Predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	19	2	these	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	19	3	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	19	4	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	19	5	help	help	VB
work_nu5xpentbrcxdlcdt3lmatbur4	19	6	answer	answer	VB
work_nu5xpentbrcxdlcdt3lmatbur4	19	7	questions	question	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	19	8	about	about	IN
work_nu5xpentbrcxdlcdt3lmatbur4	19	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	19	10	subject	subject	NN
work_nu5xpentbrcxdlcdt3lmatbur4	19	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	19	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	19	13	book	book	NN
work_nu5xpentbrcxdlcdt3lmatbur4	19	14	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	19	15	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	19	16	rec-	rec-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	19	17	ognize	ognize	VB
work_nu5xpentbrcxdlcdt3lmatbur4	19	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	19	19	entailment	entailment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	19	20	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	19	21	sentences	sentence	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	19	22	like	like	IN
work_nu5xpentbrcxdlcdt3lmatbur4	19	23	Prof.	Prof.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	19	24	Alexan-	Alexan-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	19	25	der	der	NN
work_nu5xpentbrcxdlcdt3lmatbur4	19	26	has	have	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	19	27	written	write	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	19	28	about	about	IN
work_nu5xpentbrcxdlcdt3lmatbur4	19	29	primary	primary	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	19	30	school	school	NN
work_nu5xpentbrcxdlcdt3lmatbur4	19	31	methods	method	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	19	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	20	1	Being	be	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	20	2	highly	highly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	20	3	polysemous	polysemous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	20	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	20	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	20	6	same	same	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	20	7	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	20	8	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	20	9	indicate	indicate	VB
work_nu5xpentbrcxdlcdt3lmatbur4	20	10	different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	20	11	kinds	kind	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	20	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	20	13	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	20	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	20	15	depending	depend	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	20	16	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	20	17	its	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	20	18	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	20	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	20	20	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	20	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	21	1	Furthermore	furthermore	RB
work_nu5xpentbrcxdlcdt3lmatbur4	21	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	21	3	several	several	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	21	4	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	21	5	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	21	6	indicate	indicate	VB
work_nu5xpentbrcxdlcdt3lmatbur4	21	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	21	8	same	same	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	21	9	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	21	10	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	21	11	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	22	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	22	2	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	22	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	22	4	consider	consider	VB
work_nu5xpentbrcxdlcdt3lmatbur4	22	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	22	6	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	22	7	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	22	8	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	22	9	2	2	LS
work_nu5xpentbrcxdlcdt3lmatbur4	22	10	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	22	11	Poor	poor	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	22	12	care	care	NN
work_nu5xpentbrcxdlcdt3lmatbur4	22	13	led	lead	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	22	14	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	22	15	her	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	22	16	death	death	NN
work_nu5xpentbrcxdlcdt3lmatbur4	22	17	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	22	18	pneumonia	pneumonia	NN
work_nu5xpentbrcxdlcdt3lmatbur4	22	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	23	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	23	2	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	23	3	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	23	4	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	23	5	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	23	6	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	23	7	expresses	express	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	23	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	23	9	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	23	10	Cause(death	Cause(death	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	23	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	23	12	pneumonia	pneumonia	NN
work_nu5xpentbrcxdlcdt3lmatbur4	23	13	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	23	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	24	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	24	2	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	24	3	differ-	differ-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	24	4	ent	ent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	24	5	context	context	NN
work_nu5xpentbrcxdlcdt3lmatbur4	24	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	24	7	it	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	24	8	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	24	9	denote	denote	VB
work_nu5xpentbrcxdlcdt3lmatbur4	24	10	other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	24	11	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	24	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	24	13	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	24	14	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	24	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	24	16	phrases	phrase	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	24	17	copied	copy	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	24	18	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	24	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	24	20	film	film	NN
work_nu5xpentbrcxdlcdt3lmatbur4	24	21	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	24	22	Source	source	NN
work_nu5xpentbrcxdlcdt3lmatbur4	24	23	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	24	24	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	24	25	recog-	recog-	VB
work_nu5xpentbrcxdlcdt3lmatbur4	24	26	nized	nize	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	24	27	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	24	28	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	24	29	start	start	NN
work_nu5xpentbrcxdlcdt3lmatbur4	24	30	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	24	31	Temporal	Temporal	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	24	32	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	24	33	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	25	1	On	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	25	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	25	3	other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	25	4	hand	hand	NN
work_nu5xpentbrcxdlcdt3lmatbur4	25	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	25	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	25	7	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	25	8	Cause	Cause	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	25	9	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	25	10	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	25	11	expressed	express	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	25	12	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	25	13	sev-	sev-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	25	14	eral	eral	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	25	15	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	25	16	;	;	,
work_nu5xpentbrcxdlcdt3lmatbur4	25	17	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	25	18	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	25	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	25	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	25	21	following	follow	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	25	22	phrases	phrase	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	25	23	express	express	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	25	24	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	25	25	Cause	Cause	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	25	26	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	25	27	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	25	28	died	die	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	25	29	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	25	30	pneumonia	pneumonia	NN
work_nu5xpentbrcxdlcdt3lmatbur4	25	31	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	25	32	tired	tired	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	25	33	after	after	IN
work_nu5xpentbrcxdlcdt3lmatbur4	25	34	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	25	35	surgery	surgery	NN
work_nu5xpentbrcxdlcdt3lmatbur4	25	36	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	26	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	26	2	characterize	characterize	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	26	3	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	26	4	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	26	5	expressed	express	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	26	6	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	26	7	transitive	transitive	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	26	8	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	26	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	26	10	develop	develop	VB
work_nu5xpentbrcxdlcdt3lmatbur4	26	11	accurate	accurate	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	26	12	models	model	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	26	13	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	26	14	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	26	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	26	16	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	26	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	26	18	identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	26	19	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	26	20	argu-	argu-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	26	21	ments	ment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	26	22	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	26	23	recognizing	recognize	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	26	24	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	26	25	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	26	26	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	26	27	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	26	28	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	26	29	ar-	ar-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	26	30	guments	gument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	26	31	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	27	1	Building	build	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	27	2	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	27	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	27	4	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	27	5	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	27	6	disambigua-	disambigua-	FW
work_nu5xpentbrcxdlcdt3lmatbur4	27	7	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	27	8	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	27	9	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	27	10	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	27	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	27	12	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	27	13	collapse	collapse	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	27	14	semantically	semantically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	27	15	related	relate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	27	16	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	27	17	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	27	18	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	27	19	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	27	20	derive	derive	VB
work_nu5xpentbrcxdlcdt3lmatbur4	27	21	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	27	22	re-	re-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	27	23	lation	lation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	27	24	inventory	inventory	NN
work_nu5xpentbrcxdlcdt3lmatbur4	27	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	28	1	These	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	28	2	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	28	3	act	act	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	28	4	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	28	5	predicates	predicate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	28	6	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	28	7	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	28	8	predicate	predicate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	28	9	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	28	10	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	28	11	representation	representation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	28	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	28	13	where	where	WRB
work_nu5xpentbrcxdlcdt3lmatbur4	28	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	28	15	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	28	16	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	28	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	28	18	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	28	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	28	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	28	21	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	28	22	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	28	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	28	24	231	231	CD
work_nu5xpentbrcxdlcdt3lmatbur4	28	25	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	28	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	29	1	While	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	29	2	ascertaining	ascertain	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	29	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	29	4	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	29	5	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	29	6	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	29	7	largely	largely	RB
work_nu5xpentbrcxdlcdt3lmatbur4	29	8	syntactic	syntactic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	29	9	decision	decision	NN
work_nu5xpentbrcxdlcdt3lmatbur4	29	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	29	11	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	29	12	point	point	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	29	13	out	out	RP
work_nu5xpentbrcxdlcdt3lmatbur4	29	14	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	29	15	syn-	syn-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	29	16	tactic	tactic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	29	17	parsers	parser	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	29	18	do	do	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	29	19	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	29	20	always	always	RB
work_nu5xpentbrcxdlcdt3lmatbur4	29	21	make	make	VB
work_nu5xpentbrcxdlcdt3lmatbur4	29	22	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	29	23	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	29	24	correctly	correctly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	29	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	30	1	However	however	RB
work_nu5xpentbrcxdlcdt3lmatbur4	30	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	30	3	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	30	4	illustrated	illustrate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	30	5	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	30	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	30	7	examples	example	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	30	8	above	above	RB
work_nu5xpentbrcxdlcdt3lmatbur4	30	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	30	10	identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	30	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	30	12	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	30	13	depends	depend	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	30	14	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	30	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	30	16	gov-	gov-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	30	17	ernor	ernor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	30	18	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	30	19	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	30	20	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	30	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	30	22	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	30	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	31	1	Given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	31	2	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	31	3	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	31	4	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	31	5	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	31	6	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	31	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	31	8	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	31	9	goal	goal	NN
work_nu5xpentbrcxdlcdt3lmatbur4	31	10	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	31	11	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	31	12	model	model	VB
work_nu5xpentbrcxdlcdt3lmatbur4	31	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	31	14	predicate	predicate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	31	15	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	31	16	i.e.	i.e.	FW
work_nu5xpentbrcxdlcdt3lmatbur4	32	1	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	32	2	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	32	3	rela-	rela-	IN
work_nu5xpentbrcxdlcdt3lmatbur4	32	4	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	32	5	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	32	6	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	32	7	its	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	32	8	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	32	9	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	32	10	i.e.	i.e.	FW
work_nu5xpentbrcxdlcdt3lmatbur4	33	1	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	33	2	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	33	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	33	4	ob-	ob-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	33	5	ject	ject	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	33	6	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	33	7	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	34	1	Very	very	RB
work_nu5xpentbrcxdlcdt3lmatbur4	34	2	often	often	RB
work_nu5xpentbrcxdlcdt3lmatbur4	34	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	34	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	34	5	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	34	6	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	34	7	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	34	8	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	34	9	influenced	influence	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	34	10	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	34	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	34	12	surface	surface	NN
work_nu5xpentbrcxdlcdt3lmatbur4	34	13	form	form	NN
work_nu5xpentbrcxdlcdt3lmatbur4	34	14	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	34	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	34	16	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	34	17	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	34	18	rather	rather	RB
work_nu5xpentbrcxdlcdt3lmatbur4	34	19	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	34	20	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	34	21	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	34	22	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	34	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	35	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	35	2	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	35	3	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	35	4	2	2	LS
work_nu5xpentbrcxdlcdt3lmatbur4	35	5	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	35	6	above	above	RB
work_nu5xpentbrcxdlcdt3lmatbur4	35	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	35	8	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	35	9	want	want	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	35	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	35	11	predicate	predicate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	35	12	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	35	13	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	35	14	Cause	cause	IN
work_nu5xpentbrcxdlcdt3lmatbur4	35	15	when	when	WRB
work_nu5xpentbrcxdlcdt3lmatbur4	35	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	35	17	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	35	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	35	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	35	20	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	35	21	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	35	22	any	any	DT
work_nu5xpentbrcxdlcdt3lmatbur4	35	23	illness	illness	NN
work_nu5xpentbrcxdlcdt3lmatbur4	35	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	36	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	36	2	thus	thus	RB
work_nu5xpentbrcxdlcdt3lmatbur4	36	3	suggest	suggest	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	36	4	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	36	5	model	model	VB
work_nu5xpentbrcxdlcdt3lmatbur4	36	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	36	7	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	36	8	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	36	9	along	along	IN
work_nu5xpentbrcxdlcdt3lmatbur4	36	10	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	36	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	36	12	preposi-	preposi-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	36	13	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	36	14	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	36	15	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	36	16	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	36	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	36	18	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	36	19	different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	36	20	notions	notion	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	36	21	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	36	22	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	36	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	37	1	These	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	37	2	three	three	CD
work_nu5xpentbrcxdlcdt3lmatbur4	37	3	related	related	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	37	4	aspects	aspect	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	37	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	37	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	37	7	rela-	rela-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	37	8	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	37	9	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	37	10	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	37	11	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	37	12	further	further	RB
work_nu5xpentbrcxdlcdt3lmatbur4	37	13	explained	explain	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	37	14	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	37	15	Section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	37	16	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	37	17	leading	lead	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	37	18	up	up	IN
work_nu5xpentbrcxdlcdt3lmatbur4	37	19	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	37	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	37	21	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	37	22	definition	definition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	37	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	38	1	Though	though	IN
work_nu5xpentbrcxdlcdt3lmatbur4	38	2	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	38	3	wish	wish	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	38	4	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	38	5	predict	predict	VB
work_nu5xpentbrcxdlcdt3lmatbur4	38	6	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	38	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	38	8	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	38	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	38	10	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	38	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	38	12	there	there	EX
work_nu5xpentbrcxdlcdt3lmatbur4	38	13	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	38	14	no	no	DT
work_nu5xpentbrcxdlcdt3lmatbur4	38	15	corpus	corpus	NN
work_nu5xpentbrcxdlcdt3lmatbur4	38	16	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	38	17	annotates	annotate	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	38	18	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	38	19	three	three	CD
work_nu5xpentbrcxdlcdt3lmatbur4	38	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	39	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	39	2	SemEval	SemEval	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	39	3	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	39	4	shared	share	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	39	5	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	39	6	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	39	7	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	39	8	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	39	9	disambiguation	disambiguation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	39	10	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	39	11	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	39	12	provides	provide	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	39	13	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	39	14	an-	an-	IN
work_nu5xpentbrcxdlcdt3lmatbur4	39	15	notations	notation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	39	16	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	39	17	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	39	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	40	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	40	2	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	40	3	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	40	4	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	40	5	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	40	6	gen-	gen-	VB
work_nu5xpentbrcxdlcdt3lmatbur4	40	7	erate	erate	VB
work_nu5xpentbrcxdlcdt3lmatbur4	40	8	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	40	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	40	10	test	test	NN
work_nu5xpentbrcxdlcdt3lmatbur4	40	11	corpora	corpora	NN
work_nu5xpentbrcxdlcdt3lmatbur4	40	12	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	40	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	40	14	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	40	15	la-	la-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	40	16	bels	bel	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	40	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	41	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	41	2	Section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	41	3	4	4	CD
work_nu5xpentbrcxdlcdt3lmatbur4	41	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	41	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	41	6	present	present	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	41	7	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	41	8	models	model	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	41	9	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	41	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	41	11	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	41	12	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	41	13	identification	identification	NN
work_nu5xpentbrcxdlcdt3lmatbur4	41	14	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	41	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	42	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	42	2	first	first	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	42	3	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	42	4	considers	consider	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	42	5	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	42	6	possible	possible	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	42	7	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	42	8	candi-	candi-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	42	9	dates	date	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	42	10	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	42	11	various	various	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	42	12	sources	source	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	42	13	along	along	IN
work_nu5xpentbrcxdlcdt3lmatbur4	42	14	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	42	15	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	42	16	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	42	17	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	42	18	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	42	19	predict	predict	VB
work_nu5xpentbrcxdlcdt3lmatbur4	42	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	42	21	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	42	22	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	42	23	label	label	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	42	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	43	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	43	2	second	second	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	43	3	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	43	4	treats	treat	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	43	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	43	6	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	43	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	43	8	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	43	9	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	43	10	la-	la-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	43	11	tent	tent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	43	12	variables	variable	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	43	13	during	during	IN
work_nu5xpentbrcxdlcdt3lmatbur4	43	14	learning	learn	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	43	15	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	43	16	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	43	17	generalization	generalization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	43	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	43	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	43	20	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	43	21	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	43	22	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	43	23	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	43	24	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	43	25	Yu	Yu	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	43	26	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	43	27	Joachims	Joachims	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	43	28	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	43	29	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	43	30	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	43	31	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	44	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	44	2	show	show	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	44	3	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	44	4	Section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	44	5	5	5	CD
work_nu5xpentbrcxdlcdt3lmatbur4	44	6	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	44	7	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	44	8	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	44	9	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	44	10	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	44	11	predicts	predict	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	44	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	44	13	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	44	14	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	44	15	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	44	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	44	17	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	44	18	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	44	19	im-	im-	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	44	20	proves	prove	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	44	21	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	44	22	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	44	23	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	44	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	45	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	45	2	primary	primary	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	45	3	contributions	contribution	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	45	4	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	45	5	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	45	6	paper	paper	NN
work_nu5xpentbrcxdlcdt3lmatbur4	45	7	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	45	8	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	45	9	1	1	LS
work_nu5xpentbrcxdlcdt3lmatbur4	45	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	46	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	46	2	introduce	introduce	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	46	3	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	46	4	new	new	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	46	5	inventory	inventory	NN
work_nu5xpentbrcxdlcdt3lmatbur4	46	6	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	46	7	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	46	8	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	46	9	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	46	10	covers	cover	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	46	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	46	12	34	34	CD
work_nu5xpentbrcxdlcdt3lmatbur4	46	13	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	46	14	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	46	15	formed	form	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	46	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	46	17	basis	basis	NN
work_nu5xpentbrcxdlcdt3lmatbur4	46	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	46	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	46	20	SemEval	SemEval	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	46	21	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	46	22	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	46	23	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	46	24	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	46	25	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	46	26	disambiguation	disambiguation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	46	27	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	47	1	2	2	LS
work_nu5xpentbrcxdlcdt3lmatbur4	47	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	48	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	48	2	model	model	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	48	3	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	48	4	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	48	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	48	6	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	48	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	48	8	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	48	9	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	48	10	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	48	11	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	48	12	propose	propose	VB
work_nu5xpentbrcxdlcdt3lmatbur4	48	13	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	48	14	learning	learn	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	48	15	algo-	algo-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	48	16	rithm	rithm	NN
work_nu5xpentbrcxdlcdt3lmatbur4	48	17	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	48	18	learns	learn	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	48	19	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	48	20	predict	predict	VB
work_nu5xpentbrcxdlcdt3lmatbur4	48	21	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	48	22	three	three	CD
work_nu5xpentbrcxdlcdt3lmatbur4	48	23	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	48	24	train-	train-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	48	25	ing	ing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	48	26	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	48	27	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	48	28	annotates	annotate	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	48	29	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	48	30	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	48	31	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	48	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	49	1	3	3	LS
work_nu5xpentbrcxdlcdt3lmatbur4	49	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	50	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	50	2	show	show	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	50	3	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	50	4	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	50	5	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	50	6	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	50	7	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	50	8	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	50	9	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	50	10	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	50	11	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	50	12	improves	improve	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	50	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	50	14	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	50	15	pre-	pre-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	50	16	dictor	dictor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	50	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	50	18	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	50	19	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	50	20	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	50	21	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	50	22	significant	significant	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	50	23	improvement	improvement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	50	24	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	50	25	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	50	26	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	50	27	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	51	1	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	51	2	Prepositions	Prepositions	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	51	3	&	&	CC
work_nu5xpentbrcxdlcdt3lmatbur4	51	4	Predicate	Predicate	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	51	5	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	51	6	Argument	Argument	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	51	7	Semantics	Semantics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	51	8	Semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	51	9	role	role	NN
work_nu5xpentbrcxdlcdt3lmatbur4	51	10	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	51	11	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	51	12	cf	cf	NN
work_nu5xpentbrcxdlcdt3lmatbur4	51	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	52	1	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	52	2	Gildea	Gildea	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	52	4	Jurafsky	Jurafsky	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	52	6	2002	2002	CD
work_nu5xpentbrcxdlcdt3lmatbur4	52	7	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	52	8	Palmer	Palmer	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	9	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	10	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	11	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	52	13	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	52	14	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	52	15	Punyakanok	Punyakanok	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	16	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	17	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	18	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	52	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	52	20	2008	2008	CD
work_nu5xpentbrcxdlcdt3lmatbur4	52	21	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	52	22	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	52	23	others	other	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	52	24	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	52	25	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	52	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	52	27	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	52	28	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	52	29	converting	convert	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	52	30	text	text	NN
work_nu5xpentbrcxdlcdt3lmatbur4	52	31	into	into	IN
work_nu5xpentbrcxdlcdt3lmatbur4	52	32	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	52	33	predicate	predicate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	52	34	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	52	35	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	52	36	representation	representation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	52	37	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	53	1	Given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	53	2	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	53	3	trigger	trigger	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	4	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	5	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	53	6	phrase	phrase	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	7	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	53	8	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	53	9	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	53	11	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	53	12	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	13	solves	solve	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	53	14	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	53	15	related	related	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	53	16	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	17	problems	problem	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	53	18	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	53	19	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	53	20	a	a	LS
work_nu5xpentbrcxdlcdt3lmatbur4	53	21	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	53	22	identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	53	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	53	24	rela-	rela-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	53	25	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	26	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	27	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	53	28	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	53	29	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	53	30	b	b	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	31	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	53	32	identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	53	33	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	53	34	labeling	label	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	53	35	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	53	36	argu-	argu-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	37	ments	ment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	53	38	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	53	39	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	53	40	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	53	41	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	54	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	54	2	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	54	3	has	have	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	54	4	been	be	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	54	5	studied	study	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	54	6	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	54	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	54	8	con-	con-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	54	9	text	text	NN
work_nu5xpentbrcxdlcdt3lmatbur4	54	10	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	54	11	verb	verb	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	54	12	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	54	13	nominal	nominal	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	54	14	triggers	trigger	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	54	15	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	54	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	54	17	Prop-	Prop-	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	54	18	Bank	Bank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	19	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	54	20	Palmer	Palmer	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	21	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	22	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	23	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	24	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	54	25	2005	2005	CD
work_nu5xpentbrcxdlcdt3lmatbur4	54	26	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	54	27	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	54	28	NomBank	NomBank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	29	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	54	30	Meyers	Meyers	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	31	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	32	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	33	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	34	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	54	35	2004	2004	CD
work_nu5xpentbrcxdlcdt3lmatbur4	54	36	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	54	37	annotations	annotation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	54	38	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	54	39	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	54	40	Penn	Penn	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	41	Treebank	Treebank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	42	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	54	43	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	54	44	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	54	45	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	54	46	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	54	47	FrameNet	FrameNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	48	lexicon	lexicon	NN
work_nu5xpentbrcxdlcdt3lmatbur4	54	49	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	54	50	Fillmore	Fillmore	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	51	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	52	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	53	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	54	54	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	54	55	2003	2003	CD
work_nu5xpentbrcxdlcdt3lmatbur4	54	56	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	54	57	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	54	58	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	54	59	allows	allow	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	54	60	arbitrary	arbitrary	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	54	61	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	54	62	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	54	63	trigger	trigger	VB
work_nu5xpentbrcxdlcdt3lmatbur4	54	64	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	54	65	frames	frame	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	54	66	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	55	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	55	2	paper	paper	NN
work_nu5xpentbrcxdlcdt3lmatbur4	55	3	focuses	focus	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	55	4	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	55	5	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	55	6	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	55	7	ex-	ex-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	55	8	pressed	press	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	55	9	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	55	10	transitive	transitive	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	55	11	prepositions1	prepositions1	NN
work_nu5xpentbrcxdlcdt3lmatbur4	55	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	56	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	56	2	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	56	3	define	define	VB
work_nu5xpentbrcxdlcdt3lmatbur4	56	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	56	5	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	56	6	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	56	7	tasks	task	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	56	8	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	56	9	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	56	10	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	56	11	follows	follow	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	56	12	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	56	13	identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	56	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	56	15	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	56	16	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	56	17	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	56	18	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	56	19	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	56	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	56	21	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	56	22	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	56	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	56	24	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	56	25	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	56	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	56	27	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	56	28	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	57	1	Preposi-	Preposi-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	57	2	tions	tion	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	57	3	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	57	4	mark	mark	VB
work_nu5xpentbrcxdlcdt3lmatbur4	57	5	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	57	6	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	57	7	both	both	CC
work_nu5xpentbrcxdlcdt3lmatbur4	57	8	core	core	NN
work_nu5xpentbrcxdlcdt3lmatbur4	57	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	57	10	adjunct	adjunct	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	57	11	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	57	12	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	57	13	verbal	verbal	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	57	14	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	57	15	nominal	nominal	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	57	16	predicates	predicate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	57	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	58	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	58	2	addition	addition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	58	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	58	4	they	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	58	5	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	58	6	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	58	7	trigger	trigger	VB
work_nu5xpentbrcxdlcdt3lmatbur4	58	8	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	58	9	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	58	10	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	58	11	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	58	12	part	part	NN
work_nu5xpentbrcxdlcdt3lmatbur4	58	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	58	14	other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	58	15	predicates	predicate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	58	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	59	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	59	2	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	59	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	59	4	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	59	5	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	59	6	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	59	7	3	3	LS
work_nu5xpentbrcxdlcdt3lmatbur4	59	8	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	59	9	below	below	RB
work_nu5xpentbrcxdlcdt3lmatbur4	59	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	59	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	59	12	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	59	13	phrase	phrase	NN
work_nu5xpentbrcxdlcdt3lmatbur4	59	14	starting	start	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	59	15	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	59	16	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	59	17	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	59	18	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	59	19	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	59	20	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	59	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	59	22	verb	verb	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	59	23	visit	visit	NN
work_nu5xpentbrcxdlcdt3lmatbur4	59	24	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	59	25	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	59	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	59	27	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	59	28	triggers	trigger	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	59	29	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	59	30	independent	independent	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	59	31	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	59	32	indicating	indicate	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	59	33	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	59	34	location	location	NN
work_nu5xpentbrcxdlcdt3lmatbur4	59	35	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	59	36	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	59	37	aquarium	aquarium	NN
work_nu5xpentbrcxdlcdt3lmatbur4	59	38	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	60	1	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	60	2	3	3	LS
work_nu5xpentbrcxdlcdt3lmatbur4	60	3	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	60	4	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	60	5	children	child	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	60	6	enjoyed	enjoy	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	60	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	60	8	visit	visit	NN
work_nu5xpentbrcxdlcdt3lmatbur4	60	9	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	60	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	60	11	aquarium	aquarium	NN
work_nu5xpentbrcxdlcdt3lmatbur4	60	12	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	60	13	Coney	Coney	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	60	14	Island	Island	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	60	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	61	1	FrameNet	FrameNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	61	2	covers	cover	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	61	3	some	some	DT
work_nu5xpentbrcxdlcdt3lmatbur4	61	4	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	61	5	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	61	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	61	7	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	61	8	allows	allow	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	61	9	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	61	10	temporal	temporal	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	61	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	61	12	locative	locative	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	61	13	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	61	14	directional	directional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	61	15	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	61	16	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	61	17	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	61	18	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	61	19	evoke	evoke	VB
work_nu5xpentbrcxdlcdt3lmatbur4	61	20	frames	frame	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	61	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	61	22	accounting	account	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	61	23	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	61	24	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	61	25	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	61	26	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	61	27	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	61	28	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	61	29	targets	target	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	61	30	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	61	31	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	61	32	SemEval	SemEval	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	61	33	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	61	34	shared	share	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	61	35	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	61	36	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	61	37	FrameNet	FrameNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	61	38	parsing	parsing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	61	39	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	62	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	62	2	fact	fact	NN
work_nu5xpentbrcxdlcdt3lmatbur4	62	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	62	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	62	5	state	state	NN
work_nu5xpentbrcxdlcdt3lmatbur4	62	6	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	62	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	62	8	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	62	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	62	10	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	62	11	art	art	NN
work_nu5xpentbrcxdlcdt3lmatbur4	62	12	FrameNet	FrameNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	62	13	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	62	14	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	62	15	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	62	16	Das	Das	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	62	17	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	62	18	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	62	19	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	62	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	62	21	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	62	22	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	62	23	does	do	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	62	24	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	62	25	con-	con-	VB
work_nu5xpentbrcxdlcdt3lmatbur4	62	26	sider	sider	VB
work_nu5xpentbrcxdlcdt3lmatbur4	62	27	any	any	DT
work_nu5xpentbrcxdlcdt3lmatbur4	62	28	frame	frame	NN
work_nu5xpentbrcxdlcdt3lmatbur4	62	29	inducing	induce	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	62	30	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	62	31	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	63	1	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	63	2	Baldwin	Baldwin	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	63	3	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	63	4	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	63	5	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	63	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	63	7	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	63	8	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	63	9	highlights	highlight	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	63	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	63	11	importance	importance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	63	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	63	13	studying	study	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	63	14	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	63	15	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	63	16	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	63	17	complete	complete	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	63	18	linguistic	linguistic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	63	19	1By	1by	CD
work_nu5xpentbrcxdlcdt3lmatbur4	63	20	transitive	transitive	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	63	21	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	63	22	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	63	23	refer	refer	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	63	24	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	63	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	63	26	standard	standard	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	63	27	usage	usage	NN
work_nu5xpentbrcxdlcdt3lmatbur4	63	28	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	63	29	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	63	30	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	63	31	take	take	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	63	32	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	63	33	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	63	34	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	64	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	64	2	particular	particular	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	64	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	64	4	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	64	5	do	do	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	64	6	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	64	7	con-	con-	VB
work_nu5xpentbrcxdlcdt3lmatbur4	64	8	sider	sider	VB
work_nu5xpentbrcxdlcdt3lmatbur4	64	9	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	64	10	particles	particle	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	64	11	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	64	12	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	64	13	analysis	analysis	NN
work_nu5xpentbrcxdlcdt3lmatbur4	64	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	65	1	232	232	CD
work_nu5xpentbrcxdlcdt3lmatbur4	65	2	analysis	analysis	NN
work_nu5xpentbrcxdlcdt3lmatbur4	65	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	65	4	sentences	sentence	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	65	5	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	65	6	surveys	survey	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	65	7	work	work	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	65	8	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	65	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	65	10	NLP	NLP	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	65	11	literature	literature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	65	12	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	65	13	addresses	address	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	65	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	65	15	syntax	syntax	NN
work_nu5xpentbrcxdlcdt3lmatbur4	65	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	65	17	semantics	semantic	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	65	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	65	19	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	65	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	66	1	One	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	66	2	line	line	NN
work_nu5xpentbrcxdlcdt3lmatbur4	66	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	66	4	work	work	NN
work_nu5xpentbrcxdlcdt3lmatbur4	66	5	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	66	6	Ye	Ye	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	66	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	66	8	Bald-	Bald-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	66	9	win	win	NN
work_nu5xpentbrcxdlcdt3lmatbur4	66	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	66	11	2006	2006	CD
work_nu5xpentbrcxdlcdt3lmatbur4	66	12	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	66	13	addressed	address	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	66	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	66	15	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	66	16	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	66	17	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	66	18	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	66	19	role	role	NN
work_nu5xpentbrcxdlcdt3lmatbur4	66	20	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	66	21	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	66	22	considering	consider	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	66	23	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	66	24	phrases	phrase	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	66	25	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	66	26	act	act	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	66	27	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	66	28	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	66	29	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	66	30	verbs	verb	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	66	31	according	accord	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	66	32	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	66	33	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	66	34	PropBank	PropBank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	66	35	annotation	annotation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	66	36	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	67	1	They	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	67	2	built	build	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	67	3	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	67	4	system	system	NN
work_nu5xpentbrcxdlcdt3lmatbur4	67	5	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	67	6	predicts	predict	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	67	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	67	8	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	67	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	67	10	these	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	67	11	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	67	12	phrases	phrase	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	67	13	alone	alone	RB
work_nu5xpentbrcxdlcdt3lmatbur4	67	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	68	1	However	however	RB
work_nu5xpentbrcxdlcdt3lmatbur4	68	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	68	3	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	68	4	definition	definition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	68	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	68	6	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	68	7	covered	cover	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	68	8	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	68	9	verb	verb	NN
work_nu5xpentbrcxdlcdt3lmatbur4	68	10	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	68	11	attached	attach	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	68	12	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	68	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	69	1	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	69	2	Zapirain	Zapirain	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	69	3	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	69	4	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	69	5	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	69	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	69	7	2012	2012	CD
work_nu5xpentbrcxdlcdt3lmatbur4	69	8	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	69	9	studied	study	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	69	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	69	11	impact	impact	NN
work_nu5xpentbrcxdlcdt3lmatbur4	69	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	69	13	automatically	automatically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	69	14	learned	learn	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	69	15	selec-	selec-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	69	16	tional	tional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	69	17	preferences	preference	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	69	18	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	69	19	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	69	20	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	69	21	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	69	22	verbs	verb	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	69	23	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	69	24	showed	show	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	69	25	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	69	26	modeling	model	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	69	27	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	69	28	phrases	phrase	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	69	29	sep-	sep-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	69	30	arately	arately	RB
work_nu5xpentbrcxdlcdt3lmatbur4	69	31	improves	improve	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	69	32	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	69	33	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	69	34	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	69	35	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	69	36	pre-	pre-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	69	37	diction	diction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	69	38	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	70	1	Preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	70	2	semantics	semantic	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	70	3	has	have	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	70	4	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	70	5	been	be	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	70	6	studied	study	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	70	7	via	via	IN
work_nu5xpentbrcxdlcdt3lmatbur4	70	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	70	9	Preposition	Preposition	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	70	10	Project	Project	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	70	11	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	70	12	Litkowski	Litkowski	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	70	13	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	70	14	Har-	Har-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	70	15	graves	grave	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	70	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	70	17	2005	2005	CD
work_nu5xpentbrcxdlcdt3lmatbur4	70	18	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	70	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	70	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	70	21	related	related	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	70	22	SemEval	SemEval	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	70	23	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	70	24	shared	share	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	70	25	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	70	26	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	70	27	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	70	28	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	70	29	disambiguation	disambiguation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	70	30	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	70	31	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	70	32	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	70	33	Litkowski	Litkowski	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	70	34	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	70	35	Hargraves	Hargraves	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	70	36	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	70	37	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	70	38	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	70	39	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	71	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	71	2	Preposi-	Preposi-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	71	3	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	71	4	Project	Project	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	71	5	identifies	identify	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	71	6	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	71	7	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	71	8	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	71	9	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	71	10	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	71	11	definitions	definition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	71	12	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	71	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	71	14	Oxford	Oxford	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	71	15	Dictionary	Dictionary	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	71	16	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	71	17	English	English	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	71	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	72	1	There	there	EX
work_nu5xpentbrcxdlcdt3lmatbur4	72	2	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	72	3	332	332	CD
work_nu5xpentbrcxdlcdt3lmatbur4	72	4	different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	72	5	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	72	6	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	72	7	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	72	8	predicted	predict	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	72	9	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	72	10	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	72	11	wide	wide	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	72	12	variance	variance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	72	13	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	72	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	72	15	number	number	NN
work_nu5xpentbrcxdlcdt3lmatbur4	72	16	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	72	17	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	72	18	per	per	IN
work_nu5xpentbrcxdlcdt3lmatbur4	72	19	preposi-	preposi-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	72	20	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	72	21	ranging	range	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	72	22	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	72	23	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	72	24	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	72	25	during	during	IN
work_nu5xpentbrcxdlcdt3lmatbur4	72	26	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	72	27	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	72	28	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	72	29	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	72	30	25	25	CD
work_nu5xpentbrcxdlcdt3lmatbur4	72	31	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	72	32	on	on	RB
work_nu5xpentbrcxdlcdt3lmatbur4	72	33	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	72	34	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	73	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	73	2	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	73	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	73	4	according	accord	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	73	5	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	73	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	73	7	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	73	8	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	73	9	inven-	inven-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	73	10	tory	tory	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	73	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	73	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	73	13	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	73	14	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	73	15	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	73	16	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	73	17	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	73	18	2	2	LS
work_nu5xpentbrcxdlcdt3lmatbur4	73	19	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	73	20	above	above	RB
work_nu5xpentbrcxdlcdt3lmatbur4	73	21	will	will	MD
work_nu5xpentbrcxdlcdt3lmatbur4	73	22	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	73	23	labeled	label	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	73	24	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	73	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	73	26	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	73	27	from:12(9	from:12(9	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	73	28	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	73	29	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	73	30	indicate	indicate	VB
work_nu5xpentbrcxdlcdt3lmatbur4	73	31	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	73	32	cause	cause	NN
work_nu5xpentbrcxdlcdt3lmatbur4	73	33	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	74	1	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	74	2	Dahlmeier	dahlmei	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	74	3	et	et	FW
work_nu5xpentbrcxdlcdt3lmatbur4	74	4	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	74	5	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	74	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	74	7	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	74	8	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	74	9	added	add	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	74	10	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	74	11	anno-	anno-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	74	12	tation	tation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	74	13	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	74	14	seven	seven	CD
work_nu5xpentbrcxdlcdt3lmatbur4	74	15	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	74	16	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	74	17	four	four	CD
work_nu5xpentbrcxdlcdt3lmatbur4	74	18	sections	section	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	74	19	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	74	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	74	21	Penn	Penn	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	74	22	Treebank	Treebank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	74	23	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	74	24	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	74	25	goal	goal	NN
work_nu5xpentbrcxdlcdt3lmatbur4	74	26	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	74	27	studying	study	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	74	28	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	74	29	inter-	inter-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	74	30	action	action	NN
work_nu5xpentbrcxdlcdt3lmatbur4	74	31	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	74	32	verb	verb	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	74	33	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	74	34	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	75	1	Using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	75	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	75	3	SemEval	SemEval	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	75	4	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	75	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	75	6	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	75	7	Tratz	Tratz	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	75	8	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	75	9	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	75	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	75	11	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	75	12	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	75	13	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	75	14	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	75	15	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	75	16	et	et	FW
work_nu5xpentbrcxdlcdt3lmatbur4	75	17	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	75	18	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	75	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	75	20	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	75	21	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	75	22	showed	show	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	75	23	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	75	24	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	75	25	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	75	26	offer	offer	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	75	27	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	75	28	important	important	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	75	29	cue	cue	NN
work_nu5xpentbrcxdlcdt3lmatbur4	75	30	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	75	31	identify	identify	VB
work_nu5xpentbrcxdlcdt3lmatbur4	75	32	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	75	33	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	75	34	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	75	35	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	75	36	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	75	37	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	75	38	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	75	39	Tratz	tratz	NN
work_nu5xpentbrcxdlcdt3lmatbur4	75	40	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	75	41	2011	2011	CD
work_nu5xpentbrcxdlcdt3lmatbur4	75	42	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	75	43	showed	show	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	75	44	further	further	RB
work_nu5xpentbrcxdlcdt3lmatbur4	75	45	im-	im-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	75	46	provements	provement	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	75	47	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	75	48	refining	refine	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	75	49	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	75	50	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	75	51	inventory	inventory	NN
work_nu5xpentbrcxdlcdt3lmatbur4	75	52	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	76	1	How-	How-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	76	2	ever	ever	RB
work_nu5xpentbrcxdlcdt3lmatbur4	76	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	76	4	though	though	IN
work_nu5xpentbrcxdlcdt3lmatbur4	76	5	these	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	76	6	works	work	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	76	7	used	use	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	76	8	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	76	9	dependency	dependency	NN
work_nu5xpentbrcxdlcdt3lmatbur4	76	10	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	76	11	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	76	12	identify	identify	VB
work_nu5xpentbrcxdlcdt3lmatbur4	76	13	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	76	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	76	15	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	76	16	order	order	NN
work_nu5xpentbrcxdlcdt3lmatbur4	76	17	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	76	18	overcome	overcome	VB
work_nu5xpentbrcxdlcdt3lmatbur4	76	19	parsing	parsing	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	76	20	errors	error	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	76	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	76	22	they	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	76	23	augment	augment	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	76	24	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	76	25	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	76	26	’s	’s	POS
work_nu5xpentbrcxdlcdt3lmatbur4	76	27	predictions	prediction	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	76	28	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	76	29	part	part	NN
work_nu5xpentbrcxdlcdt3lmatbur4	76	30	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	76	31	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	76	32	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	76	33	speech	speech	NN
work_nu5xpentbrcxdlcdt3lmatbur4	76	34	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	76	35	heuristics	heuristics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	76	36	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	77	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	77	2	argue	argue	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	77	3	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	77	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	77	5	while	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	77	6	disambiguating	disambiguate	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	77	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	77	8	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	77	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	77	10	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	77	11	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	77	12	does	do	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	77	13	indeed	indeed	RB
work_nu5xpentbrcxdlcdt3lmatbur4	77	14	reveal	reveal	VB
work_nu5xpentbrcxdlcdt3lmatbur4	77	15	nuances	nuance	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	77	16	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	77	17	its	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	77	18	meaning	meaning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	77	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	77	20	it	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	77	21	leads	lead	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	77	22	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	77	23	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	77	24	proliferation	proliferation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	77	25	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	77	26	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	77	27	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	77	28	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	77	29	predicted	predict	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	77	30	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	78	1	Most	most	RBS
work_nu5xpentbrcxdlcdt3lmatbur4	78	2	importantly	importantly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	78	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	78	4	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	78	5	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	78	6	do	do	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	78	7	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	78	8	transfer	transfer	VB
work_nu5xpentbrcxdlcdt3lmatbur4	78	9	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	78	10	other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	78	11	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	78	12	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	78	13	express	express	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	78	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	78	15	same	same	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	78	16	meaning	meaning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	78	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	79	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	79	2	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	79	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	79	4	both	both	DT
work_nu5xpentbrcxdlcdt3lmatbur4	79	5	finish	finish	NN
work_nu5xpentbrcxdlcdt3lmatbur4	79	6	lunch	lunch	NN
work_nu5xpentbrcxdlcdt3lmatbur4	79	7	before	before	IN
work_nu5xpentbrcxdlcdt3lmatbur4	79	8	noon	noon	NN
work_nu5xpentbrcxdlcdt3lmatbur4	79	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	79	10	finish	finish	NN
work_nu5xpentbrcxdlcdt3lmatbur4	79	11	lunch	lunch	NN
work_nu5xpentbrcxdlcdt3lmatbur4	79	12	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	79	13	noon	noon	NN
work_nu5xpentbrcxdlcdt3lmatbur4	79	14	express	express	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	79	15	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	79	16	Temporal	Temporal	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	79	17	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	79	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	80	1	According	accord	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	80	2	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	80	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	80	4	Preposition	Preposition	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	80	5	Project	Project	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	80	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	80	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	80	8	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	80	9	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	80	10	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	80	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	80	12	first	first	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	80	13	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	80	14	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	80	15	before:1(1	before:1(1	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	80	16	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	80	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	80	18	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	80	19	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	80	20	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	80	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	80	22	second	second	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	80	23	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	80	24	by:17(4	by:17(4	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	80	25	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	80	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	81	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	81	2	both	both	DT
work_nu5xpentbrcxdlcdt3lmatbur4	81	3	de-	de-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	81	4	feats	feat	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	81	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	81	6	purpose	purpose	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	81	8	identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	81	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	81	10	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	81	11	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	81	12	aid	aid	VB
work_nu5xpentbrcxdlcdt3lmatbur4	81	13	natural	natural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	81	14	language	language	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	15	understanding	understanding	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	81	17	makes	make	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	81	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	81	19	pre-	pre-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	20	diction	diction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	21	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	22	harder	hard	RBR
work_nu5xpentbrcxdlcdt3lmatbur4	81	23	than	than	IN
work_nu5xpentbrcxdlcdt3lmatbur4	81	24	it	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	81	25	should	should	MD
work_nu5xpentbrcxdlcdt3lmatbur4	81	26	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	81	27	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	81	28	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	81	29	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	81	30	stan-	stan-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	81	31	dard	dard	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	81	32	word	word	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	81	33	sense	sense	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	81	34	classification	classification	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	35	approach	approach	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	36	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	81	37	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	81	38	need	need	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	81	39	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	81	40	train	train	VB
work_nu5xpentbrcxdlcdt3lmatbur4	81	41	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	81	42	separate	separate	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	81	43	classifier	classifier	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	44	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	81	45	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	81	46	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	47	because	because	IN
work_nu5xpentbrcxdlcdt3lmatbur4	81	48	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	81	49	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	81	50	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	81	51	defined	define	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	81	52	per	per	IN
work_nu5xpentbrcxdlcdt3lmatbur4	81	53	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	81	54	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	81	55	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	82	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	82	2	other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	82	3	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	82	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	82	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	82	6	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	82	7	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	82	8	share	share	VB
work_nu5xpentbrcxdlcdt3lmatbur4	82	9	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	82	10	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	82	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	82	12	different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	82	13	prepo-	prepo-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	82	14	sitions	sition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	82	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	83	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	83	2	motivates	motivate	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	83	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	83	4	need	need	NN
work_nu5xpentbrcxdlcdt3lmatbur4	83	5	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	83	6	combine	combine	VB
work_nu5xpentbrcxdlcdt3lmatbur4	83	7	such	such	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	83	8	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	83	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	83	10	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	83	11	into	into	IN
work_nu5xpentbrcxdlcdt3lmatbur4	83	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	83	13	same	same	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	83	14	class	class	NN
work_nu5xpentbrcxdlcdt3lmatbur4	83	15	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	83	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	84	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	84	2	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	84	3	direction	direction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	84	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	84	5	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	84	6	O’Hara	O’Hara	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	84	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	84	8	Wiebe	Wiebe	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	84	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	84	10	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	84	11	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	84	12	de-	de-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	84	13	scribes	scribe	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	84	14	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	84	15	inventory	inventory	NN
work_nu5xpentbrcxdlcdt3lmatbur4	84	16	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	84	17	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	84	18	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	84	19	ob-	ob-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	84	20	tained	taine	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	84	21	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	84	22	Penn	Penn	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	84	23	Treebank	Treebank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	84	24	function	function	NN
work_nu5xpentbrcxdlcdt3lmatbur4	84	25	tags	tag	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	84	26	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	84	27	frame	frame	NN
work_nu5xpentbrcxdlcdt3lmatbur4	84	28	elements	element	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	84	29	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	84	30	FrameNet	FrameNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	84	31	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	85	1	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	85	2	Srikumar	Srikumar	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	85	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	85	4	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	85	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	85	6	2011	2011	CD
work_nu5xpentbrcxdlcdt3lmatbur4	85	7	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	85	8	merged	merge	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	85	9	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	85	10	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	85	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	85	12	seven	seven	CD
work_nu5xpentbrcxdlcdt3lmatbur4	85	13	preposi-	preposi-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	85	14	tions	tion	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	85	15	into	into	IN
work_nu5xpentbrcxdlcdt3lmatbur4	85	16	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	85	17	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	85	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	86	1	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	86	2	Litkowski	Litkowski	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	86	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	86	4	2012	2012	CD
work_nu5xpentbrcxdlcdt3lmatbur4	86	5	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	86	6	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	86	7	suggests	suggest	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	86	8	collapsing	collapse	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	86	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	86	10	definitions	definition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	86	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	86	12	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	86	13	into	into	IN
work_nu5xpentbrcxdlcdt3lmatbur4	86	14	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	86	15	smaller	small	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	86	16	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	86	17	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	86	18	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	86	19	classes	class	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	86	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	87	1	To	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	87	2	aid	aid	VB
work_nu5xpentbrcxdlcdt3lmatbur4	87	3	bet-	bet-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	87	4	ter	ter	NN
work_nu5xpentbrcxdlcdt3lmatbur4	87	5	generalization	generalization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	87	6	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	87	7	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	87	8	reduce	reduce	VB
work_nu5xpentbrcxdlcdt3lmatbur4	87	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	87	10	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	87	11	complex-	complex-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	87	12	ity	ity	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	87	13	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	87	14	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	87	15	follow	follow	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	87	16	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	87	17	line	line	NN
work_nu5xpentbrcxdlcdt3lmatbur4	87	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	87	19	work	work	NN
work_nu5xpentbrcxdlcdt3lmatbur4	87	20	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	87	21	define	define	VB
work_nu5xpentbrcxdlcdt3lmatbur4	87	22	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	87	23	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	87	24	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	87	25	rela-	rela-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	87	26	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	87	27	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	87	28	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	87	29	abstract	abstract	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	87	30	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	87	31	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	87	32	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	87	33	prepo-	prepo-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	87	34	sitions2	sitions2	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	87	35	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	88	1	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	88	2	Preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	88	3	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	88	4	triggered	trigger	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	88	5	Relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	88	6	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	88	7	section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	88	8	describes	describe	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	88	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	88	10	inventory	inventory	NN
work_nu5xpentbrcxdlcdt3lmatbur4	88	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	88	12	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	88	13	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	88	14	introduced	introduce	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	88	15	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	88	16	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	88	17	paper	paper	NN
work_nu5xpentbrcxdlcdt3lmatbur4	88	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	88	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	88	20	then	then	RB
work_nu5xpentbrcxdlcdt3lmatbur4	88	21	identifies	identify	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	88	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	88	23	components	component	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	88	24	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	88	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	88	26	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	88	27	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	88	28	extraction	extraction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	88	29	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	88	30	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	89	1	3.1	3.1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	89	2	Preposition	Preposition	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	89	3	Relation	Relation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	89	4	Inventory	Inventory	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	89	5	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	89	6	build	build	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	89	7	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	89	8	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	89	9	inventory	inventory	NN
work_nu5xpentbrcxdlcdt3lmatbur4	89	10	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	89	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	89	12	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	89	13	an-	an-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	89	14	notation	notation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	89	15	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	89	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	89	17	Preposition	Preposition	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	89	18	Project	Project	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	89	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	89	20	focusing	focus	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	89	21	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	89	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	89	23	34	34	CD
work_nu5xpentbrcxdlcdt3lmatbur4	89	24	prepositions3	prepositions3	NN
work_nu5xpentbrcxdlcdt3lmatbur4	89	25	annotated	annotate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	89	26	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	89	27	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	89	28	SemEval-2007	semeval-2007	NN
work_nu5xpentbrcxdlcdt3lmatbur4	89	29	shared	share	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	89	30	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	89	31	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	89	32	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	89	33	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	89	34	disambiguation	disambiguation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	89	35	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	90	1	As	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	90	2	discussed	discuss	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	90	3	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	90	4	Section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	90	5	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	90	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	90	7	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	90	8	construct	construct	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	90	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	90	10	in-	in-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	90	11	ventory	ventory	NN
work_nu5xpentbrcxdlcdt3lmatbur4	90	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	90	13	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	90	14	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	90	15	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	90	16	collapsing	collapse	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	90	17	se-	se-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	90	18	mantically	mantically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	90	19	related	related	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	90	20	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	90	21	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	90	22	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	90	23	differ-	differ-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	90	24	2Since	2since	CD
work_nu5xpentbrcxdlcdt3lmatbur4	90	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	90	26	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	90	27	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	90	28	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	90	29	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	90	30	annotated	annotate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	90	31	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	90	32	FrameNet	FrameNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	90	33	sentences	sentence	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	90	34	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	90	35	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	90	36	annotation	annotation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	90	37	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	90	38	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	90	39	used	use	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	90	40	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	90	41	extend	extend	VB
work_nu5xpentbrcxdlcdt3lmatbur4	90	42	FrameNet	FrameNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	90	43	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	90	44	Litkowski	Litkowski	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	90	45	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	90	46	2012	2012	CD
work_nu5xpentbrcxdlcdt3lmatbur4	90	47	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	90	48	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	91	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	91	2	believe	believe	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	91	3	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	91	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	91	5	abstract	abstract	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	91	6	la-	la-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	91	7	bels	bel	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	91	8	proposed	propose	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	91	9	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	91	10	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	91	11	paper	paper	NN
work_nu5xpentbrcxdlcdt3lmatbur4	91	12	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	91	13	further	further	RB
work_nu5xpentbrcxdlcdt3lmatbur4	91	14	help	help	VB
work_nu5xpentbrcxdlcdt3lmatbur4	91	15	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	91	16	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	91	17	effort	effort	NN
work_nu5xpentbrcxdlcdt3lmatbur4	91	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	92	1	3We	3we	CD
work_nu5xpentbrcxdlcdt3lmatbur4	92	2	consider	consider	VB
work_nu5xpentbrcxdlcdt3lmatbur4	92	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	92	4	following	follow	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	92	5	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	92	6	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	92	7	about	about	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	9	above	above	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	11	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	13	after	after	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	15	against	against	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	17	along	along	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	19	among	among	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	21	around	around	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	22	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	23	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	24	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	25	at	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	26	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	27	before	before	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	28	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	29	be-	be-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	92	30	hind	hind	NN
work_nu5xpentbrcxdlcdt3lmatbur4	92	31	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	32	beneath	beneath	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	33	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	34	beside	beside	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	35	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	36	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	37	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	38	by	by	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	39	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	40	down	down	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	41	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	42	during	during	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	43	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	44	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	45	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	46	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	47	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	48	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	49	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	50	inside	inside	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	51	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	52	into	into	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	53	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	54	like	like	UH
work_nu5xpentbrcxdlcdt3lmatbur4	92	55	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	56	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	57	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	58	off	off	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	59	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	60	on	on	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	61	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	62	onto	onto	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	63	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	64	over	over	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	65	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	66	round	round	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	92	67	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	68	through	through	RB
work_nu5xpentbrcxdlcdt3lmatbur4	92	69	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	70	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	92	71	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	72	to-	to-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	92	73	wards	ward	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	92	74	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	92	75	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	92	76	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	92	77	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	93	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	93	2	does	do	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	93	3	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	93	4	include	include	VB
work_nu5xpentbrcxdlcdt3lmatbur4	93	5	multi	multi	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	93	6	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	93	7	word	word	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	93	8	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	93	9	such	such	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	93	10	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	93	11	because	because	IN
work_nu5xpentbrcxdlcdt3lmatbur4	93	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	93	13	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	93	14	due	due	IN
work_nu5xpentbrcxdlcdt3lmatbur4	93	15	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	93	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	94	1	233	233	CD
work_nu5xpentbrcxdlcdt3lmatbur4	94	2	ent	ent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	94	3	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	94	4	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	95	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	95	2	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	95	3	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	95	4	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	95	5	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	95	6	defined	define	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	95	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	95	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	95	9	Preposition	Preposition	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	95	10	Project	Project	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	95	11	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	95	12	specifies	specify	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	95	13	related	related	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	95	14	prepo-	prepo-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	95	15	sitions	sition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	95	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	96	1	These	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	96	2	definitions	definition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	96	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	96	4	related	related	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	96	5	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	96	6	provide	provide	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	96	7	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	96	8	starting	starting	NN
work_nu5xpentbrcxdlcdt3lmatbur4	96	9	point	point	NN
work_nu5xpentbrcxdlcdt3lmatbur4	96	10	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	96	11	identify	identify	VB
work_nu5xpentbrcxdlcdt3lmatbur4	96	12	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	96	13	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	96	14	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	96	15	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	96	16	merged	merge	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	96	17	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	96	18	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	96	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	97	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	97	2	followed	follow	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	97	3	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	97	4	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	97	5	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	97	6	manual	manual	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	97	7	cleanup	cleanup	NN
work_nu5xpentbrcxdlcdt3lmatbur4	97	8	phase	phase	NN
work_nu5xpentbrcxdlcdt3lmatbur4	97	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	98	1	Some	some	DT
work_nu5xpentbrcxdlcdt3lmatbur4	98	2	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	98	3	do	do	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	98	4	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	98	5	cleanly	cleanly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	98	6	align	align	VB
work_nu5xpentbrcxdlcdt3lmatbur4	98	7	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	98	8	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	98	9	single	single	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	98	10	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	98	11	because	because	IN
work_nu5xpentbrcxdlcdt3lmatbur4	98	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	98	13	def-	def-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	98	14	initions	inition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	98	15	include	include	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	98	16	idiomatic	idiomatic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	98	17	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	98	18	figurative	figurative	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	98	19	usage	usage	NN
work_nu5xpentbrcxdlcdt3lmatbur4	98	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	99	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	99	2	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	99	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	99	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	99	5	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	99	6	in:7(5	in:7(5	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	99	7	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	99	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	99	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	99	10	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	99	11	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	99	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	99	13	ac-	ac-	DT
work_nu5xpentbrcxdlcdt3lmatbur4	99	14	cording	cord	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	99	15	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	99	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	99	17	definition	definition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	99	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	99	19	includes	include	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	99	20	both	both	CC
work_nu5xpentbrcxdlcdt3lmatbur4	99	21	spatial	spatial	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	99	22	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	99	23	figurative	figurative	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	99	24	notions	notion	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	99	25	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	99	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	99	27	spatial	spatial	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	99	28	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	99	29	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	99	30	that	that	RB
work_nu5xpentbrcxdlcdt3lmatbur4	99	31	is	is	RB
work_nu5xpentbrcxdlcdt3lmatbur4	99	32	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	99	33	both	both	CC
work_nu5xpentbrcxdlcdt3lmatbur4	99	34	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	99	35	London	London	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	99	36	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	99	37	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	99	38	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	99	39	film	film	NN
work_nu5xpentbrcxdlcdt3lmatbur4	99	40	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	99	41	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	100	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	100	2	such	such	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	100	3	cases	case	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	100	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	100	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	100	6	sam-	sam-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	100	7	pled	plead	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	100	8	20	20	CD
work_nu5xpentbrcxdlcdt3lmatbur4	100	9	examples	example	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	100	10	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	100	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	100	12	SemEval	SemEval	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	100	13	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	100	14	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	100	15	set	set	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	100	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	100	17	assigned	assign	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	100	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	100	19	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	100	20	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	100	21	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	100	22	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	100	23	majority	majority	NN
work_nu5xpentbrcxdlcdt3lmatbur4	100	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	101	1	If	if	IN
work_nu5xpentbrcxdlcdt3lmatbur4	101	2	sufficient	sufficient	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	101	3	examples	example	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	101	4	could	could	MD
work_nu5xpentbrcxdlcdt3lmatbur4	101	5	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	101	6	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	101	7	sampled	sample	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	101	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	101	9	these	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	101	10	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	101	11	were	be	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	101	12	added	add	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	101	13	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	101	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	101	15	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	101	16	Other	Other	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	101	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	101	18	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	101	19	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	101	20	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	101	21	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	101	22	semantically	semantically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	101	23	coherent	coherent	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	101	24	category	category	NN
work_nu5xpentbrcxdlcdt3lmatbur4	101	25	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	101	26	represents	represent	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	101	27	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	101	28	‘	'	``
work_nu5xpentbrcxdlcdt3lmatbur4	101	29	overflow	overflow	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	101	30	’	'	''
work_nu5xpentbrcxdlcdt3lmatbur4	101	31	case	case	NN
work_nu5xpentbrcxdlcdt3lmatbur4	101	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	102	1	Overall	overall	RB
work_nu5xpentbrcxdlcdt3lmatbur4	102	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	102	3	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	102	4	have	have	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	102	5	32	32	CD
work_nu5xpentbrcxdlcdt3lmatbur4	102	6	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	102	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	102	8	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	102	9	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	102	10	listed	list	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	102	11	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	102	12	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	102	13	14	14	CD
work_nu5xpentbrcxdlcdt3lmatbur4	102	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	103	1	A	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	103	2	companion	companion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	103	3	publication	publication	NN
work_nu5xpentbrcxdlcdt3lmatbur4	103	4	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	103	5	available	available	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	103	6	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	103	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	103	8	authors	author	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	103	9	’	’	POS
work_nu5xpentbrcxdlcdt3lmatbur4	103	10	website	website	NN
work_nu5xpentbrcxdlcdt3lmatbur4	103	11	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	103	12	provides	provide	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	103	13	detailed	detailed	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	103	14	definitions	definition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	103	15	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	103	16	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	103	17	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	103	18	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	103	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	103	20	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	103	21	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	103	22	were	be	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	103	23	merged	merge	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	103	24	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	103	25	create	create	VB
work_nu5xpentbrcxdlcdt3lmatbur4	103	26	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	103	27	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	103	28	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	104	1	Since	since	IN
work_nu5xpentbrcxdlcdt3lmatbur4	104	2	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	104	3	define	define	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	104	4	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	104	5	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	104	6	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	104	7	groups	group	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	104	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	104	9	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	104	10	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	104	11	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	104	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	104	13	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	104	14	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	104	15	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	104	16	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	104	17	uniquely	uniquely	RB
work_nu5xpentbrcxdlcdt3lmatbur4	104	18	mapped	map	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	104	19	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	104	20	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	104	21	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	104	22	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	104	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	105	1	Hence	hence	RB
work_nu5xpentbrcxdlcdt3lmatbur4	105	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	105	3	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	105	4	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	105	5	use	use	VB
work_nu5xpentbrcxdlcdt3lmatbur4	105	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	105	7	annotated	annotated	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	105	8	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	105	9	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	105	10	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	105	11	SemEval	SemEval	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	105	12	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	105	13	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	105	14	obtain	obtain	VB
work_nu5xpentbrcxdlcdt3lmatbur4	105	15	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	105	16	corpus	corpus	NN
work_nu5xpentbrcxdlcdt3lmatbur4	105	17	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	105	18	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	105	19	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	105	20	labeled	label	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	105	21	sentences	sentence	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	105	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	106	1	To	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	106	2	validate	validate	VB
work_nu5xpentbrcxdlcdt3lmatbur4	106	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	106	4	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	106	5	scheme	scheme	NN
work_nu5xpentbrcxdlcdt3lmatbur4	106	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	106	7	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	106	8	native	native	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	106	9	speak-	speak-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	106	10	ers	er	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	106	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	106	12	English	English	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	106	13	annotated	annotate	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	106	14	200	200	CD
work_nu5xpentbrcxdlcdt3lmatbur4	106	15	sentences	sentence	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	106	16	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	106	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	106	18	SemEval	SemEval	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	106	19	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	106	20	corpus	corpus	NN
work_nu5xpentbrcxdlcdt3lmatbur4	106	21	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	106	22	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	106	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	106	24	definitions	definition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	106	25	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	106	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	106	27	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	106	28	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	106	29	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	106	30	annotation	annotation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	106	31	guidelines	guideline	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	106	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	107	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	107	2	mea-	mea-	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	107	3	sured	sure	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	107	4	Cohen	Cohen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	107	5	’s	’s	POS
work_nu5xpentbrcxdlcdt3lmatbur4	107	6	kappa	kappa	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	107	7	coefficient	coefficient	NN
work_nu5xpentbrcxdlcdt3lmatbur4	107	8	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	107	9	Cohen	Cohen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	107	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	107	11	1960	1960	CD
work_nu5xpentbrcxdlcdt3lmatbur4	107	12	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	107	13	be-	be-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	107	14	tween	tween	NN
work_nu5xpentbrcxdlcdt3lmatbur4	107	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	107	16	annotators	annotator	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	107	17	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	107	18	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	107	19	0.75	0.75	CD
work_nu5xpentbrcxdlcdt3lmatbur4	107	20	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	107	21	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	107	22	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	107	23	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	107	24	annotator	annotator	NN
work_nu5xpentbrcxdlcdt3lmatbur4	107	25	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	107	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	107	27	original	original	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	107	28	corpus	corpus	NN
work_nu5xpentbrcxdlcdt3lmatbur4	107	29	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	107	30	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	107	31	0.76	0.76	CD
work_nu5xpentbrcxdlcdt3lmatbur4	107	32	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	107	33	0.74	0.74	CD
work_nu5xpentbrcxdlcdt3lmatbur4	107	34	respectively	respectively	RB
work_nu5xpentbrcxdlcdt3lmatbur4	107	35	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	108	1	3.2	3.2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	108	2	Preposition	Preposition	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	108	3	Relation	Relation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	108	4	Extraction	Extraction	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	108	5	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	6	input	input	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	7	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	108	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	9	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	10	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	11	consists	consist	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	108	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	108	13	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	14	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	15	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	108	16	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	17	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	18	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	108	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	20	goal	goal	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	21	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	108	22	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	108	23	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	108	24	model	model	VB
work_nu5xpentbrcxdlcdt3lmatbur4	108	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	26	following	follow	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	108	27	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	108	28	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	108	29	i	i	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	30	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	108	31	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	32	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	33	expressed	express	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	108	34	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	108	35	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	36	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	37	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	108	38	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	108	39	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	108	40	ii	ii	LS
work_nu5xpentbrcxdlcdt3lmatbur4	108	41	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	108	42	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	43	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	108	44	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	108	45	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	46	rela-	rela-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	108	47	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	48	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	108	49	namely	namely	RB
work_nu5xpentbrcxdlcdt3lmatbur4	108	50	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	51	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	52	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	108	53	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	108	54	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	108	55	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	109	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	109	2	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	109	3	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	109	4	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	109	5	2	2	LS
work_nu5xpentbrcxdlcdt3lmatbur4	109	6	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	109	7	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	109	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	109	9	introduction	introduction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	109	10	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	109	11	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	109	12	run-	run-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	109	13	ning	ning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	109	14	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	109	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	109	16	following	follow	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	109	17	discussion	discussion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	109	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	110	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	110	2	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	110	3	run-	run-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	110	4	4Note	4note	CD
work_nu5xpentbrcxdlcdt3lmatbur4	110	5	that	that	DT
work_nu5xpentbrcxdlcdt3lmatbur4	110	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	110	7	even	even	RB
work_nu5xpentbrcxdlcdt3lmatbur4	110	8	though	though	IN
work_nu5xpentbrcxdlcdt3lmatbur4	110	9	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	110	10	do	do	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	110	11	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	110	12	consider	consider	VB
work_nu5xpentbrcxdlcdt3lmatbur4	110	13	intransitive	intransitive	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	110	14	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	110	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	110	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	110	17	definitions	definition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	110	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	110	19	some	some	DT
work_nu5xpentbrcxdlcdt3lmatbur4	110	20	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	110	21	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	110	22	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	110	23	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	110	24	could	could	MD
work_nu5xpentbrcxdlcdt3lmatbur4	110	25	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	110	26	extended	extend	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	110	27	apply	apply	NN
work_nu5xpentbrcxdlcdt3lmatbur4	110	28	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	110	29	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	110	30	particles	particle	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	110	31	such	such	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	110	32	drive	drive	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	110	33	down	down	RP
work_nu5xpentbrcxdlcdt3lmatbur4	110	34	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	110	35	Direction	direction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	110	36	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	110	37	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	110	38	run	run	VB
work_nu5xpentbrcxdlcdt3lmatbur4	110	39	about	about	IN
work_nu5xpentbrcxdlcdt3lmatbur4	110	40	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	110	41	Manner	Manner	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	110	42	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	110	43	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	111	1	Relation	Relation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	2	Name	Name	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	3	Example	Example	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	4	Activity	Activity	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	5	good	good	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	111	6	at	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	7	boxing	box	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	111	8	Agent	Agent	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	9	opened	open	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	111	10	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	11	Annie	Annie	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	12	Attribute	Attribute	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	13	walls	wall	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	14	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	15	stone	stone	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	16	Beneficiary	Beneficiary	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	17	fight	fight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	18	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	19	Napoleon	Napoleon	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	20	Cause	Cause	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	21	died	die	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	111	22	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	23	cancer	cancer	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	24	Co	Co	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	25	-	-	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	26	Participants	participant	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	27	pick	pick	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	111	28	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	111	29	among	among	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	30	these	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	31	Destination	Destination	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	32	leaving	leave	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	111	33	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	34	London	London	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	35	Direction	Direction	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	36	drove	drive	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	111	37	towards	towards	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	38	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	39	border	border	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	40	EndState	endstate	RB
work_nu5xpentbrcxdlcdt3lmatbur4	111	41	driven	drive	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	111	42	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	43	tears	tear	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	44	Experiencer	Experiencer	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	45	warm	warm	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	111	46	towards	towards	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	47	her	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	111	48	Instrument	Instrument	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	49	cut	cut	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	111	50	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	51	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	52	knife	knife	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	53	Journey	Journey	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	54	travel	travel	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	55	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	56	road	road	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	57	Location	location	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	58	living	live	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	111	59	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	60	London	London	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	61	Manner	Manner	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	62	scream	scream	VB
work_nu5xpentbrcxdlcdt3lmatbur4	111	63	like	like	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	64	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	65	animal	animal	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	66	MediumOfCommunication	MediumOfCommunication	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	67	new	new	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	111	68	show	show	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	69	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	70	TV	tv	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	71	Numeric	Numeric	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	72	increase	increase	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	73	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	74	10	10	CD
work_nu5xpentbrcxdlcdt3lmatbur4	111	75	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	76	ObjectOfVerb	objectofverb	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	77	murder	murder	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	78	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	79	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	80	boys	boy	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	81	Opponent	Opponent	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	82	/	/	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	111	83	Contrast	Contrast	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	84	fight	fight	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	111	85	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	86	him	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	111	87	Other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	111	88	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	89	others	other	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	90	Participant	Participant	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	91	/	/	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	111	92	Accompanier	Accompanier	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	93	steak	steak	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	94	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	95	wine	wine	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	96	PartWhole	PartWhole	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	97	member	member	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	98	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	99	gang	gang	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	100	PhysicalSupport	PhysicalSupport	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	101	lean	lean	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	111	102	against	against	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	103	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	104	wall	wall	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	105	Possessor	Possessor	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	106	son	son	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	107	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	108	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	109	friend	friend	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	110	ProfessionalAspect	ProfessionalAspect	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	111	works	work	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	111	112	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	113	publishing	publish	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	111	114	Purpose	purpose	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	115	tools	tool	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	116	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	117	making	make	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	111	118	it	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	111	119	Recipient	recipient	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	111	120	unkind	unkind	RB
work_nu5xpentbrcxdlcdt3lmatbur4	111	121	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	122	her	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	111	123	Separation	separation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	124	ousted	oust	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	111	125	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	126	power	power	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	127	Source	Source	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	128	purchased	purchase	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	111	129	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	130	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	131	shop	shop	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	132	Species	Species	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	133	city	city	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	134	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	135	Prague	Prague	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	136	StartState	StartState	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	137	recover	recover	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	138	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	139	illness	illness	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	140	Temporal	Temporal	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	141	arrived	arrive	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	111	142	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	143	Monday	Monday	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	144	Topic	Topic	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	145	books	book	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	146	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	147	Shakespeare	Shakespeare	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	148	Table	Table	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	149	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	111	150	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	111	151	List	list	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	152	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	111	153	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	154	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	111	155	ning	ne	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	111	156	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	157	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	111	158	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	111	159	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	160	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	111	161	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	111	162	Cause	Cause	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	111	163	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	112	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	112	2	rep-	rep-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	112	3	resent	resent	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	112	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	112	5	predicted	predict	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	112	6	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	7	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	8	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	112	9	r.	r.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	112	10	Arguments	Arguments	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	112	11	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	112	12	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	13	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	14	crucially	crucially	RB
work_nu5xpentbrcxdlcdt3lmatbur4	112	15	depends	depend	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	112	16	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	112	17	correctly	correctly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	112	18	identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	112	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	112	20	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	112	21	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	112	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	112	23	prepo-	prepo-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	24	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	25	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	112	26	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	112	27	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	112	28	death	death	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	29	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	112	30	pneumonia	pneumonia	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	31	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	112	32	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	112	33	run-	run-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	112	34	ning	ning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	35	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	112	36	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	113	1	While	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	113	2	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	113	3	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	113	4	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	113	5	identify	identify	VB
work_nu5xpentbrcxdlcdt3lmatbur4	113	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	113	7	argu-	argu-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	113	8	ments	ment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	113	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	113	10	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	113	11	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	113	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	113	13	simply	simply	RB
work_nu5xpentbrcxdlcdt3lmatbur4	113	14	relying	rely	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	113	15	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	113	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	113	17	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	113	18	may	may	MD
work_nu5xpentbrcxdlcdt3lmatbur4	113	19	impose	impose	VB
work_nu5xpentbrcxdlcdt3lmatbur4	113	20	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	113	21	upper	upper	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	113	22	limit	limit	NN
work_nu5xpentbrcxdlcdt3lmatbur4	113	23	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	113	24	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	113	25	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	113	26	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	113	27	rela-	rela-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	113	28	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	113	29	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	113	30	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	114	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	114	2	build	build	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	114	3	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	114	4	oracle	oracle	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	114	5	experiment	experiment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	114	6	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	114	7	highlight	highlight	VB
work_nu5xpentbrcxdlcdt3lmatbur4	114	8	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	114	9	limitation	limitation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	114	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	115	1	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	115	2	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	115	3	shows	show	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	115	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	115	5	recall	recall	NN
work_nu5xpentbrcxdlcdt3lmatbur4	115	6	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	115	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	115	8	easy	easy	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	115	9	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	115	10	first	first	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	115	11	dependency	dependency	NN
work_nu5xpentbrcxdlcdt3lmatbur4	115	12	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	115	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	115	14	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	115	15	Goldberg	Goldberg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	115	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	115	17	Elhadad	Elhadad	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	115	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	115	19	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	115	20	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	115	21	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	115	22	Section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	115	23	23	23	CD
work_nu5xpentbrcxdlcdt3lmatbur4	115	24	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	115	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	115	26	Penn	Penn	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	115	27	Treebank	Treebank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	115	28	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	115	29	identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	115	30	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	115	31	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	115	32	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	115	33	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	115	34	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	115	35	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	115	36	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	116	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	116	2	define	define	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	116	3	heuristics	heuristic	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	116	4	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	116	5	generate	generate	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	116	6	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	116	7	candidate	candidate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	116	8	governors	governor	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	116	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	116	10	objects	object	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	116	11	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	116	12	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	116	13	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	116	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	117	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	117	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	117	3	gov-	gov-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	117	4	234	234	CD
work_nu5xpentbrcxdlcdt3lmatbur4	117	5	ernor	ernor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	117	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	117	7	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	117	8	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	117	9	includes	include	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	117	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	117	11	previous	previous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	117	12	verb	verb	NN
work_nu5xpentbrcxdlcdt3lmatbur4	117	13	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	117	14	noun	noun	NN
work_nu5xpentbrcxdlcdt3lmatbur4	117	15	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	117	16	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	117	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	117	18	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	117	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	117	20	it	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	117	21	includes	include	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	117	22	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	117	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	117	24	next	next	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	117	25	noun	noun	NN
work_nu5xpentbrcxdlcdt3lmatbur4	117	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	118	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	118	2	row	row	NN
work_nu5xpentbrcxdlcdt3lmatbur4	118	3	labeled	label	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	118	4	Best(Parser	Best(Parser	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	118	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	118	6	Heuristics	Heuristics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	118	7	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	118	8	shows	show	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	118	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	118	10	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	118	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	118	12	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	118	13	oracle	oracle	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	118	14	predictor	predictor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	118	15	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	118	16	selects	select	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	118	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	118	18	true	true	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	118	19	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	118	20	/	/	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	118	21	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	118	22	if	if	IN
work_nu5xpentbrcxdlcdt3lmatbur4	118	23	present	present	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	118	24	among	among	IN
work_nu5xpentbrcxdlcdt3lmatbur4	118	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	118	26	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	118	27	’s	’s	POS
work_nu5xpentbrcxdlcdt3lmatbur4	118	28	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	118	29	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	118	30	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	118	31	heuristics	heuristics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	118	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	119	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	119	2	see	see	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	119	3	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	119	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	119	5	even	even	RB
work_nu5xpentbrcxdlcdt3lmatbur4	119	6	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	119	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	119	8	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	119	9	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	119	10	domain	domain	NN
work_nu5xpentbrcxdlcdt3lmatbur4	119	11	case	case	NN
work_nu5xpentbrcxdlcdt3lmatbur4	119	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	119	13	if	if	IN
work_nu5xpentbrcxdlcdt3lmatbur4	119	14	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	119	15	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	119	16	able	able	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	119	17	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	119	18	re	re	VB
work_nu5xpentbrcxdlcdt3lmatbur4	119	19	-	-	VB
work_nu5xpentbrcxdlcdt3lmatbur4	119	20	rank	rank	VB
work_nu5xpentbrcxdlcdt3lmatbur4	119	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	119	22	can-	can-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	119	23	didates	didate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	119	24	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	119	25	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	119	26	could	could	MD
work_nu5xpentbrcxdlcdt3lmatbur4	119	27	achieve	achieve	VB
work_nu5xpentbrcxdlcdt3lmatbur4	119	28	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	119	29	big	big	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	119	30	improvement	improvement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	119	31	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	119	32	ar-	ar-	CD
work_nu5xpentbrcxdlcdt3lmatbur4	119	33	gument	gument	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	119	34	identification	identification	NN
work_nu5xpentbrcxdlcdt3lmatbur4	119	35	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	120	1	Recall	Recall	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	120	2	Governor	Governor	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	120	3	Object	Object	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	120	4	Parser	Parser	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	120	5	88.88	88.88	CD
work_nu5xpentbrcxdlcdt3lmatbur4	120	6	92.37	92.37	CD
work_nu5xpentbrcxdlcdt3lmatbur4	120	7	Best(Parser	Best(Parser	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	120	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	120	9	Heuristics	Heuristics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	120	10	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	120	11	92.50	92.50	CD
work_nu5xpentbrcxdlcdt3lmatbur4	120	12	93.06	93.06	CD
work_nu5xpentbrcxdlcdt3lmatbur4	120	13	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	120	14	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	120	15	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	120	16	Identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	120	17	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	120	18	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	120	19	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	120	20	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	120	21	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	120	22	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	120	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	120	24	Penn	Penn	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	120	25	Treebank	Treebank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	120	26	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	120	27	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	121	1	Here	here	RB
work_nu5xpentbrcxdlcdt3lmatbur4	121	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	121	3	Best(Parser	Best(Parser	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	121	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	121	5	Heuris-	Heuris-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	121	6	tics	tic	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	121	7	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	121	8	reports	report	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	121	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	121	10	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	121	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	121	12	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	121	13	oracle	oracle	NN
work_nu5xpentbrcxdlcdt3lmatbur4	121	14	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	121	15	picks	pick	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	121	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	121	17	true	true	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	121	18	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	121	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	121	20	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	121	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	121	22	if	if	IN
work_nu5xpentbrcxdlcdt3lmatbur4	121	23	present	present	NN
work_nu5xpentbrcxdlcdt3lmatbur4	121	24	among	among	IN
work_nu5xpentbrcxdlcdt3lmatbur4	121	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	121	26	candidates	candidate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	121	27	presented	present	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	121	28	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	121	29	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	121	30	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	121	31	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	121	32	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	121	33	heuristic	heuristic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	121	34	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	122	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	122	2	presents	present	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	122	3	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	122	4	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	122	5	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	122	6	domain	domain	NN
work_nu5xpentbrcxdlcdt3lmatbur4	122	7	upper	upper	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	122	8	bound	bound	NN
work_nu5xpentbrcxdlcdt3lmatbur4	122	9	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	122	10	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	122	11	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	122	12	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	122	13	detec-	detec-	FW
work_nu5xpentbrcxdlcdt3lmatbur4	122	14	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	122	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	123	1	See	see	VB
work_nu5xpentbrcxdlcdt3lmatbur4	123	2	text	text	NN
work_nu5xpentbrcxdlcdt3lmatbur4	123	3	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	123	4	further	further	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	123	5	details	detail	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	123	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	124	1	To	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	124	2	overcome	overcome	VB
work_nu5xpentbrcxdlcdt3lmatbur4	124	3	erroneous	erroneous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	124	4	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	124	5	decisions	decision	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	124	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	124	7	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	124	8	en-	en-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	124	9	tertain	tertain	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	124	10	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	124	11	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	124	12	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	124	13	candidates	candidate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	124	14	proposed	propose	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	124	15	both	both	CC
work_nu5xpentbrcxdlcdt3lmatbur4	124	16	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	124	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	124	18	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	124	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	124	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	124	21	heuristics	heuristics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	124	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	125	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	125	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	125	3	follow-	follow-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	125	4	ing	ing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	125	5	discussion	discussion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	125	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	125	7	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	125	8	denote	denote	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	125	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	125	10	chosen	choose	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	125	11	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	125	12	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	125	13	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	125	14	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	125	15	g	g	NN
work_nu5xpentbrcxdlcdt3lmatbur4	125	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	125	17	o	o	NN
work_nu5xpentbrcxdlcdt3lmatbur4	125	18	respectively	respectively	RB
work_nu5xpentbrcxdlcdt3lmatbur4	125	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	126	1	Argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	126	2	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	126	3	While	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	126	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	126	5	primary	primary	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	126	6	purpose	purpose	NN
work_nu5xpentbrcxdlcdt3lmatbur4	126	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	126	8	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	126	9	work	work	NN
work_nu5xpentbrcxdlcdt3lmatbur4	126	10	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	126	11	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	126	12	model	model	VB
work_nu5xpentbrcxdlcdt3lmatbur4	126	13	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	126	14	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	126	15	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	126	16	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	126	17	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	126	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	126	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	126	20	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	126	21	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	126	22	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	126	23	strongly	strongly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	126	24	depen-	depen-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	126	25	dent	dent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	126	26	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	126	27	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	126	28	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	126	29	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	126	30	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	126	31	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	126	32	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	126	33	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	127	1	To	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	127	2	il-	il-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	127	3	lustrate	lustrate	VB
work_nu5xpentbrcxdlcdt3lmatbur4	127	4	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	127	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	127	6	consider	consider	VB
work_nu5xpentbrcxdlcdt3lmatbur4	127	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	127	8	following	follow	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	127	9	incomplete	incomplete	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	127	10	sen-	sen-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	127	11	tence	tence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	127	12	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	127	13	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	127	14	message	message	NN
work_nu5xpentbrcxdlcdt3lmatbur4	127	15	was	be	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	127	16	delivered	deliver	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	127	17	at	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	127	18	·	·	NFP
work_nu5xpentbrcxdlcdt3lmatbur4	127	19	·	·	NFP
work_nu5xpentbrcxdlcdt3lmatbur4	127	20	·	·	NFP
work_nu5xpentbrcxdlcdt3lmatbur4	127	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	128	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	128	2	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	128	3	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	128	4	express	express	VB
work_nu5xpentbrcxdlcdt3lmatbur4	128	5	both	both	CC
work_nu5xpentbrcxdlcdt3lmatbur4	128	6	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	128	7	Temporal	Temporal	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	128	8	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	128	9	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	128	10	Location	location	NN
work_nu5xpentbrcxdlcdt3lmatbur4	128	11	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	128	12	depending	depend	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	128	13	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	128	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	128	15	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	128	16	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	128	17	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	128	18	ex-	ex-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	128	19	ample	ample	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	128	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	128	21	noon	noon	NN
work_nu5xpentbrcxdlcdt3lmatbur4	128	22	vs.	vs.	IN
work_nu5xpentbrcxdlcdt3lmatbur4	128	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	128	24	doorstep	doorstep	NN
work_nu5xpentbrcxdlcdt3lmatbur4	128	25	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	128	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	129	1	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	129	2	Agirre	Agirre	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	129	3	et	et	FW
work_nu5xpentbrcxdlcdt3lmatbur4	129	4	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	129	5	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	129	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	129	7	2008	2008	CD
work_nu5xpentbrcxdlcdt3lmatbur4	129	8	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	129	9	shows	show	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	129	10	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	129	11	modeling	model	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	129	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	129	13	se-	se-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	129	14	mantic	mantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	129	15	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	129	16	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	129	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	129	18	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	129	19	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	129	20	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	129	21	attachment	attachment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	129	22	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	129	23	improve	improve	VB
work_nu5xpentbrcxdlcdt3lmatbur4	129	24	PP	PP	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	129	25	attachment	attachment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	129	26	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	129	27	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	130	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	130	2	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	130	3	work	work	NN
work_nu5xpentbrcxdlcdt3lmatbur4	130	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	130	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	130	6	point	point	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	130	7	out	out	RP
work_nu5xpentbrcxdlcdt3lmatbur4	130	8	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	130	9	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	130	10	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	130	11	should	should	MD
work_nu5xpentbrcxdlcdt3lmatbur4	130	12	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	130	13	modeled	model	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	130	14	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	130	15	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	130	16	both	both	DT
work_nu5xpentbrcxdlcdt3lmatbur4	130	17	aspects	aspect	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	130	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	130	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	130	20	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	130	21	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	130	22	preposi-	preposi-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	130	23	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	130	24	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	130	25	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	130	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	131	1	Types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	131	2	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	131	3	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	131	4	abstraction	abstraction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	131	5	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	131	6	capture	capture	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	131	7	common	common	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	131	8	properties	property	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	131	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	131	10	groups	group	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	131	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	131	12	entities	entity	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	131	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	132	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	132	2	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	132	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	132	4	Word-	Word-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	132	5	Net	Net	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	132	6	provides	provide	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	132	7	generalizations	generalization	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	132	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	132	9	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	132	10	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	132	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	132	12	form	form	NN
work_nu5xpentbrcxdlcdt3lmatbur4	132	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	132	14	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	132	15	hypernyms	hypernyms	NN
work_nu5xpentbrcxdlcdt3lmatbur4	132	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	133	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	133	2	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	133	3	running	run	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	133	4	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	133	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	133	6	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	133	7	wish	wish	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	133	8	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	133	9	generalize	generalize	VB
work_nu5xpentbrcxdlcdt3lmatbur4	133	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	133	11	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	133	12	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	133	13	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	133	14	death	death	NN
work_nu5xpentbrcxdlcdt3lmatbur4	133	15	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	133	16	pneu-	pneu-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	133	17	monia	monia	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	133	18	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	133	19	include	include	VB
work_nu5xpentbrcxdlcdt3lmatbur4	133	20	cases	case	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	133	21	such	such	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	133	22	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	133	23	suffering	suffer	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	133	24	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	133	25	flu	flu	NN
work_nu5xpentbrcxdlcdt3lmatbur4	133	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	134	1	Figure	figure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	134	2	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	134	3	shows	show	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	134	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	134	5	hypernym	hypernym	NN
work_nu5xpentbrcxdlcdt3lmatbur4	134	6	hierarchy	hierarchy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	134	7	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	134	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	134	9	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	134	10	pneumonia	pneumonia	NN
work_nu5xpentbrcxdlcdt3lmatbur4	134	11	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	135	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	135	2	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	135	3	case	case	NN
work_nu5xpentbrcxdlcdt3lmatbur4	135	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	135	5	synsets	synset	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	135	6	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	135	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	135	8	hypernym	hypernym	NN
work_nu5xpentbrcxdlcdt3lmatbur4	135	9	hierarchy	hierarchy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	135	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	135	11	like	like	IN
work_nu5xpentbrcxdlcdt3lmatbur4	135	12	pathological	pathological	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	135	13	state	state	NN
work_nu5xpentbrcxdlcdt3lmatbur4	135	14	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	135	15	physical	physical	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	135	16	condi-	condi-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	135	17	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	135	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	135	19	would	would	MD
work_nu5xpentbrcxdlcdt3lmatbur4	135	20	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	135	21	include	include	VB
work_nu5xpentbrcxdlcdt3lmatbur4	135	22	ailments	ailment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	135	23	like	like	IN
work_nu5xpentbrcxdlcdt3lmatbur4	135	24	flu	flu	NN
work_nu5xpentbrcxdlcdt3lmatbur4	135	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	136	1	pneumonia	pneumonia	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	136	2	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	3	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	4	respiratory	respiratory	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	136	5	disease	disease	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	6	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	7	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	8	disease	disease	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	9	=	=	NFP
work_nu5xpentbrcxdlcdt3lmatbur4	136	10	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	11	illness	illness	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	12	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	13	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	14	ill	ill	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	136	15	health	health	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	16	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	17	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	18	pathological	pathological	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	136	19	state	state	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	20	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	21	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	22	physical	physical	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	136	23	condition	condition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	24	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	25	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	26	condition	condition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	27	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	28	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	29	state	state	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	30	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	31	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	32	attribute	attribute	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	33	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	34	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	35	abstraction	abstraction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	36	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	136	37	>	>	XX
work_nu5xpentbrcxdlcdt3lmatbur4	136	38	entity	entity	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	39	Figure	figure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	40	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	136	41	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	136	42	Hypernym	hypernym	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	43	hierarchy	hierarchy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	44	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	136	45	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	136	46	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	47	pneumonia	pneumonia	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	48	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	136	49	define	define	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	136	50	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	136	51	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	136	52	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	53	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	136	54	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	136	55	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	136	56	cluster	cluster	NN
work_nu5xpentbrcxdlcdt3lmatbur4	136	57	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	136	58	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	136	59	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	137	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	137	2	addition	addition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	137	3	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	137	4	WordNet	WordNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	137	5	hypernyms	hypernyms	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	137	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	137	7	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	137	8	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	137	9	cluster	cluster	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	137	10	verbs	verb	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	137	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	137	12	nouns	noun	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	137	13	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	137	14	adjectives	adjective	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	137	15	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	137	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	137	17	dependency-	dependency-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	137	18	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	137	19	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	137	20	similarity	similarity	NN
work_nu5xpentbrcxdlcdt3lmatbur4	137	21	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	137	22	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	137	23	Lin	Lin	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	137	24	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	137	25	1998	1998	CD
work_nu5xpentbrcxdlcdt3lmatbur4	137	26	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	137	27	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	137	28	treat	treat	VB
work_nu5xpentbrcxdlcdt3lmatbur4	137	29	cluster	cluster	NN
work_nu5xpentbrcxdlcdt3lmatbur4	137	30	membership	membership	NN
work_nu5xpentbrcxdlcdt3lmatbur4	137	31	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	137	32	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	137	33	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	138	1	These	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	138	2	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	138	3	described	describe	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	138	4	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	138	5	detail	detail	NN
work_nu5xpentbrcxdlcdt3lmatbur4	138	6	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	138	7	Section	Section	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	138	8	5.1	5.1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	138	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	139	1	Relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	139	2	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	139	3	involves	involve	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	139	4	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	139	5	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	139	6	identifying	identify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	139	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	139	8	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	139	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	139	10	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	139	11	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	139	12	selecting	select	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	139	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	139	14	right	right	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	139	15	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	139	16	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	139	17	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	139	18	them	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	139	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	139	20	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	139	21	together	together	RB
work_nu5xpentbrcxdlcdt3lmatbur4	139	22	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	139	23	help	help	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	139	24	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	139	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	139	26	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	139	27	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	139	28	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	140	1	Given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	140	2	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	140	3	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	140	4	candidate	candidate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	140	5	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	140	6	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	140	7	collection	collection	NN
work_nu5xpentbrcxdlcdt3lmatbur4	140	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	140	9	possible	possible	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	140	10	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	140	11	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	140	12	given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	140	13	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	140	14	WordNet	WordNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	140	15	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	140	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	140	17	similarity	similarity	NN
work_nu5xpentbrcxdlcdt3lmatbur4	140	18	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	140	19	clusters	cluster	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	140	20	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	140	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	140	22	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	140	23	need	need	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	140	24	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	140	25	select	select	VB
work_nu5xpentbrcxdlcdt3lmatbur4	140	26	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	140	27	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	140	28	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	140	29	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	140	30	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	141	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	141	2	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	141	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	141	4	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	141	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	141	6	WordNet	WordNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	141	7	case	case	NN
work_nu5xpentbrcxdlcdt3lmatbur4	141	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	141	9	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	141	10	need	need	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	141	11	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	141	12	pick	pick	VB
work_nu5xpentbrcxdlcdt3lmatbur4	141	13	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	141	14	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	141	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	141	16	hypernyms	hypernyms	NN
work_nu5xpentbrcxdlcdt3lmatbur4	141	17	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	141	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	141	19	hypernym	hypernym	NN
work_nu5xpentbrcxdlcdt3lmatbur4	141	20	hierarchy	hierarchy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	141	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	142	1	Thus	thus	RB
work_nu5xpentbrcxdlcdt3lmatbur4	142	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	142	3	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	142	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	142	5	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	142	6	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	142	7	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	142	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	142	9	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	142	10	have	have	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	142	11	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	142	12	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	142	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	142	14	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	142	15	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	142	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	142	17	comprised	comprise	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	142	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	142	19	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	142	20	element	element	NN
work_nu5xpentbrcxdlcdt3lmatbur4	142	21	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	142	22	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	142	23	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	142	24	category	category	NN
work_nu5xpentbrcxdlcdt3lmatbur4	142	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	143	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	143	2	denote	denote	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	143	3	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	143	4	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	143	5	tg	tg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	143	6	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	143	7	gover-	gover-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	143	8	nor	nor	CC
work_nu5xpentbrcxdlcdt3lmatbur4	143	9	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	143	10	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	143	11	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	143	12	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	143	13	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	143	14	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	143	15	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	143	16	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	143	17	respectively	respectively	RB
work_nu5xpentbrcxdlcdt3lmatbur4	143	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	144	1	3.3	3.3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	144	2	Problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	144	3	definition	definition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	144	4	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	144	5	input	input	NN
work_nu5xpentbrcxdlcdt3lmatbur4	144	6	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	144	7	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	144	8	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	144	9	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	144	10	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	144	11	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	144	12	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	144	13	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	144	14	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	144	15	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	144	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	145	1	Our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	145	2	goal	goal	NN
work_nu5xpentbrcxdlcdt3lmatbur4	145	3	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	145	4	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	145	5	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	145	6	model	model	VB
work_nu5xpentbrcxdlcdt3lmatbur4	145	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	145	8	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	145	9	it	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	145	10	expresses	express	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	145	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	145	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	145	13	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	145	14	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	145	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	145	16	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	145	17	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	145	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	145	19	rela-	rela-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	145	20	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	145	21	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	145	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	145	23	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	145	24	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	145	25	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	145	26	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	145	27	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	145	28	both	both	CC
work_nu5xpentbrcxdlcdt3lmatbur4	145	29	WordNet	WordNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	145	30	hypernyms	hypernyms	NN
work_nu5xpentbrcxdlcdt3lmatbur4	145	31	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	145	32	cluster	cluster	NN
work_nu5xpentbrcxdlcdt3lmatbur4	145	33	membership	membership	NN
work_nu5xpentbrcxdlcdt3lmatbur4	145	34	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	145	35	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	146	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	146	2	denote	denote	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	146	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	146	4	input	input	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	5	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	146	6	x	x	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	146	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	146	8	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	146	9	consists	consist	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	146	10	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	146	11	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	146	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	146	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	146	14	prepo-	prepo-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	15	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	16	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	146	17	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	146	18	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	146	19	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	20	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	146	21	candidates	candidate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	146	22	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	146	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	146	24	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	25	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	146	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	146	27	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	28	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	146	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	146	30	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	146	31	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	146	32	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	33	category	category	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	34	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	146	35	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	146	36	list	list	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	37	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	146	38	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	146	39	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	146	40	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	146	41	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	42	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	146	43	candidate	candidate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	146	44	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	147	1	235	235	CD
work_nu5xpentbrcxdlcdt3lmatbur4	147	2	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	147	3	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	147	5	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	147	6	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	147	7	denote	denote	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	147	8	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	147	9	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	147	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	147	11	consists	consist	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	147	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	147	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	147	14	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	15	r	r	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	147	17	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	147	18	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	147	19	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	147	20	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	147	21	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	147	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	147	23	valid	valid	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	147	24	re-	re-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	147	25	lation	lation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	26	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	147	27	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	147	28	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	29	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	147	30	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	147	31	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	147	32	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	33	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	147	34	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	35	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	147	36	denoted	denote	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	147	37	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	147	38	g	g	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	39	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	147	40	o	o	XX
work_nu5xpentbrcxdlcdt3lmatbur4	147	41	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	147	42	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	147	43	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	147	44	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	147	45	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	147	46	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	147	47	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	147	48	text	text	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	49	seg-	seg-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	50	ments	ment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	147	51	proposed	propose	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	147	52	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	147	53	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	147	54	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	147	55	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	147	56	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	147	57	heuristics	heuristics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	147	58	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	148	1	Ad-	ad-	CC
work_nu5xpentbrcxdlcdt3lmatbur4	148	2	ditionally	ditionally	RB
work_nu5xpentbrcxdlcdt3lmatbur4	148	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	148	4	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	148	5	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	148	6	consists	consist	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	148	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	148	8	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	148	9	predictions	prediction	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	148	10	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	148	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	148	12	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	148	13	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	148	14	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	148	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	148	16	denoted	denote	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	148	17	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	148	18	tg	tg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	148	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	148	20	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	148	21	respec-	respec-	VB
work_nu5xpentbrcxdlcdt3lmatbur4	148	22	tively	tively	RB
work_nu5xpentbrcxdlcdt3lmatbur4	148	23	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	148	24	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	148	25	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	148	26	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	148	27	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	148	28	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	148	29	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	148	30	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	148	31	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	148	32	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	148	33	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	148	34	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	148	35	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	148	36	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	148	37	category	category	NN
work_nu5xpentbrcxdlcdt3lmatbur4	148	38	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	149	1	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	149	2	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	149	3	summarizes	summarize	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	149	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	149	5	nota-	nota-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	149	6	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	149	7	described	describe	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	149	8	above	above	RB
work_nu5xpentbrcxdlcdt3lmatbur4	149	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	150	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	150	2	refer	refer	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	150	3	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	150	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	150	5	ith	ith	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	150	6	element	element	NN
work_nu5xpentbrcxdlcdt3lmatbur4	150	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	150	8	vectors	vector	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	150	9	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	150	10	subscripts	subscript	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	150	11	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	150	12	use	use	VB
work_nu5xpentbrcxdlcdt3lmatbur4	150	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	150	14	superscript	superscript	NN
work_nu5xpentbrcxdlcdt3lmatbur4	150	15	∗	∗	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	150	16	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	150	17	denote	denote	VB
work_nu5xpentbrcxdlcdt3lmatbur4	150	18	gold	gold	NN
work_nu5xpentbrcxdlcdt3lmatbur4	150	19	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	150	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	151	1	Recall	recall	VB
work_nu5xpentbrcxdlcdt3lmatbur4	151	2	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	151	3	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	151	4	have	have	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	151	5	gold	gold	NN
work_nu5xpentbrcxdlcdt3lmatbur4	151	6	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	151	7	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	151	8	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	151	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	151	10	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	151	11	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	151	12	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	151	13	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	151	14	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	151	15	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	151	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	151	17	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	151	18	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	151	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	152	1	Symbol	Symbol	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	152	2	Meaning	Meaning	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	152	3	x	x	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	152	4	Input	Input	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	152	5	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	152	6	pre	pre	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	152	7	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	152	8	processed	processed	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	152	9	sentence	sentence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	152	11	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	12	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	152	13	r	r	LS
work_nu5xpentbrcxdlcdt3lmatbur4	152	14	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	15	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	16	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	152	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	152	18	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	19	g	g	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	152	21	o	o	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	22	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	23	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	152	24	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	25	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	152	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	152	27	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	28	tg	tg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	152	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	152	30	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	152	31	vectors	vector	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	152	32	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	152	33	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	34	assignments	assignment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	152	35	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	152	36	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	37	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	152	38	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	39	respectively	respectively	RB
work_nu5xpentbrcxdlcdt3lmatbur4	152	40	y	y	DT
work_nu5xpentbrcxdlcdt3lmatbur4	152	41	Full	full	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	152	42	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	43	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	152	44	r	r	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	45	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	152	46	g	g	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	47	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	152	48	o	o	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	49	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	152	50	tg	tg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	152	51	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	152	52	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	152	53	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	152	54	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	55	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	152	56	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	152	57	Summary	summary	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	58	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	152	59	notation	notation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	60	4	4	CD
work_nu5xpentbrcxdlcdt3lmatbur4	152	61	Learning	Learning	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	152	62	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	63	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	152	64	A	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	152	65	key	key	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	152	66	challenge	challenge	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	67	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	152	68	modeling	model	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	152	69	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	70	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	152	71	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	152	72	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	152	73	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	152	74	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	75	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	152	76	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	152	77	annotates	annotate	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	152	78	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	152	79	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	152	80	la-	la-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	152	81	bels	bel	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	152	82	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	152	83	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	152	84	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	152	85	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	152	86	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	152	87	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	152	88	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	153	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	153	2	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	153	3	section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	153	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	153	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	153	6	introduce	introduce	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	153	7	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	153	8	approaches	approach	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	153	9	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	153	10	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	153	11	preposi-	preposi-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	153	12	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	153	13	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	153	14	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	153	15	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	153	16	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	153	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	154	1	4.1	4.1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	154	2	Feature	Feature	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	154	3	Representation	representation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	154	4	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	154	5	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	154	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	154	7	notation	notation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	154	8	Φ(x	Φ(x	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	154	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	154	10	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	154	11	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	154	12	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	154	13	indicate	indicate	VB
work_nu5xpentbrcxdlcdt3lmatbur4	154	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	154	15	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	154	16	function	function	NN
work_nu5xpentbrcxdlcdt3lmatbur4	154	17	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	154	18	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	154	19	input	input	NN
work_nu5xpentbrcxdlcdt3lmatbur4	154	20	x	x	NN
work_nu5xpentbrcxdlcdt3lmatbur4	154	21	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	154	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	154	23	full	full	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	154	24	output	output	NN
work_nu5xpentbrcxdlcdt3lmatbur4	154	25	y.	y.	.
work_nu5xpentbrcxdlcdt3lmatbur4	155	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	155	2	build	build	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	155	3	Φ	Φ	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	155	4	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	155	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	155	6	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	155	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	155	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	155	9	components	component	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	155	10	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	155	11	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	155	12	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	155	13	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	155	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	156	1	Arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	156	2	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	156	3	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	156	4	g	g	NN
work_nu5xpentbrcxdlcdt3lmatbur4	156	5	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	156	6	o	o	NN
work_nu5xpentbrcxdlcdt3lmatbur4	156	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	156	8	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	156	9	represent	represent	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	156	10	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	156	11	assignment	assignment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	156	12	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	156	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	156	14	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	156	15	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	156	16	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	156	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	156	18	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	156	19	de-	de-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	156	20	note	note	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	156	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	156	22	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	156	23	extracted	extract	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	156	24	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	156	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	156	26	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	156	27	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	156	28	φA(x	φa(x	NN
work_nu5xpentbrcxdlcdt3lmatbur4	156	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	156	30	g	g	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	156	31	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	156	32	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	156	33	φA(x	φa(x	NN
work_nu5xpentbrcxdlcdt3lmatbur4	156	34	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	156	35	o	o	LS
work_nu5xpentbrcxdlcdt3lmatbur4	156	36	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	156	37	respectively	respectively	RB
work_nu5xpentbrcxdlcdt3lmatbur4	156	38	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	157	1	2	2	LS
work_nu5xpentbrcxdlcdt3lmatbur4	157	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	158	1	Types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	158	2	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	158	3	Given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	158	4	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	158	5	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	158	6	assignment	assignment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	158	7	tgi	tgi	NN
work_nu5xpentbrcxdlcdt3lmatbur4	158	8	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	158	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	158	10	i	i	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	158	11	th	th	DT
work_nu5xpentbrcxdlcdt3lmatbur4	158	12	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	158	13	category	category	NN
work_nu5xpentbrcxdlcdt3lmatbur4	158	14	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	158	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	158	16	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	158	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	158	18	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	158	19	define	define	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	158	20	fea-	fea-	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	158	21	tures	ture	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	158	22	φT	φT	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	158	23	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	158	24	x	x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	158	25	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	158	26	g	g	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	158	27	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	158	28	t	t	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	158	29	g	g	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	158	30	i	i	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	158	31	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	158	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	159	1	Similarly	similarly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	159	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	159	3	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	159	4	define	define	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	159	5	features	feature	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	159	6	φT	φT	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	159	7	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	159	8	x	x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	159	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	159	10	o	o	NN
work_nu5xpentbrcxdlcdt3lmatbur4	159	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	159	12	t	t	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	159	13	o	o	XX
work_nu5xpentbrcxdlcdt3lmatbur4	159	14	i	i	NN
work_nu5xpentbrcxdlcdt3lmatbur4	159	15	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	159	16	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	159	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	159	18	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	159	19	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	159	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	159	21	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	159	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	160	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	160	2	combine	combine	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	160	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	160	4	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	5	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	160	6	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	160	7	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	8	features	feature	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	160	9	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	160	10	define	define	VB
work_nu5xpentbrcxdlcdt3lmatbur4	160	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	160	12	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	160	13	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	160	14	classifying	classify	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	160	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	160	16	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	160	18	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	160	19	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	160	20	denote	denote	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	160	21	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	160	22	φ(x	φ(x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	160	23	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	160	24	g	g	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	25	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	160	26	o	o	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	27	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	160	28	tg	tg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	160	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	160	30	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	160	31	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	160	32	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	160	33	φ	φ	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	34	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	160	35	∑	∑	.
work_nu5xpentbrcxdlcdt3lmatbur4	160	36	a∈{g	a∈{g	RB
work_nu5xpentbrcxdlcdt3lmatbur4	160	37	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	160	38	o	o	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	39	}	}	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	160	40	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	160	41	φA(x	φa(x	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	42	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	160	43	a	a	LS
work_nu5xpentbrcxdlcdt3lmatbur4	160	44	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	160	45	+	+	ADD
work_nu5xpentbrcxdlcdt3lmatbur4	160	46	∑	∑	.
work_nu5xpentbrcxdlcdt3lmatbur4	160	47	i	i	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	48	φT	φT	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	160	49	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	160	50	x	x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	160	51	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	160	52	a	a	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	160	53	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	160	54	t	t	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	160	55	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	160	56	i	i	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	57	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	160	58	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	160	59	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	160	60	1	1	LS
work_nu5xpentbrcxdlcdt3lmatbur4	160	61	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	160	62	Section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	160	63	5	5	CD
work_nu5xpentbrcxdlcdt3lmatbur4	160	64	describes	describe	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	160	65	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	160	66	actual	actual	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	160	67	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	160	68	used	use	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	160	69	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	160	70	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	160	71	experiments	experiment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	160	72	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	161	1	Observe	observe	VB
work_nu5xpentbrcxdlcdt3lmatbur4	161	2	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	161	3	given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	161	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	161	5	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	161	6	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	161	7	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	161	8	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	161	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	161	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	161	11	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	161	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	161	13	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	161	14	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	161	15	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	161	16	simply	simply	RB
work_nu5xpentbrcxdlcdt3lmatbur4	161	17	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	161	18	multiclass	multiclass	NN
work_nu5xpentbrcxdlcdt3lmatbur4	161	19	classification	classification	NN
work_nu5xpentbrcxdlcdt3lmatbur4	161	20	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	161	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	162	1	Thus	thus	RB
work_nu5xpentbrcxdlcdt3lmatbur4	162	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	162	3	following	follow	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	162	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	162	5	standard	standard	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	162	6	convention	convention	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	7	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	162	8	multiclass	multiclass	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	9	classification	classification	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	162	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	162	12	overall	overall	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	162	13	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	14	representation	representation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	15	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	162	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	162	17	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	18	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	162	19	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	20	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	21	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	162	22	defined	define	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	162	23	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	162	24	conjoining	conjoin	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	162	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	162	26	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	27	r	r	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	28	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	162	29	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	162	30	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	162	31	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	162	32	corresponding	correspond	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	162	33	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	162	34	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	162	35	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	162	36	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	162	37	φ	φ	NN
work_nu5xpentbrcxdlcdt3lmatbur4	162	38	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	163	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	163	2	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	163	3	us	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	163	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	163	5	full	full	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	163	6	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	163	7	representa-	representa-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	163	8	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	163	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	163	10	Φ(x	Φ(x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	163	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	163	12	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	163	13	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	163	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	164	1	4.2	4.2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	164	2	Model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	164	3	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	164	4	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	164	5	Predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	164	6	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	164	7	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	164	8	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	164	9	first	first	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	164	10	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	164	11	aims	aim	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	164	12	at	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	164	13	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	164	14	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	164	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	164	16	rela-	rela-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	164	17	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	164	18	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	164	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	164	20	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	164	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	164	22	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	164	23	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	164	24	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	164	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	165	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	165	2	falls	fall	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	165	3	into	into	IN
work_nu5xpentbrcxdlcdt3lmatbur4	165	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	165	5	standard	standard	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	165	6	multiclass	multiclass	NN
work_nu5xpentbrcxdlcdt3lmatbur4	165	7	classification	classification	NN
work_nu5xpentbrcxdlcdt3lmatbur4	165	8	set-	set-	IN
work_nu5xpentbrcxdlcdt3lmatbur4	165	9	ting	ting	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	165	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	165	11	where	where	WRB
work_nu5xpentbrcxdlcdt3lmatbur4	165	12	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	165	13	wish	wish	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	165	14	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	165	15	predict	predict	VB
work_nu5xpentbrcxdlcdt3lmatbur4	165	16	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	165	17	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	165	18	32	32	CD
work_nu5xpentbrcxdlcdt3lmatbur4	165	19	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	165	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	166	1	To	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	166	2	do	do	VB
work_nu5xpentbrcxdlcdt3lmatbur4	166	3	so	so	RB
work_nu5xpentbrcxdlcdt3lmatbur4	166	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	166	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	166	6	sum	sum	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	166	7	over	over	RP
work_nu5xpentbrcxdlcdt3lmatbur4	166	8	all	all	PDT
work_nu5xpentbrcxdlcdt3lmatbur4	166	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	166	10	possible	possible	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	166	11	assignments	assignment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	166	12	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	166	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	166	14	rest	rest	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	15	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	166	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	166	17	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	18	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	166	19	define	define	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	20	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	166	21	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	166	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	166	23	inputs	input	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	166	24	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	166	25	φ̂(x	φ̂(x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	166	26	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	166	27	=	=	NFP
work_nu5xpentbrcxdlcdt3lmatbur4	166	28	∑	∑	.
work_nu5xpentbrcxdlcdt3lmatbur4	166	29	g	g	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	30	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	166	31	o	o	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	32	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	166	33	tg	tg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	166	34	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	166	35	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	166	36	φ(x	φ(x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	166	37	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	166	38	g	g	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	39	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	166	40	o	o	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	41	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	166	42	tg	tg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	166	43	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	166	44	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	166	45	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	166	46	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	166	47	2	2	LS
work_nu5xpentbrcxdlcdt3lmatbur4	166	48	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	166	49	Effectively	effectively	RB
work_nu5xpentbrcxdlcdt3lmatbur4	166	50	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	166	51	doing	do	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	166	52	so	so	RB
work_nu5xpentbrcxdlcdt3lmatbur4	166	53	uses	use	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	166	54	all	all	PDT
work_nu5xpentbrcxdlcdt3lmatbur4	166	55	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	166	56	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	57	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	166	58	ob-	ob-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	166	59	ject	ject	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	166	60	candidates	candidate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	166	61	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	166	62	all	all	PDT
work_nu5xpentbrcxdlcdt3lmatbur4	166	63	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	166	64	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	166	65	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	166	66	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	166	67	get	get	VB
work_nu5xpentbrcxdlcdt3lmatbur4	166	68	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	166	69	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	70	representation	representation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	71	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	166	72	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	166	73	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	74	classifica-	classifica-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	75	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	76	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	166	77	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	167	1	Once	once	RB
work_nu5xpentbrcxdlcdt3lmatbur4	167	2	again	again	RB
work_nu5xpentbrcxdlcdt3lmatbur4	167	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	167	4	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	167	5	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	167	6	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	167	7	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	167	8	r	r	NN
work_nu5xpentbrcxdlcdt3lmatbur4	167	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	167	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	167	11	overall	overall	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	167	12	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	167	13	representation	representation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	167	14	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	167	15	defined	define	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	167	16	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	167	17	conjoin-	conjoin-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	167	18	ing	ing	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	167	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	167	20	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	167	21	r	r	NN
work_nu5xpentbrcxdlcdt3lmatbur4	167	22	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	167	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	167	24	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	167	25	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	167	26	that	that	DT
work_nu5xpentbrcxdlcdt3lmatbur4	167	27	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	167	28	φ̂	φ̂	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	167	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	167	30	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	167	31	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	167	32	write	write	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	167	33	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	167	34	φR(x	φr(x	NN
work_nu5xpentbrcxdlcdt3lmatbur4	167	35	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	167	36	r	r	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	167	37	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	167	38	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	168	1	Note	note	VB
work_nu5xpentbrcxdlcdt3lmatbur4	168	2	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	168	3	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	168	4	sum-	sum-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	168	5	mation	mation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	168	6	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	168	7	computationally	computationally	RB
work_nu5xpentbrcxdlcdt3lmatbur4	168	8	inexpensive	inexpensive	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	168	9	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	168	10	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	168	11	case	case	NN
work_nu5xpentbrcxdlcdt3lmatbur4	168	12	because	because	IN
work_nu5xpentbrcxdlcdt3lmatbur4	168	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	168	14	sum	sum	NN
work_nu5xpentbrcxdlcdt3lmatbur4	168	15	decomposes	decompose	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	168	16	according	accord	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	168	17	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	168	18	equation	equation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	168	19	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	168	20	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	168	21	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	168	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	169	1	With	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	169	2	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	169	3	learned	learned	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	169	4	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	169	5	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	169	6	w	w	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	169	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	169	9	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	169	10	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	169	11	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	169	12	predicted	predict	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	169	13	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	169	14	r	r	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	15	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	169	16	arg	arg	NN
work_nu5xpentbrcxdlcdt3lmatbur4	169	17	max	max	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	18	r′	r′	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	19	wT	wT	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	20	φR(x	φR(x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	169	22	r	r	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	23	′	′	XX
work_nu5xpentbrcxdlcdt3lmatbur4	169	24	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	169	25	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	169	26	3	3	LS
work_nu5xpentbrcxdlcdt3lmatbur4	169	27	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	169	28	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	169	29	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	169	30	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	169	31	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	169	32	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	33	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	169	34	Tsochantaridis	Tsochantaridis	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	35	et	et	FW
work_nu5xpentbrcxdlcdt3lmatbur4	169	36	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	37	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	38	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	169	39	2004	2004	CD
work_nu5xpentbrcxdlcdt3lmatbur4	169	40	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	169	41	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	169	42	train	train	VB
work_nu5xpentbrcxdlcdt3lmatbur4	169	43	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	169	44	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	169	45	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	169	46	w	w	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	169	47	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	169	48	predicts	predict	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	169	49	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	169	50	re-	re-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	169	51	lation	lation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	169	52	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	169	53	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	169	54	above	above	RB
work_nu5xpentbrcxdlcdt3lmatbur4	169	55	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	170	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	170	2	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	170	3	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	170	4	parameterized	parameterize	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	170	5	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	170	6	C	C	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	170	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	170	8	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	170	9	represents	represent	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	170	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	170	11	tradeoff	tradeoff	NN
work_nu5xpentbrcxdlcdt3lmatbur4	170	12	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	170	13	gener-	gener-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	170	14	alization	alization	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	170	15	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	170	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	170	17	hinge	hinge	NN
work_nu5xpentbrcxdlcdt3lmatbur4	170	18	loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	170	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	171	1	236	236	CD
work_nu5xpentbrcxdlcdt3lmatbur4	171	2	4.3	4.3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	171	3	Model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	171	4	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	171	5	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	171	6	Learning	learn	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	171	7	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	171	8	partial	partial	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	171	9	annotations	annotation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	171	10	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	171	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	171	12	second	second	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	171	13	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	171	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	171	15	even	even	RB
work_nu5xpentbrcxdlcdt3lmatbur4	171	16	though	though	IN
work_nu5xpentbrcxdlcdt3lmatbur4	171	17	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	171	18	annotation	annotation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	171	19	does	do	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	171	20	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	171	21	provide	provide	VB
work_nu5xpentbrcxdlcdt3lmatbur4	171	22	gold	gold	NN
work_nu5xpentbrcxdlcdt3lmatbur4	171	23	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	171	24	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	171	25	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	171	26	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	171	27	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	171	28	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	171	29	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	171	30	goal	goal	NN
work_nu5xpentbrcxdlcdt3lmatbur4	171	31	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	171	32	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	171	33	predict	predict	VB
work_nu5xpentbrcxdlcdt3lmatbur4	171	34	them	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	171	35	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	172	1	At	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	172	2	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	3	time	time	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	172	5	if	if	IN
work_nu5xpentbrcxdlcdt3lmatbur4	172	6	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	172	7	had	have	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	172	8	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	172	9	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	10	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	11	w	w	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	172	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	172	13	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	172	14	could	could	MD
work_nu5xpentbrcxdlcdt3lmatbur4	172	15	predict	predict	VB
work_nu5xpentbrcxdlcdt3lmatbur4	172	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	172	17	full	full	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	172	18	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	19	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	172	20	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	21	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	172	22	follows	follow	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	172	23	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	172	24	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	172	25	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	172	26	arg	arg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	172	27	max	max	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	172	28	y′	y′	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	172	29	wT	wT	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	172	30	Φ(x	Φ(x	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	172	31	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	172	32	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	172	33	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	172	34	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	172	35	4	4	LS
work_nu5xpentbrcxdlcdt3lmatbur4	172	36	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	172	37	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	172	38	propose	propose	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	172	39	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	172	40	iterative	iterative	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	172	41	learning	learning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	42	algorithm	algorithm	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	43	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	172	44	learn	learn	VB
work_nu5xpentbrcxdlcdt3lmatbur4	172	45	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	172	46	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	47	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	172	48	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	173	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	173	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	173	3	following	follow	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	173	4	discussion	discussion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	173	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	173	6	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	173	7	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	173	8	labeled	label	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	173	9	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	173	10	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	173	11	x	x	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	173	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	173	13	y∗	y∗	ADD
work_nu5xpentbrcxdlcdt3lmatbur4	173	14	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	173	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	173	16	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	173	17	refer	refer	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	173	18	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	173	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	173	20	missing	miss	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	173	21	part	part	NN
work_nu5xpentbrcxdlcdt3lmatbur4	173	22	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	173	23	its	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	173	24	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	173	25	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	173	26	h(y∗	h(y∗	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	173	27	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	173	28	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	174	1	That	that	RB
work_nu5xpentbrcxdlcdt3lmatbur4	174	2	is	is	RB
work_nu5xpentbrcxdlcdt3lmatbur4	174	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	174	4	h(y∗	h(y∗	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	174	5	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	174	6	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	174	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	174	8	assignment	assignment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	174	9	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	174	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	174	11	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	174	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	174	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	174	14	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	174	15	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	174	16	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	174	17	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	174	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	175	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	175	2	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	175	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	175	4	notation	notation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	175	5	r(y	r(y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	175	6	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	175	7	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	175	8	denote	denote	VB
work_nu5xpentbrcxdlcdt3lmatbur4	175	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	175	10	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	175	11	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	175	12	specified	specify	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	175	13	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	175	14	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	175	15	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	175	16	y.	y.	NN
work_nu5xpentbrcxdlcdt3lmatbur4	176	1	Our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	176	2	learning	learning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	176	3	algorithm	algorithm	NN
work_nu5xpentbrcxdlcdt3lmatbur4	176	4	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	176	5	closely	closely	RB
work_nu5xpentbrcxdlcdt3lmatbur4	176	6	related	relate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	176	7	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	176	8	re-	re-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	176	9	cently	cently	RB
work_nu5xpentbrcxdlcdt3lmatbur4	176	10	developed	develop	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	176	11	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	176	12	variable	variable	NN
work_nu5xpentbrcxdlcdt3lmatbur4	176	13	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	176	14	frameworks	framework	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	176	15	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	176	16	Yu	Yu	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	176	17	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	176	18	Joachims	Joachims	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	176	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	176	20	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	176	21	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	176	22	Chang	Chang	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	176	23	et	et	FW
work_nu5xpentbrcxdlcdt3lmatbur4	176	24	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	176	25	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	176	26	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	176	27	2010a	2010a	CD
work_nu5xpentbrcxdlcdt3lmatbur4	176	28	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	176	29	Chang	Chang	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	176	30	et	et	FW
work_nu5xpentbrcxdlcdt3lmatbur4	176	31	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	176	32	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	176	33	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	176	34	2010b	2010b	CD
work_nu5xpentbrcxdlcdt3lmatbur4	176	35	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	176	36	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	176	37	where	where	WRB
work_nu5xpentbrcxdlcdt3lmatbur4	176	38	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	176	39	supervision	supervision	NN
work_nu5xpentbrcxdlcdt3lmatbur4	176	40	provides	provide	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	176	41	only	only	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	176	42	partial	partial	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	176	43	annotation	annotation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	176	44	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	177	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	177	2	begin	begin	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	177	3	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	177	4	defining	define	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	177	5	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	177	6	addi-	addi-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	177	7	tional	tional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	177	8	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	177	9	procedures	procedure	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	177	10	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	177	11	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	177	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	178	1	Latent	Latent	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	178	2	Inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	178	3	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	178	4	Given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	178	5	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	178	6	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	178	7	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	178	8	w	w	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	178	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	178	10	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	178	11	partially	partially	RB
work_nu5xpentbrcxdlcdt3lmatbur4	178	12	labeled	label	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	178	13	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	178	14	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	178	15	x	x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	178	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	178	17	y∗	y∗	ADD
work_nu5xpentbrcxdlcdt3lmatbur4	178	18	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	178	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	178	20	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	178	21	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	178	22	‘	'	``
work_nu5xpentbrcxdlcdt3lmatbur4	178	23	complete	complete	VB
work_nu5xpentbrcxdlcdt3lmatbur4	178	24	’	'	''
work_nu5xpentbrcxdlcdt3lmatbur4	178	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	178	26	rest	rest	NN
work_nu5xpentbrcxdlcdt3lmatbur4	178	27	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	178	28	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	178	29	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	178	30	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	178	31	infer-	infer-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	178	32	ring	re	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	178	33	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	178	34	highest	high	JJS
work_nu5xpentbrcxdlcdt3lmatbur4	178	35	scoring	score	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	178	36	assignment	assignment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	178	37	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	178	38	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	178	39	miss-	miss-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	178	40	ing	ing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	178	41	parts	part	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	178	42	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	179	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	179	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	179	3	algorithm	algorithm	NN
work_nu5xpentbrcxdlcdt3lmatbur4	179	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	179	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	179	6	call	call	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	179	7	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	179	8	pro-	pro-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	179	9	cedure	cedure	VB
work_nu5xpentbrcxdlcdt3lmatbur4	179	10	LatentInf(w	latentinf(w	NN
work_nu5xpentbrcxdlcdt3lmatbur4	179	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	179	12	x	x	NN
work_nu5xpentbrcxdlcdt3lmatbur4	179	13	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	179	14	y∗	y∗	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	179	15	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	179	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	179	17	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	179	18	solves	solve	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	179	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	179	20	following	follow	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	179	21	maximization	maximization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	179	22	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	179	23	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	179	24	ŷ	ŷ	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	179	25	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	179	26	arg	arg	NN
work_nu5xpentbrcxdlcdt3lmatbur4	179	27	maxy	maxy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	179	28	w	w	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	179	29	T	T	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	179	30	Φ(x	Φ(x	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	179	31	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	179	32	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	179	33	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	179	34	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	179	35	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	179	36	5	5	LS
work_nu5xpentbrcxdlcdt3lmatbur4	179	37	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	179	38	s.t	s.t	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	179	39	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	180	1	r(y	r(y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	180	2	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	180	3	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	180	4	r(y∗	r(y∗	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	180	5	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	180	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	181	1	2	2	LS
work_nu5xpentbrcxdlcdt3lmatbur4	181	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	182	1	Loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	2	augmented	augment	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	182	3	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	4	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	182	5	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	182	6	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	182	7	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	182	8	variant	variant	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	182	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	182	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	182	12	standard	standard	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	182	13	loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	14	augmented	augment	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	182	15	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	16	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	182	17	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	182	18	SVMs	svm	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	182	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	182	20	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	182	21	solves	solve	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	182	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	182	23	follow-	follow-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	182	24	ing	ing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	25	maximization	maximization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	26	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	27	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	182	28	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	182	29	given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	182	30	x	x	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	31	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	182	32	fully	fully	RB
work_nu5xpentbrcxdlcdt3lmatbur4	182	33	labeled	label	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	182	34	y∗	y∗	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	182	35	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	182	36	arg	arg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	182	37	max	max	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	182	38	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	182	39	wT	wT	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	182	40	Φ(x	Φ(x	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	182	41	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	182	42	y	y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	182	43	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	182	44	+	+	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	182	45	∆(y	∆(y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	182	46	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	182	47	y∗	y∗	ADD
work_nu5xpentbrcxdlcdt3lmatbur4	182	48	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	182	49	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	182	50	6	6	CD
work_nu5xpentbrcxdlcdt3lmatbur4	182	51	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	182	52	Here	here	RB
work_nu5xpentbrcxdlcdt3lmatbur4	182	53	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	182	54	∆(y	∆(y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	182	55	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	182	56	y∗	y∗	ADD
work_nu5xpentbrcxdlcdt3lmatbur4	182	57	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	182	58	denotes	denote	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	182	59	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	182	60	loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	61	function	function	NN
work_nu5xpentbrcxdlcdt3lmatbur4	182	62	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	183	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	183	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	183	3	standard	standard	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	183	4	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	183	5	SVMs	svm	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	183	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	183	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	183	8	loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	183	9	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	183	10	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	183	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	183	12	entire	entire	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	183	13	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	183	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	184	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	184	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	184	3	Latent	Latent	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	184	4	Structural	Structural	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	184	5	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	184	6	formulation	formulation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	184	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	184	8	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	184	9	Yu	Yu	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	184	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	184	11	Joachims	Joachims	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	184	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	184	13	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	184	14	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	184	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	184	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	184	17	loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	184	18	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	184	19	defined	define	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	184	20	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	184	21	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	184	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	184	23	part	part	NN
work_nu5xpentbrcxdlcdt3lmatbur4	184	24	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	184	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	184	26	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	184	27	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	184	28	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	184	29	gold	gold	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	184	30	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	184	31	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	185	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	185	2	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	185	3	work	work	NN
work_nu5xpentbrcxdlcdt3lmatbur4	185	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	185	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	185	6	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	185	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	185	8	standard	standard	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	185	9	Hamming	Hamming	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	185	10	loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	185	11	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	185	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	185	13	entire	entire	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	185	14	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	185	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	185	16	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	185	17	scale	scale	VB
work_nu5xpentbrcxdlcdt3lmatbur4	185	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	185	19	loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	185	20	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	185	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	185	22	elements	element	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	185	23	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	185	24	h(y	h(y	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	185	25	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	185	26	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	185	27	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	185	28	parameter	parameter	NN
work_nu5xpentbrcxdlcdt3lmatbur4	185	29	α	α	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	185	30	<	<	XX
work_nu5xpentbrcxdlcdt3lmatbur4	185	31	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	185	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	186	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	186	2	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	186	3	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	186	4	gen-	gen-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	186	5	eralization	eralization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	186	6	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	186	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	186	8	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	186	9	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	186	10	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	186	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	186	12	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	186	13	corresponds	correspond	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	186	14	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	186	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	186	16	setting	setting	NN
work_nu5xpentbrcxdlcdt3lmatbur4	186	17	α	α	XX
work_nu5xpentbrcxdlcdt3lmatbur4	186	18	=	=	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	186	19	0	0	CD
work_nu5xpentbrcxdlcdt3lmatbur4	186	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	187	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	187	2	intu-	intu-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	187	3	ition	ition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	187	4	behind	behind	IN
work_nu5xpentbrcxdlcdt3lmatbur4	187	5	having	have	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	187	6	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	187	7	non	non	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	187	8	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	187	9	zero	zero	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	187	10	α	α	NN
work_nu5xpentbrcxdlcdt3lmatbur4	187	11	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	187	12	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	187	13	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	187	14	ad-	ad-	CC
work_nu5xpentbrcxdlcdt3lmatbur4	187	15	dition	dition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	187	16	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	187	17	penalizing	penalize	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	187	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	187	19	learning	learning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	187	20	algorithm	algorithm	RB
work_nu5xpentbrcxdlcdt3lmatbur4	187	21	if	if	IN
work_nu5xpentbrcxdlcdt3lmatbur4	187	22	it	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	187	23	violates	violate	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	187	24	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	187	25	annotated	annotate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	187	26	part	part	NN
work_nu5xpentbrcxdlcdt3lmatbur4	187	27	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	187	28	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	187	29	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	187	30	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	187	31	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	187	32	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	187	33	incorporate	incorporate	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	187	34	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	187	35	small	small	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	187	36	penalty	penalty	NN
work_nu5xpentbrcxdlcdt3lmatbur4	187	37	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	187	38	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	187	39	rest	rest	NN
work_nu5xpentbrcxdlcdt3lmatbur4	187	40	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	187	41	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	187	42	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	187	43	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	188	1	Using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	188	2	these	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	188	3	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	188	4	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	188	5	procedures	procedure	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	188	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	188	7	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	188	8	define	define	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	188	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	188	10	learning	learn	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	188	11	algorithm	algorithm	NN
work_nu5xpentbrcxdlcdt3lmatbur4	188	12	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	188	13	Algorithm	Algorithm	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	188	14	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	188	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	189	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	189	2	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	189	3	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	189	4	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	189	5	initialized	initialize	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	189	6	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	189	7	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	189	8	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	189	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	190	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	190	2	algorithm	algorithm	NN
work_nu5xpentbrcxdlcdt3lmatbur4	190	3	then	then	RB
work_nu5xpentbrcxdlcdt3lmatbur4	190	4	finds	find	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	190	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	190	6	best	good	JJS
work_nu5xpentbrcxdlcdt3lmatbur4	190	7	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	190	8	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	190	9	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	190	10	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	190	11	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	190	12	ex-	ex-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	190	13	amples	ample	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	190	14	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	190	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	190	16	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	190	17	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	190	18	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	190	19	steps	step	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	190	20	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	190	21	-	-	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	190	22	5	5	CD
work_nu5xpentbrcxdlcdt3lmatbur4	190	23	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	190	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	191	1	Doing	do	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	191	2	so	so	RB
work_nu5xpentbrcxdlcdt3lmatbur4	191	3	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	191	4	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	191	5	estimate	estimate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	191	6	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	191	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	191	8	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	191	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	191	10	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	191	11	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	191	12	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	191	13	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	191	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	191	15	giving	give	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	191	16	us	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	191	17	‘	'	``
work_nu5xpentbrcxdlcdt3lmatbur4	191	18	fully	fully	RB
work_nu5xpentbrcxdlcdt3lmatbur4	191	19	labeled	label	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	191	20	’	'	''
work_nu5xpentbrcxdlcdt3lmatbur4	191	21	structured	structured	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	191	22	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	191	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	192	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	192	2	algorithm	algorithm	NN
work_nu5xpentbrcxdlcdt3lmatbur4	192	3	then	then	RB
work_nu5xpentbrcxdlcdt3lmatbur4	192	4	proceeds	proceed	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	192	5	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	192	6	use	use	VB
work_nu5xpentbrcxdlcdt3lmatbur4	192	7	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	192	8	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	192	9	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	192	10	train	train	VB
work_nu5xpentbrcxdlcdt3lmatbur4	192	11	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	192	12	new	new	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	192	13	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	192	14	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	192	15	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	192	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	192	17	standard	standard	NN
work_nu5xpentbrcxdlcdt3lmatbur4	192	18	struc-	struc-	.
work_nu5xpentbrcxdlcdt3lmatbur4	192	19	tural	tural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	192	20	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	192	21	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	192	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	192	23	loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	192	24	augmented	augment	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	192	25	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	192	26	listed	list	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	192	27	above	above	IN
work_nu5xpentbrcxdlcdt3lmatbur4	192	28	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	192	29	step	step	NN
work_nu5xpentbrcxdlcdt3lmatbur4	192	30	6	6	CD
work_nu5xpentbrcxdlcdt3lmatbur4	192	31	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	192	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	193	1	These	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	193	2	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	193	3	steps	step	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	193	4	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	193	5	repeated	repeat	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	193	6	several	several	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	193	7	times	time	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	193	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	194	1	Note	note	VB
work_nu5xpentbrcxdlcdt3lmatbur4	194	2	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	194	3	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	194	4	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	194	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	194	6	summation	summation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	194	7	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	194	8	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	194	9	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	194	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	194	11	solving	solve	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	194	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	194	13	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	194	14	problems	problem	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	194	15	described	describe	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	194	16	above	above	RB
work_nu5xpentbrcxdlcdt3lmatbur4	194	17	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	194	18	computationally	computationally	RB
work_nu5xpentbrcxdlcdt3lmatbur4	194	19	inexpensive	inexpensive	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	194	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	195	1	Algorithm	Algorithm	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	195	2	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	195	3	Algorithm	Algorithm	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	195	4	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	195	5	learning	learn	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	195	6	Model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	195	7	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	195	8	Input	input	NN
work_nu5xpentbrcxdlcdt3lmatbur4	195	9	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	195	10	Examples	example	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	195	11	D	d	NN
work_nu5xpentbrcxdlcdt3lmatbur4	195	12	=	=	NFP
work_nu5xpentbrcxdlcdt3lmatbur4	195	13	{	{	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	195	14	xi	xi	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	195	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	195	16	r(y∗i	r(y∗i	LS
work_nu5xpentbrcxdlcdt3lmatbur4	195	17	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	195	18	}	}	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	195	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	195	20	where	where	WRB
work_nu5xpentbrcxdlcdt3lmatbur4	195	21	exam-	exam-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	195	22	ples	ple	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	195	23	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	195	24	labeled	label	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	195	25	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	195	26	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	195	27	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	195	28	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	195	29	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	195	30	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	196	1	1	1	LS
work_nu5xpentbrcxdlcdt3lmatbur4	196	2	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	196	3	Initialize	initialize	VB
work_nu5xpentbrcxdlcdt3lmatbur4	196	4	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	196	5	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	196	6	w	w	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	7	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	196	8	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	9	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	196	10	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	196	11	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	196	12	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	196	13	t	t	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	14	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	196	15	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	196	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	196	17	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	196	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	196	19	·	·	NFP
work_nu5xpentbrcxdlcdt3lmatbur4	196	20	·	·	NFP
work_nu5xpentbrcxdlcdt3lmatbur4	196	21	·	·	NFP
work_nu5xpentbrcxdlcdt3lmatbur4	196	22	do	do	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	196	23	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	196	24	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	196	25	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	196	26	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	196	27	xi	xi	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	28	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	196	29	y∗i	y∗i	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	30	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	196	31	∈	∈	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	32	D	D	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	33	do	do	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	196	34	4	4	CD
work_nu5xpentbrcxdlcdt3lmatbur4	196	35	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	196	36	ŷi	ŷi	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	37	←	←	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	38	LatentInf(w	latentinf(w	CD
work_nu5xpentbrcxdlcdt3lmatbur4	196	39	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	196	40	xi	xi	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	41	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	196	42	y∗i	y∗i	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	43	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	196	44	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	196	45	Eq	Eq	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	196	46	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	197	1	5	5	LS
work_nu5xpentbrcxdlcdt3lmatbur4	197	2	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	197	3	5	5	CD
work_nu5xpentbrcxdlcdt3lmatbur4	197	4	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	197	5	end	end	NN
work_nu5xpentbrcxdlcdt3lmatbur4	197	6	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	197	7	6	6	CD
work_nu5xpentbrcxdlcdt3lmatbur4	197	8	:	:	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	197	9	w	w	NN
work_nu5xpentbrcxdlcdt3lmatbur4	197	10	←	←	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	197	11	LearnSSV	LearnSSV	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	197	12	M({xi	m({xi	NN
work_nu5xpentbrcxdlcdt3lmatbur4	197	13	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	197	14	ŷi	ŷi	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	197	15	}	}	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	197	16	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	197	17	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	197	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	197	19	loss	loss	NN
work_nu5xpentbrcxdlcdt3lmatbur4	197	20	augmented	augment	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	197	21	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	197	22	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	197	23	Eq	Eq	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	197	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	198	1	6	6	CD
work_nu5xpentbrcxdlcdt3lmatbur4	198	2	7	7	CD
work_nu5xpentbrcxdlcdt3lmatbur4	198	3	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	198	4	end	end	NN
work_nu5xpentbrcxdlcdt3lmatbur4	198	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	198	6	8	8	CD
work_nu5xpentbrcxdlcdt3lmatbur4	198	7	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	198	8	return	return	NN
work_nu5xpentbrcxdlcdt3lmatbur4	198	9	w	w	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	198	10	Algorithm	Algorithm	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	198	11	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	198	12	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	198	13	parameterized	parameterize	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	198	14	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	198	15	C	C	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	198	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	198	17	α	α	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	198	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	199	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	199	2	parameter	parameter	NN
work_nu5xpentbrcxdlcdt3lmatbur4	199	3	α	α	LS
work_nu5xpentbrcxdlcdt3lmatbur4	199	4	controls	control	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	199	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	199	6	extent	extent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	199	7	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	199	8	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	199	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	199	10	hypoth-	hypoth-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	199	11	esized	esize	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	199	12	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	199	13	according	accord	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	199	14	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	199	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	199	16	previous	previous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	199	17	iteration	iteration	NN
work_nu5xpentbrcxdlcdt3lmatbur4	199	18	’s	’s	POS
work_nu5xpentbrcxdlcdt3lmatbur4	199	19	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	199	20	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	199	21	influence	influence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	199	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	199	23	learning	learning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	199	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	200	1	237	237	CD
work_nu5xpentbrcxdlcdt3lmatbur4	200	2	4.4	4.4	CD
work_nu5xpentbrcxdlcdt3lmatbur4	200	3	Joint	Joint	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	200	4	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	200	5	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	200	6	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	200	7	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	200	8	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	200	9	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	200	10	By	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	200	11	defining	define	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	200	12	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	200	13	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	200	14	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	200	15	disjoint	disjoint	NN
work_nu5xpentbrcxdlcdt3lmatbur4	200	16	sets	set	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	200	17	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	200	18	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	200	19	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	200	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	200	21	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	200	22	effectively	effectively	RB
work_nu5xpentbrcxdlcdt3lmatbur4	200	23	have	have	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	200	24	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	200	25	hierarchi-	hierarchi-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	200	26	cal	cal	NN
work_nu5xpentbrcxdlcdt3lmatbur4	200	27	relationship	relationship	NN
work_nu5xpentbrcxdlcdt3lmatbur4	200	28	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	200	29	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	200	30	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	200	31	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	200	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	201	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	201	2	suggests	suggest	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	201	3	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	201	4	joint	joint	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	201	5	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	201	6	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	201	7	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	201	8	employed	employ	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	201	9	be-	be-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	201	10	tween	tween	NN
work_nu5xpentbrcxdlcdt3lmatbur4	201	11	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	201	12	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	201	13	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	201	14	predictions	prediction	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	201	15	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	201	16	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	201	17	validity	validity	NN
work_nu5xpentbrcxdlcdt3lmatbur4	201	18	constraint	constraint	NN
work_nu5xpentbrcxdlcdt3lmatbur4	201	19	connecting	connect	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	201	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	201	21	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	201	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	202	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	202	2	idea	idea	NN
work_nu5xpentbrcxdlcdt3lmatbur4	202	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	202	4	employ-	employ-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	202	5	ing	e	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	202	6	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	202	7	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	202	8	combine	combine	VB
work_nu5xpentbrcxdlcdt3lmatbur4	202	9	independently	independently	RB
work_nu5xpentbrcxdlcdt3lmatbur4	202	10	trained	train	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	202	11	pre-	pre-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	202	12	dictors	dictor	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	202	13	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	202	14	obtain	obtain	VB
work_nu5xpentbrcxdlcdt3lmatbur4	202	15	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	202	16	coherent	coherent	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	202	17	output	output	NN
work_nu5xpentbrcxdlcdt3lmatbur4	202	18	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	202	19	has	have	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	202	20	been	be	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	202	21	used	use	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	202	22	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	202	23	various	various	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	202	24	NLP	NLP	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	202	25	tasks	task	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	202	26	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	202	27	recent	recent	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	202	28	years	year	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	202	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	202	30	starting	start	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	202	31	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	202	32	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	202	33	work	work	NN
work_nu5xpentbrcxdlcdt3lmatbur4	202	34	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	202	35	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	202	36	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	202	37	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	202	38	Yih	Yih	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	202	39	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	202	40	2004	2004	CD
work_nu5xpentbrcxdlcdt3lmatbur4	202	41	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	202	42	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	202	43	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	202	44	Yih	Yih	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	202	45	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	202	46	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	202	47	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	202	48	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	203	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	203	2	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	203	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	203	4	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	203	5	defined	define	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	203	6	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	203	7	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	203	8	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	9	et	et	FW
work_nu5xpentbrcxdlcdt3lmatbur4	203	10	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	11	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	203	13	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	203	14	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	203	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	203	16	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	203	17	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	203	18	write	write	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	203	19	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	203	20	φs(x	φs(x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	203	22	s	s	LS
work_nu5xpentbrcxdlcdt3lmatbur4	203	23	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	203	24	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	203	25	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	203	26	given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	203	27	input	input	NN
work_nu5xpentbrcxdlcdt3lmatbur4	203	28	x	x	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	203	29	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	203	30	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	203	31	label	label	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	32	s	s	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	33	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	203	34	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	203	35	train	train	VB
work_nu5xpentbrcxdlcdt3lmatbur4	203	36	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	203	37	separate	separate	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	203	38	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	203	39	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	203	40	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	203	41	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	203	42	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	203	43	SemEval	SemEval	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	44	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	203	45	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	203	46	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	203	47	φs(x	φs(x	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	48	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	203	49	s	s	LS
work_nu5xpentbrcxdlcdt3lmatbur4	203	50	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	203	51	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	203	52	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	203	53	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	203	54	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	55	algorithm	algorithm	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	203	56	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	204	1	Thus	thus	RB
work_nu5xpentbrcxdlcdt3lmatbur4	204	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	204	3	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	204	4	have	have	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	204	5	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	204	6	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	204	7	vectors	vector	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	204	8	–	–	:
work_nu5xpentbrcxdlcdt3lmatbur4	204	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	204	10	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	204	11	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	204	12	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	204	13	preposi-	preposi-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	204	14	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	204	15	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	204	16	described	describe	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	204	17	earlier	early	RBR
work_nu5xpentbrcxdlcdt3lmatbur4	204	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	204	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	204	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	204	21	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	204	22	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	204	23	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	204	24	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	204	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	205	1	At	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	205	2	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	3	time	time	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	205	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	205	6	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	7	given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	205	8	input	input	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	205	10	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	205	11	find	find	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	205	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	13	highest	high	JJS
work_nu5xpentbrcxdlcdt3lmatbur4	205	14	scoring	score	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	205	15	joint	joint	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	205	16	assignment	assignment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	17	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	205	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	19	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	205	21	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	205	22	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	205	23	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	205	24	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	205	25	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	26	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	27	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	205	28	sub-	sub-	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	29	ject	ject	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	30	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	205	31	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	32	constraint	constraint	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	33	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	205	34	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	35	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	36	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	205	37	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	38	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	39	agree	agree	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	205	40	according	accord	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	205	41	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	205	42	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	43	definition	definition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	205	44	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	205	45	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	205	46	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	205	47	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	206	1	5	5	CD
work_nu5xpentbrcxdlcdt3lmatbur4	206	2	Experiments	Experiments	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	206	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	206	4	Results	result	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	206	5	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	206	6	primary	primary	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	206	7	research	research	NN
work_nu5xpentbrcxdlcdt3lmatbur4	206	8	goal	goal	NN
work_nu5xpentbrcxdlcdt3lmatbur4	206	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	206	10	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	206	11	experiments	experiment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	206	12	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	206	13	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	206	14	evaluate	evaluate	VB
work_nu5xpentbrcxdlcdt3lmatbur4	206	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	206	16	different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	206	17	models	model	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	206	18	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	206	19	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	206	20	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	206	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	206	22	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	206	23	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	206	24	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	206	25	joint	joint	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	206	26	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	206	27	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	206	28	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	206	29	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	206	30	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	206	31	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	206	32	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	206	33	prepo-	prepo-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	206	34	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	206	35	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	206	36	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	207	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	207	2	additional	additional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	207	3	analysis	analysis	NN
work_nu5xpentbrcxdlcdt3lmatbur4	207	4	experiments	experiment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	207	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	207	6	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	207	7	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	207	8	show	show	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	207	9	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	207	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	207	11	definition	definition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	207	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	207	13	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	207	14	rela-	rela-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	207	15	tions	tion	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	207	16	indeed	indeed	RB
work_nu5xpentbrcxdlcdt3lmatbur4	207	17	captures	capture	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	207	18	cross	cross	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	207	19	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	207	20	preposition	preposition	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	207	21	semantics	semantic	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	207	22	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	207	23	taking	take	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	207	24	advantage	advantage	NN
work_nu5xpentbrcxdlcdt3lmatbur4	207	25	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	207	26	shared	shared	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	207	27	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	207	28	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	207	29	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	207	30	high-	high-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	207	31	light	light	NN
work_nu5xpentbrcxdlcdt3lmatbur4	207	32	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	207	33	need	need	NN
work_nu5xpentbrcxdlcdt3lmatbur4	207	34	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	207	35	going	go	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	207	36	beyond	beyond	IN
work_nu5xpentbrcxdlcdt3lmatbur4	207	37	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	207	38	syntactic	syntactic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	207	39	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	207	40	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	208	1	5.1	5.1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	208	2	Types	Types	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	208	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	208	4	Features	feature	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	208	5	Types	Types	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	208	6	As	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	208	7	described	describe	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	208	8	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	208	9	Section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	208	10	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	208	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	208	12	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	208	13	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	208	14	WordNet	WordNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	208	15	hypernyms	hypernyms	NN
work_nu5xpentbrcxdlcdt3lmatbur4	208	16	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	208	17	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	208	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	208	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	208	20	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	208	21	categories	category	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	208	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	209	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	209	2	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	209	3	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	209	4	hypernyms	hypernyms	NN
work_nu5xpentbrcxdlcdt3lmatbur4	209	5	within	within	IN
work_nu5xpentbrcxdlcdt3lmatbur4	209	6	four	four	CD
work_nu5xpentbrcxdlcdt3lmatbur4	209	7	levels	level	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	209	8	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	209	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	209	10	hypernym	hypernym	NN
work_nu5xpentbrcxdlcdt3lmatbur4	209	11	hier-	hier-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	209	12	archy	archy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	209	13	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	209	14	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	209	15	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	209	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	210	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	210	2	second	second	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	210	3	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	210	4	category	category	NN
work_nu5xpentbrcxdlcdt3lmatbur4	210	5	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	210	6	defined	define	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	210	7	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	210	8	word-	word-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	210	9	similarity	similarity	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	210	10	driven	drive	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	210	11	clusters	cluster	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	210	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	211	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	211	2	briefly	briefly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	211	3	describe	describe	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	211	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	211	5	clustering	clustering	NN
work_nu5xpentbrcxdlcdt3lmatbur4	211	6	process	process	NN
work_nu5xpentbrcxdlcdt3lmatbur4	211	7	here	here	RB
work_nu5xpentbrcxdlcdt3lmatbur4	211	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	212	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	212	2	thesaurus	thesaurus	NN
work_nu5xpentbrcxdlcdt3lmatbur4	212	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	212	4	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	212	5	Lin	Lin	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	212	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	212	7	1998	1998	CD
work_nu5xpentbrcxdlcdt3lmatbur4	212	8	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	212	9	specifies	specify	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	212	10	similar	similar	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	212	11	lexical	lexical	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	212	12	items	item	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	212	13	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	212	14	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	212	15	given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	212	16	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	212	17	along	along	IN
work_nu5xpentbrcxdlcdt3lmatbur4	212	18	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	212	19	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	212	20	similarity	similarity	NN
work_nu5xpentbrcxdlcdt3lmatbur4	212	21	score	score	NN
work_nu5xpentbrcxdlcdt3lmatbur4	212	22	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	212	23	0	0	CD
work_nu5xpentbrcxdlcdt3lmatbur4	212	24	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	212	25	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	212	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	213	1	It	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	213	2	treats	treat	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	213	3	nouns	noun	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	213	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	213	5	verbs	verb	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	213	6	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	213	7	adjectives	adjective	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	213	8	separately	separately	RB
work_nu5xpentbrcxdlcdt3lmatbur4	213	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	214	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	214	2	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	214	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	214	4	score	score	NN
work_nu5xpentbrcxdlcdt3lmatbur4	214	5	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	214	6	cluster	cluster	VB
work_nu5xpentbrcxdlcdt3lmatbur4	214	7	groups	group	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	214	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	214	9	similar	similar	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	214	10	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	214	11	us-	us-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	214	12	ing	e	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	214	13	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	214	14	greedy	greedy	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	214	15	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	214	16	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	214	17	covering	cover	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	214	18	approach	approach	NN
work_nu5xpentbrcxdlcdt3lmatbur4	214	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	215	1	Specifically	specifically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	215	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	215	3	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	215	4	randomly	randomly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	215	5	select	select	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	215	6	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	215	7	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	215	8	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	215	9	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	215	10	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	215	11	yet	yet	RB
work_nu5xpentbrcxdlcdt3lmatbur4	215	12	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	215	13	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	215	14	cluster	cluster	NN
work_nu5xpentbrcxdlcdt3lmatbur4	215	15	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	215	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	215	17	center	center	NN
work_nu5xpentbrcxdlcdt3lmatbur4	215	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	215	19	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	215	20	new	new	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	215	21	cluster	cluster	NN
work_nu5xpentbrcxdlcdt3lmatbur4	215	22	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	215	23	add	add	VB
work_nu5xpentbrcxdlcdt3lmatbur4	215	24	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	215	25	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	215	26	whose	whose	WP$
work_nu5xpentbrcxdlcdt3lmatbur4	215	27	score	score	NN
work_nu5xpentbrcxdlcdt3lmatbur4	215	28	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	215	29	greater	great	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	215	30	than	than	IN
work_nu5xpentbrcxdlcdt3lmatbur4	215	31	σ	σ	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	215	32	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	215	33	it	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	215	34	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	216	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	216	2	re-	re-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	216	3	peat	peat	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	216	4	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	216	5	process	process	NN
work_nu5xpentbrcxdlcdt3lmatbur4	216	6	till	till	IN
work_nu5xpentbrcxdlcdt3lmatbur4	216	7	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	216	8	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	216	9	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	216	10	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	216	11	some	some	DT
work_nu5xpentbrcxdlcdt3lmatbur4	216	12	clus-	clus-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	216	13	ter	ter	NN
work_nu5xpentbrcxdlcdt3lmatbur4	216	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	217	1	A	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	217	2	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	217	3	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	217	4	appear	appear	VB
work_nu5xpentbrcxdlcdt3lmatbur4	217	5	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	217	6	more	more	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	217	7	than	than	IN
work_nu5xpentbrcxdlcdt3lmatbur4	217	8	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	217	9	cluster	cluster	NN
work_nu5xpentbrcxdlcdt3lmatbur4	217	10	because	because	IN
work_nu5xpentbrcxdlcdt3lmatbur4	217	11	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	217	12	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	217	13	similar	similar	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	217	14	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	217	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	217	16	cluster	cluster	NN
work_nu5xpentbrcxdlcdt3lmatbur4	217	17	center	center	NN
work_nu5xpentbrcxdlcdt3lmatbur4	217	18	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	217	19	added	add	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	217	20	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	217	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	217	22	cluster	cluster	NN
work_nu5xpentbrcxdlcdt3lmatbur4	217	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	218	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	218	2	repeat	repeat	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	218	3	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	218	4	process	process	NN
work_nu5xpentbrcxdlcdt3lmatbur4	218	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	218	6	σ	σ	NN
work_nu5xpentbrcxdlcdt3lmatbur4	218	7	∈{0.1	∈{0.1	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	218	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	218	9	0.125	0.125	CD
work_nu5xpentbrcxdlcdt3lmatbur4	218	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	218	11	0.15	0.15	CD
work_nu5xpentbrcxdlcdt3lmatbur4	218	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	218	13	0.175	0.175	CD
work_nu5xpentbrcxdlcdt3lmatbur4	218	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	218	15	0.2	0.2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	218	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	218	17	0.25	0.25	CD
work_nu5xpentbrcxdlcdt3lmatbur4	218	18	}	}	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	218	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	219	1	By	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	219	2	increas-	increas-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	219	3	ing	e	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	219	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	219	5	value	value	NN
work_nu5xpentbrcxdlcdt3lmatbur4	219	6	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	219	7	σ	σ	NN
work_nu5xpentbrcxdlcdt3lmatbur4	219	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	219	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	219	10	clusters	cluster	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	219	11	become	become	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	219	12	more	more	RBR
work_nu5xpentbrcxdlcdt3lmatbur4	219	13	selec-	selec-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	219	14	tive	tive	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	219	15	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	219	16	hence	hence	RB
work_nu5xpentbrcxdlcdt3lmatbur4	219	17	smaller	small	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	219	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	220	1	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	220	2	4	4	CD
work_nu5xpentbrcxdlcdt3lmatbur4	220	3	shows	show	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	220	4	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	220	5	noun	noun	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	220	6	clusters	cluster	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	220	7	created	create	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	220	8	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	220	9	σ	σ	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	220	10	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	220	11	0.15	0.15	CD
work_nu5xpentbrcxdlcdt3lmatbur4	220	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	221	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	221	2	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	221	3	given	give	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	221	4	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	221	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	221	6	identifiers	identifier	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	221	7	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	221	8	clusters	cluster	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	221	9	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	221	10	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	221	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	221	12	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	221	13	belongs	belong	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	221	14	serve	serve	VB
work_nu5xpentbrcxdlcdt3lmatbur4	221	15	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	221	16	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	221	17	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	221	18	candidates	candidate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	221	19	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	221	20	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	221	21	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	221	22	category5	category5	NN
work_nu5xpentbrcxdlcdt3lmatbur4	221	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	222	1	Features	feature	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	222	2	Our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	222	3	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	222	4	features	feature	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	222	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	222	6	denoted	denote	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	222	7	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	222	8	φA	φA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	222	9	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	222	10	Section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	222	11	4.1	4.1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	222	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	222	13	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	222	14	derived	derive	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	222	15	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	222	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	222	17	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	222	18	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	222	19	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	222	20	set	set	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	222	21	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	222	22	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	222	23	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	222	24	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	222	25	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	222	26	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	222	27	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	222	28	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	222	29	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	222	30	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	222	31	extract	extract	VB
work_nu5xpentbrcxdlcdt3lmatbur4	222	32	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	222	33	following	following	NN
work_nu5xpentbrcxdlcdt3lmatbur4	222	34	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	222	35	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	222	36	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	222	37	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	222	38	1	1	LS
work_nu5xpentbrcxdlcdt3lmatbur4	222	39	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	223	1	Word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	223	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	223	3	part	part	NN
work_nu5xpentbrcxdlcdt3lmatbur4	223	4	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	223	5	of-	of-	CD
work_nu5xpentbrcxdlcdt3lmatbur4	223	6	speech	speech	NN
work_nu5xpentbrcxdlcdt3lmatbur4	223	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	223	8	lemma	lemma	NN
work_nu5xpentbrcxdlcdt3lmatbur4	223	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	223	10	capitalization	capitalization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	223	11	indicator	indicator	NN
work_nu5xpentbrcxdlcdt3lmatbur4	223	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	223	13	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	223	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	224	1	Con-	Con-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	224	2	flated	flate	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	224	3	part	part	NN
work_nu5xpentbrcxdlcdt3lmatbur4	224	4	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	224	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	224	6	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	224	7	speech	speech	NN
work_nu5xpentbrcxdlcdt3lmatbur4	224	8	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	224	9	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	224	10	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	224	11	Noun	Noun	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	224	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	224	13	Verb	Verb	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	224	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	224	15	Adjective	Adjective	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	224	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	224	17	Adverb	Adverb	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	224	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	224	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	224	20	Other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	224	21	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	224	22	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	224	23	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	224	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	225	1	Indicator	indicator	NN
work_nu5xpentbrcxdlcdt3lmatbur4	225	2	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	225	3	existence	existence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	225	4	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	225	5	WordNet	WordNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	225	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	225	7	4	4	CD
work_nu5xpentbrcxdlcdt3lmatbur4	225	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	226	1	WordNet	wordnet	NN
work_nu5xpentbrcxdlcdt3lmatbur4	226	2	synsets	synset	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	226	3	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	226	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	226	5	first	first	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	226	6	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	226	7	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	226	8	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	226	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	226	10	5	5	CD
work_nu5xpentbrcxdlcdt3lmatbur4	226	11	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	227	1	WordNet	WordNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	227	2	lemma	lemma	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	227	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	227	4	lexicographer	lexicographer	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	227	5	file	file	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	227	6	names	name	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	227	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	227	8	part	part	NN
work_nu5xpentbrcxdlcdt3lmatbur4	227	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	227	10	member	member	NN
work_nu5xpentbrcxdlcdt3lmatbur4	227	11	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	227	12	substance	substance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	227	13	holonyms	holonyms	NN
work_nu5xpentbrcxdlcdt3lmatbur4	227	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	227	15	6	6	CD
work_nu5xpentbrcxdlcdt3lmatbur4	227	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	228	1	Roget	roget	VB
work_nu5xpentbrcxdlcdt3lmatbur4	228	2	thesaurus	thesaurus	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	228	3	divisions	division	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	228	4	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	228	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	228	6	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	228	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	228	8	7	7	CD
work_nu5xpentbrcxdlcdt3lmatbur4	228	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	229	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	229	2	first	first	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	229	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	229	4	last	last	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	229	5	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	229	6	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	229	7	three	three	CD
work_nu5xpentbrcxdlcdt3lmatbur4	229	8	letters	letter	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	229	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	229	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	229	11	8	8	CD
work_nu5xpentbrcxdlcdt3lmatbur4	229	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	230	1	Indicators	indicator	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	230	2	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	230	3	known	known	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	230	4	af-	af-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	230	5	fixes	fix	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	230	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	231	1	Our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	231	2	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	231	3	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	231	4	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	231	5	φT	φt	NN
work_nu5xpentbrcxdlcdt3lmatbur4	231	6	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	231	7	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	231	8	simply	simply	RB
work_nu5xpentbrcxdlcdt3lmatbur4	231	9	indicators	indicator	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	231	10	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	231	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	231	12	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	231	13	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	231	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	231	15	conjoined	conjoin	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	231	16	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	231	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	231	18	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	231	19	category	category	NN
work_nu5xpentbrcxdlcdt3lmatbur4	231	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	232	1	One	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	232	2	advantage	advantage	NN
work_nu5xpentbrcxdlcdt3lmatbur4	232	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	232	4	abstracting	abstract	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	232	5	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	232	6	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	232	7	into	into	IN
work_nu5xpentbrcxdlcdt3lmatbur4	232	8	re-	re-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	232	9	lations	lation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	232	10	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	232	11	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	232	12	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	232	13	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	232	14	share	share	VB
work_nu5xpentbrcxdlcdt3lmatbur4	232	15	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	232	16	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	232	17	different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	232	18	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	232	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	233	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	233	2	base	base	NN
work_nu5xpentbrcxdlcdt3lmatbur4	233	3	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	233	4	set	set	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	233	5	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	233	6	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	233	7	both	both	DT
work_nu5xpentbrcxdlcdt3lmatbur4	233	8	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	233	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	233	10	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	233	11	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	233	12	defined	define	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	233	13	above	above	RB
work_nu5xpentbrcxdlcdt3lmatbur4	233	14	does	do	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	233	15	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	233	16	encode	encode	VB
work_nu5xpentbrcxdlcdt3lmatbur4	233	17	in-	in-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	233	18	formation	formation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	233	19	about	about	IN
work_nu5xpentbrcxdlcdt3lmatbur4	233	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	233	21	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	233	22	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	233	23	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	233	24	classified	classify	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	233	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	234	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	234	2	do	do	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	234	3	so	so	RB
work_nu5xpentbrcxdlcdt3lmatbur4	234	4	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	234	5	conjoining	conjoin	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	234	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	234	7	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	234	8	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	234	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	234	10	preposi-	preposi-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	234	11	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	234	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	235	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	235	2	addition	addition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	235	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	235	4	since	since	IN
work_nu5xpentbrcxdlcdt3lmatbur4	235	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	235	6	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	235	7	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	235	8	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	235	9	shared	share	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	235	10	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	235	11	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	235	12	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	235	13	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	235	14	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	235	15	include	include	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	235	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	235	17	base	base	NN
work_nu5xpentbrcxdlcdt3lmatbur4	235	18	features	feature	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	235	19	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	235	20	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	235	21	shared	share	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	235	22	representation	representation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	235	23	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	235	24	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	235	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	236	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	236	2	consider	consider	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	236	3	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	236	4	variants	variant	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	236	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	236	6	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	236	7	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	236	8	sets	set	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	236	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	237	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	237	2	refer	refer	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	237	3	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	237	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	237	5	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	237	6	described	describe	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	237	7	above	above	RB
work_nu5xpentbrcxdlcdt3lmatbur4	237	8	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	237	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	237	10	typed	type	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	237	11	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	237	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	238	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	238	2	addition	addition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	238	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	238	4	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	238	5	define	define	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	238	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	238	7	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	238	8	features	feature	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	238	9	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	238	10	conjoining	conjoin	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	238	11	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	238	12	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	238	13	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	238	14	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	238	15	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	238	16	typed	type	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	238	17	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	238	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	238	19	name	name	NN
work_nu5xpentbrcxdlcdt3lmatbur4	238	20	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	238	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	238	22	genera-	genera-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	238	23	tor	tor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	238	24	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	238	25	proposes	propose	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	238	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	238	27	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	238	28	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	239	1	Recall	recall	VB
work_nu5xpentbrcxdlcdt3lmatbur4	239	2	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	239	3	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	239	4	candidates	candidate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	239	5	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	239	6	proposed	propose	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	239	7	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	239	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	239	9	dependency	dependency	NN
work_nu5xpentbrcxdlcdt3lmatbur4	239	10	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	239	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	239	12	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	239	13	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	239	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	239	15	heuristics	heuristic	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	239	16	described	describe	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	239	17	earlier	early	RBR
work_nu5xpentbrcxdlcdt3lmatbur4	239	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	240	1	Hence	hence	RB
work_nu5xpentbrcxdlcdt3lmatbur4	240	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	240	3	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	240	4	5The	5the	CD
work_nu5xpentbrcxdlcdt3lmatbur4	240	5	clusters	cluster	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	240	6	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	240	7	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	240	8	downloaded	download	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	240	9	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	240	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	240	11	authors	author	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	240	12	’	’	POS
work_nu5xpentbrcxdlcdt3lmatbur4	240	13	website	website	NN
work_nu5xpentbrcxdlcdt3lmatbur4	240	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	241	1	238	238	CD
work_nu5xpentbrcxdlcdt3lmatbur4	241	2	Jimmy	Jimmy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	3	Carter	Carter	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	4	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	5	Ronald	Ronald	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	6	Reagan	Reagan	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	7	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	8	richard	richard	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	9	nixon	nixon	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	10	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	11	George	George	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	12	Bush	Bush	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	13	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	14	Lyndon	Lyndon	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	15	Johnson	Johnson	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	16	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	17	Richard	Richard	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	18	M.	M.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	19	Nixon	Nixon	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	20	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	21	Gerald	Gerald	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	22	Ford	Ford	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	23	metalwork	metalwork	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	24	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	25	porcelain	porcelain	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	26	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	27	handicraft	handicraft	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	28	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	29	jade	jade	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	30	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	31	bronzeware	bronzeware	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	32	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	33	carving	carving	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	34	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	35	pottery	pottery	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	36	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	37	ceramic	ceramic	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	38	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	39	earthenware	earthenware	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	40	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	41	jewelry	jewelry	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	42	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	43	stoneware	stoneware	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	44	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	45	lacquerware	lacquerware	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	46	degradation	degradation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	47	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	48	erosion	erosion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	49	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	50	pollution	pollution	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	51	;	;	,
work_nu5xpentbrcxdlcdt3lmatbur4	241	52	logging	log	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	241	53	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	54	desertification	desertification	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	55	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	56	siltation	siltation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	57	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	58	urbanization	urbanization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	59	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	60	felling	felling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	61	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	62	poaching	poaching	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	63	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	64	soil	soil	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	65	erosion	erosion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	66	;	;	,
work_nu5xpentbrcxdlcdt3lmatbur4	241	67	depletion	depletion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	68	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	69	water	water	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	70	pollution	pollution	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	71	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	72	deforestation	deforestation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	73	expert	expert	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	74	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	75	Wall	Wall	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	76	Street	Street	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	77	analyst	analyst	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	78	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	79	analyst	analyst	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	80	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	81	economist	economist	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	82	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	83	telecommunications	telecommunications	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	84	analyst	analyst	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	85	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	86	strategist	strategist	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	87	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	88	media	medium	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	241	89	analyst	analyst	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	90	fox	fox	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	91	news	news	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	92	channel	channel	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	93	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	94	NBC	NBC	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	95	News	News	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	96	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	97	MSNBC	MSNBC	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	98	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	99	Fox	Fox	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	100	News	News	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	101	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	102	CNBC	CNBC	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	103	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	104	CNNfn	CNNfn	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	105	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	106	C	C	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	107	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	241	108	Span	Span	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	109	Tuesdays	Tuesdays	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	241	110	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	111	Wednesdays	wednesday	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	241	112	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	113	weekday	weekday	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	114	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	115	Mondays	Mondays	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	116	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	117	Fridays	friday	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	241	118	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	119	Thursdays	thursday	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	241	120	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	121	sundays	sunday	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	241	122	;	;	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	123	Saturdays	Saturdays	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	124	Table	Table	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	125	4	4	CD
work_nu5xpentbrcxdlcdt3lmatbur4	241	126	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	241	127	Examples	example	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	241	128	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	241	129	noun	noun	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	241	130	clusters	cluster	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	241	131	generated	generate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	241	132	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	241	133	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	241	134	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	135	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	241	136	covering	cover	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	241	137	approach	approach	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	138	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	241	139	σ	σ	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	140	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	241	141	0.15	0.15	CD
work_nu5xpentbrcxdlcdt3lmatbur4	241	142	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	241	143	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	144	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	241	145	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	241	146	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	241	147	features	feature	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	241	148	would	would	MD
work_nu5xpentbrcxdlcdt3lmatbur4	241	149	conjoin	conjoin	VB
work_nu5xpentbrcxdlcdt3lmatbur4	241	150	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	241	151	corresponding	correspond	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	241	152	typed	type	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	241	153	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	241	154	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	241	155	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	241	156	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	241	157	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	158	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	241	159	previous	previous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	241	160	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	241	161	verb	verb	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	241	162	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	241	163	previous	previous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	241	164	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	241	165	noun	noun	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	166	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	241	167	previous	previous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	241	168	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	241	169	adjective	adjective	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	170	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	241	171	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	241	172	previous	previous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	241	173	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	241	174	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	241	175	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	242	1	5.2	5.2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	242	2	Experimental	experimental	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	242	3	setup	setup	NN
work_nu5xpentbrcxdlcdt3lmatbur4	242	4	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	242	5	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	242	6	All	all	PDT
work_nu5xpentbrcxdlcdt3lmatbur4	242	7	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	242	8	experiments	experiment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	242	9	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	242	10	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	242	11	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	242	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	242	13	Sem-	Sem-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	242	14	Eval	eval	NN
work_nu5xpentbrcxdlcdt3lmatbur4	242	15	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	242	16	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	242	17	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	242	18	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	242	19	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	242	20	disambigua-	disambigua-	FW
work_nu5xpentbrcxdlcdt3lmatbur4	242	21	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	242	22	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	242	23	Litkowski	Litkowski	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	242	24	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	242	25	Hargraves	Hargraves	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	242	26	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	242	27	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	242	28	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	242	29	comprising	comprise	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	242	30	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	242	31	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	242	32	annotation	annotation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	242	33	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	242	34	16176	16176	CD
work_nu5xpentbrcxdlcdt3lmatbur4	242	35	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	242	36	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	242	37	8058	8058	CD
work_nu5xpentbrcxdlcdt3lmatbur4	242	38	examples	example	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	242	39	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	242	40	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	242	41	labeled	label	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	242	42	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	242	43	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	242	44	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	242	45	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	243	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	243	2	pre	pre	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	243	3	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	243	4	processed	process	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	243	5	sentences	sentence	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	243	6	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	243	7	part	part	NN
work_nu5xpentbrcxdlcdt3lmatbur4	243	8	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	243	9	of-	of-	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	243	10	speech	speech	NN
work_nu5xpentbrcxdlcdt3lmatbur4	243	11	tags	tag	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	243	12	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	243	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	243	14	Illinois	Illinois	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	243	15	POS	POS	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	243	16	tagger	tagger	NN
work_nu5xpentbrcxdlcdt3lmatbur4	243	17	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	243	18	de-	de-	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	243	19	pendency	pendency	NN
work_nu5xpentbrcxdlcdt3lmatbur4	243	20	graphs	graph	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	243	21	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	243	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	243	23	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	243	24	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	243	25	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	243	26	Goldberg	Goldberg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	243	27	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	243	28	Elhadad	Elhadad	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	243	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	243	30	2010)6	2010)6	CD
work_nu5xpentbrcxdlcdt3lmatbur4	243	31	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	244	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	244	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	244	3	experiments	experiment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	244	4	described	describe	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	244	5	be-	be-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	244	6	low	low	NN
work_nu5xpentbrcxdlcdt3lmatbur4	244	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	244	8	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	244	9	used	use	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	244	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	244	11	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	244	12	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	244	13	annotated	annotate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	244	14	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	244	15	set	set	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	244	16	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	244	17	train	train	VB
work_nu5xpentbrcxdlcdt3lmatbur4	244	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	244	19	models	model	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	244	20	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	244	21	evaluate	evaluate	VB
work_nu5xpentbrcxdlcdt3lmatbur4	244	22	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	244	23	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	244	24	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	244	25	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	244	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	244	27	test	test	NN
work_nu5xpentbrcxdlcdt3lmatbur4	244	28	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	244	29	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	245	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	245	2	chose	choose	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	245	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	245	4	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	245	5	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	245	6	parameter	parameter	NN
work_nu5xpentbrcxdlcdt3lmatbur4	245	7	C	C	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	245	8	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	245	9	five	five	CD
work_nu5xpentbrcxdlcdt3lmatbur4	245	10	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	245	11	fold	fold	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	245	12	cross	cross	NN
work_nu5xpentbrcxdlcdt3lmatbur4	245	13	-	-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	245	14	validation	validation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	245	15	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	245	16	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	245	17	1000	1000	CD
work_nu5xpentbrcxdlcdt3lmatbur4	245	18	random	random	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	245	19	exam-	exam-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	245	20	ples	ple	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	245	21	chosen	choose	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	245	22	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	245	23	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	245	24	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	245	25	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	245	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	246	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	246	2	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	246	3	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	246	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	246	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	246	6	picked	pick	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	246	7	α	α	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	246	8	=	=	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	246	9	0.1	0.1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	246	10	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	246	11	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	246	12	validation	validation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	246	13	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	246	14	consisting	consist	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	246	15	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	246	16	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	246	17	separate	separate	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	246	18	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	246	19	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	246	20	1000	1000	CD
work_nu5xpentbrcxdlcdt3lmatbur4	246	21	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	246	22	examples	example	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	246	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	247	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	247	2	ran	run	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	247	3	Algorithm	Algorithm	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	247	4	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	247	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	247	6	20	20	CD
work_nu5xpentbrcxdlcdt3lmatbur4	247	7	rounds	round	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	247	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	248	1	Predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	248	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	248	3	most	most	RBS
work_nu5xpentbrcxdlcdt3lmatbur4	248	4	frequent	frequent	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	248	5	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	248	6	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	248	7	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	248	8	prepo-	prepo-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	248	9	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	248	10	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	248	11	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	248	12	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	248	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	248	14	21.18	21.18	CD
work_nu5xpentbrcxdlcdt3lmatbur4	248	15	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	248	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	249	1	Even	even	RB
work_nu5xpentbrcxdlcdt3lmatbur4	249	2	though	though	IN
work_nu5xpentbrcxdlcdt3lmatbur4	249	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	249	4	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	249	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	249	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	249	7	most	most	RBS
work_nu5xpentbrcxdlcdt3lmatbur4	249	8	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	249	9	frequent	frequent	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	249	10	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	249	11	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	249	12	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	249	13	poor	poor	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	249	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	249	15	it	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	249	16	does	do	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	249	17	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	249	18	represent	represent	VB
work_nu5xpentbrcxdlcdt3lmatbur4	249	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	249	20	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	249	21	’s	’s	POS
work_nu5xpentbrcxdlcdt3lmatbur4	249	22	difficulty	difficulty	NN
work_nu5xpentbrcxdlcdt3lmatbur4	249	23	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	249	24	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	249	25	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	249	26	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	249	27	good	good	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	249	28	baseline	baseline	NN
work_nu5xpentbrcxdlcdt3lmatbur4	249	29	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	250	1	To	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	250	2	compare	compare	VB
work_nu5xpentbrcxdlcdt3lmatbur4	250	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	250	4	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	250	5	prepo-	prepo-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	6	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	7	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	250	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	250	9	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	250	10	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	250	11	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	250	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	250	13	neighboring	neighboring	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	14	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	250	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	250	16	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	250	17	Ye	Ye	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	250	18	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	250	19	Baldwin	Baldwin	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	250	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	250	21	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	250	22	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	250	23	obtained	obtain	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	250	24	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	250	25	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	26	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	250	27	69.3	69.3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	250	28	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	250	30	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	250	31	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	250	32	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	250	33	designed	design	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	250	34	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	250	35	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	250	36	prepo-	prepo-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	250	37	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	38	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	39	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	40	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	250	41	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	250	42	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	250	43	et	et	FW
work_nu5xpentbrcxdlcdt3lmatbur4	250	44	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	250	45	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	250	46	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	250	47	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	250	48	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	250	49	get	get	VB
work_nu5xpentbrcxdlcdt3lmatbur4	250	50	up	up	IN
work_nu5xpentbrcxdlcdt3lmatbur4	250	51	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	250	52	84.8	84.8	CD
work_nu5xpentbrcxdlcdt3lmatbur4	250	53	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	54	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	55	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	250	56	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	250	57	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	250	58	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	251	1	Our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	251	2	re	re	NN
work_nu5xpentbrcxdlcdt3lmatbur4	251	3	-	-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	251	4	implementation	implementation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	251	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	251	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	251	7	latter	latter	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	251	8	system	system	NN
work_nu5xpentbrcxdlcdt3lmatbur4	251	9	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	251	10	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	251	11	different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	251	12	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	251	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	251	14	pre	pre	NN
work_nu5xpentbrcxdlcdt3lmatbur4	251	15	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	251	16	processing	processing	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	251	17	tools	tool	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	251	18	gets	get	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	251	19	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	251	20	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	251	21	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	251	22	83.53	83.53	CD
work_nu5xpentbrcxdlcdt3lmatbur4	251	23	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	251	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	252	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	252	2	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	252	3	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	252	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	252	5	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	252	6	baseline	baseline	NN
work_nu5xpentbrcxdlcdt3lmatbur4	252	7	system	system	NN
work_nu5xpentbrcxdlcdt3lmatbur4	252	8	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	252	9	6We	6we	CD
work_nu5xpentbrcxdlcdt3lmatbur4	252	10	used	use	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	252	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	252	12	Curator	Curator	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	252	13	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	252	14	Clarke	Clarke	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	252	15	et	et	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	252	16	al	al	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	252	17	.	.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	252	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	252	19	2012	2012	CD
work_nu5xpentbrcxdlcdt3lmatbur4	252	20	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	252	21	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	252	22	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	252	23	pre-	pre-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	252	24	processing	processing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	252	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	253	1	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	253	2	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	253	3	uses	use	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	253	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	253	5	typed	type	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	253	6	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	253	7	set	set	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	253	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	253	9	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	253	10	with-	with-	VB
work_nu5xpentbrcxdlcdt3lmatbur4	253	11	out	out	RP
work_nu5xpentbrcxdlcdt3lmatbur4	253	12	any	any	DT
work_nu5xpentbrcxdlcdt3lmatbur4	253	13	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	253	14	information	information	NN
work_nu5xpentbrcxdlcdt3lmatbur4	253	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	254	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	254	2	produces	produce	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	254	3	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	254	4	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	254	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	254	6	88.01	88.01	CD
work_nu5xpentbrcxdlcdt3lmatbur4	254	7	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	254	8	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	254	9	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	254	10	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	254	11	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	254	12	88.64	88.64	CD
work_nu5xpentbrcxdlcdt3lmatbur4	254	13	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	254	14	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	254	15	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	254	16	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	254	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	255	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	255	2	report	report	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	255	3	statistical	statistical	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	255	4	significance	significance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	255	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	255	6	results	result	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	255	7	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	255	8	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	255	9	implementation	implementation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	255	10	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	255	11	Dan	Dan	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	255	12	Bikel	Bikel	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	255	13	’s	’s	POS
work_nu5xpentbrcxdlcdt3lmatbur4	255	14	stratified	stratify	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	255	15	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	255	16	shuffling	shuffling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	255	17	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	255	18	statistical	statistical	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	255	19	significance	significance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	255	20	tester7	tester7	NN
work_nu5xpentbrcxdlcdt3lmatbur4	255	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	256	1	5.3	5.3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	256	2	Main	main	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	256	3	results	result	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	256	4	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	256	5	Relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	256	6	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	256	7	Our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	256	8	main	main	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	256	9	result	result	NN
work_nu5xpentbrcxdlcdt3lmatbur4	256	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	256	11	presented	present	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	256	12	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	256	13	Table	Table	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	256	14	5	5	CD
work_nu5xpentbrcxdlcdt3lmatbur4	256	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	256	16	compares	compare	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	256	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	256	18	baseline	baseline	NN
work_nu5xpentbrcxdlcdt3lmatbur4	256	19	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	256	20	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	256	21	without	without	IN
work_nu5xpentbrcxdlcdt3lmatbur4	256	22	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	256	23	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	256	24	against	against	IN
work_nu5xpentbrcxdlcdt3lmatbur4	256	25	other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	256	26	sys-	sys-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	256	27	tems	tem	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	256	28	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	256	29	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	256	30	both	both	DT
work_nu5xpentbrcxdlcdt3lmatbur4	256	31	models	model	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	256	32	described	describe	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	256	33	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	256	34	Section	section	NN
work_nu5xpentbrcxdlcdt3lmatbur4	256	35	4	4	CD
work_nu5xpentbrcxdlcdt3lmatbur4	256	36	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	257	1	First	first	RB
work_nu5xpentbrcxdlcdt3lmatbur4	257	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	257	3	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	257	4	see	see	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	257	5	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	257	6	adding	add	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	257	7	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	257	8	information	information	NN
work_nu5xpentbrcxdlcdt3lmatbur4	257	9	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	257	10	typed	type	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	257	11	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	257	12	improves	improve	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	257	13	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	257	14	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	257	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	257	16	baseline	baseline	NN
work_nu5xpentbrcxdlcdt3lmatbur4	257	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	258	1	Expand-	Expand-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	258	2	ing	e	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	258	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	258	4	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	258	5	space	space	NN
work_nu5xpentbrcxdlcdt3lmatbur4	258	6	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	258	7	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	258	8	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	258	9	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	258	10	further	further	RB
work_nu5xpentbrcxdlcdt3lmatbur4	258	11	im-	im-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	258	12	provements	provement	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	258	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	259	1	Finally	finally	RB
work_nu5xpentbrcxdlcdt3lmatbur4	259	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	259	3	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	259	4	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	259	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	259	6	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	259	7	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	259	8	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	259	9	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	259	10	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	259	11	another	another	DT
work_nu5xpentbrcxdlcdt3lmatbur4	259	12	improvement	improvement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	259	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	260	1	Setting	set	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	260	2	Accuracy	Accuracy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	260	3	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	260	4	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	5	Model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	260	6	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	7	No	no	DT
work_nu5xpentbrcxdlcdt3lmatbur4	260	8	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	260	9	88.01	88.01	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	10	88.64	88.64	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	11	typed	type	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	260	12	88.77	88.77	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	13	89.14	89.14	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	14	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	260	15	89.90∗	89.90∗	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	16	89.43∗	89.43∗	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	17	Joint	Joint	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	260	18	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	260	19	&	&	CC
work_nu5xpentbrcxdlcdt3lmatbur4	260	20	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	260	21	89.99∗	89.99∗	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	260	22	90.26∗†	90.26∗†	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	23	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	260	24	5	5	CD
work_nu5xpentbrcxdlcdt3lmatbur4	260	25	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	260	26	Main	main	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	260	27	results	result	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	260	28	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	260	29	Accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	260	30	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	260	31	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	260	32	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	260	33	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	261	1	Results	result	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	261	2	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	261	3	bold	bold	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	261	4	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	261	5	statistically	statistically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	261	6	significant	significant	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	261	7	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	261	8	p	p	NN
work_nu5xpentbrcxdlcdt3lmatbur4	261	9	<	<	XX
work_nu5xpentbrcxdlcdt3lmatbur4	261	10	0.01	0.01	XX
work_nu5xpentbrcxdlcdt3lmatbur4	261	11	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	261	12	improvements	improvement	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	261	13	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	261	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	261	15	system	system	NN
work_nu5xpentbrcxdlcdt3lmatbur4	261	16	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	261	17	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	261	18	unaware	unaware	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	261	19	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	261	20	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	261	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	262	1	Superscripts	Superscripts	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	262	2	∗	∗	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	262	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	262	4	†	†	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	262	5	indicate	indicate	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	262	6	significant	significant	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	262	7	improvements	improvement	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	262	8	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	262	9	typed	typed	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	262	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	262	11	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	262	12	respectively	respectively	RB
work_nu5xpentbrcxdlcdt3lmatbur4	262	13	at	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	262	14	p	p	NN
work_nu5xpentbrcxdlcdt3lmatbur4	262	15	<	<	XX
work_nu5xpentbrcxdlcdt3lmatbur4	262	16	0.01	0.01	XX
work_nu5xpentbrcxdlcdt3lmatbur4	262	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	263	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	263	2	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	263	3	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	263	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	263	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	263	6	improvement	improvement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	263	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	263	8	typed	type	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	263	9	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	263	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	263	11	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	263	12	with-	with-	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	263	13	out	out	RP
work_nu5xpentbrcxdlcdt3lmatbur4	263	14	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	263	15	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	263	16	significant	significant	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	263	17	at	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	263	18	p	p	NN
work_nu5xpentbrcxdlcdt3lmatbur4	263	19	<	<	XX
work_nu5xpentbrcxdlcdt3lmatbur4	263	20	0.05	0.05	CD
work_nu5xpentbrcxdlcdt3lmatbur4	263	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	264	1	Our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	264	2	objective	objective	NN
work_nu5xpentbrcxdlcdt3lmatbur4	264	3	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	264	4	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	264	5	predicting	predict	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	264	6	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	264	7	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	264	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	265	1	However	however	RB
work_nu5xpentbrcxdlcdt3lmatbur4	265	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	265	3	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	265	4	observe	observe	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	265	5	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	265	6	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	265	7	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	265	8	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	265	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	265	10	jointly	jointly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	265	11	pre-	pre-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	265	12	dicting	dicte	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	265	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	265	14	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	265	15	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	265	16	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	265	17	improves	improve	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	265	18	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	265	19	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	265	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	265	21	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	265	22	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	265	23	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	265	24	identification	identification	NN
work_nu5xpentbrcxdlcdt3lmatbur4	265	25	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	265	26	but	but	CC
work_nu5xpentbrcxdlcdt3lmatbur4	265	27	via	via	IN
work_nu5xpentbrcxdlcdt3lmatbur4	265	28	joint	joint	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	265	29	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	265	30	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	265	31	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	265	32	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	265	33	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	265	34	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	265	35	leads	lead	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	265	36	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	265	37	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	265	38	large	large	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	265	39	improvement	improvement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	265	40	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	265	41	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	265	42	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	265	43	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	265	44	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	266	1	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	266	2	6	6	CD
work_nu5xpentbrcxdlcdt3lmatbur4	266	3	shows	show	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	266	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	266	5	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	266	6	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	266	7	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	266	8	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	266	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	267	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	267	2	7http://www.cis.upenn.edu/∼dbikel/software.html	7http://www.cis.upenn.edu/∼dbikel/software.html	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	267	3	239	239	CD
work_nu5xpentbrcxdlcdt3lmatbur4	267	4	see	see	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	267	5	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	267	6	while	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	267	7	Model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	267	8	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	267	9	does	do	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	267	10	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	267	11	lead	lead	VB
work_nu5xpentbrcxdlcdt3lmatbur4	267	12	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	267	13	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	267	14	significant	significant	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	267	15	improvement	improvement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	267	16	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	267	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	267	18	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	267	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	267	20	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	267	21	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	267	22	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	267	23	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	267	24	ab-	ab-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	267	25	solute	solute	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	267	26	improvement	improvement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	267	27	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	267	28	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	267	29	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	267	30	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	267	31	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	268	1	Setting	set	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	268	2	Sense	Sense	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	268	3	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	268	4	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	268	5	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	268	6	re	re	NN
work_nu5xpentbrcxdlcdt3lmatbur4	268	7	-	-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	268	8	implementation	implementation	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	268	9	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	268	10	83.53	83.53	CD
work_nu5xpentbrcxdlcdt3lmatbur4	268	11	Joint	Joint	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	268	12	+	+	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	268	13	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	268	14	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	268	15	83.78	83.78	CD
work_nu5xpentbrcxdlcdt3lmatbur4	268	16	Joint	Joint	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	268	17	+	+	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	268	18	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	268	19	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	268	20	84.78∗	84.78∗	CD
work_nu5xpentbrcxdlcdt3lmatbur4	268	21	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	268	22	6	6	CD
work_nu5xpentbrcxdlcdt3lmatbur4	268	23	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	268	24	Sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	268	25	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	268	26	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	268	27	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	269	1	Joint	joint	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	269	2	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	269	3	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	269	4	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	269	5	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	269	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	269	7	while	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	269	8	improving	improve	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	269	9	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	269	10	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	269	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	269	12	does	do	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	269	13	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	269	14	help	help	VB
work_nu5xpentbrcxdlcdt3lmatbur4	269	15	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	269	16	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	269	17	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	269	18	comparison	comparison	NN
work_nu5xpentbrcxdlcdt3lmatbur4	269	19	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	269	20	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	269	21	re-	re-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	269	22	implementation	implementation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	269	23	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	269	24	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	269	25	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	269	26	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	269	27	disambiguation	disambiguation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	269	28	sys-	sys-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	269	29	tem	tem	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	269	30	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	270	1	However	however	RB
work_nu5xpentbrcxdlcdt3lmatbur4	270	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	270	3	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	270	4	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	270	5	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	270	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	270	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	270	8	improvement	improvement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	270	9	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	270	10	statis-	statis-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	270	11	tically	tically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	270	12	significant	significant	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	270	13	at	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	270	14	p	p	NN
work_nu5xpentbrcxdlcdt3lmatbur4	270	15	<	<	XX
work_nu5xpentbrcxdlcdt3lmatbur4	270	16	0.01	0.01	XX
work_nu5xpentbrcxdlcdt3lmatbur4	270	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	271	1	5.4	5.4	CD
work_nu5xpentbrcxdlcdt3lmatbur4	271	2	Ablation	ablation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	271	3	experiments	experiment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	271	4	Feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	271	5	sharing	share	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	271	6	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	271	7	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	271	8	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	271	9	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	271	10	first	first	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	271	11	analysis	analysis	NN
work_nu5xpentbrcxdlcdt3lmatbur4	271	12	experiment	experiment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	271	13	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	271	14	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	271	15	seek	seek	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	271	16	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	271	17	highlight	highlight	VB
work_nu5xpentbrcxdlcdt3lmatbur4	271	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	271	19	utility	utility	NN
work_nu5xpentbrcxdlcdt3lmatbur4	271	20	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	271	21	sharing	share	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	271	22	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	271	23	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	271	24	different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	271	25	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	271	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	272	1	To	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	272	2	do	do	VB
work_nu5xpentbrcxdlcdt3lmatbur4	272	3	so	so	RB
work_nu5xpentbrcxdlcdt3lmatbur4	272	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	272	5	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	272	6	compare	compare	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	272	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	272	8	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	272	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	272	10	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	272	11	sys-	sys-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	272	12	tem	tem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	272	13	trained	train	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	272	14	without	without	IN
work_nu5xpentbrcxdlcdt3lmatbur4	272	15	shared	shared	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	272	16	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	272	17	against	against	IN
work_nu5xpentbrcxdlcdt3lmatbur4	272	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	272	19	type-	type-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	272	20	independent	independent	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	272	21	system	system	NN
work_nu5xpentbrcxdlcdt3lmatbur4	272	22	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	272	23	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	272	24	uses	use	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	272	25	shared	shared	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	272	26	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	272	27	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	273	1	To	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	273	2	discount	discount	VB
work_nu5xpentbrcxdlcdt3lmatbur4	273	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	273	4	influence	influence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	273	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	273	6	other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	273	7	factors	factor	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	273	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	273	9	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	273	10	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	273	11	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	273	12	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	273	13	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	273	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	273	15	typed	type	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	273	16	setting	set	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	273	17	without	without	IN
work_nu5xpentbrcxdlcdt3lmatbur4	273	18	any	any	DT
work_nu5xpentbrcxdlcdt3lmatbur4	273	19	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	273	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	274	1	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	274	2	7	7	CD
work_nu5xpentbrcxdlcdt3lmatbur4	274	3	reports	report	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	274	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	274	5	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	274	6	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	274	7	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	274	8	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	274	9	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	274	10	these	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	274	11	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	274	12	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	274	13	sets	set	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	274	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	275	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	275	2	observed	observe	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	275	3	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	275	4	similar	similar	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	275	5	improve-	improve-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	275	6	ment	ment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	275	7	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	275	8	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	275	9	even	even	RB
work_nu5xpentbrcxdlcdt3lmatbur4	275	10	when	when	WRB
work_nu5xpentbrcxdlcdt3lmatbur4	275	11	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	275	12	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	275	13	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	275	14	added	add	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	275	15	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	275	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	275	17	setting	setting	NN
work_nu5xpentbrcxdlcdt3lmatbur4	275	18	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	275	19	changed	change	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	275	20	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	275	21	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	275	22	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	275	23	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	275	24	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	275	25	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	275	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	276	1	Setting	set	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	276	2	Accuracy	Accuracy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	276	3	Independent	Independent	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	276	4	87.17	87.17	CD
work_nu5xpentbrcxdlcdt3lmatbur4	276	5	+	+	CD
work_nu5xpentbrcxdlcdt3lmatbur4	276	6	Shared	share	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	276	7	88.01	88.01	CD
work_nu5xpentbrcxdlcdt3lmatbur4	276	8	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	276	9	7	7	CD
work_nu5xpentbrcxdlcdt3lmatbur4	276	10	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	276	11	Comparing	compare	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	276	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	276	13	effect	effect	NN
work_nu5xpentbrcxdlcdt3lmatbur4	276	14	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	276	15	feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	276	16	sharing	share	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	276	17	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	276	18	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	276	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	277	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	277	2	see	see	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	277	3	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	277	4	having	have	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	277	5	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	277	6	shared	share	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	277	7	representation	representation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	277	8	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	277	9	goes	go	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	277	10	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	277	11	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	277	12	improves	improve	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	277	13	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	277	14	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	277	15	rela-	rela-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	277	16	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	277	17	prediction	prediction	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	277	18	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	277	19	p	p	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	277	20	<	<	XX
work_nu5xpentbrcxdlcdt3lmatbur4	277	21	0.01	0.01	XX
work_nu5xpentbrcxdlcdt3lmatbur4	277	22	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	277	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	278	1	Different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	278	2	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	278	3	candidate	candidate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	278	4	generators	generator	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	278	5	Our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	278	6	second	second	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	278	7	ablation	ablation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	278	8	study	study	NN
work_nu5xpentbrcxdlcdt3lmatbur4	278	9	looks	look	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	278	10	at	at	IN
work_nu5xpentbrcxdlcdt3lmatbur4	278	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	278	12	effect	effect	NN
work_nu5xpentbrcxdlcdt3lmatbur4	278	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	278	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	278	15	var-	var-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	278	16	ious	ious	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	278	17	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	278	18	candidate	candidate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	278	19	generators	generator	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	278	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	279	1	Recall	recall	VB
work_nu5xpentbrcxdlcdt3lmatbur4	279	2	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	279	3	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	279	4	addition	addition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	5	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	279	6	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	279	7	dependency	dependency	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	8	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	279	10	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	279	12	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	279	13	models	model	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	279	14	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	279	15	use	use	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	279	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	279	17	previous	previous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	279	18	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	279	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	279	21	previous	previous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	279	22	noun	noun	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	23	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	279	24	adjective	adjective	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	279	25	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	279	26	verb	verb	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	279	27	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	279	28	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	29	candidates	candidate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	279	30	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	279	31	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	279	32	next	next	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	279	33	noun	noun	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	34	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	279	35	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	36	candidate	candidate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	279	37	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	280	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	280	2	refer	refer	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	280	3	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	280	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	280	5	candidates	candidate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	280	6	generated	generate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	280	7	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	280	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	280	9	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	280	10	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	280	11	Parser	Parser	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	280	12	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	280	13	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	280	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	280	15	others	other	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	280	16	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	280	17	Heuristics	Heuristics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	280	18	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	280	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	281	1	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	281	2	8	8	CD
work_nu5xpentbrcxdlcdt3lmatbur4	281	3	compares	compare	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	281	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	281	5	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	281	6	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	281	7	these	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	281	8	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	281	9	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	281	10	candidate	candidate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	281	11	generators	generator	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	281	12	against	against	IN
work_nu5xpentbrcxdlcdt3lmatbur4	281	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	281	14	full	full	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	281	15	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	281	16	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	281	17	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	281	18	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	281	19	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	281	20	both	both	CC
work_nu5xpentbrcxdlcdt3lmatbur4	281	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	281	22	typed	typed	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	281	23	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	281	24	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	281	25	settings	setting	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	281	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	282	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	282	2	see	see	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	282	3	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	282	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	282	5	heuristics	heuristic	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	282	6	give	give	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	282	7	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	282	8	better	well	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	282	9	accu-	accu-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	282	10	racy	racy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	282	11	than	than	IN
work_nu5xpentbrcxdlcdt3lmatbur4	282	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	282	13	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	282	14	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	282	15	system	system	NN
work_nu5xpentbrcxdlcdt3lmatbur4	282	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	283	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	283	2	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	283	3	because	because	IN
work_nu5xpentbrcxdlcdt3lmatbur4	283	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	283	5	heuristics	heuristic	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	283	6	often	often	RB
work_nu5xpentbrcxdlcdt3lmatbur4	283	7	contain	contain	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	283	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	283	9	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	283	10	/	/	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	283	11	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	283	12	pre-	pre-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	283	13	dicted	dicte	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	283	14	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	283	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	283	16	dependency	dependency	NN
work_nu5xpentbrcxdlcdt3lmatbur4	283	17	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	284	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	284	2	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	284	3	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	284	4	always	always	RB
work_nu5xpentbrcxdlcdt3lmatbur4	284	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	284	6	case	case	NN
work_nu5xpentbrcxdlcdt3lmatbur4	284	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	284	8	though	though	RB
work_nu5xpentbrcxdlcdt3lmatbur4	284	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	284	10	because	because	IN
work_nu5xpentbrcxdlcdt3lmatbur4	284	11	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	284	12	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	284	13	generators	generator	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	284	14	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	284	15	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	284	16	slightly	slightly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	284	17	better	well	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	284	18	performing	perform	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	284	19	system	system	NN
work_nu5xpentbrcxdlcdt3lmatbur4	284	20	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	284	21	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	284	22	statistically	statistically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	284	23	significant	significant	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	284	24	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	284	25	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	285	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	285	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	285	3	overall	overall	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	285	4	system	system	NN
work_nu5xpentbrcxdlcdt3lmatbur4	285	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	285	6	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	285	7	retain	retain	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	285	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	285	9	de-	de-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	285	10	pendency	pendency	NN
work_nu5xpentbrcxdlcdt3lmatbur4	285	11	parser	parser	NN
work_nu5xpentbrcxdlcdt3lmatbur4	285	12	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	285	13	one	one	CD
work_nu5xpentbrcxdlcdt3lmatbur4	285	14	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	285	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	285	16	generators	generator	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	285	17	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	285	18	order	order	NN
work_nu5xpentbrcxdlcdt3lmatbur4	285	19	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	285	20	capture	capture	VB
work_nu5xpentbrcxdlcdt3lmatbur4	285	21	long	long	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	285	22	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	285	23	range	range	NN
work_nu5xpentbrcxdlcdt3lmatbur4	285	24	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	285	25	/	/	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	285	26	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	285	27	candidates	candidate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	285	28	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	285	29	may	may	MD
work_nu5xpentbrcxdlcdt3lmatbur4	285	30	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	285	31	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	285	32	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	285	33	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	285	34	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	285	35	selected	select	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	285	36	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	285	37	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	285	38	heuristics	heuristics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	285	39	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	286	1	Feature	feature	NN
work_nu5xpentbrcxdlcdt3lmatbur4	286	2	sets	set	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	286	3	Generator	generator	NN
work_nu5xpentbrcxdlcdt3lmatbur4	286	4	typed	type	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	286	5	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	286	6	Parser	Parser	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	286	7	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	286	8	87.12	87.12	CD
work_nu5xpentbrcxdlcdt3lmatbur4	286	9	87.12	87.12	CD
work_nu5xpentbrcxdlcdt3lmatbur4	286	10	Heuristics	Heuristics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	286	11	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	286	12	87.63	87.63	CD
work_nu5xpentbrcxdlcdt3lmatbur4	286	13	88.84	88.84	CD
work_nu5xpentbrcxdlcdt3lmatbur4	286	14	All	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	286	15	88.01	88.01	CD
work_nu5xpentbrcxdlcdt3lmatbur4	286	16	89.12	89.12	CD
work_nu5xpentbrcxdlcdt3lmatbur4	286	17	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	286	18	8	8	CD
work_nu5xpentbrcxdlcdt3lmatbur4	286	19	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	286	20	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	286	21	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	286	22	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	286	23	different	different	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	286	24	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	286	25	candi-	candi-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	286	26	date	date	NN
work_nu5xpentbrcxdlcdt3lmatbur4	286	27	generators	generator	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	286	28	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	287	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	287	2	see	see	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	287	3	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	287	4	considering	consider	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	287	5	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	287	6	larger	large	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	287	7	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	287	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	287	9	candidate	candidate	NN
work_nu5xpentbrcxdlcdt3lmatbur4	287	10	generators	generator	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	287	11	gives	give	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	287	12	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	287	13	big	big	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	287	14	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	287	15	improvement	improvement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	287	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	288	1	6	6	CD
work_nu5xpentbrcxdlcdt3lmatbur4	288	2	Discussion	Discussion	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	288	3	There	there	EX
work_nu5xpentbrcxdlcdt3lmatbur4	288	4	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	288	5	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	288	6	key	key	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	288	7	differences	difference	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	288	8	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	288	9	Model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	288	10	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	288	11	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	288	12	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	288	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	289	1	First	first	RB
work_nu5xpentbrcxdlcdt3lmatbur4	289	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	289	3	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	289	4	former	former	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	289	5	predicts	predict	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	289	6	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	289	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	289	8	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	289	9	label	label	NN
work_nu5xpentbrcxdlcdt3lmatbur4	289	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	289	11	while	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	289	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	289	13	latter	latter	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	289	14	predicts	predict	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	289	15	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	289	16	entire	entire	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	289	17	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	289	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	290	1	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	290	2	9	9	CD
work_nu5xpentbrcxdlcdt3lmatbur4	290	3	shows	show	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	290	4	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	290	5	predictions	prediction	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	290	6	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	290	7	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	290	8	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	290	9	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	290	10	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	290	11	label	label	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	290	12	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	290	13	WordNet	WordNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	290	14	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	290	15	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	290	16	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	291	1	These	these	DT
work_nu5xpentbrcxdlcdt3lmatbur4	291	2	examples	example	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	291	3	show	show	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	291	4	how	how	WRB
work_nu5xpentbrcxdlcdt3lmatbur4	291	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	291	6	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	291	7	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	291	8	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	291	9	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	291	10	thought	think	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	291	11	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	291	12	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	291	13	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	291	14	explanation	explanation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	291	15	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	291	16	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	291	17	choice	choice	NN
work_nu5xpentbrcxdlcdt3lmatbur4	291	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	291	19	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	291	20	label	label	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	291	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	292	1	Input	Input	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	292	2	Relation	Relation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	292	3	Hypernyms	Hypernyms	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	292	4	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	5	object	object	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	6	died	die	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	292	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	292	8	pneumonia	pneumonia	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	9	Cause	cause	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	10	experience	experience	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	11	disease	disease	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	12	suffered	suffer	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	292	13	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	292	14	flu	flu	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	15	Cause	Cause	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	292	16	experience	experience	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	17	disease	disease	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	18	recovered	recover	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	292	19	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	292	20	flu	flu	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	292	21	StartState	StartState	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	292	22	change	change	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	23	disease	disease	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	24	Table	table	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	25	9	9	CD
work_nu5xpentbrcxdlcdt3lmatbur4	292	26	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	292	27	Example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	292	28	predictions	prediction	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	292	29	according	accord	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	292	30	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	292	31	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	292	32	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	292	33	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	293	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	293	2	hypernyms	hypernyms	NN
work_nu5xpentbrcxdlcdt3lmatbur4	293	3	column	column	NN
work_nu5xpentbrcxdlcdt3lmatbur4	293	4	shows	show	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	293	5	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	293	6	representative	representative	NN
work_nu5xpentbrcxdlcdt3lmatbur4	293	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	293	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	293	9	synset	synset	NN
work_nu5xpentbrcxdlcdt3lmatbur4	293	10	chosen	choose	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	293	11	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	293	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	293	13	WordNet	WordNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	293	14	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	293	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	294	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	294	2	see	see	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	294	3	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	294	4	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	294	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	294	6	com-	com-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	294	7	bination	bination	NN
work_nu5xpentbrcxdlcdt3lmatbur4	294	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	294	9	experience	experience	NN
work_nu5xpentbrcxdlcdt3lmatbur4	294	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	294	11	disease	disease	NN
work_nu5xpentbrcxdlcdt3lmatbur4	294	12	suggests	suggest	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	294	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	294	14	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	294	15	Cause	Cause	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	294	16	while	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	294	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	294	18	change	change	NN
work_nu5xpentbrcxdlcdt3lmatbur4	294	19	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	294	20	disease	disease	NN
work_nu5xpentbrcxdlcdt3lmatbur4	294	21	indicate	indicate	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	294	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	294	23	rela-	rela-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	294	24	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	294	25	StartState	StartState	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	294	26	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	295	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	295	2	main	main	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	295	3	difference	difference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	295	4	between	between	IN
work_nu5xpentbrcxdlcdt3lmatbur4	295	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	295	6	two	two	CD
work_nu5xpentbrcxdlcdt3lmatbur4	295	7	models	model	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	295	8	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	295	9	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	295	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	295	11	treatment	treatment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	295	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	295	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	295	14	unlabeled	unlabeled	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	295	15	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	295	16	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	295	17	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	295	18	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	295	19	parts	part	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	295	20	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	295	21	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	295	22	structure	structure	NN
work_nu5xpentbrcxdlcdt3lmatbur4	295	23	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	295	24	namely	namely	RB
work_nu5xpentbrcxdlcdt3lmatbur4	295	25	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	295	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	295	27	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	295	28	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	295	29	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	295	30	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	295	31	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	295	32	during	during	IN
work_nu5xpentbrcxdlcdt3lmatbur4	295	33	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	295	34	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	295	35	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	295	36	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	296	1	During	during	IN
work_nu5xpentbrcxdlcdt3lmatbur4	296	2	training	training	NN
work_nu5xpentbrcxdlcdt3lmatbur4	296	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	296	4	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	296	5	240	240	CD
work_nu5xpentbrcxdlcdt3lmatbur4	296	6	each	each	DT
work_nu5xpentbrcxdlcdt3lmatbur4	296	7	example	example	NN
work_nu5xpentbrcxdlcdt3lmatbur4	296	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	296	9	Model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	296	10	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	296	11	aggregates	aggregate	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	296	12	features	feature	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	296	13	from	from	IN
work_nu5xpentbrcxdlcdt3lmatbur4	296	14	all	all	DT
work_nu5xpentbrcxdlcdt3lmatbur4	296	15	governors	governor	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	296	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	296	17	objects	object	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	296	18	even	even	RB
work_nu5xpentbrcxdlcdt3lmatbur4	296	19	if	if	IN
work_nu5xpentbrcxdlcdt3lmatbur4	296	20	they	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	296	21	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	296	22	possibly	possibly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	296	23	ir-	ir-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	296	24	relevant	relevant	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	296	25	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	296	26	which	which	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	296	27	may	may	MD
work_nu5xpentbrcxdlcdt3lmatbur4	296	28	lead	lead	VB
work_nu5xpentbrcxdlcdt3lmatbur4	296	29	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	296	30	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	296	31	much	much	RB
work_nu5xpentbrcxdlcdt3lmatbur4	296	32	bigger	big	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	296	33	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	296	34	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	296	35	terms	term	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	296	36	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	296	37	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	296	38	number	number	NN
work_nu5xpentbrcxdlcdt3lmatbur4	296	39	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	296	40	active	active	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	296	41	weights	weight	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	296	42	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	297	1	On	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	297	2	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	297	3	other	other	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	297	4	hand	hand	NN
work_nu5xpentbrcxdlcdt3lmatbur4	297	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	297	6	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	297	7	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	297	8	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	297	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	297	10	Algorithm	Algorithm	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	297	11	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	297	12	uses	use	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	297	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	297	14	single	single	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	297	15	high-	high-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	297	16	est	est	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	297	17	scoring	score	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	297	18	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	297	19	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	297	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	297	21	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	297	22	variables	variable	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	297	23	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	297	24	accord-	accord-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	297	25	ing	e	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	297	26	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	297	27	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	297	28	current	current	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	297	29	parameters	parameter	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	297	30	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	297	31	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	297	32	refine	refine	VB
work_nu5xpentbrcxdlcdt3lmatbur4	297	33	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	297	34	parame-	parame-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	297	35	ters	ter	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	297	36	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	298	1	Indeed	indeed	RB
work_nu5xpentbrcxdlcdt3lmatbur4	298	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	298	3	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	298	4	our	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	298	5	experiments	experiment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	298	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	298	7	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	298	8	observed	observe	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	298	9	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	298	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	298	11	number	number	NN
work_nu5xpentbrcxdlcdt3lmatbur4	298	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	298	13	non	non	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	298	14	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	298	15	zero	zero	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	298	16	weights	weight	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	298	17	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	298	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	298	19	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	298	20	vec-	vec-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	298	21	tor	tor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	298	22	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	298	23	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	298	24	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	298	25	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	298	26	much	much	RB
work_nu5xpentbrcxdlcdt3lmatbur4	298	27	smaller	small	JJR
work_nu5xpentbrcxdlcdt3lmatbur4	298	28	than	than	IN
work_nu5xpentbrcxdlcdt3lmatbur4	298	29	that	that	DT
work_nu5xpentbrcxdlcdt3lmatbur4	298	30	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	298	31	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	298	32	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	298	33	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	299	1	For	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	299	2	instance	instance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	299	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	299	4	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	299	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	299	6	typed	type	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	299	7	setting	setting	NN
work_nu5xpentbrcxdlcdt3lmatbur4	299	8	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	299	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	299	10	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	299	11	vec-	vec-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	299	12	tor	tor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	299	13	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	299	14	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	299	15	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	299	16	had	have	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	299	17	2.57	2.57	CD
work_nu5xpentbrcxdlcdt3lmatbur4	299	18	million	million	CD
work_nu5xpentbrcxdlcdt3lmatbur4	299	19	elements	element	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	299	20	while	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	299	21	that	that	DT
work_nu5xpentbrcxdlcdt3lmatbur4	299	22	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	299	23	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	299	24	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	299	25	had	have	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	299	26	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	299	27	1.0	1.0	CD
work_nu5xpentbrcxdlcdt3lmatbur4	299	28	million	million	CD
work_nu5xpentbrcxdlcdt3lmatbur4	299	29	weights	weight	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	299	30	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	300	1	Similarly	similarly	RB
work_nu5xpentbrcxdlcdt3lmatbur4	300	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	300	3	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	300	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	300	5	typed+gen	typed+gen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	300	6	setting	set	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	300	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	300	8	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	300	9	1	1	CD
work_nu5xpentbrcxdlcdt3lmatbur4	300	10	had	have	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	300	11	5.41	5.41	CD
work_nu5xpentbrcxdlcdt3lmatbur4	300	12	million	million	CD
work_nu5xpentbrcxdlcdt3lmatbur4	300	13	non	non	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	300	14	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	300	15	zero	zero	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	300	16	elements	element	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	300	17	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	300	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	300	19	weight	weight	NN
work_nu5xpentbrcxdlcdt3lmatbur4	300	20	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	300	21	while	while	IN
work_nu5xpentbrcxdlcdt3lmatbur4	300	22	Model	Model	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	300	23	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	300	24	had	have	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	300	25	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	300	26	2.21	2.21	CD
work_nu5xpentbrcxdlcdt3lmatbur4	300	27	million	million	CD
work_nu5xpentbrcxdlcdt3lmatbur4	300	28	non	non	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	300	29	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	300	30	zero	zero	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	300	31	elements	element	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	300	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	301	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	301	2	learning	learning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	301	3	algorithm	algorithm	NN
work_nu5xpentbrcxdlcdt3lmatbur4	301	4	itself	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	301	5	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	301	6	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	301	7	generalization	generalization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	301	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	301	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	301	10	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	301	11	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	301	12	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	301	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	301	14	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	301	15	Yu	Yu	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	301	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	301	17	Joachims	Joachims	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	301	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	301	19	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	301	20	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	301	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	302	1	By	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	302	2	setting	set	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	302	3	α	α	XX
work_nu5xpentbrcxdlcdt3lmatbur4	302	4	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	302	5	zero	zero	CD
work_nu5xpentbrcxdlcdt3lmatbur4	302	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	302	7	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	302	8	get	get	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	302	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	302	10	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	302	11	struc-	struc-	FW
work_nu5xpentbrcxdlcdt3lmatbur4	302	12	ture	ture	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	302	13	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	302	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	303	1	However	however	RB
work_nu5xpentbrcxdlcdt3lmatbur4	303	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	303	3	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	303	4	found	find	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	303	5	via	via	IN
work_nu5xpentbrcxdlcdt3lmatbur4	303	6	cross	cross	NN
work_nu5xpentbrcxdlcdt3lmatbur4	303	7	-	-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	303	8	validation	validation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	303	9	that	that	IN
work_nu5xpentbrcxdlcdt3lmatbur4	303	10	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	303	11	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	303	12	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	303	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	303	14	best	good	JJS
work_nu5xpentbrcxdlcdt3lmatbur4	303	15	setting	setting	NN
work_nu5xpentbrcxdlcdt3lmatbur4	303	16	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	303	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	303	18	parameter	parameter	NN
work_nu5xpentbrcxdlcdt3lmatbur4	303	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	304	1	A	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	304	2	theoretical	theoretical	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	304	3	understanding	understanding	NN
work_nu5xpentbrcxdlcdt3lmatbur4	304	4	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	304	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	304	6	sparsity	sparsity	NN
work_nu5xpentbrcxdlcdt3lmatbur4	304	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	304	8	weights	weight	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	304	9	learned	learn	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	304	10	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	304	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	304	12	algorithm	algorithm	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	304	13	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	304	14	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	304	15	study	study	NN
work_nu5xpentbrcxdlcdt3lmatbur4	304	16	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	304	17	its	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	304	18	conver-	conver-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	304	19	gence	gence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	304	20	properties	property	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	304	21	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	304	22	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	304	23	avenue	avenue	NN
work_nu5xpentbrcxdlcdt3lmatbur4	304	24	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	304	25	future	future	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	304	26	research	research	NN
work_nu5xpentbrcxdlcdt3lmatbur4	304	27	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	305	1	7	7	LS
work_nu5xpentbrcxdlcdt3lmatbur4	305	2	Conclusion	Conclusion	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	305	3	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	305	4	addressed	address	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	305	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	305	6	problem	problem	NN
work_nu5xpentbrcxdlcdt3lmatbur4	305	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	305	8	modeling	model	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	305	9	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	305	10	re-	re-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	305	11	lations	lation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	305	12	expressed	express	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	305	13	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	305	14	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	305	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	306	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	306	2	approached	approach	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	306	3	this	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	306	4	task	task	NN
work_nu5xpentbrcxdlcdt3lmatbur4	306	5	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	306	6	defining	define	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	306	7	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	306	8	set	set	NN
work_nu5xpentbrcxdlcdt3lmatbur4	306	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	306	10	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	306	11	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	306	12	that	that	WDT
work_nu5xpentbrcxdlcdt3lmatbur4	306	13	combine	combine	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	306	14	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	306	15	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	306	16	across	across	IN
work_nu5xpentbrcxdlcdt3lmatbur4	306	17	prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	306	18	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	307	1	Doing	do	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	307	2	so	so	RB
work_nu5xpentbrcxdlcdt3lmatbur4	307	3	allowed	allow	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	307	4	us	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	307	5	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	307	6	leverage	leverage	VB
work_nu5xpentbrcxdlcdt3lmatbur4	307	7	existing	exist	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	307	8	annotated	annotate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	307	9	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	307	10	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	307	11	data	datum	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	307	12	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	307	13	induce	induce	VB
work_nu5xpentbrcxdlcdt3lmatbur4	307	14	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	307	15	corpus	corpus	NN
work_nu5xpentbrcxdlcdt3lmatbur4	307	16	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	307	17	prepo-	prepo-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	307	18	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	307	19	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	307	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	308	1	We	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	308	2	modeled	model	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	308	3	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	308	4	relations	relation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	308	5	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	308	6	terms	term	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	308	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	308	8	its	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	308	9	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	308	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	308	11	namely	namely	RB
work_nu5xpentbrcxdlcdt3lmatbur4	308	12	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	308	13	governor	governor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	308	14	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	308	15	ob-	ob-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	308	16	ject	ject	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	308	17	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	308	18	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	308	19	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	308	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	308	21	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	308	22	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	308	23	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	308	24	types	type	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	308	25	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	308	26	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	308	27	arguments	argument	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	308	28	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	309	1	Using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	309	2	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	309	3	generalization	generalization	NN
work_nu5xpentbrcxdlcdt3lmatbur4	309	4	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	309	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	309	6	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	309	7	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	309	8	SVM	SVM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	309	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	309	10	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	309	11	trained	train	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	309	12	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	309	13	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	309	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	309	15	argument	argument	NN
work_nu5xpentbrcxdlcdt3lmatbur4	309	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	309	17	type	type	NN
work_nu5xpentbrcxdlcdt3lmatbur4	309	18	predictor	predictor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	309	19	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	309	20	only	only	RB
work_nu5xpentbrcxdlcdt3lmatbur4	309	21	annotated	annotate	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	309	22	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	309	23	labels	label	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	309	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	310	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	310	2	allowed	allow	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	310	3	us	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	310	4	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	310	5	get	get	VB
work_nu5xpentbrcxdlcdt3lmatbur4	310	6	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	310	7	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	310	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	310	9	89.43	89.43	CD
work_nu5xpentbrcxdlcdt3lmatbur4	310	10	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	310	11	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	310	12	re-	re-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	310	13	lation	lation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	310	14	prediction	prediction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	310	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	311	1	By	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	311	2	employing	employ	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	311	3	joint	joint	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	311	4	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	311	5	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	311	6	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	311	7	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	311	8	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	311	9	predictor	predictor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	311	10	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	311	11	we	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	311	12	further	further	RB
work_nu5xpentbrcxdlcdt3lmatbur4	311	13	improved	improve	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	311	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	311	15	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	311	16	accuracy	accuracy	NN
work_nu5xpentbrcxdlcdt3lmatbur4	311	17	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	311	18	90.23	90.23	CD
work_nu5xpentbrcxdlcdt3lmatbur4	311	19	%	%	NN
work_nu5xpentbrcxdlcdt3lmatbur4	311	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	312	1	Acknowledgments	acknowledgment	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	312	2	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	312	3	authors	author	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	312	4	wish	wish	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	312	5	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	312	6	thank	thank	VB
work_nu5xpentbrcxdlcdt3lmatbur4	312	7	Martha	Martha	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	312	8	Palmer	Palmer	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	312	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	312	10	Nathan	Nathan	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	312	11	Schneider	Schneider	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	312	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	312	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	312	14	anonymous	anonymous	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	312	15	reviewers	reviewer	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	312	16	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	312	17	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	312	18	editor	editor	NN
work_nu5xpentbrcxdlcdt3lmatbur4	312	19	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	312	20	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	312	21	valuable	valuable	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	312	22	feed-	feed-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	312	23	back	back	RB
work_nu5xpentbrcxdlcdt3lmatbur4	312	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	313	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	313	2	authors	author	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	313	3	gratefully	gratefully	RB
work_nu5xpentbrcxdlcdt3lmatbur4	313	4	acknowledge	acknowledge	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	313	5	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	313	6	support	support	NN
work_nu5xpentbrcxdlcdt3lmatbur4	313	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	313	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	313	9	Defense	Defense	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	10	Advanced	Advanced	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	11	Research	Research	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	12	Projects	Projects	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	313	13	Agency	Agency	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	14	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	313	15	DARPA	DARPA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	16	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	313	17	Ma-	ma-	IN
work_nu5xpentbrcxdlcdt3lmatbur4	313	18	chine	chine	NN
work_nu5xpentbrcxdlcdt3lmatbur4	313	19	Reading	Reading	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	20	Program	Program	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	21	under	under	IN
work_nu5xpentbrcxdlcdt3lmatbur4	313	22	Air	Air	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	23	Force	Force	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	24	Research	Research	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	25	Laboratory	Laboratory	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	26	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	313	27	AFRL	AFRL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	313	28	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	313	29	prime	prime	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	313	30	contract	contract	NN
work_nu5xpentbrcxdlcdt3lmatbur4	313	31	no	no	NN
work_nu5xpentbrcxdlcdt3lmatbur4	313	32	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	314	1	FA8750	FA8750	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	314	2	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	314	3	09-C-0181	09-c-0181	NN
work_nu5xpentbrcxdlcdt3lmatbur4	314	4	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	315	1	This	this	DT
work_nu5xpentbrcxdlcdt3lmatbur4	315	2	material	material	NN
work_nu5xpentbrcxdlcdt3lmatbur4	315	3	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	315	4	also	also	RB
work_nu5xpentbrcxdlcdt3lmatbur4	315	5	based	base	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	315	6	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	315	7	research	research	NN
work_nu5xpentbrcxdlcdt3lmatbur4	315	8	sponsored	sponsor	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	315	9	by	by	IN
work_nu5xpentbrcxdlcdt3lmatbur4	315	10	DARPA	DARPA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	315	11	under	under	IN
work_nu5xpentbrcxdlcdt3lmatbur4	315	12	agree-	agree-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	315	13	ment	ment	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	315	14	number	number	NN
work_nu5xpentbrcxdlcdt3lmatbur4	315	15	FA8750	FA8750	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	315	16	-	-	:
work_nu5xpentbrcxdlcdt3lmatbur4	315	17	13	13	CD
work_nu5xpentbrcxdlcdt3lmatbur4	315	18	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	315	19	2	2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	315	20	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	315	21	0008	0008	CD
work_nu5xpentbrcxdlcdt3lmatbur4	315	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	316	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	316	2	U.S.	U.S.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	316	3	Government	Government	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	316	4	is	be	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	316	5	au-	au-	RB
work_nu5xpentbrcxdlcdt3lmatbur4	316	6	thorized	thorize	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	316	7	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	316	8	reproduce	reproduce	VB
work_nu5xpentbrcxdlcdt3lmatbur4	316	9	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	316	10	distribute	distribute	VB
work_nu5xpentbrcxdlcdt3lmatbur4	316	11	reprints	reprint	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	316	12	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	316	13	Governmental	governmental	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	316	14	purposes	purpose	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	316	15	notwithstanding	notwithstanding	IN
work_nu5xpentbrcxdlcdt3lmatbur4	316	16	any	any	DT
work_nu5xpentbrcxdlcdt3lmatbur4	316	17	copyright	copyright	NN
work_nu5xpentbrcxdlcdt3lmatbur4	316	18	notation	notation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	316	19	thereon	thereon	NN
work_nu5xpentbrcxdlcdt3lmatbur4	316	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	317	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	317	2	views	view	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	317	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	317	4	conclusions	conclusion	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	317	5	contained	contain	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	317	6	herein	herein	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	317	7	are	be	VBP
work_nu5xpentbrcxdlcdt3lmatbur4	317	8	those	those	DT
work_nu5xpentbrcxdlcdt3lmatbur4	317	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	317	10	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	317	11	authors	author	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	317	12	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	317	13	should	should	MD
work_nu5xpentbrcxdlcdt3lmatbur4	317	14	not	not	RB
work_nu5xpentbrcxdlcdt3lmatbur4	317	15	be	be	VB
work_nu5xpentbrcxdlcdt3lmatbur4	317	16	interpreted	interpret	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	317	17	as	as	IN
work_nu5xpentbrcxdlcdt3lmatbur4	317	18	necessarily	necessarily	RB
work_nu5xpentbrcxdlcdt3lmatbur4	317	19	representing	represent	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	317	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	317	21	of-	of-	CD
work_nu5xpentbrcxdlcdt3lmatbur4	317	22	ficial	ficial	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	317	23	policies	policy	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	317	24	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	317	25	endorsements	endorsement	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	317	26	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	317	27	either	either	CC
work_nu5xpentbrcxdlcdt3lmatbur4	317	28	expressed	express	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	317	29	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	317	30	implied	imply	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	317	31	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	317	32	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	317	33	DARPA	DARPA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	317	34	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	317	35	AFRL	AFRL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	317	36	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	317	37	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	317	38	U.S.	U.S.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	317	39	Government	Government	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	317	40	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	318	1	References	References	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	318	2	E.	E.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	318	3	Agirre	Agirre	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	318	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	318	5	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	318	6	Baldwin	Baldwin	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	318	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	318	8	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	318	9	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	318	10	Martinez	Martinez	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	318	11	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	319	1	2008	2008	CD
work_nu5xpentbrcxdlcdt3lmatbur4	319	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	320	1	Improv-	Improv-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	320	2	ing	e	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	320	3	parsing	parsing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	320	4	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	320	5	PP	PP	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	320	6	attachment	attachment	NN
work_nu5xpentbrcxdlcdt3lmatbur4	320	7	performance	performance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	320	8	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	320	9	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	320	10	information	information	NN
work_nu5xpentbrcxdlcdt3lmatbur4	320	11	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	321	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	321	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	321	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	321	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	321	5	Annual	Annual	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	321	6	Meeting	Meeting	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	321	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	321	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	321	9	Association	Association	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	321	10	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	321	11	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	321	12	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	321	13	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	321	14	ACL	ACL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	321	15	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	321	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	321	17	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	321	18	317–325	317–325	CD
work_nu5xpentbrcxdlcdt3lmatbur4	321	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	321	20	Columbus	Columbus	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	321	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	321	22	USA	USA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	321	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	322	1	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	322	2	Baldwin	Baldwin	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	322	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	322	4	V.	V.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	322	5	Kordoni	Kordoni	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	322	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	322	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	322	8	A.	A.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	322	9	Villavicencio	Villavicencio	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	322	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	323	1	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	323	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	324	1	Prepositions	preposition	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	324	2	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	324	3	applications	application	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	324	4	:	:	.
work_nu5xpentbrcxdlcdt3lmatbur4	324	5	A	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	324	6	survey	survey	NN
work_nu5xpentbrcxdlcdt3lmatbur4	324	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	324	8	introduc-	introduc-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	324	9	tion	tion	NN
work_nu5xpentbrcxdlcdt3lmatbur4	324	10	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	324	11	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	324	12	special	special	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	324	13	issue	issue	NN
work_nu5xpentbrcxdlcdt3lmatbur4	324	14	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	325	1	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	325	2	Linguistics	Linguistics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	325	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	325	4	35(2):119–149	35(2):119–149	CD
work_nu5xpentbrcxdlcdt3lmatbur4	325	5	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	326	1	M.	M.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	326	2	Chang	Chang	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	326	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	326	4	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	326	5	Goldwasser	Goldwasser	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	326	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	326	7	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	326	8	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	326	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	326	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	326	11	V.	V.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	326	12	Srikumar	Srikumar	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	326	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	327	1	2010a	2010a	CD
work_nu5xpentbrcxdlcdt3lmatbur4	327	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	328	1	Discriminative	discriminative	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	328	2	learning	learn	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	328	3	over	over	IN
work_nu5xpentbrcxdlcdt3lmatbur4	328	4	constrained	constrained	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	328	5	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	328	6	representations	representation	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	328	7	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	329	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	329	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	329	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	329	5	Annual	Annual	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	6	Meet-	Meet-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	7	ing	ing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	329	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	329	9	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	329	10	North	North	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	11	American	American	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	12	Association	Association	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	329	14	Computa-	Computa-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	15	tional	tional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	329	16	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	17	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	329	18	NAACL	NAACL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	19	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	329	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	329	21	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	329	22	429–437	429–437	CD
work_nu5xpentbrcxdlcdt3lmatbur4	329	23	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	329	24	Los	Los	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	25	An-	An-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	26	geles	gele	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	329	27	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	329	28	USA	USA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	329	29	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	330	1	M.	M.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	330	2	Chang	Chang	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	330	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	330	4	V.	V.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	330	5	Srikumar	Srikumar	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	330	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	330	7	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	330	8	Goldwasser	Goldwasser	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	330	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	330	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	330	11	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	330	12	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	330	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	331	1	2010b	2010b	CD
work_nu5xpentbrcxdlcdt3lmatbur4	331	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	332	1	Structured	structured	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	332	2	output	output	NN
work_nu5xpentbrcxdlcdt3lmatbur4	332	3	learning	learn	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	332	4	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	332	5	indirect	indirect	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	332	6	super-	super-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	332	7	vision	vision	NN
work_nu5xpentbrcxdlcdt3lmatbur4	332	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	333	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	333	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	333	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	333	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	333	5	International	International	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	333	6	Conference	Conference	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	333	7	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	333	8	Machine	Machine	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	333	9	Learning	Learning	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	333	10	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	333	11	ICML	ICML	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	333	12	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	333	13	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	333	14	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	333	15	199–206	199–206	CD
work_nu5xpentbrcxdlcdt3lmatbur4	333	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	333	17	Haifa	Haifa	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	333	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	333	19	Israel	Israel	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	333	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	334	1	J.	J.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	334	2	Clarke	Clarke	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	334	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	334	4	V.	V.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	334	5	Srikumar	Srikumar	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	334	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	334	7	M.	M.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	334	8	Sammons	Sammons	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	334	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	334	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	334	11	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	334	12	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	334	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	335	1	2012	2012	CD
work_nu5xpentbrcxdlcdt3lmatbur4	335	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	336	1	An	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	336	2	NLP	NLP	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	336	3	Curator	Curator	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	336	4	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	336	5	or	or	CC
work_nu5xpentbrcxdlcdt3lmatbur4	336	6	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	336	7	How	how	WRB
work_nu5xpentbrcxdlcdt3lmatbur4	336	8	I	-PRON-	PRP
work_nu5xpentbrcxdlcdt3lmatbur4	336	9	Learned	learn	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	336	10	to	to	TO
work_nu5xpentbrcxdlcdt3lmatbur4	336	11	Stop	stop	VB
work_nu5xpentbrcxdlcdt3lmatbur4	336	12	Wor-	Wor-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	336	13	rying	rye	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	336	14	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	336	15	Love	Love	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	336	16	NLP	NLP	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	336	17	Pipelines	Pipelines	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	336	18	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	336	19	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	337	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	337	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	337	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	337	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	337	5	International	International	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	337	6	Conference	Conference	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	337	7	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	337	8	Language	Language	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	337	9	Resources	Resources	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	337	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	337	11	Evaluation	Evaluation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	337	12	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	337	13	LREC	LREC	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	337	14	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	337	15	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	337	16	pages	page	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	337	17	3276–3283	3276–3283	CD
work_nu5xpentbrcxdlcdt3lmatbur4	337	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	337	19	Istanbul	Istanbul	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	337	20	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	337	21	Turkey	Turkey	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	337	22	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	338	1	J.	J.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	338	2	Cohen	Cohen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	338	3	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	339	1	1960	1960	CD
work_nu5xpentbrcxdlcdt3lmatbur4	339	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	340	1	A	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	340	2	coefficient	coefficient	NN
work_nu5xpentbrcxdlcdt3lmatbur4	340	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	340	4	agreement	agreement	NN
work_nu5xpentbrcxdlcdt3lmatbur4	340	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	340	6	nominal	nominal	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	340	7	scales	scale	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	340	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	341	1	Educational	educational	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	341	2	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	341	3	Psychological	Psychological	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	341	4	Measurement	Measurement	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	341	5	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	341	6	20:37–46	20:37–46	CD
work_nu5xpentbrcxdlcdt3lmatbur4	341	7	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	342	1	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	342	2	Dahlmeier	Dahlmeier	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	342	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	342	4	H.	H.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	342	5	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	342	6	Ng	Ng	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	342	7	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	342	8	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	342	9	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	342	10	Schultz	Schultz	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	342	11	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	343	1	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	343	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	344	1	Joint	joint	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	344	2	learning	learning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	344	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	344	4	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	344	5	senses	sense	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	344	6	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	344	7	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	344	8	roles	role	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	344	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	344	10	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	344	11	phrases	phrase	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	344	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	345	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	345	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	345	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	345	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	345	5	Confer-	Confer-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	345	6	ence	ence	NN
work_nu5xpentbrcxdlcdt3lmatbur4	345	7	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	345	8	Empirical	Empirical	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	345	9	Methods	Methods	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	345	10	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	345	11	Natural	Natural	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	345	12	Language	Language	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	345	13	Pro-	Pro-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	345	14	cessing	cesse	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	345	15	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	345	16	EMNLP	EMNLP	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	345	17	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	345	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	345	19	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	345	20	450–458	450–458	CD
work_nu5xpentbrcxdlcdt3lmatbur4	345	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	345	22	Singapore	Singapore	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	345	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	346	1	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	346	2	Das	Das	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	346	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	346	4	N.	N.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	346	5	Schneider	Schneider	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	346	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	346	7	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	346	8	Chen	Chen	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	346	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	346	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	346	11	N.	N.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	346	12	Smith	Smith	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	346	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	347	1	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	347	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	348	1	Probabilistic	probabilistic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	348	2	frame	frame	NN
work_nu5xpentbrcxdlcdt3lmatbur4	348	3	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	348	4	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	348	5	parsing	parsing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	348	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	349	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	349	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	349	4	Human	Human	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	5	Language	Language	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	6	Technologies	Technologies	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	349	7	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	349	8	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	349	9	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	349	10	Annual	Annual	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	11	241	241	CD
work_nu5xpentbrcxdlcdt3lmatbur4	349	12	Conference	Conference	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	349	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	349	15	North	north	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	349	16	American	american	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	349	17	Chapter	Chapter	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	349	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	349	20	As-	As-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	21	sociation	sociation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	349	22	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	349	23	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	24	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	25	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	349	26	pages	page	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	349	27	948	948	CD
work_nu5xpentbrcxdlcdt3lmatbur4	349	28	–	–	:
work_nu5xpentbrcxdlcdt3lmatbur4	349	29	956	956	CD
work_nu5xpentbrcxdlcdt3lmatbur4	349	30	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	349	31	Los	Los	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	32	Angeles	Angeles	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	33	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	349	34	USA	USA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	349	35	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	350	1	C.	C.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	350	2	Fillmore	Fillmore	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	350	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	350	4	C.	C.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	350	5	Johnson	Johnson	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	350	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	350	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	350	8	M.	M.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	350	9	Petruck	Petruck	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	350	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	351	1	2003	2003	CD
work_nu5xpentbrcxdlcdt3lmatbur4	351	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	352	1	Back-	Back-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	352	2	ground	grind	VBD
work_nu5xpentbrcxdlcdt3lmatbur4	352	3	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	352	4	FrameNet	FrameNet	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	352	5	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	353	1	International	International	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	353	2	Journal	Journal	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	353	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	353	4	Lexi-	Lexi-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	353	5	cography	cography	NN
work_nu5xpentbrcxdlcdt3lmatbur4	353	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	353	7	16(3):235–250	16(3):235–250	CD
work_nu5xpentbrcxdlcdt3lmatbur4	353	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	354	1	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	354	2	Gildea	Gildea	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	354	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	354	4	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	354	5	Jurafsky	Jurafsky	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	354	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	355	1	2002	2002	CD
work_nu5xpentbrcxdlcdt3lmatbur4	355	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	356	1	Automatic	automatic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	356	2	label-	label-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	356	3	ing	ing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	356	4	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	356	5	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	356	6	roles	role	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	356	7	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	357	1	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	357	2	Linguistics	Linguistics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	357	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	357	4	28(3):245–288	28(3):245–288	CD
work_nu5xpentbrcxdlcdt3lmatbur4	357	5	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	358	1	Y.	Y.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	358	2	Goldberg	Goldberg	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	358	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	358	4	M.	M.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	358	5	Elhadad	Elhadad	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	358	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	359	1	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	359	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	360	1	An	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	360	2	efficient	efficient	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	360	3	algo-	algo-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	360	4	rithm	rithm	NN
work_nu5xpentbrcxdlcdt3lmatbur4	360	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	360	6	easy	easy	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	360	7	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	360	8	first	first	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	360	9	non	non	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	360	10	-	-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	360	11	directional	directional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	360	12	dependency	dependency	NN
work_nu5xpentbrcxdlcdt3lmatbur4	360	13	pars-	pars-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	360	14	ing	ing	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	360	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	361	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	361	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	361	4	Human	Human	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	5	Language	Language	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	6	Technolo-	Technolo-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	7	gies	gi	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	361	8	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	361	9	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	361	10	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	361	11	Annual	Annual	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	12	Conference	Conference	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	13	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	361	14	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	361	15	North	North	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	16	Ameri-	Ameri-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	17	can	can	MD
work_nu5xpentbrcxdlcdt3lmatbur4	361	18	Chapter	chapter	NN
work_nu5xpentbrcxdlcdt3lmatbur4	361	19	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	361	20	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	361	21	Association	Association	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	22	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	361	23	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	24	Lin-	Lin-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	25	guistics	guistic	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	361	26	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	361	27	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	361	28	742–750	742–750	CD
work_nu5xpentbrcxdlcdt3lmatbur4	361	29	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	361	30	Los	Los	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	31	Angeles	Angeles	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	32	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	361	33	USA	USA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	361	34	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	362	1	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	362	2	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	362	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	362	4	S.	S.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	362	5	Tratz	Tratz	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	362	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	362	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	362	8	E.	E.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	362	9	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	362	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	363	1	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	363	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	364	1	What	what	WP
work_nu5xpentbrcxdlcdt3lmatbur4	364	2	’s	’	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	364	3	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	364	4	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	364	5	prepo-	prepo-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	364	6	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	364	7	?	?	.
work_nu5xpentbrcxdlcdt3lmatbur4	365	1	dimensions	dimension	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	365	2	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	365	3	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	365	4	disambiguation	disambiguation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	365	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	365	6	an	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	365	7	in-	in-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	365	8	teresting	teresting	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	365	9	word	word	NN
work_nu5xpentbrcxdlcdt3lmatbur4	365	10	class	class	NN
work_nu5xpentbrcxdlcdt3lmatbur4	365	11	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	366	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	366	2	Coling	Coling	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	366	3	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	366	4	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	366	5	Posters	poster	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	366	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	366	7	pages	page	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	366	8	454–462	454–462	CD
work_nu5xpentbrcxdlcdt3lmatbur4	366	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	366	10	Beijing	Beijing	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	366	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	366	12	China	China	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	366	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	367	1	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	367	2	Lin	Lin	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	367	3	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	368	1	1998	1998	CD
work_nu5xpentbrcxdlcdt3lmatbur4	368	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	369	1	Automatic	automatic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	369	2	retrieval	retrieval	NN
work_nu5xpentbrcxdlcdt3lmatbur4	369	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	369	4	clustering	clustering	NN
work_nu5xpentbrcxdlcdt3lmatbur4	369	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	369	6	sim-	sim-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	369	7	ilar	ilar	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	369	8	words	word	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	369	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	370	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	370	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	370	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	370	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	370	5	Annual	Annual	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	370	6	Meeting	Meeting	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	370	7	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	370	8	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	370	9	Association	Association	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	370	10	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	370	11	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	370	12	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	370	13	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	370	14	ACL	ACL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	370	15	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	370	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	370	17	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	370	18	768–774	768–774	CD
work_nu5xpentbrcxdlcdt3lmatbur4	370	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	370	20	Montreal	Montreal	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	370	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	370	22	Canada	Canada	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	370	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	371	1	K.	K.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	371	2	Litkowski	Litkowski	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	371	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	371	4	O.	O.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	371	5	Hargraves	Hargraves	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	371	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	372	1	2005	2005	CD
work_nu5xpentbrcxdlcdt3lmatbur4	372	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	373	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	373	2	Preposition	Preposition	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	373	3	Project	Project	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	373	4	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	374	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	374	2	ACL	ACL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	374	3	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	374	4	SIGSEM	SIGSEM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	374	5	Workshop	Workshop	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	374	6	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	374	7	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	374	8	Linguistic	linguistic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	374	9	Dimensions	Dimensions	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	374	10	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	374	11	Prepositions	Prepositions	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	374	12	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	374	13	their	-PRON-	PRP$
work_nu5xpentbrcxdlcdt3lmatbur4	374	14	Use	Use	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	374	15	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	374	16	Computa-	Computa-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	374	17	tional	tional	NN
work_nu5xpentbrcxdlcdt3lmatbur4	374	18	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	374	19	Formalisms	Formalisms	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	374	20	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	374	21	Applications	Applications	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	374	22	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	374	23	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	374	24	171–179	171–179	CD
work_nu5xpentbrcxdlcdt3lmatbur4	374	25	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	374	26	Colchester	Colchester	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	374	27	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	374	28	UK	UK	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	374	29	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	375	1	K.	K.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	375	2	Litkowski	Litkowski	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	375	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	375	4	O.	O.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	375	5	Hargraves	Hargraves	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	375	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	376	1	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	376	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	377	1	SemEval-2007	SemEval-2007	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	377	2	Task	Task	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	377	3	06	06	CD
work_nu5xpentbrcxdlcdt3lmatbur4	377	4	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	377	5	Word	Word	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	377	6	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	377	7	Sense	Sense	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	377	8	Disambiguation	Disambiguation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	377	9	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	377	10	Prepositions	Prepositions	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	377	11	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	378	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	378	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	378	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	378	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	378	5	Fourth	Fourth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	378	6	International	International	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	378	7	Workshop	Workshop	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	378	8	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	378	9	Semantic	Semantic	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	378	10	Evaluations	Evaluations	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	378	11	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	378	12	SemEval-2007	semeval-2007	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	378	13	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	378	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	378	15	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	378	16	24	24	CD
work_nu5xpentbrcxdlcdt3lmatbur4	378	17	–	–	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	378	18	29	29	CD
work_nu5xpentbrcxdlcdt3lmatbur4	378	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	378	20	Prague	Prague	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	378	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	378	22	Czech	Czech	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	378	23	Republic	Republic	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	378	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	379	1	K.	K.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	379	2	Litkowski	Litkowski	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	379	3	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	380	1	2012	2012	CD
work_nu5xpentbrcxdlcdt3lmatbur4	380	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	381	1	Proposed	propose	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	381	2	Next	Next	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	381	3	Steps	Steps	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	381	4	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	381	5	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	381	6	Prepo-	Prepo-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	381	7	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	381	8	Project	Project	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	381	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	382	1	Technical	Technical	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	382	2	Report	Report	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	382	3	12	12	CD
work_nu5xpentbrcxdlcdt3lmatbur4	382	4	-	-	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	382	5	01	01	CD
work_nu5xpentbrcxdlcdt3lmatbur4	382	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	382	7	CL	CL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	382	8	Research	Research	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	382	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	383	1	A.	A.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	2	Meyers	Meyers	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	383	4	R.	R.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	5	Reeves	Reeves	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	383	7	C.	C.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	8	Macleod	Macleod	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	383	10	R.	R.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	11	Szekely	Szekely	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	383	13	V.	V.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	14	Zielin-	Zielin-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	383	15	ska	ska	NN
work_nu5xpentbrcxdlcdt3lmatbur4	383	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	383	17	B.	B.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	384	1	Young	young	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	384	2	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	384	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	384	4	R.	R.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	384	5	Grishman	Grishman	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	384	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	385	1	2004	2004	CD
work_nu5xpentbrcxdlcdt3lmatbur4	385	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	386	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	386	2	Nom-	Nom-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	386	3	Bank	Bank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	386	4	project	project	NN
work_nu5xpentbrcxdlcdt3lmatbur4	386	5	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	386	6	An	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	386	7	interim	interim	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	386	8	report	report	NN
work_nu5xpentbrcxdlcdt3lmatbur4	386	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	387	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	387	2	HLT	HLT	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	387	3	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	387	4	NAACL	NAACL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	387	5	2004	2004	CD
work_nu5xpentbrcxdlcdt3lmatbur4	387	6	Workshop	workshop	NN
work_nu5xpentbrcxdlcdt3lmatbur4	387	7	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	387	8	Frontiers	frontier	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	387	9	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	387	10	Corpus	Corpus	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	387	11	Annotation	Annotation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	387	12	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	387	13	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	387	14	24	24	CD
work_nu5xpentbrcxdlcdt3lmatbur4	387	15	–	–	SYM
work_nu5xpentbrcxdlcdt3lmatbur4	387	16	31	31	CD
work_nu5xpentbrcxdlcdt3lmatbur4	387	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	387	18	Boston	Boston	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	387	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	387	20	USA	USA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	387	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	388	1	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	388	2	O’Hara	O’Hara	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	388	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	388	4	J.	J.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	388	5	Wiebe	Wiebe	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	388	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	389	1	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	389	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	390	1	Exploiting	exploit	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	390	2	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	390	3	role	role	NN
work_nu5xpentbrcxdlcdt3lmatbur4	390	4	resources	resource	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	390	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	390	6	preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	390	7	disambiguation	disambiguation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	390	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	391	1	Computa-	Computa-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	391	2	tional	tional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	391	3	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	391	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	391	5	35(2):151–184	35(2):151–184	CD
work_nu5xpentbrcxdlcdt3lmatbur4	391	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	392	1	M.	M.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	392	2	Palmer	Palmer	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	392	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	392	4	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	392	5	Gildea	Gildea	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	392	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	392	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	392	8	P.	P.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	392	9	Kingsbury	Kingsbury	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	392	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	393	1	2005	2005	CD
work_nu5xpentbrcxdlcdt3lmatbur4	393	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	394	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	394	2	Proposition	Proposition	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	394	3	Bank	Bank	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	394	4	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	394	5	An	an	DT
work_nu5xpentbrcxdlcdt3lmatbur4	394	6	Annotated	Annotated	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	394	7	Corpus	Corpus	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	394	8	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	394	9	Semantic	Semantic	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	394	10	Roles	Roles	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	394	11	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	395	1	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	395	2	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	395	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	395	4	31(1):71–106	31(1):71–106	CD
work_nu5xpentbrcxdlcdt3lmatbur4	395	5	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	396	1	M.	M.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	396	2	Palmer	Palmer	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	396	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	396	4	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	396	5	Gildea	Gildea	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	396	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	396	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	396	8	N.	N.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	396	9	Xue	Xue	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	396	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	397	1	2010	2010	CD
work_nu5xpentbrcxdlcdt3lmatbur4	397	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	398	1	Semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	398	2	Role	Role	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	398	3	Labeling	Labeling	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	398	4	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	398	5	volume	volume	NN
work_nu5xpentbrcxdlcdt3lmatbur4	398	6	3	3	CD
work_nu5xpentbrcxdlcdt3lmatbur4	398	7	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	399	1	Morgan	Morgan	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	399	2	&	&	CC
work_nu5xpentbrcxdlcdt3lmatbur4	399	3	Claypool	Claypool	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	399	4	Publishers	Publishers	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	399	5	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	400	1	V.	V.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	400	2	Punyakanok	Punyakanok	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	400	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	400	4	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	400	5	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	400	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	400	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	400	8	W.	W.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	400	9	Yih	Yih	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	400	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	401	1	2008	2008	CD
work_nu5xpentbrcxdlcdt3lmatbur4	401	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	402	1	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	402	2	impor-	impor-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	402	3	tance	tance	NN
work_nu5xpentbrcxdlcdt3lmatbur4	402	4	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	402	5	syntactic	syntactic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	402	6	parsing	parsing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	402	7	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	402	8	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	402	9	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	402	10	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	402	11	role	role	NN
work_nu5xpentbrcxdlcdt3lmatbur4	402	12	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	402	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	403	1	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	403	2	Linguistics	Linguistics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	403	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	403	4	34(2	34(2	CD
work_nu5xpentbrcxdlcdt3lmatbur4	403	5	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	403	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	404	1	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	404	2	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	404	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	404	4	W.	W.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	404	5	Yih	Yih	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	404	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	405	1	2004	2004	CD
work_nu5xpentbrcxdlcdt3lmatbur4	405	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	406	1	A	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	406	2	linear	linear	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	406	3	programming	programming	NN
work_nu5xpentbrcxdlcdt3lmatbur4	406	4	formu-	formu-	FW
work_nu5xpentbrcxdlcdt3lmatbur4	406	5	lation	lation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	406	6	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	406	7	global	global	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	406	8	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	406	9	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	406	10	natural	natural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	406	11	language	language	NN
work_nu5xpentbrcxdlcdt3lmatbur4	406	12	tasks	task	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	406	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	407	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	407	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	407	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	407	5	Annual	Annual	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	6	Conference	Conference	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	7	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	407	8	Compu-	Compu-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	9	tational	tational	NN
work_nu5xpentbrcxdlcdt3lmatbur4	407	10	Natural	Natural	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	11	Language	Language	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	12	Learning	Learning	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	13	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	407	14	CoNLL	CoNLL	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	15	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	407	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	407	17	pages	page	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	407	18	1–8	1–8	CD
work_nu5xpentbrcxdlcdt3lmatbur4	407	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	407	20	Boston	Boston	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	407	22	USA	USA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	407	23	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	408	1	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	408	2	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	408	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	408	4	W.	W.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	408	5	Yih	Yih	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	408	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	409	1	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	409	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	410	1	Global	global	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	410	2	inference	inference	NN
work_nu5xpentbrcxdlcdt3lmatbur4	410	3	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	410	4	en-	en-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	410	5	tity	tity	NN
work_nu5xpentbrcxdlcdt3lmatbur4	410	6	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	410	7	relation	relation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	410	8	identification	identification	NN
work_nu5xpentbrcxdlcdt3lmatbur4	410	9	via	via	IN
work_nu5xpentbrcxdlcdt3lmatbur4	410	10	a	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	410	11	linear	linear	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	410	12	program-	program-	NN
work_nu5xpentbrcxdlcdt3lmatbur4	410	13	ming	ming	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	410	14	formulation	formulation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	410	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	411	1	Introduction	introduction	NN
work_nu5xpentbrcxdlcdt3lmatbur4	411	2	to	to	IN
work_nu5xpentbrcxdlcdt3lmatbur4	411	3	Statistical	statistical	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	411	4	Rela-	Rela-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	411	5	tional	tional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	411	6	Learning	learning	NN
work_nu5xpentbrcxdlcdt3lmatbur4	411	7	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	412	1	V.	V.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	412	2	Srikumar	Srikumar	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	412	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	412	4	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	412	5	Roth	Roth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	412	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	413	1	2011	2011	CD
work_nu5xpentbrcxdlcdt3lmatbur4	413	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	414	1	A	a	DT
work_nu5xpentbrcxdlcdt3lmatbur4	414	2	joint	joint	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	414	3	model	model	NN
work_nu5xpentbrcxdlcdt3lmatbur4	414	4	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	414	5	extended	extended	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	414	6	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	414	7	role	role	NN
work_nu5xpentbrcxdlcdt3lmatbur4	414	8	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	414	9	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	415	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	415	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	415	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	415	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	415	5	Conference	Conference	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	415	6	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	415	7	Empirical	Empirical	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	415	8	Methods	Methods	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	415	9	in	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	415	10	Natural	Natural	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	415	11	Lan-	Lan-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	415	12	guage	guage	NN
work_nu5xpentbrcxdlcdt3lmatbur4	415	13	Processing	processing	NN
work_nu5xpentbrcxdlcdt3lmatbur4	415	14	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	415	15	EMNLP	EMNLP	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	415	16	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	415	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	415	18	Edinburgh	Edinburgh	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	415	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	415	20	Scotland	Scotland	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	415	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	416	1	S.	S.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	416	2	Tratz	Tratz	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	416	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	416	4	D.	D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	416	5	Hovy	Hovy	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	416	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	417	1	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	417	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	418	1	Disambiguation	Disambiguation	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	418	2	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	418	3	prepo-	prepo-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	418	4	sition	sition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	418	5	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	418	6	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	418	7	linguistically	linguistically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	418	8	motivated	motivated	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	418	9	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	418	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	419	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	419	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	419	4	Human	Human	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	5	Language	Language	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	6	Technologies	Technologies	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	419	7	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	419	8	The	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	419	9	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	419	10	Annual	Annual	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	11	Conference	Conference	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	12	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	419	13	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	419	14	North	North	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	15	American	American	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	16	Chap-	Chap-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	17	ter	ter	NN
work_nu5xpentbrcxdlcdt3lmatbur4	419	18	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	419	19	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	419	20	Association	Association	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	21	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	419	22	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	23	Linguistics	Linguistics	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	419	24	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	419	25	Companion	Companion	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	26	Volume	volume	NN
work_nu5xpentbrcxdlcdt3lmatbur4	419	27	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	419	28	Student	Student	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	29	Research	Research	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	30	Workshop	Workshop	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	31	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	419	32	Doctoral	Doctoral	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	33	Consortium	Consortium	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	34	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	419	35	pages	page	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	419	36	96–100	96–100	CD
work_nu5xpentbrcxdlcdt3lmatbur4	419	37	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	419	38	Boulder	Boulder	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	39	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	419	40	USA	USA	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	419	41	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	420	1	S.	S.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	420	2	Tratz	Tratz	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	420	3	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	421	1	2011	2011	CD
work_nu5xpentbrcxdlcdt3lmatbur4	421	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	422	1	Semantically	semantically	RB
work_nu5xpentbrcxdlcdt3lmatbur4	422	2	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	422	3	enriched	enrich	VBN
work_nu5xpentbrcxdlcdt3lmatbur4	422	4	Parsing	parse	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	422	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	422	6	Natu-	Natu-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	422	7	ral	ral	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	422	8	Language	Language	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	422	9	Understanding	Understanding	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	422	10	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	423	1	Ph.D.	Ph.D.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	423	2	thesis	thesis	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	423	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	423	4	University	University	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	423	5	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	423	6	Southern	Southern	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	423	7	California	California	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	423	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	424	1	I.	I.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	424	2	Tsochantaridis	Tsochantaridis	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	424	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	424	4	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	424	5	Hofmann	Hofmann	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	424	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	424	7	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	424	8	Joachims	Joachims	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	424	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	424	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	424	11	Y.	Y.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	424	12	Altun	Altun	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	424	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	425	1	2004	2004	CD
work_nu5xpentbrcxdlcdt3lmatbur4	425	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	426	1	Support	support	NN
work_nu5xpentbrcxdlcdt3lmatbur4	426	2	vector	vector	NN
work_nu5xpentbrcxdlcdt3lmatbur4	426	3	machine	machine	NN
work_nu5xpentbrcxdlcdt3lmatbur4	426	4	learning	learn	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	426	5	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	426	6	interde-	interde-	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	426	7	pendent	pendent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	426	8	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	426	9	structured	structured	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	426	10	output	output	NN
work_nu5xpentbrcxdlcdt3lmatbur4	426	11	spaces	space	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	426	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	427	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	427	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	427	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	427	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	427	5	International	International	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	427	6	Conference	Conference	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	427	7	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	427	8	Machine	Machine	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	427	9	Learning	Learning	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	427	10	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	427	11	ICML	ICML	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	427	12	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	427	13	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	427	14	pages	page	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	427	15	104–111	104–111	CD
work_nu5xpentbrcxdlcdt3lmatbur4	427	16	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	427	17	Banff	Banff	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	427	18	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	427	19	Canada	Canada	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	427	20	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	428	1	P.	P.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	428	2	Ye	Ye	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	428	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	428	4	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	428	5	Baldwin	Baldwin	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	428	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	429	1	2006	2006	CD
work_nu5xpentbrcxdlcdt3lmatbur4	429	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	430	1	Semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	430	2	role	role	NN
work_nu5xpentbrcxdlcdt3lmatbur4	430	3	labeling	labeling	NN
work_nu5xpentbrcxdlcdt3lmatbur4	430	4	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	430	5	prepositional	prepositional	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	430	6	phrases	phrase	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	430	7	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	431	1	ACM	ACM	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	431	2	Transactions	transaction	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	431	3	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	431	4	Asian	Asian	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	431	5	Language	Language	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	431	6	Information	Information	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	431	7	Processing	Processing	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	431	8	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	431	9	TALIP	TALIP	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	431	10	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	431	11	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	431	12	5(3):228	5(3):228	NN
work_nu5xpentbrcxdlcdt3lmatbur4	431	13	–	–	:
work_nu5xpentbrcxdlcdt3lmatbur4	431	14	244	244	CD
work_nu5xpentbrcxdlcdt3lmatbur4	431	15	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	432	1	P.	P.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	432	2	Ye	Ye	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	432	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	432	4	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	432	5	Baldwin	Baldwin	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	432	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	433	1	2007	2007	CD
work_nu5xpentbrcxdlcdt3lmatbur4	433	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	434	1	MELB	melb	NN
work_nu5xpentbrcxdlcdt3lmatbur4	434	2	-	-	HYPH
work_nu5xpentbrcxdlcdt3lmatbur4	434	3	YB	yb	NN
work_nu5xpentbrcxdlcdt3lmatbur4	434	4	:	:	:
work_nu5xpentbrcxdlcdt3lmatbur4	434	5	Preposition	preposition	NN
work_nu5xpentbrcxdlcdt3lmatbur4	434	6	sense	sense	NN
work_nu5xpentbrcxdlcdt3lmatbur4	434	7	disambiguation	disambiguation	NN
work_nu5xpentbrcxdlcdt3lmatbur4	434	8	using	use	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	434	9	rich	rich	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	434	10	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	434	11	features	feature	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	434	12	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	435	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	435	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	435	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	435	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	435	5	Fourth	Fourth	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	435	6	International	International	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	435	7	Workshop	Workshop	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	435	8	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	435	9	Semantic	Semantic	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	435	10	Evaluations	Evaluations	NNPS
work_nu5xpentbrcxdlcdt3lmatbur4	435	11	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	435	12	SemEval-2007	semeval-2007	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	435	13	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	435	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	435	15	pages	page	VBZ
work_nu5xpentbrcxdlcdt3lmatbur4	435	16	241	241	CD
work_nu5xpentbrcxdlcdt3lmatbur4	435	17	–	–	:
work_nu5xpentbrcxdlcdt3lmatbur4	435	18	244	244	CD
work_nu5xpentbrcxdlcdt3lmatbur4	435	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	435	20	Prague	Prague	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	435	21	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	435	22	Czech	Czech	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	435	23	Republic	Republic	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	435	24	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	436	1	C.	C.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	436	2	Yu	Yu	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	436	3	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	436	4	T.	T.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	436	5	Joachims	joachim	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	436	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	437	1	2009	2009	CD
work_nu5xpentbrcxdlcdt3lmatbur4	437	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	438	1	Learning	learn	VBG
work_nu5xpentbrcxdlcdt3lmatbur4	438	2	structural	structural	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	438	3	SVMs	svm	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	438	4	with	with	IN
work_nu5xpentbrcxdlcdt3lmatbur4	438	5	latent	latent	NN
work_nu5xpentbrcxdlcdt3lmatbur4	438	6	variables	variable	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	438	7	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	439	1	In	in	IN
work_nu5xpentbrcxdlcdt3lmatbur4	439	2	Proceedings	Proceedings	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	439	3	of	of	IN
work_nu5xpentbrcxdlcdt3lmatbur4	439	4	the	the	DT
work_nu5xpentbrcxdlcdt3lmatbur4	439	5	Inter-	Inter-	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	439	6	national	national	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	439	7	Conference	Conference	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	439	8	on	on	IN
work_nu5xpentbrcxdlcdt3lmatbur4	439	9	Machine	Machine	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	439	10	Learning	Learning	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	439	11	(	(	-LRB-
work_nu5xpentbrcxdlcdt3lmatbur4	439	12	ICML	ICML	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	439	13	)	)	-RRB-
work_nu5xpentbrcxdlcdt3lmatbur4	439	14	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	439	15	pages	page	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	439	16	1–8	1–8	CD
work_nu5xpentbrcxdlcdt3lmatbur4	439	17	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	439	18	Montreal	Montreal	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	439	19	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	439	20	Canada	Canada	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	439	21	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	440	1	B.	B.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	440	2	Zapirain	Zapirain	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	440	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	440	4	E.	E.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	440	5	Agirre	Agirre	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	440	6	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	440	7	L.	L.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	440	8	Màrquez	Màrquez	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	440	9	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	440	10	and	and	CC
work_nu5xpentbrcxdlcdt3lmatbur4	440	11	M.	M.	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	440	12	Surdeanu	Surdeanu	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	440	13	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	441	1	2012	2012	CD
work_nu5xpentbrcxdlcdt3lmatbur4	441	2	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	442	1	Selectional	Selectional	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	442	2	preferences	preference	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	442	3	for	for	IN
work_nu5xpentbrcxdlcdt3lmatbur4	442	4	semantic	semantic	JJ
work_nu5xpentbrcxdlcdt3lmatbur4	442	5	role	role	NN
work_nu5xpentbrcxdlcdt3lmatbur4	442	6	classi-	classi-	XX
work_nu5xpentbrcxdlcdt3lmatbur4	442	7	fication	fication	NN
work_nu5xpentbrcxdlcdt3lmatbur4	442	8	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	443	1	Computational	Computational	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	443	2	Linguistics	Linguistics	NNP
work_nu5xpentbrcxdlcdt3lmatbur4	443	3	,	,	,
work_nu5xpentbrcxdlcdt3lmatbur4	443	4	pages	page	NNS
work_nu5xpentbrcxdlcdt3lmatbur4	443	5	1–33	1–33	CD
work_nu5xpentbrcxdlcdt3lmatbur4	443	6	.	.	.
work_nu5xpentbrcxdlcdt3lmatbur4	444	1	242	242	CD