id	sid	tid	token	lemma	pos
work_vbzp3iewy5h7rhfiok5o6i3du4	1	1	Transactions	transaction	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	1	2	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	1	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	1	4	Association	Association	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	1	5	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	1	6	Computational	Computational	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	1	7	Linguistics	Linguistics	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	1	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	1	9	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	1	10	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	1	11	2013	2013	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	1	12	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	1	13	243–254	243–254	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	1	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	2	1	Action	Action	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	2	2	Editor	Editor	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	2	3	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	2	4	Philipp	Philipp	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	2	5	Koehn	Koehn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	2	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	3	1	Submitted	submit	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	3	2	12/2012	12/2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	3	3	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	3	4	Revised	Revised	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	3	5	3/2013	3/2013	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	3	6	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	3	7	Published	publish	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	3	8	5/2013	5/2013	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	3	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	4	1	c	c	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	4	2	©	©	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	4	3	2013	2013	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	4	4	Association	Association	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	4	5	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	4	6	Computational	Computational	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	4	7	Linguistics	Linguistics	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	4	8	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	5	1	Unsupervised	Unsupervised	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	2	Tree	Tree	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	3	Induction	Induction	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	4	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	5	Tree	Tree	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	5	7	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	8	Translation	Translation	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	9	Feifei	Feifei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	10	Zhai	Zhai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	5	12	Jiajun	Jiajun	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	13	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	5	15	Yu	Yu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	16	Zhou	Zhou	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	17	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	5	18	Chengqing	Chengqing	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	19	Zong	Zong	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	20	National	National	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	21	Laboratory	Laboratory	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	22	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	23	Pattern	Pattern	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	24	Recognition	Recognition	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	25	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	5	26	Institute	Institute	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	27	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	28	Automation	Automation	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	29	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	5	30	Chinese	Chinese	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	31	Academy	Academy	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	32	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	33	Sciences	Sciences	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	5	34	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	5	35	Beijing	Beijing	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	36	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	5	37	China	China	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	38	{	{	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	5	39	ffzhai,jjzhang,yzhou,cqzong}@nlpr.ia.ac.cn	ffzhai,jjzhang,yzhou,cqzong}@nlpr.ia.ac.cn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	40	Abstract	Abstract	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	41	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	42	current	current	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	5	43	research	research	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	44	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	5	45	most	most	JJS
work_vbzp3iewy5h7rhfiok5o6i3du4	5	46	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	47	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	5	48	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	49	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	50	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	5	51	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	5	52	built	build	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	53	directly	directly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	5	54	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	55	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	5	56	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	5	57	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	6	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	2	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	6	3	study	study	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	6	5	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	6	6	go	go	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	6	7	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	8	another	another	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	6	9	direction	direction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	10	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	6	11	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	6	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	6	13	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	14	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	15	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	16	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	6	17	unsupervised	unsupervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	6	18	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	19	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	20	derived	derive	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	21	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	22	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	6	23	novel	novel	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	6	24	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	6	25	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	6	26	parametric	parametric	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	6	27	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	6	28	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	6	29	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	7	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	7	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	7	3	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	7	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	7	5	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	7	6	utilize	utilize	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	7	7	synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	7	8	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	7	9	substitution	substitution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	7	10	grammars	grammar	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	7	11	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	7	12	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	7	13	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	7	14	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	7	15	capture	capture	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	7	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	7	17	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	7	18	mapping	mapping	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	7	19	between	between	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	7	20	language	language	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	7	21	pairs	pair	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	7	22	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	8	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	8	2	train	train	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	8	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	8	4	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	8	5	efficiently	efficiently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	8	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	8	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	8	8	develop	develop	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	8	9	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	8	10	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	8	11	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	8	12	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	8	13	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	8	14	novel	novel	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	8	15	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	8	16	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	8	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	9	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	9	2	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	9	3	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	9	4	capable	capable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	9	5	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	9	6	exploring	explore	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	9	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	9	8	infinite	infinite	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	9	9	space	space	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	9	10	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	9	11	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	9	12	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	9	13	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	9	14	performing	perform	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	9	15	local	local	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	9	16	changes	change	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	9	17	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	9	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	9	19	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	9	20	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	9	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	10	1	Experimental	experimental	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	10	2	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	10	3	show	show	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	10	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	10	6	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	10	8	to-	to-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	9	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	10	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	11	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	12	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	10	13	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	10	14	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	10	15	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	16	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	10	17	significantly	significantly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	10	18	outperforms	outperform	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	10	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	10	20	strong	strong	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	10	21	baseline	baseline	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	10	22	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	23	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	10	24	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	25	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	10	26	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	27	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	28	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	10	29	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	10	30	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	10	31	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	11	1	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	11	2	Introduction	introduction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	3	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	4	recent	recent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	11	5	years	year	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	11	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	11	7	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	11	9	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	10	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	11	models1	models1	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	12	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	11	13	drawing	draw	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	11	14	more	more	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	11	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	11	16	more	more	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	11	17	attention	attention	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	18	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	11	20	community	community	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	21	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	22	statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	11	23	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	24	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	11	25	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	11	26	SMT	SMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	11	27	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	11	28	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	12	1	Due	due	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	2	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	3	their	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	12	4	remarkable	remarkable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	12	5	ability	ability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	6	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	12	7	incorporate	incorporate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	12	8	context	context	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	9	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	10	information	information	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	12	12	long	long	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	12	13	distance	distance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	14	reordering	reorder	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	12	15	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	12	17	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	18	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	19	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	12	20	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	12	22	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	23	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	24	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	12	25	have	have	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	26	shown	show	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	27	promising	promising	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	12	28	progress	progress	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	29	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	30	improving	improve	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	12	31	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	32	quality	quality	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	12	33	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	12	34	Liu	Liu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	35	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	36	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	37	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	38	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	12	39	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	12	40	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	12	41	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	12	42	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	12	43	Quirk	Quirk	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	44	et	et	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	12	45	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	46	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	47	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	12	48	2005	2005	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	12	49	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	12	50	Galley	Galley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	51	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	52	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	53	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	54	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	12	55	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	12	56	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	12	57	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	12	58	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	12	59	Marcu	Marcu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	60	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	61	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	62	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	63	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	12	64	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	12	65	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	12	66	Shen	Shen	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	67	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	68	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	69	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	70	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	12	71	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	12	72	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	12	73	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	74	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	75	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	76	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	12	77	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	12	78	2011b	2011b	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	12	79	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	12	80	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	13	1	However	however	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	13	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	13	3	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	13	5	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	6	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	7	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	13	8	always	always	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	13	9	suffer	suffer	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	13	10	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	11	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	13	12	major	major	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	13	13	challenges	challenge	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	13	14	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	13	15	1	1	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	13	16	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	13	17	They	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	13	18	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	13	19	usually	usually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	13	20	built	build	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	21	directly	directly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	13	22	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	23	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	24	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	13	25	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	13	26	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	13	27	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	13	28	generated	generate	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	29	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	30	supervised	supervise	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	13	31	linguistic	linguistic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	13	32	parsers	parser	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	13	33	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	14	1	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	14	2	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	14	3	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	14	5	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	6	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	7	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	8	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	14	9	defined	define	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	10	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	11	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	14	12	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	13	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	14	14	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	15	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	14	16	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	17	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	14	18	side	side	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	14	19	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	14	20	both	both	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	14	21	sides	side	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	14	22	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	15	1	However	however	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	15	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	15	3	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	15	4	many	many	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	15	5	language	language	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	15	6	pairs	pair	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	15	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	15	8	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	15	9	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	15	10	difficult	difficult	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	15	11	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	15	12	acquire	acquire	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	15	13	such	such	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	15	14	corresponding	corresponding	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	15	15	linguistic	linguistic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	15	16	parsers	parser	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	15	17	due	due	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	15	18	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	15	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	15	20	lack	lack	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	15	21	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	15	22	Tree	Tree	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	15	23	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	15	24	bank	bank	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	15	25	resources	resource	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	15	26	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	15	27	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	15	28	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	16	1	2	2	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	16	2	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	16	3	Parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	16	4	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	16	5	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	16	6	actually	actually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	16	7	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	16	8	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	16	9	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	16	10	model	model	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	16	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	16	12	explain	explain	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	16	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	16	14	monolingual	monolingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	16	15	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	16	16	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	16	17	rather	rather	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	16	18	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	16	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	16	20	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	16	21	mapping	mapping	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	16	22	between	between	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	16	23	language	language	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	16	24	pairs	pair	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	16	25	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	17	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	17	2	indicates	indicate	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	17	3	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	17	4	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	17	5	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	17	6	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	17	7	usually	usually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	17	8	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	17	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	17	10	optimal	optimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	17	11	choice	choice	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	17	12	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	17	13	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	17	14	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	17	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	17	16	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	17	17	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	17	18	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	17	19	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	17	20	Wang	Wang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	17	21	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	17	22	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	17	23	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	17	24	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	17	25	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	17	26	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	17	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	18	1	Based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	2	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	18	4	above	above	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	18	5	analysis	analysis	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	18	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	18	8	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	18	9	conclude	conclude	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	18	10	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	18	12	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	13	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	14	that	that	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	18	15	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	18	16	independent	independent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	18	17	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	18	Tree	Tree	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	18	19	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	18	20	bank	bank	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	21	resources	resource	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	18	22	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	18	23	simultaneously	simultaneously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	18	24	considers	consider	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	18	25	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	18	26	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	18	27	mapping	mapping	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	28	inside	inside	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	29	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	18	30	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	18	31	sentence	sentence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	32	pairs	pair	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	18	33	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	18	34	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	18	35	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	18	36	good	good	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	18	37	choice	choice	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	38	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	39	building	build	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	18	40	tree-	tree-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	41	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	42	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	18	43	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	18	44	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	19	1	Therefore	therefore	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	19	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	19	3	complying	comply	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	19	4	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	19	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	19	6	above	above	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	19	7	conditions	condition	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	19	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	19	9	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	19	10	propose	propose	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	19	11	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	19	12	unsupervised	unsupervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	19	13	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	19	14	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	19	15	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	19	16	tree-	tree-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	19	17	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	19	18	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	19	19	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	19	20	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	19	21	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	19	22	study	study	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	19	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	20	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	20	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	20	3	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	20	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	20	5	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	20	6	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	20	7	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	20	8	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	20	9	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	20	10	combining	combine	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	20	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	20	12	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	20	13	classes	class	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	20	14	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	20	15	their	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	20	16	boundary	boundary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	20	17	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	20	18	rather	rather	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	20	19	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	20	20	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	20	21	syntactic	syntactic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	20	22	labels	label	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	20	23	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	20	24	such	such	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	20	25	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	20	26	NP	NP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	20	27	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	20	28	VP	VP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	20	29	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	21	1	Furthermore	furthermore	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	21	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	21	3	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	21	4	these	these	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	21	5	node	node	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	21	6	labels	label	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	21	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	21	8	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	21	9	design	design	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	21	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	21	11	generative	generative	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	21	12	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	21	13	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	21	14	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	21	15	infer	infer	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	21	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	21	17	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	21	18	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	21	19	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	21	20	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	21	21	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	21	22	synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	21	23	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	21	24	substitution	substitution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	21	25	grammars	grammar	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	21	26	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	21	27	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	21	28	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	21	29	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	21	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	22	1	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	22	2	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	22	3	derived	derive	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	22	4	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	22	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	22	6	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	22	7	alignments	alignment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	22	8	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	22	9	thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	22	10	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	22	11	grasp	grasp	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	22	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	22	13	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	22	14	mapping	mapping	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	22	15	effectively	effectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	22	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	23	1	Training	train	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	23	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	23	3	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	23	4	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	23	5	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	23	6	difficult	difficult	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	23	7	due	due	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	23	8	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	23	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	23	10	exponential	exponential	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	23	11	space	space	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	23	12	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	23	13	possible	possible	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	23	14	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	23	15	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	23	16	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	23	17	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	23	18	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	23	19	instance	instance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	23	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	24	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	24	2	therefore	therefore	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	24	3	develop	develop	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	24	4	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	24	5	efficient	efficient	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	24	6	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	24	7	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	24	8	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	24	9	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	24	10	novel	novel	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	24	11	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	24	12	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	24	13	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	24	14	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	24	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	25	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	25	2	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	25	3	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	25	4	capable	capable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	25	5	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	25	6	exploring	explore	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	25	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	25	8	infinite	infinite	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	25	9	space	space	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	25	10	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	25	11	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	25	12	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	25	13	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	25	14	performing	perform	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	25	15	local	local	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	25	16	changes	change	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	25	17	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	25	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	25	19	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	25	20	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	25	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	26	1	2	2	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	26	2	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	26	3	believe	believe	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	26	4	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	26	5	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	26	6	possible	possible	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	26	7	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	26	8	design	design	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	26	9	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	26	10	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	26	11	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	26	12	infer	infer	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	26	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	26	14	node	node	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	26	15	label	label	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	26	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	26	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	26	18	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	26	19	jointly	jointly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	26	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	27	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	27	2	plan	plan	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	27	3	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	27	4	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	5	future	future	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	27	6	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	27	8	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	27	9	here	here	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	27	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	27	11	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	27	12	focus	focus	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	27	13	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	27	14	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	15	inferring	infer	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	27	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	27	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	18	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	19	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	20	terms	term	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	27	21	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	22	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	27	23	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	24	labels	label	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	27	25	derived	derive	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	26	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	27	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	27	28	classes	class	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	27	29	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	28	1	243	243	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	28	2	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	28	3	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	28	4	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	28	5	formed	form	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	28	6	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	28	7	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	28	8	way	way	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	28	9	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	28	10	independent	independent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	28	11	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	28	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	28	13	Tree	Tree	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	28	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	28	15	bank	bank	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	28	16	resources	resource	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	28	17	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	28	18	simultaneously	simultaneously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	28	19	exploits	exploit	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	28	20	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	28	21	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	28	22	mapping	mapping	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	28	23	effectively	effectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	28	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	29	1	Experiments	experiment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	29	2	show	show	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	29	3	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	29	5	proposed	propose	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	6	unsupervised	unsupervise	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	7	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	8	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	29	9	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	29	11	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	12	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	29	13	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	29	14	more	more	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	29	15	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	29	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	29	17	reasonable	reasonable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	29	18	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	19	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	20	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	29	21	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	22	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	23	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	24	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	29	25	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	26	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	29	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	30	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	30	2	main	main	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	30	3	contributions	contribution	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	30	4	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	5	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	30	6	study	study	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	7	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	30	8	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	9	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	30	10	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	30	11	1	1	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	30	12	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	30	13	Instead	instead	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	30	14	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	30	16	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	18	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	30	19	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	30	20	propose	propose	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	30	21	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	30	22	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	30	23	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	24	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	30	25	induce	induce	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	30	26	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	30	27	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	28	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	30	29	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	30	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	31	tree-	tree-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	30	32	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	33	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	30	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	31	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	31	2	U	U	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	31	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	31	4	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	31	5	exploits	exploit	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	31	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	31	7	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	31	8	mapping	mapping	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	31	9	effectively	effectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	31	10	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	31	11	does	do	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	31	12	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	31	13	rely	rely	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	31	14	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	31	15	any	any	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	31	16	Tree	Tree	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	31	17	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	31	18	bank	bank	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	31	19	resources	resource	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	31	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	32	1	2	2	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	32	2	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	32	3	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	32	4	design	design	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	32	5	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	32	6	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	32	7	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	32	8	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	32	9	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	32	10	novel	novel	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	32	11	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	32	12	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	32	13	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	32	14	train	train	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	32	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	32	16	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	32	17	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	32	18	efficiently	efficiently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	32	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	33	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	33	2	remainder	remainder	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	33	3	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	33	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	33	5	paper	paper	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	33	6	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	33	7	organized	organize	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	33	8	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	33	9	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	33	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	34	1	Section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	34	2	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	34	3	introduces	introduce	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	34	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	34	5	related	related	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	34	6	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	34	7	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	35	1	Section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	35	2	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	35	3	describes	describe	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	35	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	35	5	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	35	6	generation	generation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	35	7	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	35	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	35	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	35	10	Section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	35	11	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	35	12	depicts	depict	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	35	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	35	14	adopted	adopt	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	35	15	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	35	16	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	35	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	36	1	Section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	36	2	5	5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	36	3	describes	describe	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	36	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	36	5	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	36	6	sampling	sample	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	36	7	algorithm	algorithm	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	36	8	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	36	9	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	36	10	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	36	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	37	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	37	2	Section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	37	3	6	6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	37	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	37	5	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	37	6	analyze	analyze	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	37	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	37	8	achieved	achieve	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	37	9	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	37	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	37	11	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	37	12	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	37	13	evaluate	evaluate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	37	14	their	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	37	15	effectiveness	effectiveness	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	37	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	38	1	Finally	finally	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	38	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	38	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	38	4	conclude	conclude	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	38	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	38	6	paper	paper	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	38	7	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	38	8	Section	Section	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	38	9	7	7	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	38	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	39	1	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	39	2	Related	related	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	39	3	Work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	4	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	5	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	39	6	study	study	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	39	8	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	39	9	move	move	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	39	10	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	11	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	39	12	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	39	13	direction	direction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	14	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	39	15	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	39	16	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	39	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	18	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	39	19	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	20	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	21	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	22	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	23	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	39	24	unsupervised	unsupervise	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	25	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	26	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	39	27	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	39	28	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	39	29	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	40	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	40	2	unsupervised	unsupervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	40	3	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	40	4	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	40	5	induction	induction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	40	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	40	7	DeNero	DeNero	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	40	8	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	40	9	Uszkoreit	Uszkoreit	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	40	10	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	40	11	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	40	12	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	40	13	adopted	adopt	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	40	14	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	40	15	parallel	parallel	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	40	16	parsing	parse	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	40	17	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	40	18	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	40	19	induce	induce	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	40	20	unlabeled	unlabeled	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	40	21	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	40	22	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	40	23	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	40	24	sentences	sentence	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	40	25	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	40	26	syntactic	syntactic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	40	27	pre	pre	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	40	28	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	40	29	reordering	reordering	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	40	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	41	1	Our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	41	2	previous	previous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	41	3	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	41	4	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	41	5	Zhai	Zhai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	41	6	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	41	7	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	41	8	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	41	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	41	10	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	41	11	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	41	12	designed	design	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	41	13	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	41	14	EM-	EM-	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	41	15	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	41	16	method	method	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	41	17	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	41	18	construct	construct	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	41	19	unsupervised	unsupervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	41	20	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	41	21	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	41	22	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	41	23	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	41	24	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	41	25	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	41	26	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	41	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	42	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	42	2	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	3	differs	differ	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	42	4	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	42	6	above	above	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	42	7	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	8	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	9	that	that	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	42	10	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	42	11	design	design	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	42	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	42	13	novel	novel	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	42	14	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	42	15	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	16	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	42	17	induce	induce	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	42	18	unsupervised	unsupervise	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	19	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	20	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	42	21	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	42	22	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	42	23	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	42	24	prior	prior	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	42	25	knowledge	knowledge	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	26	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	42	27	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	42	28	encoded	encode	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	29	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	30	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	42	31	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	42	32	more	more	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	42	33	freely	freely	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	42	34	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	42	35	effectively	effectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	42	36	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	43	1	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	43	2	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	43	3	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	43	4	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	44	1	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	44	2	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	44	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	44	4	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	44	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	44	6	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	44	7	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	44	8	utilized	utilize	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	44	9	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	44	10	methods	method	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	44	11	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	44	12	learn	learn	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	44	13	synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	44	14	context	context	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	44	15	free	free	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	44	16	grammars	grammar	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	44	17	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	44	18	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	44	19	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	44	20	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	44	21	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	44	22	parallel	parallel	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	44	23	corpus	corpus	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	44	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	45	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	45	2	obtained	obtained	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	45	3	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	45	4	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	45	5	further	far	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	45	6	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	45	7	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	45	8	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	45	9	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	45	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	45	11	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	45	12	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	45	13	hierarchical	hierarchical	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	45	14	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	45	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	45	16	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	45	17	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	45	18	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	45	19	Chiang	Chiang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	45	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	45	21	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	45	22	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	45	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	46	1	Levenberg	Levenberg	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	46	2	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	46	3	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	46	4	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	47	1	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	47	2	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	47	3	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	47	4	employed	employ	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	47	5	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	47	6	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	47	7	method	method	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	47	8	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	47	9	learn	learn	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	47	10	discontinuous	discontinuous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	47	11	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	47	12	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	47	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	48	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	48	2	study	study	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	3	differs	differ	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	48	4	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	5	their	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	48	6	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	7	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	8	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	48	9	concentrate	concentrate	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	48	10	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	11	constructing	construct	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	48	12	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	13	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	48	14	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	15	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	16	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	48	17	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	18	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	48	19	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	48	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	49	1	Our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	49	2	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	49	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	49	4	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	49	5	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	49	6	learned	learn	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	49	7	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	49	8	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	49	9	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	49	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	49	11	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	49	12	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	49	13	more	more	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	49	14	appropriate	appropriate	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	49	15	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	49	16	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	49	17	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	49	18	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	49	19	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	49	20	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	49	21	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	49	22	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	49	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	50	1	Burkett	Burkett	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	50	2	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	50	3	Klein	Klein	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	50	4	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	50	5	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	50	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	50	7	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	50	8	Burkett	Burkett	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	50	9	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	50	10	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	50	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	51	1	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	51	2	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	51	3	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	51	4	focused	focus	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	51	5	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	51	6	joint	joint	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	51	7	parsing	parsing	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	51	8	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	51	9	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	51	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	52	1	They	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	52	2	utilized	utilize	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	52	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	52	4	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	52	5	Tree	Tree	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	52	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	52	7	bank	bank	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	52	8	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	52	9	train	train	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	52	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	52	11	joint	joint	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	52	12	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	52	13	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	52	14	both	both	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	52	15	parsing	parsing	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	52	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	52	17	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	52	18	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	52	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	53	1	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	53	2	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	53	3	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	53	4	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	53	5	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	53	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	53	7	adopted	adopt	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	53	8	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	53	9	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	53	10	method	method	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	53	11	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	53	12	infer	infer	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	53	13	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	53	14	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	53	15	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	53	16	exploring	explore	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	53	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	53	18	space	space	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	53	19	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	53	20	alignments	alignment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	53	21	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	53	22	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	53	23	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	53	24	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	53	25	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	54	1	Liu	Liu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	54	2	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	54	3	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	54	4	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	55	1	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	55	2	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	55	3	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	55	4	re	re	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	55	5	-	-	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	55	6	trained	train	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	55	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	55	8	linguistic	linguistic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	55	9	parsers	parser	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	55	10	bilingually	bilingually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	55	11	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	55	12	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	55	13	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	55	14	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	55	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	56	1	Burkett	Burkett	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	56	2	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	56	3	Klein	Klein	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	56	4	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	56	5	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	56	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	56	7	utilized	utilize	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	56	8	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	56	9	transformation	transformation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	56	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	56	11	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	56	12	method	method	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	56	13	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	56	14	learn	learn	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	56	15	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	56	16	sequence	sequence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	56	17	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	56	18	monolingual	monolingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	56	19	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	56	20	transformations	transformation	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	56	21	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	56	22	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	56	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	57	1	Compared	compare	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	2	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	3	their	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	57	4	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	57	6	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	57	7	do	do	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	57	8	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	57	9	rely	rely	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	57	10	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	11	any	any	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	57	12	Tree	Tree	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	57	13	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	57	14	bank	bank	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	15	resources	resource	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	57	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	57	17	focus	focus	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	57	18	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	19	generating	generate	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	57	20	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	57	21	unsupervised	unsupervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	57	22	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	23	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	57	24	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	25	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	26	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	57	27	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	28	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	57	29	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	57	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	58	1	Zollmann	Zollmann	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	58	2	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	58	3	Venugopal	Venugopal	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	58	4	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	58	5	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	58	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	58	7	substituted	substitute	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	58	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	58	9	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	58	10	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	58	11	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	58	12	X	x	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	58	13	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	58	14	hierarchical	hierarchical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	58	15	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	58	16	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	58	17	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	58	18	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	58	19	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	58	20	extended	extended	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	58	21	syntactic	syntactic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	58	22	categories	category	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	58	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	59	1	Zollmann	Zollmann	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	59	2	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	59	3	Vogel	Vogel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	59	4	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	59	5	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	59	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	59	7	further	further	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	59	8	labeled	label	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	59	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	59	10	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	59	11	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	59	12	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	59	13	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	59	14	tags	tag	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	59	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	59	16	unsupervised	unsupervise	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	59	17	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	59	18	classes	class	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	59	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	60	1	Our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	60	2	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	3	differs	differ	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	60	4	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	5	theirs	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	60	6	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	7	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	8	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	60	9	present	present	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	60	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	60	11	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	60	12	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	13	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	60	14	learn	learn	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	60	15	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	60	16	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	60	17	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	18	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	60	19	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	60	20	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	60	22	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	23	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	60	24	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	25	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	26	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	60	27	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	28	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	29	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	60	30	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	60	31	rather	rather	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	60	32	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	33	designing	design	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	60	34	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	60	35	labeling	labeling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	36	strategy	strategy	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	37	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	38	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	60	39	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	60	40	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	61	1	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	61	2	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	61	3	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	61	4	Generation	Generation	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	61	5	Process	Process	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	61	6	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	7	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	61	8	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	61	10	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	61	11	induce	induce	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	61	12	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	61	13	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	61	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	61	15	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	61	16	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	61	18	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	19	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	61	20	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	61	22	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	23	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	24	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	25	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	61	26	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	61	27	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	61	28	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	29	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	30	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	61	31	synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	61	32	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	33	substitution	substitution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	34	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	35	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	61	36	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	61	37	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	61	38	between	between	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	39	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	40	strings	string	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	61	41	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	61	42	target	target	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	61	43	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	61	44	fragments	fragment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	61	45	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	62	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	62	2	take	take	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	62	3	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	62	4	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	62	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	62	6	generation	generation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	62	7	grammar	grammar	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	62	8	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	62	9	match	match	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	62	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	62	11	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	62	12	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	62	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	63	1	Typically	typically	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	63	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	63	3	such	such	PDT
work_vbzp3iewy5h7rhfiok5o6i3du4	63	4	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	63	5	STSG3	stsg3	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	6	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	63	7	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	63	8	5-tuple	5-tuple	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	63	9	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	10	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	63	11	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	63	12	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	63	13	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	63	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	63	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	63	16	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	63	17	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	63	18	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	19	t	t	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	20	t	t	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	21	tG	tG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	22	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	23	S	S	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	24	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	25	¦	¦	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	26	¦	¦	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	27	where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	63	28	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	63	29	i	i	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	30	s	s	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	63	31	¦	¦	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	32	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	63	33	t	t	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	34	¦	¦	:
work_vbzp3iewy5h7rhfiok5o6i3du4	63	35	represent	represent	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	63	36	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	63	37	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	38	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	39	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	40	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	63	41	target	target	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	63	42	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	63	43	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	63	44	respectively	respectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	63	45	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	63	46	i	i	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	47	tN	tN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	48	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	63	49	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	63	50	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	51	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	52	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	53	non	non	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	63	54	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	63	55	terminals	terminal	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	63	56	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	63	57	i	i	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	58	t	t	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	63	59	tS	tS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	60	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	61	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	62	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	63	63	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	63	64	start	start	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	65	root	root	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	66	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	63	67	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	63	68	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	63	69	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	63	70	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	63	71	i	i	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	63	72	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	73	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	63	74	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	63	75	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	76	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	77	set	set	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	63	78	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	64	1	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	64	2	Generally	generally	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	64	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	64	4	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	64	5	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	64	6	involves	involve	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	64	7	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	64	8	fragments	fragment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	64	9	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	64	10	both	both	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	64	11	sides	side	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	64	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	65	1	Here	here	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	65	2	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	65	3	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	65	4	consider	consider	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	65	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	65	6	special	special	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	65	7	case	case	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	65	8	where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	65	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	65	10	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	65	11	side	side	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	65	12	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	65	13	actually	actually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	65	14	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	65	15	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	65	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	66	1	244	244	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	66	2	Apart	apart	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	66	3	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	66	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	66	5	start	start	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	66	6	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	66	7	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	66	8	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	66	9	tS	tS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	66	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	66	11	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	66	12	define	define	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	66	13	all	all	PDT
work_vbzp3iewy5h7rhfiok5o6i3du4	66	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	66	15	other	other	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	66	16	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	66	17	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	66	18	terminals	terminal	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	66	19	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	66	20	tN	tN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	66	21	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	66	22	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	66	23	classes	class	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	66	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	67	1	Inspired	inspire	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	67	2	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	67	3	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	67	4	Zollmann	Zollmann	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	67	5	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	67	6	Vogel	Vogel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	67	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	67	8	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	67	9	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	67	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	67	11	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	67	12	divide	divide	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	67	13	these	these	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	67	14	non	non	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	67	15	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	67	16	terminals	terminal	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	67	17	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	67	18	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	67	19	categories	category	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	67	20	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	67	21	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	67	22	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	67	23	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	67	24	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	67	25	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	67	26	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	67	27	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	67	28	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	67	29	multi	multi	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	67	30	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	67	31	word	word	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	67	32	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	67	33	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	67	34	terminals	terminal	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	67	35	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	68	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	68	2	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	68	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	68	4	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	5	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	68	6	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	68	7	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	68	8	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	68	9	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	68	10	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	11	class	class	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	12	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	68	13	such	such	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	68	14	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	15	C	C	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	68	16	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	68	17	meaning	mean	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	68	18	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	19	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	68	20	dominates	dominate	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	68	21	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	68	22	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	23	whose	whose	WP$
work_vbzp3iewy5h7rhfiok5o6i3du4	68	24	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	25	class	class	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	26	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	68	27	C.	C.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	68	28	Two	Two	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	68	29	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	68	30	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	31	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	68	32	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	68	33	terminals	terminal	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	68	34	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	68	35	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	36	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	68	37	stand	stand	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	68	38	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	39	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	68	40	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	68	41	strings	string	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	68	42	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	69	1	They	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	69	2	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	69	3	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	69	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	69	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	69	6	form	form	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	69	7	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	69	8	C1+C2	C1+C2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	69	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	69	10	where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	69	11	C1	C1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	69	12	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	69	13	C2	C2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	69	14	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	69	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	69	16	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	69	17	classes	class	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	69	18	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	69	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	69	20	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	69	21	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	69	22	separately	separately	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	69	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	70	1	Accordingly	accordingly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	70	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	70	3	multi	multi	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	70	4	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	70	5	word	word	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	70	6	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	70	7	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	70	8	terminals	terminal	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	70	9	represent	represent	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	70	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	70	11	strings	string	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	70	12	containing	contain	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	70	13	more	more	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	70	14	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	70	15	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	70	16	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	70	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	71	1	They	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	71	2	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	71	3	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	71	4	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	71	5	C1	c1	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	71	6	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	71	7	Cn	cn	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	71	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	71	9	demanding	demand	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	71	10	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	71	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	71	12	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	71	13	classes	class	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	71	14	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	71	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	71	16	leftmost	leftmost	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	71	17	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	71	18	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	71	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	71	20	rightmost	rightmost	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	71	21	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	71	22	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	71	23	C1	C1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	71	24	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	71	25	Cn	Cn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	71	26	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	71	27	respectively	respectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	71	28	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	72	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	72	2	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	72	3	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	72	4	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	72	5	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	72	6	play	play	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	72	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	72	8	role	role	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	72	9	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	72	10	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	72	11	class4	class4	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	72	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	73	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	73	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	73	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	73	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	73	5	head	head	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	73	6	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	73	7	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	73	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	73	9	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	73	10	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	73	11	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	73	12	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	73	13	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	73	14	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	73	15	multi	multi	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	73	16	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	73	17	word	word	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	73	18	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	73	19	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	73	20	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	73	21	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	73	22	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	73	23	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	73	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	74	1	It	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	74	2	requires	require	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	74	3	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	74	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	74	5	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	74	6	tags	tag	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	74	7	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	74	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	74	9	leftmost	leftmost	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	74	10	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	74	11	rightmost	rightmost	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	74	12	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	74	13	must	must	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	74	14	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	74	15	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	74	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	74	17	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	74	18	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	74	19	respectively	respectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	74	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	75	1	Xiong	Xiong	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	75	2	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	75	3	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	75	4	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	76	1	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	76	2	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	76	3	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	76	4	showed	show	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	76	5	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	76	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	76	7	boundary	boundary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	76	8	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	76	9	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	76	10	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	76	11	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	76	12	indicator	indicator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	76	13	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	76	14	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	76	15	reordering	reordering	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	76	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	77	1	Thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	77	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	77	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	77	4	believe	believe	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	77	5	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	77	6	combining	combine	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	77	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	77	8	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	77	9	class	class	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	77	10	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	77	11	boundary	boundary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	77	12	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	77	13	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	77	14	denote	denote	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	77	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	77	16	whole	whole	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	77	17	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	77	18	well	well	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	77	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	78	1	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	2	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	78	3	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	4	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	5	PRP	PRP	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	78	6	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	7	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	78	8	x0	x0	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	9	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	10	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	78	11	x1	x1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	12	VBP+RB	VBP+RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	13	ᡁԜ	ᡁԜ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	14	x1	x1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	15	x0	x0	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	16	wo	wo	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	78	17	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	78	18	men	man	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	78	19	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	78	20	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	78	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	79	1	An	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	79	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	79	3	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	79	4	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	79	5	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	79	6	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	79	7	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	79	8	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	80	1	Each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	80	2	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	3	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	5	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	6	consists	consist	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	80	7	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	8	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	80	9	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	10	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	80	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	80	13	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	14	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	15	fragment	fragment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	80	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	81	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	81	3	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	4	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	5	fragment	fragment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	81	7	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	81	8	internal	internal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	81	9	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	10	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	81	11	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	12	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	13	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	81	14	non-	non-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	15	terminal	terminal	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	16	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	17	tN	tN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	81	18	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	81	19	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	81	20	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	81	21	leaf	leaf	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	22	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	81	23	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	81	24	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	25	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	26	either	either	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	81	27	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	81	28	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	29	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	30	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	31	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	32	¦	¦	:
work_vbzp3iewy5h7rhfiok5o6i3du4	81	33	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	81	34	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	81	35	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	81	36	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	81	37	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	81	38	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	81	39	tN	tN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	81	40	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	82	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	82	2	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	82	3	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	82	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	82	5	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	82	6	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	82	7	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	82	8	comprises	comprise	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	82	9	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	82	10	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	82	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	82	12	variables	variable	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	82	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	83	1	Each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	83	2	variable	variable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	83	3	corresponds	correspond	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	83	4	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	83	5	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	83	6	leaf	leaf	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	83	7	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	83	8	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	83	9	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	83	10	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	83	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	83	12	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	83	13	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	83	14	fragment	fragment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	83	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	84	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	84	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	84	3	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	84	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	84	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	84	6	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	84	7	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	84	8	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	84	9	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	84	10	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	84	11	rewrite	rewrite	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	84	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	84	13	root	root	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	84	14	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	84	15	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	84	16	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	84	17	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	84	18	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	84	19	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	84	20	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	84	21	fragment	fragment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	84	22	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	85	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	85	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	85	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	85	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	85	5	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	85	6	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	85	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	85	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	85	9	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	85	10	rewrites	rewrite	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	85	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	85	12	head	head	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	85	13	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	85	14	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	85	15	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	85	16	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	85	17	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	85	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	85	19	corresponding	correspond	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	85	20	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	85	21	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	85	22	fragment	fragment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	85	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	86	1	An	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	86	2	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	86	3	derivation	derivation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	4	refers	refer	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	86	5	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	86	7	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	8	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	9	generating	generate	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	86	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	86	11	specific	specific	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	86	12	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	13	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	14	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	86	15	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	16	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	17	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	86	18	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	86	19	demand	demand	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	20	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	21	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	86	22	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	86	23	tagger	tagger	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	24	impairs	impairs	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	86	25	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	86	26	independence	independence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	27	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	28	manual	manual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	86	29	resources	resource	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	86	30	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	31	some	some	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	86	32	extent	extent	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	86	33	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	87	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	87	2	future	future	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	87	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	87	4	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	87	5	plan	plan	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	87	6	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	87	7	design	design	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	87	8	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	87	9	method	method	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	87	10	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	87	11	learn	learn	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	87	12	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	87	13	unsupervised	unsupervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	87	14	labels	label	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	87	15	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	87	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	87	17	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	87	18	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	87	19	terminals	terminal	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	87	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	88	1	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	88	2	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	88	3	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	88	4	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	88	5	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	89	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	89	2	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	89	3	begins	begin	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	89	4	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	89	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	89	6	start	start	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	89	7	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	89	8	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	89	9	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	89	10	tS	t	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	89	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	89	12	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	89	13	empty	empty	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	89	14	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	89	15	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	89	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	90	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	90	2	repeatedly	repeatedly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	90	3	choose	choose	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	90	4	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	90	5	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	90	6	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	90	7	rewrite	rewrite	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	90	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	90	9	leaf	leaf	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	90	10	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	90	11	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	90	12	terminals	terminal	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	90	13	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	90	14	expand	expand	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	90	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	90	16	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	90	17	until	until	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	90	18	no	no	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	90	19	leaf	leaf	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	90	20	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	90	21	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	90	22	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	90	23	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	90	24	left	leave	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	90	25	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	91	1	Finally	finally	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	91	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	91	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	91	4	acquire	acquire	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	91	5	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	91	6	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	91	7	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	91	8	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	91	9	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	91	10	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	91	11	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	91	12	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	91	13	defined	define	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	91	14	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	91	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	91	16	derivation	derivation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	91	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	92	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	92	2	probability	probability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	3	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	4	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	92	5	derivation	derivation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	6	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	92	7	given	give	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	8	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	9	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	92	10	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	92	11	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	92	12	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	92	13	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	92	14	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	92	15	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	16	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	92	17	n	n	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	92	18	i	i	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	19	i	i	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	20	i	i	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	21	p	p	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	22	d	d	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	23	p	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	24	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	25	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	26	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	27	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	92	28	1	1	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	92	29	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	92	30	where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	92	31	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	92	32	derivation	derivation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	33	comprises	comprise	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	92	34	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	92	35	sequence	sequence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	36	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	37	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	92	38	d=(r1,	d=(r1,	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	39	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	40	,rn	,rn	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	41	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	92	42	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	92	43	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	92	44	Ni	Ni	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	45	represents	represent	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	92	46	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	92	47	root	root	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	48	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	92	49	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	50	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	92	51	ri	ri	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	92	52	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	93	1	Hence	hence	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	93	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	93	3	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	4	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	93	5	specific	specific	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	93	6	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	93	7	sentence	sentence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	8	pair	pair	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	93	10	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	93	11	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	93	12	generate	generate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	93	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	93	14	best	good	JJS
work_vbzp3iewy5h7rhfiok5o6i3du4	93	15	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	16	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	93	17	side	side	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	18	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	19	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	20	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	21	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	22	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	93	23	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	93	24	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	93	25	independent	independent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	93	26	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	27	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	93	28	Tree-	Tree-	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	93	29	bank	bank	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	93	30	resources	resource	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	93	31	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	94	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	94	2	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	94	3	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	94	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	94	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	94	6	above	above	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	94	7	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	94	8	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	94	9	learned	learn	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	94	10	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	94	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	94	12	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	94	13	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	94	14	that	that	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	94	15	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	94	16	detailed	detail	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	94	17	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	94	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	94	19	next	next	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	94	20	section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	94	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	95	1	Actually	actually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	95	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	95	3	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	95	4	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	95	5	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	95	6	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	95	7	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	95	8	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	95	9	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	95	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	95	11	U-	u-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	95	12	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	95	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	96	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	96	2	do	do	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	96	3	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	96	4	use	use	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	96	5	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	96	6	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	96	7	most	most	JJS
work_vbzp3iewy5h7rhfiok5o6i3du4	96	8	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	96	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	96	10	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	96	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	96	12	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	96	13	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	96	14	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	96	15	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	96	16	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	96	17	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	96	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	97	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	97	2	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	97	3	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	97	4	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	97	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	97	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	97	7	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	97	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	97	9	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	97	10	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	97	11	optimized	optimize	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	97	12	through	through	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	97	13	selecting	select	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	97	14	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	97	15	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	97	16	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	97	17	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	97	18	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	97	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	98	1	These	these	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	98	2	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	98	3	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	98	4	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	98	5	consistent	consistent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	98	6	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	98	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	98	8	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	98	9	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	98	10	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	98	11	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	98	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	98	13	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	98	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	98	15	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	98	16	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	98	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	99	1	Another	another	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	99	2	reason	reason	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	99	3	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	99	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	99	5	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	99	6	has	have	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	99	7	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	99	8	stronger	strong	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	99	9	expressive	expressive	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	99	10	power	power	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	99	11	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	99	12	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	99	13	construction	construction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	99	14	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	99	15	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	99	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	100	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	2	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	100	3	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	100	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	100	5	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	6	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	100	8	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	9	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	100	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	100	11	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	100	12	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	13	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	100	14	although	although	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	15	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	100	16	linguistically	linguistically	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	100	17	informed	inform	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	18	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	100	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	100	20	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	100	21	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	22	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	23	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	100	24	tags	tag	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	100	25	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	100	26	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	100	27	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	100	28	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	100	29	distinguishing	distinguish	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	100	30	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	100	31	ones	one	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	100	32	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	101	1	However	however	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	101	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	101	3	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	101	4	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	101	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	101	6	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	101	7	have	have	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	101	8	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	101	9	discard	discard	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	101	10	all	all	PDT
work_vbzp3iewy5h7rhfiok5o6i3du4	101	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	101	12	internal	internal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	101	13	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	101	14	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	101	15	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	101	16	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	101	17	flattening	flatten	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	101	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	101	19	U-	u-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	101	20	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	101	21	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	101	22	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	101	23	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	101	24	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	101	25	express	express	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	101	26	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	101	27	same	same	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	101	28	sequence	sequence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	101	29	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	101	30	leading	lead	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	101	31	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	101	32	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	101	33	poor	poor	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	101	34	ability	ability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	101	35	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	101	36	distinguishing	distinguish	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	101	37	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	101	38	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	101	39	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	101	40	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	101	41	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	101	42	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	101	43	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	101	44	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	102	1	Thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	102	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	102	3	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	102	4	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	102	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	102	6	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	102	7	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	102	8	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	102	9	more	more	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	102	10	specific	specific	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	102	11	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	102	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	102	13	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	102	14	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	102	15	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	102	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	103	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	103	2	addition	addition	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	103	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	103	4	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	103	5	find	find	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	103	6	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	103	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	103	8	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	103	9	SCFG	scfg	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	103	10	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	103	11	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	103	12	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	103	13	even	even	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	103	14	significantly	significantly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	103	15	outperform	outperform	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	103	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	103	17	heuristic	heuristic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	103	18	SCFG	scfg	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	103	19	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	103	20	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	103	21	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	103	22	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	103	23	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	103	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	104	1	2009)5	2009)5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	104	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	105	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	105	2	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	105	3	indicate	indicate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	105	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	105	6	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	105	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	105	8	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	9	derivation	derivation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	10	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	11	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	12	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	13	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	105	14	product	product	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	15	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	105	16	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	105	17	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	105	18	such	such	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	105	19	good	good	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	105	20	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	21	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	22	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	105	23	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	24	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	105	25	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	105	26	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	106	1	Considering	consider	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	106	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	106	3	above	above	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	106	4	reasons	reason	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	106	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	106	6	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	106	7	believe	believe	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	106	8	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	106	10	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	106	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	106	12	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	13	learning	learning	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	14	procedure	procedure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	15	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	106	16	result	result	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	106	17	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	18	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	106	19	better	well	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	106	20	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	21	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	22	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	23	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	24	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	106	25	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	106	26	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	106	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	107	1	5	5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	107	2	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	3	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	107	4	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	5	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	6	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	7	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	107	9	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	107	10	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	107	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	107	12	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	13	Chinese	Chinese	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	107	15	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	16	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	107	17	English	english	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	107	18	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	19	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	107	20	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	107	21	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	107	22	SCFG	scfg	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	23	grammar	grammar	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	24	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	107	25	outperform	outperform	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	107	26	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	107	27	heuristic	heuristic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	107	28	SCFG	scfg	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	29	grammar	grammar	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	30	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	31	0.1	0.1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	107	32	BLEU	BLEU	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	33	points	point	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	107	34	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	35	NIST	NIST	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	36	MT	MT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	37	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	107	38	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	107	39	0.6	0.6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	107	40	BLEU	BLEU	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	41	points	point	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	107	42	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	43	NIST	NIST	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	44	MT	MT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	45	2005	2005	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	107	46	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	47	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	107	48	NEWS	NEWS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	107	49	domain	domain	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	107	50	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	108	1	245	245	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	108	2	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	108	3	Bayesian	Bayesian	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	108	4	Model	Model	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	108	5	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	108	6	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	108	7	section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	108	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	108	9	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	108	10	present	present	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	108	11	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	108	12	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	108	13	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	108	14	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	108	15	learn	learn	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	108	16	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	108	17	defined	define	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	108	18	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	108	19	section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	108	20	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	108	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	109	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	109	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	109	3	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	109	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	109	5	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	109	6	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	109	7	șN	șN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	109	8	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	109	9	denote	denote	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	109	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	109	11	probability	probability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	109	12	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	109	13	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	109	14	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	109	15	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	109	16	p	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	109	17	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	109	18	N	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	109	19	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	109	20	Equation	Equation	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	109	21	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	109	22	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	109	23	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	109	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	110	1	șN	șN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	2	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	110	3	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	110	4	multinomial	multinomial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	110	5	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	110	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	8	impose	impose	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	9	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	110	10	Dirichlet	Dirichlet	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	11	prior	prior	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	110	12	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	13	DP	dp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	14	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	15	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	16	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	17	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	110	18	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	110	19	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	110	20	|	|	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	110	21	~	~	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	22	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	23	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	24	|	|	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	110	25	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	110	26	~	~	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	27	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	28	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	110	29	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	30	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	31	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	32	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	33	N	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	34	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	35	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	36	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	37	r	r	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	38	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	39	Multi	Multi	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	40	P	P	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	41	DP	dp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	42	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	43	N	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	44	T	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	45	T	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	46	D	d	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	47	D	d	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	48	<	<	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	110	49	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	50	2	2	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	110	51	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	52	where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	110	53	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	110	54	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	55	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	56	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	57	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	58	N	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	59	<	<	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	110	60	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	61	base	base	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	62	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	63	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	110	64	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	110	65	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	66	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	110	67	assign	assign	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	110	68	prior	prior	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	110	69	probabilities	probability	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	110	70	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	71	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	110	72	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	110	73	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	110	74	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	110	75	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	111	1	ĮN	įn	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	111	2	controls	control	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	111	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	111	4	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	111	5	’s	’s	POS
work_vbzp3iewy5h7rhfiok5o6i3du4	111	6	tendency	tendency	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	111	7	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	111	8	either	either	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	111	9	reuse	reuse	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	111	10	existing	exist	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	111	11	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	111	12	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	111	13	create	create	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	111	14	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	111	15	ones	one	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	111	16	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	111	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	111	18	base	base	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	111	19	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	111	20	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	111	21	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	111	22	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	111	23	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	111	24	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	111	25	N	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	111	26	<	<	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	111	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	112	1	Instead	instead	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	112	2	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	112	3	denoting	denote	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	112	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	112	5	multinomial	multinomial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	112	6	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	112	7	explicitly	explicitly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	112	8	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	112	9	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	112	10	specific	specific	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	112	11	șN	șN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	112	12	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	112	13	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	112	14	integrate	integrate	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	112	15	over	over	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	112	16	all	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	112	17	possible	possible	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	112	18	values	value	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	112	19	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	112	20	șN	șN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	112	21	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	112	22	achieve	achieve	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	112	23	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	112	24	probabilities	probability	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	112	25	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	112	26	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	112	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	113	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	113	2	integration	integration	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	3	results	result	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	113	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	113	6	following	follow	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	113	7	conditional	conditional	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	113	8	probability	probability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	9	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	10	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	11	ri	ri	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	12	given	give	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	113	14	previously	previously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	113	15	observed	observe	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	16	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	113	17	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	18	-	-	:
work_vbzp3iewy5h7rhfiok5o6i3du4	113	19	i	i	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	20	=	=	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	113	21	r1	r1	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	22	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	113	23	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	24	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	113	25	ri-1	ri-1	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	26	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	113	27	00	00	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	28	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	113	29	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	30	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	113	31	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	113	32	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	33	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	113	34	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	113	35	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	113	36	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	113	37	i	i	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	38	i	i	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	39	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	40	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	41	ii	ii	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	42	i	i	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	43	N	n	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	113	44	i	i	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	45	N	n	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	113	46	N	n	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	113	47	n	n	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	113	48	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	49	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	50	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	51	p	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	52	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	53	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	54	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	55	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	56	n	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	57	D	d	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	58	D	d	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	59	D	d	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	60	�	�	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	61	�	�	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	113	62	�	�	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	113	63	�	�	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	113	64	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	65	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	113	66	3	3	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	113	67	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	113	68	Where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	113	69	n	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	70	-	-	:
work_vbzp3iewy5h7rhfiok5o6i3du4	113	71	i	i	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	72	ri	ri	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	113	73	denotes	denote	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	113	74	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	113	75	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	76	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	77	ri	ri	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	113	78	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	79	ir	ir	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	80	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	81	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	113	82	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	113	83	n	n	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	113	84	-i	-i	:
work_vbzp3iewy5h7rhfiok5o6i3du4	113	85	N	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	86	represents	represent	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	113	87	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	113	88	total	total	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	113	89	count	count	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	90	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	91	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	113	92	rewriting	rewrite	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	113	93	non-	non-	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	94	terminal	terminal	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	95	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	96	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	113	97	ir	ir	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	98	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	113	99	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	114	1	Thanks	thank	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	114	2	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	114	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	114	4	exchangeability	exchangeability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	114	5	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	114	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	114	7	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	114	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	114	9	all	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	114	10	permutations	permutation	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	114	11	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	114	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	114	13	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	114	14	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	114	15	actually	actually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	114	16	equiprobable	equiprobable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	114	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	115	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	115	2	means	mean	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	115	3	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	115	4	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	115	5	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	115	6	compute	compute	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	115	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	115	8	probability	probability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	115	9	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	115	10	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	115	11	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	115	12	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	115	13	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	115	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	115	15	previous	previous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	115	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	115	17	subsequent	subsequent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	115	18	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	115	19	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	115	20	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	116	1	consider	consider	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	116	2	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	116	3	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	116	4	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	116	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	116	6	last	last	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	116	7	one	one	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	116	8	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	116	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	117	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	117	2	characteristic	characteristic	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	117	3	allows	allow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	117	4	us	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	117	5	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	117	6	design	design	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	117	7	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	117	8	efficient	efficient	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	117	9	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	117	10	sampling	sample	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	117	11	algorithm	algorithm	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	117	12	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	117	13	train	train	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	117	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	117	15	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	117	16	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	117	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	118	1	4.1	4.1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	118	2	Base	Base	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	118	3	Distribution	Distribution	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	118	4	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	118	5	base	base	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	118	6	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	118	7	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	118	8	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	118	9	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	118	10	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	118	11	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	118	12	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	118	13	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	118	14	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	118	15	designed	design	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	118	16	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	118	17	assign	assign	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	118	18	prior	prior	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	118	19	probabilities	probability	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	118	20	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	118	21	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	118	22	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	118	23	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	118	24	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	118	25	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	119	1	Because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	2	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	119	3	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	4	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	5	consists	consist	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	119	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	7	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	119	8	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	9	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	10	fragment	fragment	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	11	frag	frag	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	12	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	119	13	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	119	14	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	15	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	16	str	str	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	17	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	119	19	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	119	21	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	22	follow	follow	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	23	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	24	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	119	25	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	26	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	27	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	119	28	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	29	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	119	30	decompose	decompose	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	119	31	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	119	32	prior	prior	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	119	33	probability	probability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	34	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	119	35	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	36	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	37	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	38	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	39	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	40	N	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	41	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	42	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	119	43	factors	factor	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	119	44	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	45	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	119	46	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	119	47	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	119	48	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	49	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	50	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	51	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	52	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	53	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	54	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	55	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	56	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	57	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	58	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	59	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	60	P	P	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	61	frag	frag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	62	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	63	P	P	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	64	str	str	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	65	frag	frag	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	66	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	67	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	68	4	4	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	119	69	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	70	where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	119	71	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	72	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	73	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	119	74	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	75	frag	frag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	76	N	n	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	119	77	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	119	78	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	119	79	probability	probability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	80	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	81	producing	produce	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	119	82	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	119	83	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	84	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	119	85	fragment	fragment	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	86	frag	frag	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	119	87	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	120	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	120	2	generate	generate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	120	3	frag	frag	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	120	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	120	5	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	120	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	120	7	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	120	8	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	120	9	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	120	10	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	120	11	used	use	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	120	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	120	13	geometric	geometric	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	120	14	prior	prior	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	120	15	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	120	16	decide	decide	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	120	17	how	how	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	120	18	many	many	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	120	19	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	120	20	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	120	21	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	120	22	assign	assign	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	120	23	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	120	24	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	120	25	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	121	1	Differently	differently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	121	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	121	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	121	4	require	require	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	121	5	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	121	6	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	121	7	multi	multi	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	121	8	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	121	9	word	word	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	121	10	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	121	11	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	121	12	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	121	13	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	121	14	must	must	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	121	15	have	have	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	121	16	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	121	17	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	121	18	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	121	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	122	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	122	2	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	122	3	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	122	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	122	5	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	122	6	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	122	7	has	have	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	122	8	been	be	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	122	9	verified	verify	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	122	10	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	122	11	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	122	12	very	very	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	122	13	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	122	14	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	122	15	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	122	16	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	122	17	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	122	18	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	122	19	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	122	20	Wang	Wang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	122	21	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	122	22	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	122	23	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	122	24	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	122	25	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	122	26	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	122	27	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	122	28	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	122	29	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	122	30	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	122	31	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	122	32	2011a	2011a	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	122	33	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	122	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	123	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	123	2	generation	generation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	123	3	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	123	4	starts	start	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	123	5	at	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	123	6	root	root	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	123	7	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	123	8	N.	N.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	123	9	At	At	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	123	10	first	first	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	123	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	123	12	root	root	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	123	13	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	123	14	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	123	15	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	123	16	expanded	expand	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	123	17	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	123	18	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	123	19	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	123	20	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	123	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	124	1	Then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	124	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	124	3	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	124	4	newly	newly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	124	5	generated	generate	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	124	6	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	124	7	will	will	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	124	8	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	124	9	checked	check	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	124	10	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	124	11	expand	expand	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	124	12	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	124	13	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	124	14	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	124	15	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	124	16	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	124	17	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	124	18	probability	probability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	124	19	pexpand	pexpand	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	124	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	125	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	125	2	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	125	3	repeats	repeat	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	125	4	until	until	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	125	5	all	all	PDT
work_vbzp3iewy5h7rhfiok5o6i3du4	125	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	125	7	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	125	8	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	125	9	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	125	10	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	125	11	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	125	12	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	125	13	checked	check	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	125	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	126	1	Obviously	obviously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	126	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	126	3	pexpand	pexpand	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	126	4	controls	control	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	126	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	126	6	scale	scale	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	126	7	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	126	8	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	126	9	fragments	fragment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	126	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	126	11	where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	126	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	126	13	large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	126	14	pexpand	pexpand	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	126	15	corresponds	correspond	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	126	16	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	126	17	large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	126	18	fragments	fragment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	126	19	6	6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	126	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	127	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	127	2	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	3	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	4	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	127	5	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	127	6	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	127	7	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	127	8	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	127	9	drawn	draw	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	10	uniformly	uniformly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	127	11	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	127	13	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	127	15	side	side	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	16	vocabulary	vocabulary	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	17	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	127	18	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	127	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	127	20	non-	non-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	21	terminal	terminal	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	22	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	127	23	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	127	24	created	create	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	25	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	26	asking	ask	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	127	27	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	127	28	questions	question	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	127	29	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	30	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	31	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	127	32	1	1	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	127	33	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	127	34	What	what	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	127	35	type	type	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	36	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	37	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	127	38	node	node	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	39	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	127	40	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	127	41	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	127	42	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	43	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	127	44	two-	two-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	45	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	127	46	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	127	47	multi	multi	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	48	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	49	word	word	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	50	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	51	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	52	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	127	53	?	?	.
work_vbzp3iewy5h7rhfiok5o6i3du4	128	1	2	2	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	128	2	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	128	3	What	what	WP
work_vbzp3iewy5h7rhfiok5o6i3du4	128	4	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	128	5	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	128	6	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	128	7	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	128	8	label	label	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	128	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	128	10	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	128	11	?	?	.
work_vbzp3iewy5h7rhfiok5o6i3du4	129	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	129	2	answer	answer	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	129	3	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	129	4	question	question	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	129	5	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	129	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	129	7	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	129	8	chosen	choose	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	129	9	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	129	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	129	11	uniform	uniform	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	129	12	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	129	13	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	129	14	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	129	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	129	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	129	17	probability	probability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	129	18	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	129	19	1/3	1/3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	129	20	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	129	21	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	129	22	type	type	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	129	23	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	129	24	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	129	25	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	129	26	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	129	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	130	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	130	2	entire	entire	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	130	3	generation	generation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	130	4	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	130	5	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	130	6	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	130	7	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	130	8	top	top	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	130	9	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	130	10	down	down	RP
work_vbzp3iewy5h7rhfiok5o6i3du4	130	11	manner	manner	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	130	12	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	130	13	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	130	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	130	15	generating	generate	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	130	16	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	130	17	parent	parent	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	130	18	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	130	19	first	first	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	130	20	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	130	21	then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	130	22	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	130	23	children	child	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	130	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	131	1	With	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	2	respect	respect	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	3	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	4	question	question	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	5	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	131	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	131	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	131	8	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	131	10	father	father	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	131	11	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	131	12	has	have	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	131	13	determined	determine	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	131	15	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	131	16	tags	tag	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	131	17	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	18	boundary	boundary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	131	19	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	131	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	131	21	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	131	22	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	131	23	need	need	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	131	24	one	one	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	25	POS	pos	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	26	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	27	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	131	28	generate	generate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	131	29	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	131	30	label	label	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	31	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	32	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	131	33	current	current	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	131	34	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	131	35	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	132	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	132	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	5	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	6	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	132	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	132	8	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	132	10	father	father	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	11	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	132	12	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	132	13	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	132	14	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	132	15	demands	demand	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	132	16	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	132	18	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	132	19	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	20	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	21	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	132	22	rightmost	rightmost	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	132	23	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	24	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	132	25	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	132	26	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	132	27	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	132	28	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	132	29	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	30	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	31	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	132	32	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	132	33	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	132	34	must	must	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	132	35	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	132	36	satisfy	satisfy	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	132	37	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	132	38	condition	condition	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	132	39	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	133	1	Therefore	therefore	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	133	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	133	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	133	4	choose	choose	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	133	5	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	133	6	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	133	7	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	133	8	VBP	vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	133	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	133	10	obtain	obtain	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	133	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	133	12	label	label	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	133	13	VBP+RB	VBP+RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	133	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	134	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	134	2	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	134	3	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	134	4	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	134	5	drawn	draw	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	134	6	uniformly	uniformly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	134	7	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	134	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	134	9	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	134	10	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	134	11	set	set	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	134	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	135	1	If	if	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	135	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	135	3	current	current	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	135	4	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	135	5	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	135	6	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	135	7	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	135	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	135	9	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	135	10	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	135	11	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	135	12	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	135	13	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	135	14	question	question	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	135	15	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	135	16	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	135	17	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	135	18	unnecessary	unnecessary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	135	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	136	1	Similarly	similarly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	136	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	136	3	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	4	respect	respect	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	5	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	136	7	two-	two-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	136	8	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	9	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	136	10	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	136	11	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	136	12	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	13	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	136	14	questions	question	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	136	15	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	136	16	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	136	17	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	136	18	2	2	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	136	19	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	136	20	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	136	21	both	both	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	136	22	unnecessary	unnecessary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	136	23	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	24	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	136	25	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	136	26	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	27	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	136	28	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	29	they	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	136	30	have	have	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	136	31	already	already	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	136	32	been	be	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	33	defined	define	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	34	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	35	their	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	136	36	father	father	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	136	37	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	136	38	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	137	1	As	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	137	2	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	137	3	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	137	4	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	137	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	137	6	generative	generative	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	137	7	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	137	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	137	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	137	10	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	137	11	fragment	fragment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	137	12	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	137	13	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	137	14	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	137	15	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	137	16	created	create	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	137	17	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	137	18	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	137	19	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	137	20	a.	a.	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	138	1	Determine	determine	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	138	2	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	138	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	138	4	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	138	5	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	138	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	138	7	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	138	8	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	138	9	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	138	10	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	138	11	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	138	12	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	138	13	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	138	14	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	138	15	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	138	16	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	138	17	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	138	18	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	138	19	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	138	20	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	138	21	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	138	22	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	138	23	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	138	24	b.	b.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	139	1	Expand	expand	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	139	2	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	139	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	139	4	generate	generate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	139	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	139	6	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	139	7	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	139	8	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	139	9	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	139	10	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	139	11	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	139	12	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	139	13	6	6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	139	14	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	139	15	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	139	16	experiment	experiment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	139	17	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	139	18	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	139	19	set	set	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	139	20	pexpand	pexpand	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	139	21	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	139	22	1/3	1/3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	139	23	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	139	24	encourage	encourage	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	139	25	small	small	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	139	26	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	139	27	fragments	fragment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	139	28	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	140	1	246	246	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	140	2	c.	c.	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	3	Determine	Determine	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	140	6	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	140	7	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	8	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	9	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	10	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	11	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	12	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	140	13	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	140	14	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	140	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	140	16	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	17	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	140	18	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	140	19	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	140	20	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	140	21	d.	d.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	22	Utilize	utilize	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	140	23	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	140	24	predetermined	predetermine	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	25	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	26	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	140	27	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	140	28	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	29	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	30	VBP	vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	31	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	140	32	form	form	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	140	33	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	140	34	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	35	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	36	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	140	37	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	140	38	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	140	39	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	40	non-	non-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	41	terminal	terminal	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	42	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	140	43	VBP+RB	VBP+RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	44	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	140	45	e.	e.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	46	Expand	Expand	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	47	VBP+RB	VBP+RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	48	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	140	49	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	140	50	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	51	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	140	52	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	53	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	140	54	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	140	55	f.	f.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	56	Do	do	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	57	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	140	58	expand	expand	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	140	59	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	60	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	140	61	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	140	62	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	141	1	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	141	2	|	|	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	141	3	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	141	4	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	141	5	str	str	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	141	6	frag	frag	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	141	7	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	141	8	Equation	Equation	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	141	9	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	141	10	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	141	11	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	141	12	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	141	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	141	14	probability	probability	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	141	15	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	141	16	generating	generate	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	141	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	141	18	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	141	19	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	141	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	141	21	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	141	22	contains	contain	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	141	23	several	several	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	141	24	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	141	25	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	141	26	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	141	27	variables	variable	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	141	28	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	142	1	Inspired	inspire	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	2	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	3	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	4	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	5	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	6	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	7	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	142	9	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	142	10	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	142	12	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	13	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	14	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	142	15	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	16	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	142	17	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	142	18	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	19	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	142	20	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	21	define	define	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	22	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	23	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	24	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	25	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	26	str	str	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	27	frag	frag	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	28	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	29	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	142	30	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	142	31	var	var	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	32	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	142	33	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	142	34	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	142	35	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	36	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	37	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	38	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	39	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	142	40	1	1	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	142	41	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	42	|	|	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	142	43	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	44	poisson	poisson	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	45	sw	sw	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	46	sw	sw	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	142	47	sw	sw	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	48	c	c	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	49	c	c	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	50	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	51	i	i	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	52	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	53	str	str	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	54	frag	frag	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	55	P	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	56	c	c	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	57	c	c	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	58	i	i	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	59	u	u	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	60	u	u	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	61	¦	¦	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	62	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	63	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	64	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	65	5	5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	142	66	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	142	67	where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	142	68	csw	csw	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	142	69	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	142	70	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	142	71	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	72	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	73	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	142	74	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	75	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	142	76	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	77	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	142	78	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	143	1	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	143	2	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	143	3	means	mean	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	143	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	143	5	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	143	6	vocabulary	vocabulary	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	143	7	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	143	8	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	144	1	Further	further	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	144	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	144	3	cvar	cvar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	144	4	denotes	denote	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	144	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	144	6	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	144	7	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	144	8	variables	variable	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	144	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	144	10	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	144	11	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	144	12	determined	determine	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	144	13	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	144	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	144	15	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	144	16	fragment	fragment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	144	17	frag	frag	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	144	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	145	1	As	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	145	2	shown	show	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	145	3	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	145	4	Equation(5	Equation(5	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	145	5	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	145	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	145	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	145	8	first	first	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	145	9	determine	determine	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	145	10	how	how	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	145	11	many	many	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	145	12	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	145	13	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	145	14	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	145	15	generate	generate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	145	16	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	145	17	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	145	18	Poisson	Poisson	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	145	19	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	145	20	Ppoisson(csw;1	ppoisson(csw;1	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	145	21	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	145	22	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	145	23	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	145	24	imposes	impose	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	145	25	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	145	26	stable	stable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	145	27	preference	preference	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	145	28	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	145	29	short	short	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	145	30	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	145	31	strings	string	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	145	32	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	146	1	Then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	146	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	146	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	146	4	draw	draw	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	146	5	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	146	6	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	146	7	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	146	8	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	146	9	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	146	10	uniform	uniform	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	146	11	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	146	12	over	over	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	146	13	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	146	14	s.	s.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	147	1	Afterwards	afterwards	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	147	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	147	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	147	4	insert	insert	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	147	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	147	6	variables	variable	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	147	7	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	147	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	147	9	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	147	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	148	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	148	2	variables	variable	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	148	3	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	148	4	inserted	insert	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	148	5	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	148	6	at	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	148	7	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	148	8	time	time	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	148	9	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	148	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	148	11	uniform	uniform	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	148	12	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	148	13	over	over	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	148	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	148	15	possible	possible	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	148	16	positions	position	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	148	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	149	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	149	2	factor	factor	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	149	3	discourages	discourage	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	149	4	more	more	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	149	5	variables	variable	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	149	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	150	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	150	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	150	3	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	150	4	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	150	5	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	150	6	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	150	7	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	150	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	150	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	150	10	generative	generative	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	150	11	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	150	12	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	150	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	150	14	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	150	15	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	150	16	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	150	17	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	150	18	a.	a.	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	151	1	Decide	decide	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	151	2	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	151	3	generate	generate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	151	4	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	151	5	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	151	6	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	151	7	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	151	8	b.	b.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	152	1	Generate	generate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	152	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	152	3	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	4	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	5	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	152	6	ᡁԜ	ᡁԜ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	152	7	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	152	8	wo	wo	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	9	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	152	10	men	man	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	152	11	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	152	12	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	152	13	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	152	14	c.	c.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	152	15	Insert	Insert	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	152	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	152	17	first	first	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	152	18	variable	variable	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	19	after	after	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	20	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	152	21	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	22	;	;	:
work_vbzp3iewy5h7rhfiok5o6i3du4	152	23	d.	d.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	152	24	Insert	Insert	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	152	25	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	152	26	second	second	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	152	27	variable	variable	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	28	between	between	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	29	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	152	30	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	31	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	152	32	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	152	33	first	first	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	152	34	variable	variable	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	152	35	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	153	1	Intuitively	intuitively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	153	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	153	3	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	153	4	good	good	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	153	5	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	153	6	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	153	7	should	should	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	153	8	carry	carry	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	153	9	both	both	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	153	10	small	small	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	153	11	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	153	12	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	153	13	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	153	14	enough	enough	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	153	15	generality	generality	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	153	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	153	17	large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	153	18	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	153	19	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	153	20	enough	enough	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	153	21	context	context	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	153	22	information	information	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	153	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	154	1	DeNero	DeNero	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	154	2	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	154	3	Klein	Klein	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	154	4	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	154	5	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	154	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	154	7	proposed	propose	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	154	8	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	154	9	statement	statement	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	154	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	154	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	154	12	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	154	13	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	154	14	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	154	15	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	154	16	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	154	17	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	154	18	has	have	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	154	19	verified	verify	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	154	20	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	154	21	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	154	22	their	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	154	23	experiments	experiment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	154	24	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	154	25	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	154	26	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	154	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	155	1	Our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	155	2	base	base	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	155	3	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	155	4	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	155	5	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	155	6	designed	design	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	155	7	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	155	8	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	155	9	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	155	10	intuition	intuition	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	155	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	156	1	Considering	consider	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	156	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	156	3	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	156	4	factors	factor	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	156	5	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	156	6	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	156	7	base	base	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	156	8	distribution	distribution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	156	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	156	10	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	156	11	penalize	penalize	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	156	12	both	both	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	156	13	large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	156	14	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	156	15	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	156	16	fragments	fragment	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	156	17	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	156	18	many	many	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	156	19	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	156	20	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	156	21	long	long	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	156	22	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	156	23	strings	string	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	156	24	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	156	25	many	many	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	156	26	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	156	27	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	156	28	variables	variable	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	156	29	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	157	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	157	2	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	157	3	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	157	4	tends	tend	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	157	5	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	157	6	select	select	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	157	7	both	both	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	157	8	small	small	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	157	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	157	10	frequent	frequent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	157	11	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	157	12	production	production	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	157	13	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	157	14	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	157	15	construct	construct	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	157	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	157	17	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	157	18	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	157	19	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	157	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	158	1	With	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	158	2	these	these	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	158	3	types	type	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	158	4	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	158	5	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	158	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	158	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	158	8	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	158	9	extract	extract	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	158	10	small	small	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	158	11	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	158	12	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	158	13	good	good	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	158	14	generality	generality	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	158	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	158	16	simultaneously	simultaneously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	158	17	obtain	obtain	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	158	18	large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	158	19	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	158	20	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	158	21	enough	enough	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	158	22	context	context	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	158	23	information	information	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	158	24	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	158	25	composition	composition	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	158	26	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	159	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	159	2	will	will	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	159	3	show	show	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	159	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	159	5	effectiveness	effectiveness	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	159	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	159	7	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	159	8	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	159	9	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	159	10	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	159	11	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	159	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	159	13	verification	verification	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	159	14	experiment	experiment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	159	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	160	1	5	5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	160	2	Model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	160	3	Training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	160	4	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	160	5	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	160	6	Sampling	Sampling	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	160	7	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	160	8	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	160	9	section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	160	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	160	11	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	160	12	introduce	introduce	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	160	13	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	160	14	collapsed	collapse	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	160	15	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	160	16	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	160	17	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	160	18	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	160	19	enables	enable	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	160	20	us	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	160	21	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	160	22	train	train	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	160	23	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	160	24	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	160	25	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	160	26	efficiently	efficiently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	160	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	161	1	5.1	5.1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	161	2	Initialization	Initialization	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	161	3	State	State	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	161	4	At	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	161	5	first	first	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	161	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	161	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	161	8	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	161	9	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	161	10	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	161	11	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	161	12	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	161	13	initialize	initialize	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	161	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	161	15	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	161	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	162	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	162	2	get	get	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	162	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	162	4	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	162	5	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	162	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	162	7	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	162	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	162	9	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	162	10	recursively	recursively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	162	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	162	12	randomly	randomly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	162	13	segment	segment	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	162	14	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	162	15	sentence	sentence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	162	16	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	162	17	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	162	18	parts	part	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	162	19	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	162	20	simultaneously	simultaneously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	162	21	create	create	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	162	22	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	162	23	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	162	24	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	162	25	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	162	26	dominate	dominate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	162	27	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	162	28	part	part	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	162	29	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	163	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	163	2	created	create	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	163	3	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	163	4	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	163	5	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	163	6	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	163	7	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	163	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	163	9	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	163	10	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	163	11	terminals	terminal	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	163	12	described	describe	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	163	13	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	163	14	section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	163	15	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	163	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	164	1	Using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	164	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	164	3	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	164	4	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	5	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	164	7	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	164	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	164	9	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	10	sentences	sentence	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	164	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	164	12	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	13	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	164	15	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	164	16	extract	extract	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	164	17	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	164	18	GHKM	GHKM	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	164	19	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	20	rules7	rules7	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	21	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	22	terms	term	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	164	23	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	24	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	164	25	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	164	26	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	164	27	Galley	Galley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	164	28	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	164	29	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	164	30	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	164	31	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	164	32	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	164	33	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	164	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	165	1	Frontier	Frontier	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	165	2	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	165	3	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	165	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	165	5	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	165	6	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	165	7	that	that	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	165	8	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	165	9	map	map	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	165	10	onto	onto	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	165	11	contiguous	contiguous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	165	12	substrings	substring	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	165	13	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	165	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	165	15	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	165	16	side	side	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	165	17	via	via	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	165	18	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	165	19	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	165	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	166	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	166	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	166	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	166	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	166	5	bold	bold	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	166	6	italic	italic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	166	7	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	166	8	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	166	9	shadows	shadow	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	166	10	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	166	11	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	166	12	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	166	13	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	166	14	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	166	15	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	166	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	167	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	2	addition	addition	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	167	4	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	167	5	should	should	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	167	6	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	167	7	noted	note	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	8	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	167	10	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	11	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	12	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	167	13	fixed8	fixed8	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	167	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	167	16	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	167	17	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	167	18	explore	explore	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	167	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	167	20	entire	entire	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	167	21	space	space	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	22	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	23	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	24	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	167	25	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	26	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	167	27	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	167	28	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	168	1	Differently	differently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	168	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	168	3	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	168	4	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	168	5	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	168	6	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	168	7	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	168	8	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	168	9	designed	design	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	168	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	168	11	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	168	12	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	168	13	infer	infer	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	168	14	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	168	15	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	168	16	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	168	17	fixing	fix	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	168	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	168	19	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	168	20	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	168	21	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	168	22	exploring	explore	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	168	23	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	168	24	space	space	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	168	25	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	168	26	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	168	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	169	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	169	2	believe	believe	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	169	3	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	169	4	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	169	5	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	169	6	possible	possible	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	169	7	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	169	8	investigate	investigate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	169	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	169	10	space	space	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	169	11	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	169	12	both	both	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	169	13	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	169	14	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	169	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	169	16	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	169	17	simultaneously	simultaneously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	169	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	170	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	170	2	subject	subject	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	170	3	will	will	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	170	4	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	170	5	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	170	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	170	7	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	170	8	future	future	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	170	9	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	170	10	topics	topic	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	170	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	171	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	2	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	171	3	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	4	instance	instance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	5	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	171	6	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	171	7	pair	pair	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	8	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	9	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	10	sentence	sentence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	171	12	target	target	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	171	13	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	171	15	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	16	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	17	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	171	18	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	171	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	171	20	extracted	extracted	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	171	21	GHKM	GHKM	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	171	22	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	171	23	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	24	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	171	25	compose	compose	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	171	26	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	171	27	unique	unique	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	171	28	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	171	29	derivation9	derivation9	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	171	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	172	1	Moreover	moreover	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	172	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	172	3	all	all	PDT
work_vbzp3iewy5h7rhfiok5o6i3du4	172	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	172	5	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	172	6	developed	develop	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	172	7	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	172	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	172	9	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	172	10	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	172	11	constitute	constitute	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	172	12	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	172	13	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	172	14	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	172	15	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	172	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	172	17	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	172	18	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	172	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	173	1	7	7	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	173	2	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	173	3	attach	attach	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	173	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	173	5	unaligned	unaligned	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	173	6	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	173	7	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	173	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	173	9	lowest	low	JJS
work_vbzp3iewy5h7rhfiok5o6i3du4	173	10	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	173	11	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	173	12	that	that	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	173	13	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	173	14	cover	cover	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	173	15	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	173	16	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	173	17	terms	term	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	173	18	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	173	19	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	173	20	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	173	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	174	1	8	8	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	174	2	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	174	3	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	174	4	might	may	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	174	5	reinforce	reinforce	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	174	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	174	7	frequent	frequent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	174	8	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	174	9	errors	error	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	174	10	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	174	11	AE	AE	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	174	12	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	174	13	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	174	14	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	174	15	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	174	16	harm	harm	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	174	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	174	18	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	174	19	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	174	20	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	174	21	TM	TM	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	174	22	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	174	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	175	1	Actually	actually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	175	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	175	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	175	4	frequent	frequent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	175	5	AEs	ae	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	175	6	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	175	7	greatly	greatly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	175	8	impair	impair	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	175	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	175	10	conventional	conventional	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	175	11	TM	tm	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	175	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	176	1	Besides	besides	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	176	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	176	3	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	176	4	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	176	5	encourages	encourage	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	176	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	176	7	correct	correct	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	176	8	alignments	alignment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	176	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	176	10	simultaneously	simultaneously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	176	11	discourages	discourage	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	176	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	176	13	infrequent	infrequent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	176	14	AEs	ae	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	176	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	177	1	Thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	177	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	177	3	compared	compare	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	177	4	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	177	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	177	6	conventional	conventional	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	177	7	TMs	tm	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	177	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	177	9	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	177	10	believe	believe	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	177	11	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	177	12	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	177	13	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	177	14	TM	TM	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	177	15	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	177	16	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	177	17	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	177	18	worse	bad	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	177	19	due	due	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	177	20	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	177	21	AEs	ae	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	177	22	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	178	1	Our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	178	2	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	178	3	experiments	experiment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	178	4	verify	verify	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	178	5	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	178	6	point	point	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	178	7	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	178	8	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	178	9	will	will	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	178	10	conduct	conduct	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	178	11	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	178	12	much	much	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	178	13	detailed	detailed	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	178	14	analysis	analysis	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	178	15	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	178	16	future	future	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	178	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	179	1	9	9	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	179	2	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	179	3	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	179	4	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	179	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	179	6	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	179	7	GHKM	GHKM	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	179	8	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	179	9	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	179	10	Galley	Galley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	179	11	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	179	12	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	179	13	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	179	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	179	15	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	179	16	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	179	17	here	here	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	179	18	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	179	19	reduce	reduce	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	179	20	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	179	21	complexity	complexity	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	179	22	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	179	23	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	179	24	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	179	25	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	180	1	247	247	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	180	2	jin	jin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	180	4	tian	tian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	180	5	jian	jian	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	180	7	mianwo	mianwo	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	180	9	men	men	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	10	zai	zai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	180	12	ci	ci	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	13	PRP+VBP	prp+vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	180	14	today	today	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	180	15	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	16	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	17	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	18	meet	meet	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	19	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	20	again	again	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	180	21	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	22	Ӻཙ	Ӻཙ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	23	ᡁԜ	ᡁԜ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	24	⅑޽	⅑޽	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	180	25	㿱䶒	㿱䶒	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	26	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	27	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	180	28	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	29	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	30	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	180	31	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	32	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	180	33	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	180	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	181	1	Illustration	illustration	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	181	2	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	181	3	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	181	4	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	181	5	U	u	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	181	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	181	7	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	181	8	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	181	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	182	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	182	2	bold	bold	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	182	3	italic	italic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	182	4	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	182	5	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	182	6	shadows	shadow	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	182	7	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	182	8	frontier	fronti	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	182	9	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	182	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	183	1	Under	under	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	183	2	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	183	3	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	183	4	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	183	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	183	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	183	7	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	183	8	modifies	modify	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	183	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	183	10	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	183	11	U	U	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	183	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	183	13	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	183	14	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	183	15	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	183	16	sample	sample	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	183	17	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	183	18	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	183	19	create	create	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	183	20	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	183	21	series	series	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	183	22	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	183	23	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	183	24	ones	one	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	183	25	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	183	26	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	183	27	samples	sample	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	183	28	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	183	29	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	183	30	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	183	31	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	183	32	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	183	33	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	184	1	Consequently	consequently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	184	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	184	3	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	184	4	STSGs	stsg	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	184	5	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	184	6	created	create	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	184	7	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	184	8	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	184	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	184	10	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	184	11	U	U	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	184	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	184	13	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	184	14	simultaneously	simultaneously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	184	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	184	16	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	184	17	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	184	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	184	19	next	next	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	184	20	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	184	21	operation	operation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	184	22	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	185	1	Repeatedly	repeatedly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	185	2	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	185	3	after	after	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	185	4	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	185	5	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	185	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	185	7	iterations	iteration	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	185	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	185	9	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	185	10	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	185	11	obtain	obtain	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	185	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	185	13	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	185	14	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	185	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	185	16	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	185	17	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	185	18	building	building	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	185	19	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	185	20	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	185	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	186	1	5.2	5.2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	186	2	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	186	3	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	186	4	Operators	Operators	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	186	5	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	186	6	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	186	7	section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	186	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	186	9	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	186	10	develop	develop	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	186	11	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	186	12	novel	novel	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	186	13	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	186	14	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	186	15	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	186	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	186	17	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	186	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	187	1	They	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	187	2	explore	explore	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	187	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	187	4	entire	entire	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	187	5	space	space	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	187	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	187	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	187	8	U	U	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	187	9	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	187	10	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	187	11	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	187	12	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	187	13	performing	perform	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	187	14	local	local	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	187	15	changes	change	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	187	16	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	187	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	187	18	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	187	19	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	187	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	188	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	188	2	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	188	3	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	188	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	188	5	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	188	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	188	7	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	188	8	given	give	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	188	9	sentence	sentence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	188	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	188	11	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	188	12	define	define	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	188	13	s-	s-	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	188	14	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	188	15	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	188	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	188	17	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	188	18	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	188	19	root	root	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	188	20	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	188	21	covering	cover	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	188	22	at	at	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	188	23	least	least	RBS
work_vbzp3iewy5h7rhfiok5o6i3du4	188	24	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	188	25	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	188	26	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	189	1	Thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	189	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	189	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	189	4	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	189	5	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	189	6	s	s	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	189	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	189	8	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	189	9	contains	contain	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	189	10	all	all	PDT
work_vbzp3iewy5h7rhfiok5o6i3du4	189	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	189	12	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	189	13	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	189	14	except	except	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	189	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	189	16	root	root	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	189	17	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	189	18	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	189	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	189	20	pre	pre	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	189	21	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	189	22	terminal	terminal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	189	23	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	189	24	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	189	25	leaf	leaf	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	189	26	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	189	27	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	189	28	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	189	29	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	189	30	call	call	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	189	31	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	189	32	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	189	33	s	s	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	189	34	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	189	35	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	189	36	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	190	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	190	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	190	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	190	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	190	5	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	6	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	190	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	190	8	PRB	PRB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	9	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	10	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	11	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	190	12	PRP+VBP	prp+vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	190	13	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	14	s	s	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	190	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	190	16	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	190	17	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	190	18	while	while	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	190	19	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	20	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	190	21	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	22	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	23	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	24	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	190	25	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	190	26	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	190	27	s	s	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	190	28	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	190	29	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	190	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	191	1	Since	since	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	191	3	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	191	4	tag	tag	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	5	sequence	sequence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	191	8	sentence	sentence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	9	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	191	10	fixed	fix	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	191	12	all	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	191	13	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	191	14	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	191	15	s	s	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	191	16	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	191	17	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	191	18	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	191	19	stay	stay	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	191	20	unchanged	unchanged	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	191	21	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	22	all	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	191	23	possible	possible	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	191	24	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	25	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	191	26	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	191	27	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	28	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	191	29	sentence	sentence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	191	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	192	1	Based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	192	2	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	192	3	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	192	4	fact	fact	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	192	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	192	6	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	192	7	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	192	8	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	192	9	work	work	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	192	10	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	192	11	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	192	12	s	s	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	192	13	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	192	14	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	192	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	193	1	Further	further	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	193	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	193	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	193	4	assign	assign	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	193	5	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	193	6	descendant	descendant	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	193	7	candidates	candidate	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	193	8	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	193	9	DC	DC	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	193	10	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	193	11	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	193	12	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	193	13	s	s	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	193	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	193	15	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	193	16	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	193	17	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	193	18	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	193	19	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	193	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	193	21	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	193	22	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	193	23	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	193	24	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	193	25	sibling	sibling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	193	26	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	194	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	194	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	194	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	194	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	194	5	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	194	6	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	194	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	194	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	194	9	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	194	10	DCs	dc	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	194	11	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	194	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	194	13	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	194	14	-	-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	194	15	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	194	16	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	194	17	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	194	18	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	194	19	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	194	20	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	194	21	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	194	22	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	194	23	respectively	respectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	194	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	195	1	According	accord	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	195	2	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	195	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	195	4	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	195	5	DCs	dc	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	195	6	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	195	7	governs	govern	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	195	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	195	9	every	every	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	195	10	s-	s-	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	195	11	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	195	12	might	may	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	195	13	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	195	14	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	195	15	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	195	16	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	195	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	195	18	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	195	19	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	195	20	states	state	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	195	21	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	195	22	1	1	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	195	23	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	195	24	Left	leave	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	195	25	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	195	26	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	195	27	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	195	28	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	195	29	3(a	3(a	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	195	30	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	195	31	shows	show	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	195	32	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	195	33	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	195	34	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	195	35	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	195	36	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	195	37	governs	govern	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	195	38	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	195	39	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	195	40	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	195	41	DCs	dc	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	195	42	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	195	43	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	195	44	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	195	45	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	195	46	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	195	47	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	195	48	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	195	49	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	195	50	PRP+VBP	prp+vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	195	51	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	196	1	2	2	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	196	2	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	196	3	Right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	196	4	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	196	5	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	196	6	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	196	7	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	196	8	3(b	3(b	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	196	9	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	196	10	shows	show	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	196	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	196	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	196	13	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	196	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	196	15	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	196	16	governs	govern	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	196	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	196	18	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	196	19	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	196	20	DCs	dc	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	196	21	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	196	22	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	196	23	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	196	24	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	196	25	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	196	26	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	196	27	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	196	28	labeled	label	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	196	29	VBP+RB	VBP+RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	196	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	197	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	197	2	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	197	3	specific	specific	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	197	4	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	197	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	197	6	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	197	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	197	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	197	9	states	state	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	197	10	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	197	11	s	s	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	197	12	-	-	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	197	13	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	197	14	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	197	15	fixed	fix	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	197	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	198	1	Thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	198	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	198	3	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	198	4	changing	change	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	198	5	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	198	6	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	198	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	198	8	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	198	9	’s	’s	POS
work_vbzp3iewy5h7rhfiok5o6i3du4	198	10	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	198	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	198	12	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	198	13	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	198	14	easily	easily	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	198	15	transform	transform	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	198	16	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	198	17	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	198	18	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	198	19	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	198	20	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	198	21	another	another	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	198	22	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	198	23	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	198	24	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	198	25	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	198	26	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	198	27	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	198	28	current	current	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	198	29	sample	sample	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	198	30	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	198	31	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	198	32	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	198	33	one	one	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	198	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	199	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	199	2	formulate	formulate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	199	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	199	4	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	199	6	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	7	transformation	transformation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	8	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	199	10	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	11	associate	associate	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	199	13	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	199	14	variable	variable	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	15	Ȍ䌜{0,1	ȍ䌜{0,1	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	16	}	}	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	199	17	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	18	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	199	19	s	s	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	20	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	199	21	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	22	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	199	23	indicating	indicate	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	199	24	whether	whether	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	25	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	199	26	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	27	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	199	28	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	29	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	199	30	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	31	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	199	32	left	left	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	33	�	�	.
work_vbzp3iewy5h7rhfiok5o6i3du4	199	34	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	35	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	36	�	�	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	37	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	199	38	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	199	39	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	40	�	�	.
work_vbzp3iewy5h7rhfiok5o6i3du4	199	41	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	42	�	�	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	43	�	�	,
work_vbzp3iewy5h7rhfiok5o6i3du4	199	44	�	�	,
work_vbzp3iewy5h7rhfiok5o6i3du4	199	45	Then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	199	46	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	47	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	199	48	change	change	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	199	49	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	199	50	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	51	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	199	52	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	53	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	54	changing	change	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	199	55	value	value	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	56	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	199	57	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	199	58	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	199	59	parameters	parameter	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	199	60	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	200	1	Our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	200	2	first	first	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	200	3	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	200	4	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	200	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	200	6	Rotate	Rotate	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	200	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	200	8	just	just	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	200	9	works	work	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	200	10	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	200	11	sampling	sample	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	200	12	value	value	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	200	13	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	200	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	200	15	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	200	16	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	200	17	parameters	parameter	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	200	18	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	200	19	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	200	20	at	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	200	21	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	200	22	time	time	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	200	23	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	200	24	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	200	25	changing	change	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	200	26	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	200	27	U	U	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	200	28	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	200	29	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	200	30	accordingly	accordingly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	200	31	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	201	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	201	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	5	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	6	3(a	3(a	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	201	7	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	201	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	201	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	201	10	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	201	12	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	13	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	201	14	currently	currently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	201	15	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	201	17	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	201	18	VWDWH	VWDWH	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	19	�	�	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	201	20	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	21	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	22	�	�	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	23	�	�	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	201	24	�	�	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	201	25	�	�	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	26	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	201	27	H	h	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	28	�	�	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	29	VDPSOH	vdpsoh	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	30	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	31	WKH	WKH	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	32	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	33	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	34	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	35	RI	RI	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	36	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	37	WKLV	WKLV	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	38	�	�	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	201	39	QRGH	QRGH	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	201	40	�	�	,
work_vbzp3iewy5h7rhfiok5o6i3du4	201	41	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	42	DQG	DQG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	43	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	44	LI	LI	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	45	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	46	WKH	WKH	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	47	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	48	VDPSOHG	vdpsohg	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	49	�	�	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	201	50	YDOXH	YDOXH	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	51	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	52	RI	RI	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	53	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	54	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	55	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	56	LV	LV	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	57	�	�	:
work_vbzp3iewy5h7rhfiok5o6i3du4	201	58	�	�	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	59	�	�	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	201	60	�	�	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	201	61	ZH	ZH	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	62	�	�	,
work_vbzp3iewy5h7rhfiok5o6i3du4	201	63	NHHS	nhhs	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	64	�	�	:
work_vbzp3iewy5h7rhfiok5o6i3du4	201	65	WKH	WKH	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	66	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	67	VWUXFWXUH	VWUXFWXUH	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	201	68	�	�	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	201	69	unchanged	unchanged	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	201	70	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	201	71	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	201	72	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	201	73	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	74	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	201	75	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	201	76	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	201	77	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	202	1	Otherwise	otherwise	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	202	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	202	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	202	4	change	change	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	202	5	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	202	6	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	202	7	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	202	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	202	9	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	202	10	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	202	11	�	�	.
work_vbzp3iewy5h7rhfiok5o6i3du4	202	12	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	202	13	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	202	14	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	202	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	202	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	202	17	transform	transform	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	202	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	202	19	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	202	20	-	-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	202	21	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	202	22	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	202	23	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	202	24	3(b	3(b	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	202	25	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	202	26	accordingly	accordingly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	202	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	203	1	jian	jian	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	2	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	203	3	mianwo	mianwo	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	203	5	men	men	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	6	zai	zai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	203	8	ci	ci	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	9	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	203	11	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	12	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	13	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	14	meet	meet	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	15	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	16	again	again	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	203	17	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	18	ᡁԜ	ᡁԜ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	19	⅑޽	⅑޽	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	203	20	㿱䶒	㿱䶒	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	21	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	22	...	...	.
work_vbzp3iewy5h7rhfiok5o6i3du4	203	23	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	24	PRP+VBP	prp+vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	203	25	jian	jian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	203	26	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	203	27	mianwo	mianwo	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	28	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	203	29	men	men	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	30	zai	zai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	31	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	203	32	ci	ci	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	33	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	34	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	203	35	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	36	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	37	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	38	meet	meet	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	39	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	40	again	again	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	203	41	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	42	ᡁԜ	ᡁԜ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	43	⅑޽	⅑޽	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	203	44	㿱䶒	㿱䶒	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	45	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	46	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	203	47	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	48	VBP+RB	VBP+RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	49	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	203	50	b	b	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	51	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	203	52	Ȍ=1(a	Ȍ=1(a	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	53	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	203	54	Ȍ=0	Ȍ=0	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	55	Rotate	Rotate	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	56	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	203	57	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	203	58	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	204	1	Illustration	illustration	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	204	2	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	204	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	204	4	Rotate	Rotate	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	204	5	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	204	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	205	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	205	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	205	3	figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	205	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	205	5	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	205	6	a	a	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	205	7	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	205	8	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	205	9	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	205	10	b	b	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	205	11	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	205	12	denote	denote	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	205	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	205	14	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	205	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	205	16	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	205	17	’s	’s	POS
work_vbzp3iewy5h7rhfiok5o6i3du4	205	18	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	205	19	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	205	20	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	205	21	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	205	22	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	205	23	respectively	respectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	205	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	206	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	206	2	bold	bold	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	206	3	italic	italic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	206	4	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	206	5	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	206	6	shadows	shadow	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	206	7	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	206	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	206	9	figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	206	10	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	206	11	frontier	fronti	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	206	12	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	206	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	207	1	Obviously	obviously	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	207	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	207	3	towards	towards	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	207	4	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	207	5	s	s	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	207	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	207	7	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	207	8	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	207	9	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	207	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	207	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	207	12	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	207	13	values	value	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	207	14	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	207	15	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	207	16	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	207	17	define	define	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	207	18	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	207	19	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	207	20	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	207	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	207	22	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	207	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	208	1	Using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	208	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	208	3	GHKM	GHKM	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	208	4	algorithm	algorithm	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	208	5	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	208	6	Galley	Galley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	208	7	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	208	8	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	208	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	209	1	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	209	2	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	209	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	209	4	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	209	5	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	209	6	get	get	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	209	7	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	209	8	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	209	9	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	209	10	derivations	derivation	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	209	11	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	209	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	209	13	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	209	14	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	209	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	209	16	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	209	17	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	209	18	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	209	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	209	20	fixed	fix	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	209	21	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	209	22	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	209	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	210	1	Each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	210	2	derivation	derivation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	210	3	carries	carry	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	210	4	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	210	5	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	210	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	210	7	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	210	8	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	210	9	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	210	10	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	210	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	210	12	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	210	13	GHKM	GHKM	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	210	14	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	210	15	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	210	16	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	210	17	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	210	18	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	210	19	own	own	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	210	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	211	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	211	3	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	211	4	derivations	derivation	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	211	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	211	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	211	7	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	211	8	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	211	9	defined	define	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	10	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	211	12	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	211	13	states	state	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	211	14	include	include	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	211	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	211	16	one	one	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	17	rooted	root	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	18	at	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	211	20	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	211	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	211	22	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	211	23	’s	’s	POS
work_vbzp3iewy5h7rhfiok5o6i3du4	211	24	lowest	low	JJS
work_vbzp3iewy5h7rhfiok5o6i3du4	211	25	ancestor	ancestor	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	26	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	27	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	211	28	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	211	29	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	211	30	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	211	31	one	one	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	32	rooted	root	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	33	at	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	34	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	211	35	s	s	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	36	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	211	37	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	38	if	if	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	39	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	211	40	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	211	41	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	211	42	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	43	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	211	44	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	212	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	2	instance	instance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	5	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	6	3(a	3(a	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	7	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	9	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	11	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	212	13	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	14	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	15	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	212	16	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	17	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	18	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	19	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	20	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	21	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	22	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	23	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	24	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	25	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	26	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	27	defines	define	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	28	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	212	29	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	30	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	31	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	32	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	33	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	34	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	35	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	36	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	37	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	38	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	39	...	...	.
work_vbzp3iewy5h7rhfiok5o6i3du4	212	40	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	41	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	42	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	43	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	44	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	45	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	46	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	47	leftr	leftr	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	48	x	x	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	49	x	x	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	50	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	51	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	52	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	53	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	54	VBP	vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	55	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	56	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	57	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	58	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	59	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	60	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	61	o	o	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	62	�	�	.
work_vbzp3iewy5h7rhfiok5o6i3du4	212	63	Differently	differently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	212	64	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	65	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	66	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	67	3(b	3(b	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	68	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	69	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	70	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	71	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	72	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	212	73	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	74	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	75	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	76	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	77	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	78	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	212	79	thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	212	80	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	81	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	82	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	83	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	84	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	85	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	86	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	87	defines	define	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	88	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	89	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	212	90	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	91	248	248	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	92	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	93	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	94	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	95	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	96	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	97	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	98	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	99	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	100	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	101	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	102	:	:	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	103	...	...	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	104	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	105	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	106	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	107	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	108	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	109	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	110	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	111	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	112	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	113	right	right	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	212	114	right	right	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	212	115	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	116	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	117	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	118	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	119	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	120	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	121	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	122	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	123	VBP	vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	124	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	125	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	126	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	127	x	x	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	128	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	129	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	130	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	131	VBP	vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	132	x	x	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	212	133	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	134	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	135	�	�	.
work_vbzp3iewy5h7rhfiok5o6i3du4	212	136	o	o	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	137	�	�	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	138	o	o	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	139	�	�	.
work_vbzp3iewy5h7rhfiok5o6i3du4	212	140	Using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	212	141	these	these	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	142	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	143	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	212	144	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	145	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	146	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	147	derivations	derivation	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	212	148	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	149	evaluated	evaluate	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	150	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	151	follows	follow	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	152	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	153	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	154	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	155	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	156	value	value	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	157	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	158	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	159	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	212	160	denote	denote	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	212	161	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	162	corresponding	corresponding	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	163	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	164	derivation	derivation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	165	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	166	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	212	167	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	168	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	169	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	170	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	171	0	0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	172	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	173	0	0	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	174	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	175	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	176	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	177	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	178	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	179	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	180	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	181	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	182	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	183	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	184	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	185	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	186	|	|	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	187	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	188	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	189	|	|	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	190	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	191	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	212	192	left	leave	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	212	193	right	right	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	212	194	right	right	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	212	195	right	right	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	212	196	right	right	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	212	197	right	right	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	212	198	p	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	199	p	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	200	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	201	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	202	p	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	203	p	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	204	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	205	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	206	r	r	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	207	p	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	208	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	209	r	r	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	210	p	p	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	211	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	212	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	213	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	214	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	215	�	�	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	216	�	�	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	217	�	�	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	218	�	�	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	219	�	�	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	220	�	�	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	221	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	222	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	223	<	<	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	212	224	v	v	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	225	<	<	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	212	226	v	v	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	227	Where	where	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	212	228	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	229	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	212	230	refers	refer	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	231	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	232	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	233	conditional	conditional	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	234	context	context	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	235	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	236	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	212	237	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	212	238	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	239	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	240	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	241	all	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	242	other	other	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	212	243	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	212	244	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	245	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	212	246	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	212	247	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	212	248	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	213	1	All	all	PDT
work_vbzp3iewy5h7rhfiok5o6i3du4	213	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	213	3	probabilities	probability	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	213	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	213	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	213	6	above	above	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	213	7	formulas	formula	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	213	8	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	213	9	computed	compute	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	213	10	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	213	11	Equation(3	equation(3	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	213	12	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	213	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	214	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	214	2	then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	214	3	normalize	normalize	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	214	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	214	5	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	214	6	scores	score	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	214	7	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	214	8	sample	sample	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	214	9	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	214	10	value	value	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	214	11	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	214	12	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	214	13	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	214	14	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	214	15	them	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	214	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	215	1	With	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	215	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	215	3	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	215	4	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	215	5	described	describe	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	215	6	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	215	7	section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	215	8	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	215	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	215	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	215	11	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	215	12	ZLOO	ZLOO	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	13	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	14	SUHIHU	SUHIHU	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	15	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	16	WKH	WKH	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	17	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	18	Ȍ	Ȍ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	19	�	�	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	215	20	WKDW	WKDW	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	21	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	22	SURGXFHV	SURGXFHV	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	23	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	24	VPDOO	vpdoo	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	215	25	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	26	DQG	DQG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	27	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	28	IUHTXHQW	IUHTXHQW	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	29	�	�	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	215	30	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	215	31	rules	rule	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	215	32	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	216	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	216	2	tendency	tendency	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	3	results	result	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	216	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	5	more	more	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	216	6	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	7	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	216	8	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	216	10	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	216	12	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	13	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	216	14	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	216	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	216	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	216	17	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	216	18	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	216	19	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	216	20	tends	tend	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	216	21	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	216	22	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	216	23	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	24	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	216	25	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	26	that	that	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	216	27	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	216	28	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	216	29	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	30	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	31	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	216	32	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	216	33	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	216	34	will	will	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	216	35	factor	factor	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	216	36	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	216	37	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	38	instance	instance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	39	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	216	40	more	more	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	216	41	small	small	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	216	42	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	216	43	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	216	44	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	217	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	217	2	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	217	3	way	way	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	217	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	217	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	217	6	overall	overall	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	217	7	likelihood	likelihood	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	217	8	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	217	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	217	10	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	217	11	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	217	12	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	217	13	improved	improve	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	217	14	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	217	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	217	16	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	217	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	218	1	Theoretically	theoretically	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	218	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	218	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	218	4	Rotate	Rotate	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	218	5	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	218	6	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	218	7	capable	capable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	218	8	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	218	9	arriving	arrive	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	218	10	at	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	218	11	any	any	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	218	12	possible	possible	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	218	13	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	218	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	218	15	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	218	16	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	218	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	218	18	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	218	19	U-	U-	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	218	20	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	218	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	219	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	219	2	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	219	3	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	4	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	219	5	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	219	6	first	first	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	219	7	convert	convert	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	219	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	219	9	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	219	10	U	U	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	219	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	219	12	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	13	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	14	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	219	15	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	219	16	branch	branch	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	18	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	219	20	Rotate	Rotate	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	219	21	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	22	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	219	23	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	219	24	then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	219	25	transform	transform	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	219	26	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	219	27	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	28	any	any	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	219	29	other	other	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	219	30	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	31	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	219	32	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	219	33	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	220	1	However	however	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	220	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	220	3	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	220	4	may	may	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	220	5	take	take	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	220	6	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	220	7	long	long	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	220	8	time	time	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	220	9	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	220	10	do	do	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	220	11	so	so	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	220	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	221	1	Thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	221	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	221	3	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	221	4	speed	speed	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	221	5	up	up	RP
work_vbzp3iewy5h7rhfiok5o6i3du4	221	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	221	7	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	8	transformation	transformation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	9	process	process	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	221	11	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	221	12	employ	employ	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	221	13	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	221	14	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	221	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	221	16	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	17	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	221	18	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	19	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	221	21	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	221	22	takes	take	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	221	23	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	221	24	pair	pair	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	25	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	26	s-	s-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	27	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	221	28	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	29	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	221	30	parent	parent	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	31	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	221	32	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	33	relationship	relationship	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	34	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	35	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	221	36	unit	unit	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	37	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	38	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	221	39	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	222	1	Similar	similar	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	222	2	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	222	4	Rotate	Rotate	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	222	5	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	222	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	222	8	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	222	9	assign	assign	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	222	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	222	11	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	222	12	variable	variable	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	13	ȟ䌜{0,1	ȟ䌜{0,1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	222	14	}	}	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	222	15	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	16	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	222	17	unit	unit	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	18	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	222	19	update	update	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	222	20	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	222	21	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	22	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	222	23	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	24	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	25	sampling	sample	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	222	26	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	222	27	value	value	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	28	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	29	ȟ	ȟ	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	222	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	223	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	223	2	method	method	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	223	3	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	223	4	sampling	sample	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	223	5	ȟ	ȟ	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	223	6	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	223	7	similar	similar	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	223	8	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	223	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	223	10	one	one	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	223	11	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	223	12	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	223	13	Ȍ.	ȍ.	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	224	1	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	224	2	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	224	3	shows	show	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	224	4	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	224	5	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	224	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	224	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	224	8	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	224	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	225	1	As	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	225	2	shown	show	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	225	3	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	225	4	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	225	5	4(a	4(a	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	225	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	225	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	225	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	225	9	unit	unit	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	225	10	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	225	11	…	…	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	225	12	VBP	vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	225	13	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	225	14	PRP+VBP	prp+vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	225	15	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	225	16	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	225	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	225	18	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	225	19	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	225	20	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	225	21	ȟ=0	ȟ=0	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	225	22	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	225	23	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	225	24	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	225	25	governs	govern	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	225	26	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	225	27	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	225	28	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	225	29	descendants	descendant	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	225	30	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	225	31	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	225	32	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	225	33	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	225	34	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	225	35	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	225	36	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	225	37	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	226	1	By	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	226	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	226	3	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	226	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	226	5	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	226	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	226	7	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	226	8	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	226	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	226	10	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	226	11	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	226	12	convert	convert	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	226	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	226	14	unit	unit	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	226	15	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	226	16	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	226	17	4(b	4(b	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	226	18	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	226	19	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	226	20	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	226	21	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	226	22	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	226	23	ULJKW	ULJKW	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	226	24	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	226	25	VWDWH	VWDWH	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	226	26	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	226	27	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	226	28	ȟ=1	ȟ=1	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	226	29	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	226	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	227	1	Just	just	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	227	2	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	227	3	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	227	4	4(b	4(b	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	227	5	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	227	6	shows	show	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	227	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	227	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	227	9	governed	govern	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	227	10	descendants	descendant	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	227	11	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	227	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	227	13	unit	unit	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	227	14	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	227	15	turned	turn	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	227	16	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	227	17	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	227	18	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	227	19	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	227	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	227	21	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	227	22	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	227	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	228	1	It	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	228	2	may	may	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	228	3	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	228	4	confusing	confusing	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	228	5	when	when	WRB
work_vbzp3iewy5h7rhfiok5o6i3du4	228	6	choosing	choose	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	228	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	228	8	parent-	parent-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	228	9	child	child	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	228	10	s	s	POS
work_vbzp3iewy5h7rhfiok5o6i3du4	228	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	228	12	node	node	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	228	13	pair	pair	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	14	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	15	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	16	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	228	18	parent	parent	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	19	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	20	always	always	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	228	21	faces	face	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	228	22	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	228	23	choices	choice	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	228	24	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	228	25	combining	combine	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	228	26	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	228	27	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	228	28	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	29	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	228	30	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	228	31	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	32	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	33	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	228	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	229	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	229	2	avoid	avoid	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	229	3	confusion	confusion	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	229	5	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	229	6	split	split	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	229	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	229	8	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	229	9	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	229	10	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	229	12	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	13	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	14	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	15	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	229	16	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	229	17	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	229	18	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	229	19	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	229	20	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	229	22	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	229	23	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	229	24	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	25	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	26	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	229	27	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	229	28	works	work	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	229	29	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	30	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	229	31	parent	parent	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	32	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	33	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	229	34	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	229	35	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	229	36	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	37	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	229	38	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	229	39	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	229	40	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	229	41	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	42	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	229	43	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	229	44	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	229	45	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	46	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	47	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	229	48	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	229	49	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	229	50	considers	consider	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	229	51	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	229	52	parent	parent	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	53	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	54	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	229	55	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	229	56	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	229	57	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	229	58	10	10	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	229	59	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	230	1	Therefore	therefore	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	230	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	230	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	230	4	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	230	5	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	230	6	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	230	7	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	230	8	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	230	9	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	230	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	230	11	Two-	Two-	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	230	12	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	230	13	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	230	14	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	230	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	230	16	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	230	17	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	230	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	231	1	jin	jin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	2	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	3	tian	tian	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	4	jian	jian	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	6	mianwo	mianwo	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	8	men	men	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	9	zai	zai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	11	ci	ci	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	12	PRP+VBP	prp+vbp	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	231	13	Today	Today	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	14	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	15	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	16	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	17	meet	meet	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	18	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	19	again	again	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	231	20	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	21	Ӻཙ	Ӻཙ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	22	ᡁԜ	ᡁԜ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	23	⅑޽	⅑޽	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	231	24	㿱䶒	㿱䶒	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	25	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	26	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	231	27	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	28	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	29	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	231	30	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	31	jin	jin	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	231	32	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	33	tian	tian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	231	34	jian	jian	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	35	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	36	mianwo	mianwo	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	37	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	38	men	men	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	39	zai	zai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	40	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	41	ci	ci	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	231	42	VBP+RB	VBP+RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	43	Today	Today	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	44	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	45	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	46	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	47	meet	meet	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	48	VBP	VBP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	49	again	again	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	231	50	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	51	Ӻཙ	Ӻཙ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	52	ᡁԜ	ᡁԜ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	53	⅑޽	⅑޽	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	231	54	㿱䶒	㿱䶒	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	55	PRP	PRP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	56	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	231	57	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	58	NN	NN	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	59	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	231	60	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	231	61	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	231	62	a	a	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	231	63	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	231	64	ȟ=0	ȟ=0	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	231	65	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	231	66	b	b	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	231	67	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	231	68	ȟ=1	ȟ=1	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	231	69	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	231	70	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	71	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	231	72	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	73	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	231	74	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	231	75	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	231	76	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	231	77	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	231	78	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	232	1	Illustration	illustration	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	232	2	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	232	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	232	4	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	232	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	232	6	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	232	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	232	8	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	232	9	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	232	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	233	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	233	2	bold	bold	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	233	3	italic	italic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	233	4	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	233	5	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	233	6	shadows	shadow	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	233	7	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	233	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	233	9	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	233	10	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	233	11	frontier	fronti	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	233	12	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	233	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	234	1	During	during	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	2	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	234	4	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	5	each	each	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	234	6	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	7	instance	instance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	234	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	234	10	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	11	first	first	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	234	12	applies	apply	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	234	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	234	14	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	234	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	234	16	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	17	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	234	18	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	234	19	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	234	20	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	21	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	22	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	23	all	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	234	24	candidate	candidate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	25	pairs	pair	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	234	26	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	27	s	s	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	234	28	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	234	29	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	234	30	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	234	31	parent	parent	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	32	s-	s-	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	234	33	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	234	34	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	234	35	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	234	36	left	left	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	234	37	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	38	s	s	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	234	39	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	234	40	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	41	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	234	42	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	43	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	234	44	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	45	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	234	46	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	234	47	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	235	1	After	after	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	2	that	that	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	235	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	235	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	235	5	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	235	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	235	7	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	235	9	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	235	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	235	11	Rotate	rotate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	12	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	13	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	235	14	applied	apply	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	15	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	16	all	all	PDT
work_vbzp3iewy5h7rhfiok5o6i3du4	235	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	235	18	candidate	candidate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	19	pairs	pair	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	235	20	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	21	s	s	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	235	22	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	235	23	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	235	24	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	235	25	parent	parent	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	26	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	235	27	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	235	28	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	235	29	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	235	30	its	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	235	31	right	right	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	235	32	child	child	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	235	33	s	s	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	235	34	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	235	35	node	node	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	235	36	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	235	37	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	236	1	Then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	236	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	236	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	236	4	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	236	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	236	6	Rotate	Rotate	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	236	7	operator	operator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	236	8	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	236	9	every	every	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	236	10	s	s	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	236	11	-	-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	236	12	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	236	13	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	236	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	236	15	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	236	16	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	236	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	236	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	237	1	By	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	237	2	utilizing	utilize	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	237	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	237	4	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	237	5	separately	separately	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	237	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	237	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	237	8	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	237	9	guarantee	guarantee	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	237	10	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	237	11	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	237	12	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	237	13	satisfies	satisfie	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	237	14	detailed	detailed	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	237	15	balance	balance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	237	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	238	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	238	2	visit	visit	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	238	3	all	all	PDT
work_vbzp3iewy5h7rhfiok5o6i3du4	238	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	238	5	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	238	6	instances	instance	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	238	7	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	238	8	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	238	9	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	238	10	order	order	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	238	11	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	238	12	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	238	13	iteration	iteration	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	238	14	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	238	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	239	1	After	after	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	239	2	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	239	3	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	239	4	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	239	5	iterations	iteration	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	239	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	239	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	239	8	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	239	9	obtain	obtain	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	239	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	239	11	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	239	12	U	u	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	239	13	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	239	14	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	239	15	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	239	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	239	17	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	239	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	239	19	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	239	20	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	239	21	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	239	22	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	239	23	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	239	24	accordingly	accordingly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	239	25	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	240	1	6	6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	240	2	Experiments	experiment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	240	3	6.1	6.1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	240	4	Experimental	experimental	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	240	5	Setup	setup	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	240	6	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	240	7	experiments	experiment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	240	8	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	240	9	conducted	conduct	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	240	10	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	240	11	Chinese	Chinese	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	240	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	240	13	to-	to-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	240	14	English	english	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	240	15	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	240	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	241	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	241	2	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	241	3	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	241	4	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	241	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	241	6	FBIS	FBIS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	241	7	corpus	corpus	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	241	8	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	241	9	approximately	approximately	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	241	10	7.1	7.1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	241	11	million	million	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	241	12	Chinese	chinese	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	241	13	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	241	14	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	241	15	9.2	9.2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	241	16	million	million	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	241	17	English	english	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	241	18	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	241	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	242	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	242	2	obtain	obtain	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	242	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	242	4	bidirectional	bidirectional	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	242	5	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	242	6	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	242	7	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	242	8	GIZA++	GIZA++	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	242	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	242	10	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	242	11	then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	242	12	adopt	adopt	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	242	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	242	14	grow	grow	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	242	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	242	16	diag	diag	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	242	17	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	242	18	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	242	19	-	-	:
work_vbzp3iewy5h7rhfiok5o6i3du4	242	20	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	242	21	strategy	strategy	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	242	22	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	242	23	obtain	obtain	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	242	24	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	242	25	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	242	26	symmetric	symmetric	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	242	27	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	242	28	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	243	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	243	2	train	train	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	243	3	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	243	4	5-	5-	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	243	5	gram	gram	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	6	language	language	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	7	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	8	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	243	10	Xinhua	Xinhua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	243	11	portion	portion	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	12	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	243	14	English	English	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	243	15	Gigaword	Gigaword	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	243	16	corpus	corpus	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	17	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	243	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	243	19	English	english	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	243	20	part	part	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	21	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	22	10	10	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	243	23	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	243	24	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	243	25	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	243	26	take	take	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	243	27	more	more	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	243	28	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	243	29	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	30	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	243	31	unit	unit	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	32	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	33	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	34	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	243	35	but	but	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	243	36	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	243	37	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	243	38	make	make	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	243	39	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	243	40	algorithm	algorithm	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	243	41	much	much	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	243	42	more	more	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	243	43	complex	complex	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	243	44	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	244	1	249	249	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	244	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	244	3	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	244	4	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	244	5	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	245	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	245	2	tuning	tuning	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	245	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	245	4	testing	testing	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	245	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	245	6	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	245	7	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	245	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	245	9	NIST	NIST	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	245	10	MT	MT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	245	11	2003	2003	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	245	12	evaluation	evaluation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	245	13	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	245	14	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	245	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	245	16	development	development	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	245	17	set	set	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	245	18	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	245	19	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	245	20	use	use	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	245	21	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	245	22	NIST	NIST	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	245	23	MT04	MT04	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	245	24	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	245	25	MT05	MT05	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	245	26	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	245	27	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	245	28	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	245	29	test	test	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	245	30	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	245	31	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	246	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	246	2	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	246	3	MERT	MERT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	246	4	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	246	5	Och	Och	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	246	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	246	7	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	246	8	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	246	9	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	246	10	tune	tune	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	246	11	parameters	parameter	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	246	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	247	1	Since	since	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	247	2	MERT	MERT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	247	3	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	247	4	prone	prone	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	247	5	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	247	6	search	search	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	247	7	errors	error	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	247	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	247	9	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	247	10	run	run	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	247	11	MERT	MERT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	247	12	5	5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	247	13	times	time	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	247	14	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	247	15	select	select	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	247	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	247	17	best	good	JJS
work_vbzp3iewy5h7rhfiok5o6i3du4	247	18	tuning	tuning	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	247	19	parameters	parameter	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	247	20	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	247	21	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	247	22	tuning	tuning	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	247	23	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	247	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	248	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	248	2	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	248	3	quality	quality	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	248	4	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	248	5	evaluated	evaluate	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	248	6	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	248	7	case	case	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	248	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	248	9	insensitive	insensitive	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	248	10	BLEU-4	BLEU-4	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	248	11	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	248	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	248	13	shortest	short	JJS
work_vbzp3iewy5h7rhfiok5o6i3du4	248	14	length	length	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	248	15	penalty	penalty	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	248	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	249	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	249	2	statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	249	3	significance	significance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	249	4	test	test	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	249	5	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	249	6	performed	perform	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	249	7	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	249	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	249	9	re	re	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	249	10	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	249	11	sampling	sampling	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	249	12	approach	approach	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	249	13	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	249	14	Koehn	Koehn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	249	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	249	16	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	249	17	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	249	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	250	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	250	2	create	create	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	250	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	250	4	baseline	baseline	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	5	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	250	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	8	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	250	10	open-	open-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	250	11	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	12	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	13	4.0	4.0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	250	14	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	15	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	250	16	Ganitkevitch	Ganitkevitch	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	17	et	et	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	250	18	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	19	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	250	21	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	250	22	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	250	23	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	250	24	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	250	25	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	250	26	hierarchical	hierarchical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	250	27	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	28	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	250	29	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	30	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	250	31	HPB	HPB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	32	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	250	33	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	34	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	250	35	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	250	36	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	250	37	syntax	syntax	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	38	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	250	39	augmented	augment	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	40	MT	MT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	41	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	250	42	SAMT	SAMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	43	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	250	44	11	11	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	250	45	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	250	46	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	250	47	Zollmann	Zollmann	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	48	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	250	49	Venugopal	Venugopal	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	250	50	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	250	51	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	250	52	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	250	53	respectively	respectively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	250	54	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	251	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	251	2	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	3	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	4	used	use	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	5	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	6	testing	test	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	251	7	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	251	8	effectiveness	effectiveness	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	9	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	10	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	251	11	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	251	13	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	251	14	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	251	15	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	251	16	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	17	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	251	18	house	house	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	19	string-	string-	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	251	20	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	251	22	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	23	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	24	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	251	25	abbreviated	abbreviate	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	26	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	251	27	s2	s2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	251	28	t	t	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	251	29	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	251	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	252	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	252	2	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	252	3	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	252	4	implemented	implement	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	252	5	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	252	6	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	252	7	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	252	8	Galley	Galley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	252	9	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	252	10	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	252	11	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	252	12	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	252	13	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	252	14	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	252	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	252	16	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	252	17	Marcu	Marcu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	252	18	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	252	19	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	252	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	253	1	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	253	2	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	253	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	254	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	254	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	254	3	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	254	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	254	5	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	6	extract	extract	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	7	both	both	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	254	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	254	9	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	254	10	GHKM	GHKM	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	11	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	254	12	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	254	13	Galley	Galley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	14	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	15	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	16	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	17	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	254	18	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	254	19	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	254	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	254	21	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	254	22	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	254	23	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	254	24	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	254	25	SPMT	SPMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	26	Model	Model	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	27	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	254	28	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	254	29	Galley	Galley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	30	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	31	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	32	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	33	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	254	34	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	254	35	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	254	36	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	254	37	phrases	phrase	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	254	38	up	up	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	254	39	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	254	40	length	length	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	254	41	L=5	L=5	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	254	42	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	254	43	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	254	44	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	254	45	side	side	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	254	46	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	255	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	255	2	then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	255	3	obtain	obtain	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	255	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	255	5	composed	compose	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	255	6	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	255	7	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	255	8	composing	compose	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	255	9	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	255	10	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	255	11	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	255	12	adjacent	adjacent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	255	13	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	255	14	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	255	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	256	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	256	2	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	256	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	256	4	above	above	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	256	5	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	6	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	7	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	256	9	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	256	10	first	first	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	256	11	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	256	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	256	13	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	14	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	256	16	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	256	17	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	256	18	generated	generate	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	19	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	20	parsing	parse	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	256	21	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	256	22	English	english	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	256	23	side	side	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	24	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	25	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	256	26	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	256	27	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	256	28	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	29	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	256	30	Berkeley	Berkeley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	256	31	parser	parser	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	256	32	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	256	33	Petrov	Petrov	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	256	34	et	et	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	256	35	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	256	36	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	256	37	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	256	38	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	256	39	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	256	40	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	257	1	Then	then	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	257	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	257	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	257	4	binarize	binarize	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	257	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	257	6	English	english	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	257	7	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	257	8	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	257	9	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	257	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	257	11	head	head	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	257	12	binarization	binarization	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	257	13	approach	approach	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	257	14	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	257	15	Wang	Wang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	257	16	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	257	17	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	257	18	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	257	19	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	257	20	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	257	21	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	257	22	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	257	23	use	use	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	257	24	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	257	25	resulting	result	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	257	26	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	257	27	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	257	28	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	257	29	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	257	30	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	257	31	another	another	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	257	32	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	257	33	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	257	34	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	257	35	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	258	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	258	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	258	3	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	258	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	258	5	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	258	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	258	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	258	8	run	run	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	258	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	258	10	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	258	11	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	258	12	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	258	13	1000	1000	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	258	14	iterations	iteration	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	258	15	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	258	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	258	17	whole	whole	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	258	18	corpus	corpus	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	258	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	259	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	259	2	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	259	3	uses	use	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	259	4	1,087s	1,087	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	259	5	per	per	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	259	6	iteration	iteration	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	259	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	259	8	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	259	9	average	average	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	259	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	259	11	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	259	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	259	13	single	single	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	259	14	core	core	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	259	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	259	16	2.3	2.3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	259	17	GHz	GHz	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	259	18	Intel	Intel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	259	19	Xeon	Xeon	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	259	20	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	259	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	260	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	260	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	260	3	hyperparameters	hyperparameter	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	260	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	260	5	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	260	6	set	set	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	260	7	Į	į	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	260	8	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	260	9	0.1	0.1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	260	10	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	260	11	pexpand	pexpand	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	260	12	=	=	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	260	13	1/3	1/3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	260	14	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	260	15	give	give	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	260	16	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	260	17	preference	preference	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	260	18	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	260	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	260	20	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	260	21	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	260	22	small	small	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	260	23	fragments	fragment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	260	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	261	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	261	2	built	build	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	261	3	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	261	4	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	261	5	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	261	6	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	261	7	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	261	8	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	261	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	261	10	achieved	achieve	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	261	11	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	261	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	261	13	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	261	14	after	after	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	261	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	261	16	1000th	1000th	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	261	17	iteration	iteration	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	261	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	262	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	262	2	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	262	3	use	use	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	262	4	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	262	5	sample	sample	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	262	6	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	262	7	extract	extract	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	262	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	262	9	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	262	10	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	262	11	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	262	12	multiple	multiple	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	262	13	samples	sample	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	262	14	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	262	15	result	result	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	262	16	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	262	17	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	262	18	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	262	19	that	that	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	262	20	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	262	21	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	262	22	too	too	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	262	23	large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	262	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	263	1	11	11	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	263	2	From	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	263	3	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	263	4	Zollmann	Zollmann	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	263	5	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	263	6	Vogel	Vogel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	263	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	263	8	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	263	9	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	263	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	263	11	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	263	12	find	find	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	263	13	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	263	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	263	15	performance	performance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	263	16	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	263	17	SAMT	SAMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	263	18	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	263	19	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	263	20	similar	similar	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	263	21	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	263	22	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	263	23	method	method	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	263	24	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	263	25	labeling	label	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	263	26	SCFG	SCFG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	263	27	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	263	28	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	263	29	POS	POS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	263	30	tags	tag	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	263	31	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	264	1	Thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	264	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	264	3	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	264	4	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	264	5	convenient	convenient	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	264	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	264	7	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	264	8	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	264	9	conduct	conduct	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	264	10	experiments	experiment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	264	11	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	264	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	264	13	SAMT	SAMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	264	14	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	264	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	265	1	6.2	6.2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	265	2	Analysis	analysis	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	3	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	4	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	265	5	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	265	6	Sampler	Sampler	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	265	7	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	265	8	evaluate	evaluate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	265	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	265	10	effectiveness	effectiveness	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	11	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	265	13	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	265	14	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	265	16	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	265	17	explore	explore	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	265	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	265	19	change	change	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	20	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	21	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	265	22	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	23	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	265	24	’s	’s	POS
work_vbzp3iewy5h7rhfiok5o6i3du4	265	25	likelihood	likelihood	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	26	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	27	increasing	increase	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	265	28	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	265	29	iterations	iteration	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	265	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	266	1	1.239E+08	1.239e+08	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	2	1.243E+08	1.243e+08	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	3	1.247E+08	1.247e+08	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	4	1.251E+08	1.251e+08	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	5	1.255E+08	1.255e+08	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	6	1.259E+08	1.259e+08	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	7	100	100	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	8	200	200	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	9	300	300	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	10	400	400	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	11	500	500	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	12	600	600	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	13	700	700	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	14	800	800	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	15	900	900	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	16	1000	1000	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	17	Number	Number	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	18	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	266	19	Sampling	Sampling	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	20	Iterations	Iterations	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	266	21	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	22	eg	eg	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	266	23	at	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	266	24	iv	iv	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	25	e-	e-	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	266	26	Lo	Lo	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	27	g	g	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	28	Li	Li	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	29	ke	ke	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	30	lih	lih	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	31	oo	oo	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	266	32	d	d	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	266	33	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	266	34	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	35	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	266	36	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	37	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	266	38	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	39	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	266	40	5	5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	266	41	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	267	1	Histograms	histogram	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	267	2	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	267	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	267	4	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	267	5	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	267	6	’s	’s	POS
work_vbzp3iewy5h7rhfiok5o6i3du4	267	7	likelihood	likelihood	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	267	8	vs.	vs.	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	267	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	267	10	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	267	11	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	267	12	sampling	sample	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	267	13	iterations	iteration	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	267	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	268	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	268	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	268	3	figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	268	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	268	5	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	268	6	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	268	7	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	268	8	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	268	9	refers	refer	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	268	10	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	268	11	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	268	12	independent	independent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	268	13	runs	run	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	268	14	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	268	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	268	16	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	268	17	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	268	18	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	268	19	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	268	20	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	268	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	268	22	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	268	23	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	268	24	initialization	initialization	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	268	25	states	state	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	268	26	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	269	1	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	269	2	5	5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	269	3	depicts	depict	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	269	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	269	5	negative	negative	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	269	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	269	7	log	log	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	269	8	likelihood	likelihood	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	269	9	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	269	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	269	11	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	269	12	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	269	13	after	after	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	269	14	several	several	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	269	15	sampling	sample	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	269	16	iterations	iteration	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	269	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	270	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	270	2	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	270	3	show	show	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	270	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	270	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	270	6	overall	overall	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	270	7	likelihood	likelihood	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	270	8	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	270	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	270	10	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	270	11	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	270	12	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	270	13	improved	improve	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	270	14	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	270	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	270	16	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	270	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	271	1	Moreover	moreover	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	271	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	271	3	comparing	compare	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	271	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	271	5	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	271	6	independent	independent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	271	7	runs	run	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	271	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	271	9	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	271	10	see	see	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	271	11	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	271	12	although	although	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	271	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	271	14	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	271	15	begins	begin	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	271	16	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	271	17	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	271	18	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	271	19	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	271	20	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	271	21	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	271	22	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	271	23	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	271	24	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	271	25	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	271	26	’s	’s	POS
work_vbzp3iewy5h7rhfiok5o6i3du4	271	27	likelihood	likelihood	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	271	28	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	271	29	always	always	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	271	30	similar	similar	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	271	31	during	during	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	271	32	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	271	33	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	272	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	272	2	demonstrates	demonstrate	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	272	3	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	272	4	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	272	5	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	272	6	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	272	7	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	272	8	sensitive	sensitive	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	272	9	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	272	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	272	11	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	272	12	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	272	13	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	272	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	272	15	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	272	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	272	17	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	272	18	always	always	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	272	19	arrive	arrive	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	272	20	at	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	272	21	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	272	22	good	good	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	272	23	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	272	24	state	state	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	272	25	beginning	begin	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	272	26	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	272	27	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	272	28	initialization	initialization	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	272	29	states	state	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	272	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	273	1	Thus	thus	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	273	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	273	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	273	4	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	273	5	utilize	utilize	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	273	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	273	7	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	273	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	273	9	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	273	10	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	273	11	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	273	12	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	273	13	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	273	14	further	further	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	273	15	analysis	analysis	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	273	16	hereafter	hereafter	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	273	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	274	1	1.035E+07	1.035e+07	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	2	1.040E+07	1.040e+07	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	3	1.045E+07	1.045e+07	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	4	1.050E+07	1.050e+07	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	5	1.055E+07	1.055e+07	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	6	1.060E+07	1.060e+07	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	7	100	100	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	8	200	200	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	9	300	300	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	10	400	400	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	11	500	500	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	12	600	600	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	13	700	700	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	14	800	800	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	15	900	900	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	16	1000	1000	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	17	Number	Number	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	18	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	274	19	Sampling	Sampling	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	20	Iterations	Iterations	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	274	21	To	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	274	22	ta	ta	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	23	l	l	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	274	24	N	n	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	274	25	um	um	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	274	26	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	274	27	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	274	28	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	274	29	F	F	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	30	ro	ro	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	31	nt	nt	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	274	32	ie	ie	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	33	r	r	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	34	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	35	od	od	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	36	es	es	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	274	37	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	274	38	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	39	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	274	40	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	41	random	random	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	274	42	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	43	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	274	44	6	6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	274	45	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	275	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	275	2	total	total	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	275	3	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	275	4	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	275	5	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	275	6	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	275	7	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	275	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	275	9	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	275	10	independent	independent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	275	11	runs	run	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	275	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	276	1	6.3	6.3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	276	2	Analysis	Analysis	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	276	3	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	276	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	276	5	U	u	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	276	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	276	7	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	276	8	Structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	276	9	Acquiring	acquire	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	276	10	better	well	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	276	11	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	276	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	276	13	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	276	14	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	276	15	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	276	16	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	276	17	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	276	18	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	276	19	purpose	purpose	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	276	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	277	1	However	however	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	277	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	277	3	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	277	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	277	5	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	277	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	277	7	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	277	8	achieved	achieve	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	277	9	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	277	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	277	11	250	250	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	277	12	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	277	13	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	277	14	appropriate	appropriate	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	277	15	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	277	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	277	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	277	18	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	277	19	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	277	20	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	277	21	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	277	22	?	?	.
work_vbzp3iewy5h7rhfiok5o6i3du4	278	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	278	2	answer	answer	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	278	3	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	278	4	question	question	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	278	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	278	6	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	278	7	first	first	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	278	8	analyze	analyze	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	278	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	278	10	effect	effect	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	278	11	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	278	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	278	13	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	278	14	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	278	15	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	278	16	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	278	17	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	278	18	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	278	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	279	1	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	279	2	6	6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	279	3	shows	show	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	279	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	279	5	total	total	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	279	6	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	279	7	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	279	8	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	279	9	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	279	10	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	279	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	279	12	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	279	13	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	279	14	during	during	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	279	15	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	279	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	280	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	280	2	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	280	3	show	show	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	280	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	280	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	280	6	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	280	7	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	280	8	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	280	9	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	280	10	increases	increase	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	280	11	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	280	12	increased	increase	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	280	13	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	280	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	281	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	281	2	tendency	tendency	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	281	3	indicates	indicate	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	281	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	281	5	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	281	6	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	281	7	prefers	prefer	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	281	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	281	9	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	281	10	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	281	11	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	281	12	more	more	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	281	13	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	281	14	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	281	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	282	1	Consequently	consequently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	282	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	282	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	282	4	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	282	5	U	u	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	282	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	282	7	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	282	8	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	282	9	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	282	10	always	always	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	282	11	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	282	12	factored	factor	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	282	13	into	into	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	282	14	many	many	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	282	15	small	small	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	282	16	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	282	17	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	282	18	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	282	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	283	1	Just	just	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	283	2	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	283	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	283	4	have	have	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	283	5	argued	argue	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	283	6	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	283	7	section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	283	8	4.1	4.1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	283	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	283	10	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	283	11	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	283	12	beneficial	beneficial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	283	13	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	283	14	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	283	15	good	good	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	283	16	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	283	17	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	283	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	284	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	284	2	demonstrate	demonstrate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	284	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	284	4	above	above	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	284	5	analysis	analysis	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	284	7	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	8	7	7	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	284	9	shows	show	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	284	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	284	11	visual	visual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	284	12	comparison	comparison	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	13	between	between	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	14	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	284	15	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	16	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	284	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	18	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	284	19	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	20	random	random	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	284	21	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	284	22	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	284	23	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	284	24	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	284	25	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	284	26	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	27	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	28	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	284	29	found	find	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	30	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	31	head	head	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	32	binarization	binarization	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	284	33	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	284	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	285	1	Because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	285	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	285	3	traditional	traditional	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	285	4	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	285	5	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	285	6	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	285	7	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	285	8	binarized	binarized	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	285	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	285	10	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	285	11	do	do	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	285	12	not	not	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	285	13	consider	consider	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	285	14	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	285	15	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	285	16	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	285	17	analysis	analysis	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	285	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	286	1	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	2	7	7	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	286	3	shows	show	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	286	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	5	whether	whether	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	6	it	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	286	7	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	286	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	286	9	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	10	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	11	fragment	fragment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	12	or	or	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	286	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	286	14	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	15	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	16	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	286	18	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	19	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	286	20	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	286	21	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	286	22	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	286	23	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	286	24	always	always	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	286	25	tend	tend	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	286	26	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	286	27	obtain	obtain	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	286	28	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	286	29	smaller	small	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	286	30	ones12	ones12	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	286	31	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	287	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	287	2	comparison	comparison	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	287	3	verifies	verify	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	287	4	that	that	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	287	5	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	287	6	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	287	7	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	287	8	induction	induction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	287	9	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	287	10	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	287	11	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	287	12	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	287	13	shifting	shift	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	287	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	287	15	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	287	16	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	287	17	away	away	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	287	18	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	287	19	complex	complex	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	287	20	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	287	21	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	287	22	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	287	23	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	287	24	tend	tend	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	287	25	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	287	26	negatively	negatively	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	287	27	affect	affect	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	287	28	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	287	29	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	288	1	0	0	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	2	200k	200k	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	3	400k	400k	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	288	4	600k	600k	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	288	5	800k	800k	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	6	1000k	1000k	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	7	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	8	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	9	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	10	5	5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	11	6	6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	12	7	7	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	13	8	8	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	14	9	9	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	15	10	10	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	16	>	>	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	288	17	=	=	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	288	18	11	11	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	19	U	U	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	20	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	288	21	Tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	22	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	288	23	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	24	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	25	Number	Number	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	26	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	27	Nodes	Nodes	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	288	28	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	29	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	288	30	Target	Target	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	31	Tree	Tree	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	32	Fragment	Fragment	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	33	N	N	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	34	um	um	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	288	35	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	288	36	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	37	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	38	R	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	39	ul	ul	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	40	es	es	XX
work_vbzp3iewy5h7rhfiok5o6i3du4	288	41	Number	Number	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	42	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	43	Words	Words	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	288	44	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	288	45	Variables	Variables	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	288	46	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	47	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	288	48	Source	Source	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	49	String	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	50	0	0	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	51	300k	300k	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	288	52	600k	600k	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	288	53	900k	900k	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	54	1200k	1200k	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	55	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	56	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	57	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	58	4	4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	59	5	5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	60	6	6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	61	7	7	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	62	N	n	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	63	um	um	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	288	64	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	288	65	r	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	66	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	67	R	r	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	288	68	ul	ul	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	69	es	es	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	70	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	288	71	7	7	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	288	72	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	289	1	Histograms	histogram	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	289	2	over	over	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	289	3	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	289	4	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	289	5	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	289	6	statistics	statistic	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	289	7	comparing	compare	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	289	8	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	289	9	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	289	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	289	11	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	289	12	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	289	13	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	289	14	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	289	15	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	289	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	290	1	12	12	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	290	2	Binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	290	3	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	290	4	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	290	5	get	get	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	290	6	more	more	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	290	7	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	290	8	fragments	fragment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	290	9	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	290	10	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	290	11	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	290	12	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	290	13	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	290	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	290	15	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	290	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	291	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	291	2	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	291	3	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	291	4	there	there	EX
work_vbzp3iewy5h7rhfiok5o6i3du4	291	5	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	291	6	many	many	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	291	7	unary	unary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	291	8	edges	edge	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	291	9	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	291	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	291	11	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	291	12	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	291	13	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	291	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	291	15	while	while	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	291	16	no	no	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	291	17	unary	unary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	291	18	edge	edge	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	291	19	exists	exist	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	291	20	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	291	21	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	291	22	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	291	23	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	291	24	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	291	25	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	292	1	Specifically	specifically	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	292	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	292	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	292	4	show	show	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	292	5	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	292	6	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	292	7	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	292	8	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	292	9	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	292	10	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	292	11	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	292	12	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	292	13	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	292	14	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	292	15	-	-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	292	16	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	292	17	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	292	18	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	292	19	8	8	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	292	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	293	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	293	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	293	3	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	293	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	293	5	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	293	6	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	293	7	more	more	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	293	8	conducive	conducive	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	293	9	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	293	10	extracting	extract	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	293	11	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	293	12	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	293	13	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	293	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	294	1	For	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	2	example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	294	4	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	294	5	translate	translate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	294	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	294	7	Chinese	chinese	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	294	8	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	9	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	294	10	ӵ	ӵ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	294	11	Ѫ	Ѫ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	294	12	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	294	13	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	294	14	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	294	15	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	294	16	extract	extract	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	294	17	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	294	18	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	19	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	294	20	R2	R2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	294	21	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	22	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	294	23	9	9	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	294	24	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	294	25	directly	directly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	294	26	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	27	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	294	28	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	29	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	294	30	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	31	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	32	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	294	33	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	34	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	294	35	ӵ	ӵ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	294	36	Ѫ	Ѫ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	294	37	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	294	38	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	294	39	governed	govern	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	40	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	41	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	294	42	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	43	node	node	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	294	44	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	294	45	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	294	46	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	294	47	node	node	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	294	48	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	294	49	VBD+RB	VBD+RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	294	50	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	294	51	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	295	1	However	however	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	295	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	295	3	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	295	4	no	no	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	295	5	node	node	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	295	6	governs	govern	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	295	7	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	295	8	ӵ	ӵ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	295	9	Ѫ	Ѫ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	295	10	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	295	11	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	295	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	295	13	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	295	14	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	295	15	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	295	16	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	295	17	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	295	18	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	295	19	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	295	20	obtain	obtain	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	295	21	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	295	22	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	295	23	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	295	24	R1	R1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	295	25	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	295	26	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	295	27	9	9	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	295	28	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	295	29	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	295	30	many	many	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	295	31	extra	extra	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	295	32	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	295	33	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	295	34	edges	edge	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	295	35	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	295	36	such	such	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	295	37	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	295	38	node	node	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	295	39	CD	CD	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	295	40	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	295	41	R1	R1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	295	42	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	296	1	Due	due	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	296	2	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	296	3	these	these	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	296	4	extra	extra	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	296	5	things	thing	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	296	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	296	7	R1	R1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	296	8	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	296	9	too	too	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	296	10	large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	296	11	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	296	12	show	show	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	296	13	good	good	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	296	14	generality	generality	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	296	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	297	1	was	be	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	297	2	QP	QP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	3	dollarsUS1500only	dollarsUS1500only	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	4	VBD	VBD	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	5	NNSNNPCDRB	NNSNNPCDRB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	6	NP	NP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	7	NP	NP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	8	ӵ	ӵ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	9	㖾ݳаॳӄⲮѪ	㖾ݳаॳӄⲮѪ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	10	NP	NP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	297	12	COMP	COMP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	13	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	297	14	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	297	15	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	297	16	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	297	17	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	297	18	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	297	19	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	297	20	b	b	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	297	21	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	297	22	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	297	23	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	297	24	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	297	25	was	be	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	297	26	dollarsUS1500only	dollarsUS1500only	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	27	VBD	VBD	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	28	NNSNNPCDRB	NNSNNPCDRB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	29	ӵ	ӵ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	30	㖾ݳаॳӄⲮѪ	㖾ݳаॳӄⲮѪ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	31	VBD+RB	VBD+RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	32	NNP+NNS	NNP+NNS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	33	CD	cd	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	297	34	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	297	35	NNS	NNS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	36	VBD	VBD	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	37	...	...	:
work_vbzp3iewy5h7rhfiok5o6i3du4	297	38	NNS	NNS	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	39	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	297	40	8	8	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	297	41	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	298	1	Example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	298	2	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	298	3	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	298	4	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	298	5	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	298	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	299	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	299	2	node	node	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	299	3	NP	NP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	299	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	299	5	COMP	COMP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	299	6	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	299	7	achieved	achieve	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	299	8	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	299	9	head	head	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	299	10	binarization	binarization	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	299	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	300	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	300	2	bold	bold	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	300	3	italic	italic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	300	4	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	300	5	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	300	6	shadows	shadow	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	300	7	denote	denote	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	300	8	frontier	frontier	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	300	9	nodes	node	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	300	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	301	1	was	be	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	301	2	QP	QP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	3	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	301	4	VBD	VBD	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	5	CD	cd	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	301	6	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	301	7	x0RB	x0RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	8	NP	NP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	9	NP	NP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	10	NP	NP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	301	12	COMP	COMP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	13	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	301	14	x1	x1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	15	was	be	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	301	16	only	only	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	301	17	VBD	VBD	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	18	RB	RB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	19	ӵ	ӵ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	20	ѪVBD+RB	ѪVBD+RB	.
work_vbzp3iewy5h7rhfiok5o6i3du4	301	21	ӵ	ӵ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	22	x1x0Ѫ	x1x0Ѫ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	23	R1	R1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	301	24	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	301	25	R2	r2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	301	26	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	301	27	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	301	28	9	9	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	301	29	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	302	1	Example	example	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	302	2	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	302	3	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	302	4	translate	translate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	302	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	302	6	Chinese	chinese	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	302	7	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	302	8	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	302	9	ӵ	ӵ	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	302	10	Ѫ.	Ѫ.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	302	11	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	302	12	R1	R1	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	302	13	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	302	14	extracted	extract	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	302	15	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	302	16	Figure	Figure	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	302	17	8(a	8(a	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	302	18	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	302	19	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	302	20	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	302	21	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	302	22	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	302	23	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	302	24	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	302	25	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	302	26	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	303	1	R2	R2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	303	2	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	303	3	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	303	4	Figure	figure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	303	5	8(b	8(b	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	303	6	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	303	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	303	8	i.e.	i.e.	FW
work_vbzp3iewy5h7rhfiok5o6i3du4	303	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	303	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	303	11	U	U	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	303	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	303	13	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	303	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	304	1	251	251	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	304	2	Based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	304	3	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	304	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	304	5	above	above	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	304	6	analysis	analysis	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	304	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	304	8	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	304	9	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	304	10	conclude	conclude	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	304	11	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	304	12	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	304	13	proposed	propose	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	304	14	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	304	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	304	16	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	304	17	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	304	18	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	304	19	conducive	conducive	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	304	20	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	304	21	extracting	extract	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	304	22	small	small	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	304	23	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	304	24	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	304	25	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	304	26	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	304	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	305	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	305	2	indicates	indicate	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	305	3	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	305	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	305	5	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	305	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	305	7	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	305	8	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	305	9	more	more	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	305	10	consistent	consistent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	305	11	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	305	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	305	13	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	305	14	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	305	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	305	16	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	305	17	good	good	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	305	18	at	at	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	305	19	capturing	capture	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	305	20	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	305	21	mapping	mapping	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	305	22	information	information	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	305	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	306	1	Therefore	therefore	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	306	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	306	3	because	because	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	4	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	5	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	306	6	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	306	7	always	always	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	306	8	constrained	constrain	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	9	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	10	cross	cross	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	306	11	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	306	12	lingual	lingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	306	13	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	14	divergence	divergence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	306	16	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	306	17	believe	believe	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	306	18	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	306	20	proposed	propose	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	21	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	22	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	306	23	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	306	24	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	306	25	result	result	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	306	26	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	27	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	306	28	better	well	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	306	29	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	30	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	306	31	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	307	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	307	2	demonstrate	demonstrate	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	307	3	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	307	4	conclusion	conclusion	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	307	5	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	307	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	307	7	next	next	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	307	8	sub	sub	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	307	9	-	-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	307	10	section	section	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	307	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	308	1	6.4	6.4	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	308	2	Final	Final	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	308	3	Translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	308	4	Results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	308	5	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	308	6	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	308	7	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	308	8	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	308	9	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	308	10	shown	show	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	308	11	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	308	12	Table	table	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	308	13	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	308	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	309	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	309	3	table	table	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	309	5	lines	line	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	309	6	3	3	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	309	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	309	8	6	6	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	309	9	refer	refer	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	10	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	309	12	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	13	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	309	14	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	309	16	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	17	systems	system	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	309	18	built	build	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	19	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	20	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	309	21	types	type	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	309	22	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	23	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	309	24	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	309	25	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	310	1	Table	table	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	310	2	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	310	3	shows	show	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	310	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	310	5	all	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	310	6	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	310	7	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	310	8	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	310	9	systems	system	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	310	10	outperform	outperform	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	310	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	310	12	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	310	13	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	310	14	HPB	HPB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	310	15	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	310	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	310	17	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	310	18	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	310	19	SAMT	SAMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	310	20	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	310	21	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	310	22	significantly	significantly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	310	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	311	1	This	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	311	2	comparison	comparison	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	311	3	verifies	verify	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	311	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	311	5	superiority	superiority	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	311	6	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	311	7	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	311	8	in	in	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	311	9	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	311	10	house	house	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	311	11	s2	s2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	311	12	t	t	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	311	13	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	311	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	312	1	Moreover	moreover	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	312	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	312	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	312	4	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	312	5	shown	show	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	6	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	7	Table	table	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	8	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	312	9	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	312	10	demonstrate	demonstrate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	312	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	312	12	effectiveness	effectiveness	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	13	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	14	head	head	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	15	binarization	binarization	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	16	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	312	17	which	which	WDT
work_vbzp3iewy5h7rhfiok5o6i3du4	312	18	helps	help	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	312	19	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	312	20	improve	improve	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	312	21	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	312	22	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	23	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	24	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	25	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	312	26	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	27	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	312	28	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	29	all	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	312	30	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	312	31	tasks	task	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	312	32	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	313	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	313	2	test	test	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	313	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	313	4	effectiveness	effectiveness	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	313	5	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	313	6	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	313	7	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	313	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	313	9	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	313	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	313	11	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	313	12	give	give	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	313	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	313	14	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	313	15	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	313	16	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	313	17	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	313	18	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	313	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	313	20	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	313	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	313	22	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	313	23	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	313	24	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	313	25	random	random	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	313	26	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	313	27	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	313	28	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	314	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	314	2	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	314	3	show	show	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	314	4	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	314	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	314	6	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	314	7	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	314	8	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	314	9	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	314	10	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	314	11	achieves	achieve	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	314	12	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	314	13	best	good	JJS
work_vbzp3iewy5h7rhfiok5o6i3du4	314	14	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	314	15	result	result	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	314	16	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	314	17	all	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	314	18	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	314	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	314	20	systems	system	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	314	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	315	1	It	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	315	2	surpasses	surpass	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	315	3	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	315	4	s2	s2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	315	5	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	315	6	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	315	7	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	315	8	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	315	9	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	315	10	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	315	11	1.47	1.47	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	315	12	BLEU	BLEU	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	315	13	points	point	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	315	14	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	315	15	MT04	MT04	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	315	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	315	17	1.44	1.44	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	315	18	BLEU	BLEU	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	315	19	points	point	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	315	20	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	315	21	MT05	MT05	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	315	22	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	316	1	Moreover	moreover	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	316	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	316	3	even	even	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	316	4	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	316	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	316	6	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	316	7	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	8	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	316	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	316	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	316	11	achieved	achieve	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	12	s2	s2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	316	13	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	14	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	15	is	be	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	316	16	still	still	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	316	17	lower	low	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	316	18	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	19	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	316	20	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	316	22	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	23	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	316	24	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	25	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	26	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	27	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	28	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	29	0.97	0.97	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	316	30	BLEU	BLEU	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	316	31	points	point	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	316	32	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	33	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	316	34	combined	combined	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	316	35	test	test	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	36	set	set	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	316	37	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	317	1	From	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	317	3	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	4	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	317	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	317	6	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	317	7	can	can	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	317	8	validate	validate	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	317	9	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	317	10	former	former	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	317	11	analysis	analysis	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	12	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	13	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	317	14	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	317	15	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	317	16	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	317	17	generated	generate	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	18	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	19	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	317	20	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	317	21	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	22	induction	induction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	23	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	24	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	317	25	more	more	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	317	26	appropriate	appropriate	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	317	27	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	28	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	29	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	317	30	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	31	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	317	32	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	33	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	34	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	35	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	317	36	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	317	37	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	318	1	System	System	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	2	MT04	MT04	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	3	MT05	MT05	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	4	All	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	318	5	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	6	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	7	HPB	HPB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	8	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	9	31.73	31.73	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	10	28.82	28.82	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	11	30.64	30.64	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	12	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	13	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	14	SAMT	SAMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	15	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	16	32.48	32.48	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	17	29.77	29.77	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	18	31.56	31.56	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	19	s2	s2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	20	t	t	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	21	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	22	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	23	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	318	24	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	25	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	26	33.73	33.73	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	27	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	28	30.25	30.25	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	29	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	30	32.75	32.75	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	31	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	32	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	33	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	34	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	35	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	318	36	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	318	37	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	38	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	318	39	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	40	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	41	34.09	34.09	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	42	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	43	30.99	30.99	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	44	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	45	#	#	$
work_vbzp3iewy5h7rhfiok5o6i3du4	318	46	32.92	32.92	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	47	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	48	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	49	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	50	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	51	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	52	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	318	53	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	54	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	318	55	35.20	35.20	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	56	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	57	#	#	$
work_vbzp3iewy5h7rhfiok5o6i3du4	318	58	31.69	31.69	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	59	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	60	#	#	$
work_vbzp3iewy5h7rhfiok5o6i3du4	318	61	33.89	33.89	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	62	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	318	63	#	#	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	64	Table	table	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	318	65	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	318	66	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	319	1	Results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	319	2	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	319	3	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	319	4	case	case	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	319	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	319	6	insensitive	insensitive	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	319	7	BLEU-4	BLEU-4	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	319	8	scores	score	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	319	9	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	319	10	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	319	11	s2	s2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	319	12	t	t	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	319	13	systems	system	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	319	14	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	319	15	different	different	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	319	16	types	type	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	319	17	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	319	18	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	319	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	320	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	320	2	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	320	3	*	*	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	320	4	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	320	5	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	320	6	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	320	7	#	#	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	320	8	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	320	9	denote	denote	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	320	10	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	320	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	320	12	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	320	13	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	320	14	significantly	significantly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	320	15	better	well	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	320	16	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	320	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	320	18	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	320	19	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	320	20	SAMT	SAMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	320	21	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	320	22	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	320	23	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	320	24	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	320	25	s2	s2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	320	26	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	320	27	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	320	28	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	320	29	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	320	30	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	320	31	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	320	32	p<0.01	p<0.01	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	320	33	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	320	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	321	1	6.5	6.5	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	321	2	Large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	321	3	Data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	321	4	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	321	5	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	321	6	conduct	conduct	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	321	7	an	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	321	8	experiment	experiment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	321	9	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	321	10	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	321	11	larger	large	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	321	12	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	321	13	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	321	14	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	321	15	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	321	16	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	321	17	LDC	LDC	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	321	18	corpus13	corpus13	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	321	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	322	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	322	2	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	322	3	corpus	corpus	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	322	4	contains	contain	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	322	5	2.1	2.1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	322	6	M	M	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	322	7	sentence	sentence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	322	8	pairs	pair	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	322	9	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	322	10	approximately	approximately	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	322	11	27.7	27.7	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	322	12	M	m	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	322	13	Chinese	chinese	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	322	14	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	322	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	322	16	31.9	31.9	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	322	17	M	M	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	322	18	English	English	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	322	19	words	word	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	322	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	323	1	Similarly	similarly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	323	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	323	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	323	4	train	train	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	323	5	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	323	6	5-gram	5-gram	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	323	7	language	language	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	323	8	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	323	9	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	323	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	323	11	Xinhua	Xinhua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	323	12	portion	portion	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	323	13	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	323	14	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	323	15	English	English	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	323	16	Gigaword	Gigaword	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	323	17	corpus	corpus	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	323	18	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	323	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	323	20	English	english	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	323	21	part	part	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	323	22	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	323	23	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	323	24	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	323	25	corpus	corpus	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	323	26	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	324	1	With	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	324	3	same	same	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	324	4	settings	setting	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	324	5	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	6	before	before	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	324	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	324	8	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	324	9	run	run	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	324	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	324	11	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	324	12	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	13	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	14	1000	1000	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	324	15	iterations	iteration	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	324	16	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	324	17	utilize	utilize	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	324	18	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	324	19	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	324	20	U	u	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	324	21	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	324	22	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	23	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	24	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	324	25	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	324	26	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	324	27	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	28	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	324	29	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	30	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	324	31	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	32	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	33	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	324	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	325	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	325	2	final	final	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	325	3	BLEU	BLEU	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	325	4	score	score	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	325	5	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	325	6	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	325	7	shown	show	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	325	8	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	325	9	Table	Table	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	325	10	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	325	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	326	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	2	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	326	3	scenario	scenario	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	4	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	5	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	326	6	large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	326	7	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	326	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	326	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	326	10	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	11	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	326	12	to-	to-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	13	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	14	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	15	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	326	16	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	326	17	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	18	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	326	19	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	326	20	still	still	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	326	21	significantly	significantly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	326	22	outperforms	outperform	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	326	23	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	326	24	system	system	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	25	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	326	26	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	326	27	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	326	28	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	327	1	System	System	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	2	MT04	MT04	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	3	MT05	MT05	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	4	All	all	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	327	5	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	6	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	7	HPB	HPB	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	8	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	9	34.55	34.55	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	10	33.11	33.11	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	11	34.01	34.01	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	12	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	13	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	14	SAMT	SAMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	15	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	16	34.76	34.76	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	17	33.72	33.72	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	18	34.37	34.37	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	19	s2	s2	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	20	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	21	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	22	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	23	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	327	24	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	25	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	26	36.40	36.40	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	27	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	28	34.53	34.53	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	29	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	30	35.70	35.70	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	31	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	32	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	33	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	34	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	35	binary	binary	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	327	36	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	327	37	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	38	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	327	39	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	40	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	41	37.38	37.38	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	42	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	43	#	#	$
work_vbzp3iewy5h7rhfiok5o6i3du4	327	44	35.14	35.14	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	45	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	46	#	#	$
work_vbzp3iewy5h7rhfiok5o6i3du4	327	47	36.54	36.54	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	48	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	49	#	#	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	50	s2	s2	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	51	t	t	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	52	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	53	U	u	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	54	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	327	55	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	56	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	327	57	38.02	38.02	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	58	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	59	#	#	$
work_vbzp3iewy5h7rhfiok5o6i3du4	327	60	36.12	36.12	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	61	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	62	#	#	$
work_vbzp3iewy5h7rhfiok5o6i3du4	327	63	37.34	37.34	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	64	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	327	65	#	#	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	66	Table	table	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	327	67	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	327	68	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	328	1	Results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	328	2	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	328	3	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	328	4	case	case	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	328	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	328	6	insensitive	insensitive	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	328	7	BLEU-4	BLEU-4	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	328	8	scores	score	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	328	9	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	328	10	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	328	11	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	328	12	large	large	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	328	13	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	328	14	data	datum	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	328	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	329	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	329	2	meaning	meaning	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	329	3	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	329	4	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	329	5	*	*	NFP
work_vbzp3iewy5h7rhfiok5o6i3du4	329	6	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	329	7	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	329	8	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	329	9	#	#	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	329	10	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	329	11	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	329	12	similar	similar	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	329	13	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	329	14	Table	table	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	329	15	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	329	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	330	1	7	7	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	330	2	Conclusion	Conclusion	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	330	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	330	4	Future	Future	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	330	5	Work	Work	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	330	6	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	7	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	330	8	paper	paper	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	330	10	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	330	11	explored	explore	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	330	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	330	13	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	330	14	direction	direction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	15	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	330	16	build	build	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	330	17	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	330	18	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	19	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	330	20	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	21	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	22	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	23	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	24	unsupervised	unsupervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	330	25	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	330	26	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	330	27	rather	rather	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	330	28	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	29	supervised	supervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	330	30	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	330	31	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	330	32	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	331	1	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	331	2	achieve	achieve	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	331	3	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	331	4	purpose	purpose	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	331	6	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	331	7	have	have	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	331	8	made	make	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	9	two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	331	10	major	major	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	331	11	efforts	effort	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	331	12	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	13	this	this	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	331	14	paper	paper	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	15	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	331	16	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	331	17	1	1	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	331	18	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	331	19	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	331	20	have	have	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	331	21	proposed	propose	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	22	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	331	23	novel	novel	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	331	24	generative	generative	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	331	25	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	331	26	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	27	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	331	28	induce	induce	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	331	29	effective	effective	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	331	30	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	331	31	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	331	32	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	331	33	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	34	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	35	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	331	36	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	37	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	331	38	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	332	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	332	2	utilized	utilize	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	332	3	STSG	STSG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	332	4	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	332	5	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	332	6	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	332	7	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	332	8	grasp	grasp	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	332	9	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	332	10	mapping	mapping	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	332	11	information	information	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	332	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	333	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	333	2	further	further	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	333	3	imposed	impose	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	333	4	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	333	5	reasonable	reasonable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	333	6	hierarchical	hierarchical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	333	7	prior	prior	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	333	8	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	333	9	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	333	10	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	333	11	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	333	12	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	333	13	encouraging	encourage	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	333	14	small	small	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	333	15	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	333	16	frequent	frequent	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	333	17	minimal	minimal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	333	18	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	333	19	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	333	20	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	333	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	334	1	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	334	2	2	2	LS
work_vbzp3iewy5h7rhfiok5o6i3du4	334	3	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	334	4	To	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	334	5	train	train	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	334	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	334	7	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	334	8	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	334	9	induction	induction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	334	10	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	334	11	efficiently	efficiently	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	334	12	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	334	13	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	334	14	developed	develop	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	334	15	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	334	16	Gibbs	Gibbs	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	334	17	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	334	18	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	334	19	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	334	20	novel	novel	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	334	21	Gibbs	Gibbs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	334	22	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	334	23	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	335	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	335	2	operators	operator	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	335	3	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	335	4	designed	design	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	335	5	specifically	specifically	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	335	6	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	335	7	explore	explore	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	335	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	335	9	infinite	infinite	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	335	10	space	space	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	335	11	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	335	12	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	335	13	structures	structure	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	335	14	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	335	15	performing	perform	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	335	16	local	local	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	335	17	changes	change	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	335	18	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	335	19	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	335	20	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	335	21	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	335	22	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	336	1	13	13	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	336	2	LDC	LDC	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	336	3	category	category	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	336	4	number	number	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	336	5	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	336	6	LDC2000T50	ldc2000t50	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	336	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	336	8	LDC2002E18	ldc2002e18	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	336	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	336	10	LDC2003E07	ldc2003e07	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	336	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	336	12	LDC2004T07	ldc2004t07	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	336	13	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	336	14	LDC2005T06	LDC2005T06	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	336	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	336	16	LDC2002L27	LDC2002L27	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	336	17	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	336	18	LDC2005T10	LDC2005T10	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	336	19	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	336	20	LDC2005T34	LDC2005T34	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	336	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	337	1	252	252	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	337	2	Experiments	experiment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	337	3	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	337	4	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	337	5	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	337	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	337	7	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	337	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	337	9	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	337	10	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	337	11	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	337	12	demonstrated	demonstrate	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	337	13	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	337	14	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	337	15	U	U	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	337	16	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	337	17	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	337	18	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	337	19	better	well	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	337	20	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	337	21	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	337	22	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	337	23	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	337	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	338	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	338	2	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	338	3	results	result	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	338	4	verify	verify	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	338	5	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	338	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	338	7	well	well	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	338	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	338	9	designed	design	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	338	10	unsupervised	unsupervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	338	11	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	338	12	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	338	13	actually	actually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	338	14	more	more	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	338	15	appropriate	appropriate	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	338	16	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	338	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	338	18	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	338	19	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	338	20	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	338	21	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	338	22	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	338	23	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	338	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	339	1	Therefore	therefore	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	339	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	339	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	339	4	believe	believe	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	339	5	that	that	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	339	6	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	339	7	unsupervised	unsupervised	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	339	8	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	339	9	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	339	10	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	339	11	be	be	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	339	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	339	13	promising	promising	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	339	14	research	research	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	339	15	direction	direction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	339	16	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	339	17	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	339	18	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	339	19	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	339	20	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	339	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	340	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	340	2	future	future	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	340	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	340	4	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	340	5	plan	plan	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	340	6	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	340	7	testify	testify	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	340	8	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	340	9	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	340	10	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	340	11	various	various	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	340	12	initial	initial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	340	13	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	340	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	340	15	such	such	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	340	16	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	340	17	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	340	18	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	340	19	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	340	20	formed	form	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	340	21	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	340	22	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	340	23	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	340	24	et	et	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	340	25	al	al	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	340	26	.	.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	340	27	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	340	28	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	340	29	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	340	30	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	341	1	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	341	2	also	also	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	341	3	plan	plan	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	341	4	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	341	5	perform	perform	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	341	6	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	341	7	detailed	detailed	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	341	8	empirical	empirical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	341	9	comparison	comparison	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	341	10	between	between	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	341	11	STST	STST	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	341	12	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	341	13	SCFG	scfg	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	341	14	under	under	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	341	15	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	341	16	settings	setting	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	341	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	342	1	Moreover	moreover	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	342	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	342	3	we	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	342	4	will	will	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	342	5	further	far	RBR
work_vbzp3iewy5h7rhfiok5o6i3du4	342	6	conduct	conduct	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	342	7	experiments	experiment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	342	8	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	342	9	compare	compare	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	342	10	our	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	342	11	methods	method	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	342	12	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	342	13	other	other	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	342	14	relevant	relevant	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	342	15	works	work	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	342	16	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	342	17	such	such	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	342	18	as	as	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	342	19	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	342	20	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	342	21	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	342	22	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	342	23	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	342	24	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	342	25	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	342	26	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	342	27	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	342	28	Burkett	Burkett	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	342	29	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	342	30	Klein	Klein	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	342	31	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	342	32	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	342	33	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	342	34	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	343	1	Acknowledgments	acknowledgment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	343	2	We	-PRON-	PRP
work_vbzp3iewy5h7rhfiok5o6i3du4	343	3	would	would	MD
work_vbzp3iewy5h7rhfiok5o6i3du4	343	4	like	like	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	343	5	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	343	6	thank	thank	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	343	7	Philipp	Philipp	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	343	8	Koehn	Koehn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	343	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	343	10	three	three	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	343	11	anonymous	anonymous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	343	12	reviewers	reviewer	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	343	13	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	343	14	their	-PRON-	PRP$
work_vbzp3iewy5h7rhfiok5o6i3du4	343	15	valuable	valuable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	343	16	comments	comment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	343	17	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	343	18	suggestions	suggestion	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	343	19	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	344	1	The	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	344	2	research	research	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	344	3	work	work	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	344	4	has	have	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	344	5	been	be	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	344	6	funded	fund	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	344	7	by	by	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	344	8	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	344	9	Hi	Hi	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	344	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	344	11	Tech	Tech	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	344	12	Research	Research	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	344	13	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	344	14	Development	Development	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	344	15	Program	Program	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	344	16	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	344	17	“	"	``
work_vbzp3iewy5h7rhfiok5o6i3du4	344	18	863	863	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	344	19	”	"	''
work_vbzp3iewy5h7rhfiok5o6i3du4	344	20	Program	Program	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	344	21	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	344	22	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	344	23	China	China	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	344	24	under	under	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	344	25	Grant	Grant	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	344	26	No	No	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	344	27	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	345	1	2011AA01A207	2011aa01a207	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	345	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	345	3	2012AA011101	2012aa011101	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	345	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	345	5	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	345	6	2012AA011102	2012aa011102	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	345	7	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	346	1	References	reference	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	346	2	Phil	Phil	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	346	3	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	346	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	346	5	Trevor	Trevor	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	346	6	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	346	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	346	8	Miles	Miles	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	346	9	Osborne	Osborne	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	346	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	347	1	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	347	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	348	1	Bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	348	2	synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	348	3	grammar	grammar	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	348	4	induction	induction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	348	5	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	349	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	349	2	Advances	advance	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	349	3	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	349	4	Neural	Neural	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	349	5	Information	Information	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	349	6	Processing	Processing	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	349	7	Systems	Systems	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	349	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	349	9	volume	volume	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	349	10	21	21	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	349	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	349	12	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	349	13	161	161	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	349	14	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	349	15	168	168	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	349	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	350	1	Phil	Phil	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	350	2	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	350	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	350	4	Trevor	Trevor	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	350	5	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	350	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	350	7	Chris	Chris	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	350	8	Dyer	Dyer	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	350	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	350	10	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	350	11	Miles	Miles	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	350	12	Osborne	Osborne	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	350	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	351	1	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	351	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	352	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	352	2	gibbs	gibb	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	352	3	sampler	sampler	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	352	4	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	352	5	phrasal	phrasal	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	352	6	synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	352	7	grammar	grammar	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	352	8	induction	induction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	352	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	353	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	353	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	353	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	354	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	354	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	354	3	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	354	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	354	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	354	6	782	782	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	354	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	354	8	790	790	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	354	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	355	1	Phil	Phil	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	355	2	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	355	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	355	4	Trevor	Trevor	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	355	5	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	355	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	356	1	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	356	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	357	1	Inducing	induce	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	357	2	synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	357	3	grammars	grammar	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	357	4	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	357	5	slice	slice	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	357	6	sampling	sampling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	357	7	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	358	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	358	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	358	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	359	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	359	2	NAACL	NAACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	359	3	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	359	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	359	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	359	6	238	238	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	359	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	359	8	241	241	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	359	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	360	1	David	David	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	360	2	Burkett	Burkett	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	360	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	360	4	Dan	Dan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	360	5	Klein	Klein	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	360	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	361	1	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	361	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	362	1	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	362	2	languages	language	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	362	3	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	362	4	better	well	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	362	5	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	362	6	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	362	7	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	362	8	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	362	9	syntactic	syntactic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	362	10	Parsing	parsing	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	362	11	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	362	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	363	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	363	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	363	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	364	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	364	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	364	3	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	364	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	364	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	364	6	877	877	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	364	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	364	8	886	886	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	364	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	365	1	David	David	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	365	2	Burkett	Burkett	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	365	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	365	4	John	John	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	365	5	Blitzer	Blitzer	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	365	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	365	7	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	365	8	Dan	Dan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	365	9	Klein	Klein	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	365	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	366	1	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	366	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	367	1	Joint	joint	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	367	2	parsing	parsing	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	367	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	367	4	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	367	5	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	367	6	weakly	weakly	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	367	7	synchronized	synchronize	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	367	8	grammars	grammar	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	367	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	368	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	368	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	368	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	369	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	369	2	NAACL	NAACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	369	3	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	369	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	369	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	369	6	127	127	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	369	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	369	8	135	135	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	369	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	370	1	David	David	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	370	2	Burkett	Burkett	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	370	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	370	4	Dan	Dan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	370	5	Klein	Klein	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	370	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	371	1	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	371	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	372	1	Transforming	transform	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	372	2	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	372	3	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	372	4	improve	improve	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	372	5	syntactic	syntactic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	372	6	convergence	convergence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	372	7	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	373	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	373	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	373	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	374	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	374	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	374	3	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	374	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	374	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	374	6	863	863	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	374	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	374	8	872	872	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	374	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	375	1	David	David	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	375	2	Chiang	Chiang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	375	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	376	1	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	376	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	377	1	Hierarchical	hierarchical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	377	2	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	377	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	377	4	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	377	5	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	377	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	378	1	Computational	Computational	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	378	2	Linguistics	Linguistics	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	378	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	378	4	33	33	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	378	5	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	378	6	2	2	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	378	7	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	378	8	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	379	1	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	379	2	201	201	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	379	3	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	379	4	228	228	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	379	5	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	380	1	Dekai	Dekai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	380	2	Wu	Wu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	380	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	381	1	1996	1996	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	381	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	382	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	382	2	polynomial	polynomial	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	382	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	382	4	time	time	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	382	5	algorithm	algorithm	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	382	6	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	382	7	statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	382	8	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	382	9	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	382	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	383	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	383	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	383	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	384	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	384	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	384	3	1996	1996	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	384	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	384	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	384	6	152	152	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	384	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	384	8	158	158	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	384	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	385	1	Dekai	Dekai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	385	2	Wu	Wu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	385	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	386	1	1997	1997	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	386	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	387	1	Stochastic	stochastic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	387	2	inversion	inversion	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	387	3	transduction	transduction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	387	4	grammars	grammar	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	387	5	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	387	6	bilingual	bilingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	387	7	parsing	parsing	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	387	8	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	387	9	parallel	parallel	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	387	10	corpora	corpora	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	387	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	388	1	Computational	Computational	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	388	2	Linguistics	Linguistics	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	388	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	388	4	23:377	23:377	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	388	5	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	388	6	404	404	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	388	7	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	389	1	Trevor	Trevor	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	389	2	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	389	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	389	4	Phil	Phil	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	389	5	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	389	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	390	1	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	390	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	391	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	391	2	bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	391	3	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	391	4	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	391	5	syntax	syntax	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	391	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	391	7	directed	direct	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	391	8	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	391	9	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	391	10	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	391	11	grammar	grammar	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	391	12	induction	induction	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	391	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	392	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	392	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	392	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	393	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	393	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	393	3	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	393	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	393	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	393	6	352	352	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	393	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	393	8	361	361	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	393	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	394	1	Trevor	Trevor	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	394	2	Cohn	Cohn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	394	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	394	4	Phil	Phil	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	394	5	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	394	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	394	7	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	394	8	Sharon	Sharon	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	394	9	Goldwater	Goldwater	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	394	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	395	1	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	395	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	396	1	Inducing	induce	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	396	2	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	396	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	396	4	substitution	substitution	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	396	5	grammars	grammar	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	396	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	397	1	Journal	Journal	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	397	2	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	397	3	Machine	Machine	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	397	4	Learning	Learning	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	397	5	Research	Research	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	397	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	397	7	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	397	8	3053	3053	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	397	9	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	397	10	3096	3096	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	397	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	398	1	Brooke	Brooke	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	398	2	Cowan	Cowan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	398	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	398	4	Ivona	Ivona	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	398	5	Kucerova	Kucerova	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	398	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	398	7	Michael	Michael	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	398	8	Collins	Collins	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	398	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	399	1	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	399	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	400	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	400	2	discriminative	discriminative	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	400	3	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	400	4	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	400	5	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	400	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	400	7	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	400	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	400	9	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	400	10	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	400	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	401	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	401	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	401	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	402	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	402	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	402	3	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	402	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	402	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	402	6	232	232	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	402	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	402	8	241	241	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	402	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	403	1	John	John	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	403	2	DeNero	DeNero	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	403	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	403	4	Dan	Dan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	403	5	Klein	Klein	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	403	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	404	1	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	404	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	405	1	Tailoring	tailor	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	405	2	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	405	3	alignments	alignment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	405	4	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	405	5	syntactic	syntactic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	405	6	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	405	7	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	405	8	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	406	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	406	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	406	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	407	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	407	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	407	3	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	407	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	407	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	407	6	17	17	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	407	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	407	8	24	24	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	407	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	408	1	John	John	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	408	2	DeNero	DeNero	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	408	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	408	4	Jakob	Jakob	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	408	5	Uszkoreit	Uszkoreit	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	408	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	409	1	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	409	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	410	1	Inducing	induce	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	410	2	sentence	sentence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	410	3	structure	structure	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	410	4	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	410	5	parallel	parallel	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	410	6	corpora	corpora	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	410	7	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	410	8	reordering	reordering	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	410	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	411	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	411	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	411	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	412	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	412	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	412	3	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	412	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	412	5	pages	page	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	412	6	193	193	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	412	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	412	8	203	203	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	412	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	413	1	Chris	Chris	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	413	2	Dyer	Dyer	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	413	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	414	1	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	414	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	415	1	Two	two	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	415	2	monolingual	monolingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	415	3	parses	parse	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	415	4	are	be	VBP
work_vbzp3iewy5h7rhfiok5o6i3du4	415	5	better	well	JJR
work_vbzp3iewy5h7rhfiok5o6i3du4	415	6	than	than	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	415	7	one	one	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	415	8	(	(	-LRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	415	9	synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	415	10	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	415	11	)	)	-RRB-
work_vbzp3iewy5h7rhfiok5o6i3du4	415	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	416	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	416	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	416	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	417	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	417	2	NAACL	NAACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	417	3	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	417	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	417	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	417	6	263	263	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	417	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	417	8	266	266	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	417	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	418	1	Jason	Jason	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	418	2	Eisner	Eisner	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	418	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	419	1	2003	2003	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	419	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	420	1	Learning	learn	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	420	2	non	non	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	420	3	-	-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	420	4	isomorphic	isomorphic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	420	5	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	420	6	mappings	mapping	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	420	7	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	420	8	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	420	9	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	420	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	421	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	421	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	421	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	422	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	422	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	422	3	2003	2003	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	422	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	422	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	422	6	205	205	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	422	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	422	8	208	208	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	422	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	423	1	Michel	Michel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	423	2	Galley	Galley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	423	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	423	4	Mark	Mark	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	423	5	Hopkins	Hopkins	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	423	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	423	7	Kevin	Kevin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	423	8	Knight	Knight	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	423	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	423	10	Daniel	Daniel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	423	11	Marcu	Marcu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	423	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	424	1	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	424	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	425	1	What	what	WP
work_vbzp3iewy5h7rhfiok5o6i3du4	425	2	’s	’	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	425	3	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	425	4	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	425	5	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	425	6	rule	rule	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	425	7	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	426	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	426	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	426	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	427	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	427	2	HLT	HLT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	427	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	427	4	NAACL	NAACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	427	5	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	427	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	427	7	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	427	8	273–280	273–280	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	427	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	428	1	Michel	Michel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	2	Galley	Galley	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	428	4	Jonathan	Jonathan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	5	Graehl	Graehl	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	428	7	Kevin	Kevin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	8	Knight	Knight	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	428	10	Daniel	Daniel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	11	Marcu	Marcu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	12	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	428	13	Steve	Steve	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	14	DeNeefe	DeNeefe	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	15	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	428	16	Wei	Wei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	17	Wang	Wang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	18	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	428	19	Ignacio	Ignacio	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	20	Thayer	Thayer	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	428	21	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	429	1	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	429	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	430	1	Scalable	scalable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	430	2	inference	inference	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	430	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	430	4	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	430	5	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	430	6	context	context	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	430	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	430	8	rich	rich	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	430	9	syntactic	syntactic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	430	10	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	430	11	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	430	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	431	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	431	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	431	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	432	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	432	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	432	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	432	4	COLING	COLING	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	432	5	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	432	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	432	7	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	432	8	961	961	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	432	9	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	432	10	968	968	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	432	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	433	1	Jonathan	Jonathan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	2	Weese	Weese	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	433	4	Juri	Juri	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	5	Ganitkevitch	Ganitkevitch	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	433	7	Chris	Chris	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	8	Callison-	Callison-	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	9	Burch	Burch	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	433	11	Matt	Matt	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	12	Post	Post	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	13	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	433	14	Adam	Adam	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	15	Lopez	Lopez	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	433	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	434	1	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	434	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	435	1	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	435	2	3.0	3.0	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	435	3	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	435	4	syntax	syntax	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	435	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	435	6	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	435	7	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	435	8	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	435	9	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	435	10	the	the	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	435	11	thrax	thrax	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	435	12	Grammar	Grammar	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	435	13	Extractor	Extractor	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	435	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	436	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	436	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	436	3	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	436	4	WMT11	WMT11	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	436	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	436	6	pages	page	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	436	7	478-	478-	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	436	8	484	484	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	436	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	437	1	Liang	Liang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	437	2	Huang	Huang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	437	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	437	4	Kevin	Kevin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	437	5	Knight	Knight	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	437	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	437	7	Aravind	Aravind	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	437	8	Joshi	Joshi	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	437	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	438	1	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	438	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	439	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	439	2	syntax	syntax	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	439	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	439	4	directed	direct	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	439	5	translator	translator	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	439	6	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	439	7	extended	extended	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	439	8	domain	domain	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	439	9	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	439	10	locality	locality	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	439	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	440	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	440	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	440	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	441	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	441	2	AMTA	AMTA	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	441	3	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	441	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	441	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	441	6	65	65	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	441	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	441	8	73	73	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	441	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	442	1	Philipp	Philipp	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	442	2	Koehn	Koehn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	442	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	442	4	Franz	Franz	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	442	5	Och	Och	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	442	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	442	7	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	442	8	Daniel	Daniel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	442	9	Marcu	Marcu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	442	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	443	1	2003	2003	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	443	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	444	1	Statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	444	2	phrase	phrase	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	444	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	444	4	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	444	5	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	444	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	444	7	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	444	8	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	444	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	445	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	445	2	HLT	HLT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	445	3	/	/	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	445	4	NAACL	NAACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	445	5	2003	2003	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	445	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	445	7	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	445	8	48	48	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	445	9	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	445	10	54	54	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	445	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	446	1	253	253	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	446	2	Philipp	Philipp	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	446	3	Koehn	Koehn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	446	4	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	447	1	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	447	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	448	1	Statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	448	2	significance	significance	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	448	3	tests	test	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	448	4	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	448	5	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	448	6	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	448	7	evaluation	evaluation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	448	8	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	449	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	449	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	449	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	450	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	450	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	450	3	2004	2004	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	450	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	450	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	450	6	388–395	388–395	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	450	7	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	451	1	Philipp	Philipp	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	2	Koehn	Koehn	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	451	4	Hieu	Hieu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	5	Hoang	Hoang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	451	7	Alexandra	Alexandra	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	8	Birch	Birch	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	451	10	Chris	Chris	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	11	Callison	Callison	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	12	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	451	13	Burch	Burch	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	451	15	Marcello	Marcello	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	16	Federico	Federico	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	17	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	451	18	Nicola	Nicola	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	19	Bertoldi	Bertoldi	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	451	21	Brooke	Brooke	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	22	Cowan	Cowan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	23	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	451	24	Wade	Wade	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	25	Shen	Shen	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	26	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	451	27	Christine	Christine	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	28	Moran	Moran	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	29	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	451	30	RichDUG	RichDUG	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	451	31	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	32	=	=	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	451	33	HQV	HQV	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	34	�	�	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	451	35	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	36	&	&	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	451	37	KULV	KULV	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	38	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	39	'	'	POS
work_vbzp3iewy5h7rhfiok5o6i3du4	451	40	\HU	\HU	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	41	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	42	DQG	DQG	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	43	�	�	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	451	44	2QGĜHM	2qgĝhm	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	451	45	�	�	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	451	46	%	%	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	451	47	RMDU	rmdu	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	451	48	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	452	1	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	452	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	453	1	Moses	Moses	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	453	2	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	453	3	open	open	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	453	4	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	453	5	toolkit	toolkit	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	453	6	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	453	7	statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	453	8	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	453	9	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	453	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	454	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	454	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	454	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	455	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	455	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	455	3	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	455	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	455	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	455	6	177	177	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	455	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	455	8	180	180	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	455	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	456	1	Abby	Abby	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	456	2	Levenberg	Levenberg	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	456	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	456	4	Chris	Chris	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	456	5	Dyer	Dyer	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	456	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	456	7	Phil	Phil	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	456	8	Blunsom	Blunsom	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	456	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	457	1	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	457	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	458	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	458	2	bayesian	bayesian	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	458	3	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	458	4	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	458	5	learning	learn	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	458	6	SCFGs	SCFGs	NNPS
work_vbzp3iewy5h7rhfiok5o6i3du4	458	7	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	458	8	discontiguous	discontiguous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	458	9	Rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	458	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	459	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	459	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	459	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	460	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	460	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	460	3	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	460	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	460	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	460	6	223	223	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	460	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	460	8	232	232	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	460	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	461	1	Zhifei	Zhifei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	2	Li	Li	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	461	4	Chris	Chris	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	5	Callison	Callison	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	461	7	Burch	Burch	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	461	9	Chris	Chris	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	10	Dyer	Dyer	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	461	12	Juri	Juri	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	13	Ganitkevitch	Ganitkevitch	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	461	15	Sanjeev	Sanjeev	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	16	Khudanpur	Khudanpur	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	17	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	461	18	Lane	Lane	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	19	Schwartz	Schwartz	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	20	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	461	21	Wren	Wren	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	461	22	N.G.	N.G.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	462	1	Thornton	Thornton	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	462	2	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	462	3	Jonathan	Jonathan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	462	4	Weese	Weese	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	462	5	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	462	6	Omar	Omar	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	462	7	F.	F.	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	462	8	Zaidan	Zaidan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	462	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	463	1	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	463	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	464	1	Joshua	Joshua	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	464	2	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	464	3	An	an	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	464	4	open	open	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	464	5	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	464	6	toolkit	toolkit	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	464	7	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	464	8	parsing	parse	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	464	9	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	464	10	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	464	11	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	464	12	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	464	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	465	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	465	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	465	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	466	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	466	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	466	3	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	466	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	466	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	466	6	135	135	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	466	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	466	8	139	139	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	466	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	467	1	Shujie	Shujie	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	467	2	Liu	Liu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	467	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	467	4	Chi	Chi	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	467	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	467	6	Ho	Ho	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	467	7	Li	Li	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	467	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	467	9	Mu	Mu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	467	10	Li	Li	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	467	11	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	467	12	Ming	Ming	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	467	13	Zhou	Zhou	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	467	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	468	1	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	468	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	469	1	Re-	re-	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	469	2	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	469	3	monolingual	monolingual	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	469	4	parser	parser	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	469	5	bilingually	bilingually	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	469	6	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	469	7	syntactic	syntactic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	469	8	SMT	SMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	469	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	470	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	470	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	470	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	471	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	471	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	471	3	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	471	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	471	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	471	6	854	854	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	471	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	471	8	862	862	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	471	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	472	1	Yang	Yang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	472	2	Liu	Liu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	472	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	472	4	Qun	Qun	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	472	5	Liu	Liu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	472	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	472	7	Shouxun	Shouxun	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	472	8	Lin	Lin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	472	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	473	1	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	473	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	474	1	Tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	474	2	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	474	3	to-	to-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	474	4	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	474	5	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	474	6	template	template	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	474	7	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	474	8	statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	474	9	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	474	10	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	474	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	475	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	475	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	475	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	476	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	476	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	476	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	476	4	COLING	COLING	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	476	5	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	476	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	476	7	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	476	8	609	609	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	476	9	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	476	10	616	616	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	476	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	477	1	Yang	Yang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	477	2	Liu	Liu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	477	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	477	4	Yajuan	Yajuan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	477	5	Lv	Lv	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	477	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	477	7	Qun	Qun	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	477	8	Liu	Liu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	477	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	478	1	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	478	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	479	1	Improving	improve	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	479	2	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	479	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	479	4	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	479	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	479	6	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	479	7	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	479	8	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	479	9	packed	packed	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	479	10	forests	forest	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	479	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	480	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	480	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	480	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	481	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	481	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	481	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	481	4	IJCNLP	IJCNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	481	5	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	481	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	481	7	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	481	8	558	558	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	481	9	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	481	10	566	566	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	481	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	482	1	Daniel	Daniel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	482	2	Marcu	Marcu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	482	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	482	4	Wei	Wei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	482	5	Wang	Wang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	482	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	482	7	Abdessamad	Abdessamad	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	482	8	Echihabi	Echihabi	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	482	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	482	10	Kevin	Kevin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	482	11	Knight	Knight	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	482	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	483	1	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	483	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	484	1	SPMT	SPMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	484	2	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	484	3	Statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	484	4	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	484	5	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	484	6	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	484	7	syntactified	syntactified	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	484	8	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	484	9	language	language	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	484	10	phrases	phrase	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	484	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	485	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	485	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	485	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	486	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	486	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	486	3	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	486	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	486	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	486	6	44	44	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	486	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	486	8	52	52	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	486	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	487	1	Franz	Franz	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	487	2	Och	Och	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	487	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	487	4	2003	2003	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	487	5	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	488	1	Minimum	minimum	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	488	2	error	error	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	488	3	rate	rate	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	488	4	training	training	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	488	5	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	488	6	statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	488	7	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	488	8	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	488	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	489	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	489	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	489	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	490	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	490	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	490	3	2003	2003	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	490	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	490	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	490	6	160	160	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	490	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	490	8	167	167	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	490	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	491	1	Kishore	Kishore	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	491	2	Papineni	Papineni	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	491	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	491	4	Salim	Salim	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	491	5	Roukos	Roukos	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	491	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	491	7	Todd	Todd	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	491	8	Ward	Ward	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	491	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	491	10	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	491	11	Wei-	Wei-	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	491	12	Jing	Jing	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	491	13	Zhu	Zhu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	491	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	492	1	2002	2002	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	492	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	493	1	BLEU	BLEU	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	493	2	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	493	3	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	493	4	method	method	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	493	5	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	493	6	automatic	automatic	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	493	7	evaluation	evaluation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	493	8	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	493	9	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	493	10	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	493	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	494	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	494	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	494	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	495	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	495	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	495	3	2002	2002	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	495	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	495	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	495	6	311	311	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	495	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	495	8	318	318	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	495	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	496	1	Slav	Slav	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	496	2	Petrov	Petrov	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	496	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	496	4	Leon	Leon	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	496	5	Barrett	Barrett	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	496	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	496	7	Romain	Romain	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	496	8	Thibaux	Thibaux	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	496	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	496	10	Dan	Dan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	496	11	Klein	Klein	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	496	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	497	1	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	497	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	498	1	Learning	learn	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	498	2	accurate	accurate	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	498	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	498	4	compact	compact	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	498	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	498	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	498	7	interpretable	interpretable	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	498	8	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	498	9	annotation	annotation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	498	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	499	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	499	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	499	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	500	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	500	2	COLING-	COLING-	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	500	3	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	500	4	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	500	5	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	500	6	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	500	7	433	433	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	500	8	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	500	9	440	440	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	500	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	501	1	Chris	Chris	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	501	2	Quirk	Quirk	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	501	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	501	4	Arul	Arul	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	501	5	Menezes	Menezes	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	501	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	501	7	Colin	Colin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	501	8	Cherry	Cherry	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	501	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	502	1	2005	2005	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	502	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	503	1	Dependency	dependency	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	503	2	treelet	treelet	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	503	3	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	503	4	:	:	:
work_vbzp3iewy5h7rhfiok5o6i3du4	503	5	syntactically	syntactically	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	503	6	informed	inform	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	503	7	phrasal	phrasal	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	503	8	SMT	SMT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	503	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	504	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	504	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	504	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	505	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	505	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	505	3	2005	2005	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	505	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	505	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	505	6	271	271	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	505	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	505	8	279	279	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	505	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	506	1	Libin	Libin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	506	2	Shen	Shen	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	506	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	506	4	Jinxi	Jinxi	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	506	5	Xu	Xu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	506	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	506	7	Ralph	Ralph	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	506	8	Weischedel	Weischedel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	506	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	507	1	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	507	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	508	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	508	2	new	new	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	508	3	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	4	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	508	5	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	6	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	508	7	dependency	dependency	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	8	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	9	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	10	algorithm	algorithm	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	11	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	12	a	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	508	13	target	target	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	14	dependency	dependency	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	15	language	language	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	16	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	508	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	509	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	509	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	509	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	510	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	510	2	ACL-08	ACL-08	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	510	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	510	4	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	510	5	577	577	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	510	6	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	510	7	585	585	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	510	8	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	511	1	Wei	Wei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	511	2	Wang	Wang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	511	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	511	4	Kevin	Kevin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	511	5	Knight	Knight	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	511	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	511	7	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	511	8	Daniel	Daniel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	511	9	Marcu	Marcu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	511	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	512	1	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	512	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	513	1	Binarizing	binarize	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	513	2	syntax	syntax	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	513	3	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	513	4	to	to	TO
work_vbzp3iewy5h7rhfiok5o6i3du4	513	5	improve	improve	VB
work_vbzp3iewy5h7rhfiok5o6i3du4	513	6	syntax	syntax	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	513	7	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	513	8	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	513	9	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	513	10	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	513	11	accuracy	accuracy	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	513	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	514	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	514	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	514	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	515	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	515	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	515	3	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	515	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	515	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	515	6	746	746	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	515	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	515	8	754	754	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	515	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	516	1	Wei	Wei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	516	2	Wang	Wang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	516	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	516	4	Jonathan	Jonathan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	516	5	May	May	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	516	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	516	7	Kevin	Kevin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	516	8	Knight	Knight	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	516	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	516	10	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	516	11	Daniel	Daniel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	516	12	Marcu	Marcu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	516	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	517	1	2010	2010	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	517	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	518	1	Re	Re	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	518	2	-	-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	3	structuring	structuring	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	518	5	re	re	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	6	-	-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	7	labeling	labeling	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	518	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	518	10	re-	re-	RB
work_vbzp3iewy5h7rhfiok5o6i3du4	518	11	aligning	align	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	518	12	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	13	syntax	syntax	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	518	15	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	16	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	17	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	518	18	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	519	1	Computational	Computational	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	519	2	Linguistics	Linguistics	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	519	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	519	4	36(2):247–277	36(2):247–277	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	519	5	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	520	1	Feifei	Feifei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	520	2	Zhai	Zhai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	520	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	520	4	Jiajun	Jiajun	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	520	5	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	520	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	520	7	Yu	Yu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	520	8	Zhou	Zhou	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	520	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	520	10	Chengqing	Chengqing	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	520	11	Zong	Zong	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	520	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	521	1	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	521	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	522	1	Tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	522	2	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	522	3	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	522	4	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	522	5	without	without	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	522	6	using	use	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	522	7	parse	parse	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	522	8	trees	tree	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	522	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	523	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	523	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	523	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	524	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	524	2	COLING	COLING	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	524	3	2012	2012	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	524	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	524	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	524	6	3037-	3037-	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	524	7	3054	3054	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	524	8	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	525	1	Hao	Hao	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	525	2	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	525	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	525	4	Liang	Liang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	525	5	Huang	Huang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	525	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	525	7	Daniel	Daniel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	525	8	Gildea	Gildea	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	525	9	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	525	10	Kevin	Kevin	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	525	11	Knight	Knight	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	525	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	526	1	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	526	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	527	1	Synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	527	2	binarization	binarization	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	527	3	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	527	4	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	527	5	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	527	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	528	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	528	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	528	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	529	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	529	2	HLT	HLT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	529	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	529	4	NAACL	NAACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	529	5	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	529	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	529	7	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	529	8	256	256	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	529	9	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	529	10	263	263	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	529	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	530	1	Hao	Hao	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	530	2	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	530	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	530	4	Daniel	Daniel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	530	5	Gildea	Gildea	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	530	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	530	7	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	530	8	David	David	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	530	9	Chiang	Chiang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	530	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	531	1	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	531	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	532	1	Extracting	extract	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	532	2	synchronous	synchronous	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	532	3	grammars	grammar	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	532	4	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	532	5	from	from	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	532	6	word	word	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	532	7	level	level	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	532	8	alignments	alignment	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	532	9	in	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	532	10	linear	linear	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	532	11	time	time	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	532	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	533	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	533	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	533	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	534	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	534	2	COLING	COLING	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	534	3	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	534	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	534	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	534	6	1081	1081	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	534	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	534	8	1088	1088	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	534	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	535	1	Hao	Hao	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	535	2	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	535	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	535	4	Licheng	Licheng	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	535	5	Fang	Fang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	535	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	535	7	Peng	Peng	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	535	8	Xu	Xu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	535	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	535	10	Xiaoyun	Xiaoyun	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	535	11	Wu	Wu	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	535	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	536	1	2011a	2011a	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	536	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	537	1	Binarized	binarized	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	537	2	forest	forest	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	537	3	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	537	4	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	537	5	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	537	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	538	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	538	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	538	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	539	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	539	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	539	3	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	539	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	539	5	pages	pages	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	539	6	835	835	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	539	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	539	8	845	845	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	539	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	540	1	Hui	Hui	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	2	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	540	4	Min	Min	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	5	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	540	7	Haizhou	Haizhou	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	8	Li	Li	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	9	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	540	10	Aiti	Aiti	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	11	Aw	Aw	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	12	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	540	13	Chew	Chew	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	14	Lim	Lim	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	15	Tan	Tan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	540	16	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	541	1	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	541	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	542	1	Forest	forest	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	542	2	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	542	3	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	542	4	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	542	5	sequence	sequence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	542	6	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	542	7	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	542	8	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	542	9	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	542	10	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	543	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	543	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	543	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	544	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	544	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	544	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	544	4	IJCNLP	IJCNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	544	5	2009	2009	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	544	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	544	7	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	544	8	172	172	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	544	9	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	544	10	180	180	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	544	11	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	545	1	Jiajun	Jiajun	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	545	2	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	545	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	545	4	Feifei	Feifei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	545	5	Zhai	Zhai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	545	6	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	545	7	Chengqing	Chengqing	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	545	8	Zong	Zong	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	545	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	546	1	2011b	2011b	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	546	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	547	1	Augmenting	augment	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	547	2	string	string	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	547	4	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	547	6	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	7	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	8	models	model	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	547	9	with	with	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	10	fuzzy	fuzzy	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	547	11	use	use	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	12	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	13	source	source	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	14	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	547	15	side	side	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	16	syntax	syntax	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	547	17	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	548	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	548	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	548	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	549	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	549	2	EMNLP	EMNLP	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	549	3	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	549	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	549	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	549	6	204	204	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	549	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	549	8	215	215	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	549	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	550	1	Min	Min	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	2	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	3	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	550	4	Hongfei	Hongfei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	5	Jiang	Jiang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	6	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	550	7	Ai	Ai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	8	Ti	Ti	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	9	Aw	Aw	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	550	11	Jun	Jun	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	12	Sun	Sun	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	13	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	550	14	Chew	Chew	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	15	Lim	Lim	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	16	Tan	Tan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	17	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	550	18	Sheng	Sheng	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	19	Li	Li	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	550	20	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	551	1	2007	2007	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	551	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	552	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	552	2	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	552	3	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	552	4	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	552	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	552	6	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	552	7	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	552	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	552	9	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	552	10	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	552	11	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	552	12	statistical	statistical	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	552	13	Machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	552	14	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	552	15	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	553	1	MT	MT	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	553	2	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	553	3	Summit-07	Summit-07	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	553	4	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	554	1	pages	page	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	554	2	535	535	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	554	3	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	554	4	542	542	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	554	5	Min	Min	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	6	Zhang	Zhang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	7	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	554	8	Hongfei	Hongfei	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	9	Jiang	Jiang	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	10	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	554	11	Ai	Ai	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	12	ti	ti	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	13	Aw	aw	UH
work_vbzp3iewy5h7rhfiok5o6i3du4	554	14	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	554	15	Haizhou	Haizhou	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	16	Li	Li	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	17	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	554	18	Chew	Chew	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	19	Lim	Lim	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	20	Tan	Tan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	21	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	554	22	Sheng	Sheng	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	23	Li	Li	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	554	24	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	555	1	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	555	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	556	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	556	2	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	556	3	sequence	sequence	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	556	4	alignment	alignment	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	556	5	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	556	6	based	base	VBN
work_vbzp3iewy5h7rhfiok5o6i3du4	556	7	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	556	8	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	556	9	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	556	10	-	-	HYPH
work_vbzp3iewy5h7rhfiok5o6i3du4	556	11	tree	tree	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	556	12	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	556	13	model	model	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	556	14	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	557	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	557	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	557	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	558	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	558	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	558	3	2008	2008	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	558	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	558	5	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	558	6	559	559	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	558	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	558	8	567	567	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	558	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	559	1	Andreas	Andreas	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	559	2	Zollmann	Zollmann	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	559	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	559	4	Ashish	Ashish	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	559	5	Venugopal	Venugopal	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	559	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	560	1	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	560	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	561	1	Syntax	Syntax	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	561	2	augmented	augment	VBD
work_vbzp3iewy5h7rhfiok5o6i3du4	561	3	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	561	4	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	561	5	via	via	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	561	6	chart	chart	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	561	7	parsing	parse	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	561	8	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	562	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	562	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	562	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	563	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	563	2	Workshop	Workshop	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	563	3	on	on	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	563	4	Statistical	Statistical	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	563	5	Machine	Machine	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	563	6	Translation	Translation	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	563	7	2006	2006	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	563	8	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	563	9	pages	page	VBZ
work_vbzp3iewy5h7rhfiok5o6i3du4	563	10	138	138	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	563	11	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	563	12	141	141	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	563	13	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	564	1	Andreas	Andreas	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	564	2	Zollmann	Zollmann	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	564	3	and	and	CC
work_vbzp3iewy5h7rhfiok5o6i3du4	564	4	Stephan	Stephan	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	564	5	Vogel	Vogel	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	564	6	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	565	1	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	565	2	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	566	1	A	a	DT
work_vbzp3iewy5h7rhfiok5o6i3du4	566	2	word-	word-	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	566	3	class	class	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	566	4	approach	approach	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	566	5	to	to	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	566	6	labeling	label	VBG
work_vbzp3iewy5h7rhfiok5o6i3du4	566	7	PSCFG	pscfg	JJ
work_vbzp3iewy5h7rhfiok5o6i3du4	566	8	rules	rule	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	566	9	for	for	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	566	10	machine	machine	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	566	11	translation	translation	NN
work_vbzp3iewy5h7rhfiok5o6i3du4	566	12	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	567	1	In	in	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	567	2	Proc	Proc	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	567	3	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	568	1	of	of	IN
work_vbzp3iewy5h7rhfiok5o6i3du4	568	2	ACL	ACL	NNP
work_vbzp3iewy5h7rhfiok5o6i3du4	568	3	2011	2011	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	568	4	,	,	,
work_vbzp3iewy5h7rhfiok5o6i3du4	568	5	pages	page	NNS
work_vbzp3iewy5h7rhfiok5o6i3du4	568	6	1	1	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	568	7	-	-	SYM
work_vbzp3iewy5h7rhfiok5o6i3du4	568	8	11	11	CD
work_vbzp3iewy5h7rhfiok5o6i3du4	568	9	.	.	.
work_vbzp3iewy5h7rhfiok5o6i3du4	569	1	254	254	CD