id	sid	tid	token	lemma	pos
work_7jtgrdqc4zeb7lul7vifzjglr4	1	1	Transactions	transaction	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	1	2	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	1	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	1	4	Association	Association	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	1	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	1	6	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	1	7	Linguistics	Linguistics	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	1	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	1	9	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	1	10	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	1	11	2013	2013	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	1	12	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	1	13	219–230	219–230	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	1	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	2	1	Action	Action	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	2	2	Editor	Editor	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	2	3	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	2	4	Brian	Brian	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	2	5	Roark	Roark	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	2	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	3	1	Submitted	submit	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	3	2	1/2013	1/2013	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	3	3	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	3	4	Revised	Revised	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	3	5	3/2013	3/2013	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	3	6	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	3	7	Published	publish	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	3	8	5/2013	5/2013	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	3	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	4	1	c	c	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	4	2	©	©	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	4	3	2013	2013	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	4	4	Association	Association	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	4	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	4	6	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	4	7	Linguistics	Linguistics	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	4	8	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	5	1	Joint	Joint	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	2	Arc	Arc	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	3	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	5	4	factored	factored	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	5	5	Parsing	Parsing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	5	7	Syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	5	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	5	9	Semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	5	10	Dependencies	Dependencies	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	5	11	Xavier	Xavier	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	12	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	13	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	5	14	Xavier	Xavier	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	15	Carreras	Carreras	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	5	17	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	18	Màrquez	Màrquez	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	19	TALP	TALP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	20	Research	Research	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	21	Center	Center	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	22	Universitat	Universitat	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	23	Politècnica	Politècnica	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	24	de	de	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	5	25	Catalunya	Catalunya	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	26	Jordi	Jordi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	27	Girona	Girona	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	28	1–3	1–3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	5	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	5	30	08034	08034	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	5	31	Barcelona	Barcelona	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	32	{	{	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	5	33	xlluis,carreras,lluism}@lsi.upc.edu	xlluis,carreras,lluism}@lsi.upc.edu	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	34	Abstract	Abstract	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	35	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	5	36	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	5	37	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	5	38	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	39	introduce	introduce	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	5	40	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	5	41	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	5	42	arc	arc	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	5	43	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	5	44	factored	factored	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	5	45	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	5	46	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	5	47	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	5	48	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	5	49	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	5	50	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	5	51	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	5	52	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	6	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	6	2	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	6	3	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	6	4	labeler	labeler	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	6	5	predicts	predict	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	6	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	6	7	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	6	8	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	6	9	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	6	10	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	6	11	connect	connect	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	6	12	predicates	predicate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	6	13	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	6	14	their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	6	15	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	6	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	7	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	7	2	process	process	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	7	3	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	7	4	framed	frame	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	7	5	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	7	6	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	7	7	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	7	8	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	7	9	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	7	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	7	11	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	7	12	allows	allow	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	7	13	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	7	14	control	control	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	7	15	some	some	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	7	16	well	well	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	7	17	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	7	18	formedness	formedness	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	7	19	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	7	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	8	1	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	8	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	8	3	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	8	4	part	part	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	8	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	8	6	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	8	7	define	define	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	8	8	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	8	9	standard	standard	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	8	10	arc	arc	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	8	11	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	8	12	factored	factored	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	8	13	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	8	14	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	8	15	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	8	16	predicts	predict	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	8	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	8	18	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	8	19	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	8	20	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	8	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	9	1	Finally	finally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	9	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	9	3	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	9	4	employ	employ	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	9	5	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	9	6	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	9	7	techniques	technique	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	9	8	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	9	9	produce	produce	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	9	10	consis-	consis-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	9	11	tent	tent	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	9	12	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	9	13	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	9	14	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	9	15	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	9	16	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	9	17	struc-	struc-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	9	18	tures	ture	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	9	19	while	while	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	9	20	searching	search	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	9	21	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	9	22	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	9	23	large	large	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	9	24	space	space	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	9	25	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	9	26	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	9	27	configurations	configuration	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	9	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	10	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	10	2	experiments	experiment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	10	3	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	10	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	10	5	CoNLL-2009	conll-2009	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	10	6	English	English	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	10	7	benchmark	benchmark	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	10	8	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	10	9	ob-	ob-	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	10	10	serve	serve	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	10	11	very	very	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	10	12	competitive	competitive	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	10	13	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	10	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	11	1	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	11	2	Introduction	introduction	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	3	Semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	11	4	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	5	labeling	labeling	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	6	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	11	7	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	11	8	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	11	9	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	11	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	11	11	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	13	identi-	identi-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	11	14	fying	fye	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	11	15	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	11	16	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	11	17	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	18	lexical	lexical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	11	19	predicates	predicate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	11	20	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	21	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	11	22	sen-	sen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	23	tence	tence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	24	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	11	25	labeling	label	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	11	26	them	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	11	27	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	11	28	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	11	29	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	11	30	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	11	31	Gildea	Gildea	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	11	32	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	11	33	Jurafsky	Jurafsky	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	11	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	11	35	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	11	36	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	11	37	Màrquez	Màrquez	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	11	38	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	11	39	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	11	40	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	11	41	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	11	42	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	11	43	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	11	44	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	12	1	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	12	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	12	3	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	12	4	important	important	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	12	5	shallow	shallow	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	12	6	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	12	7	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	8	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	9	NLP	NLP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	12	10	since	since	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	11	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	12	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	12	13	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	14	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	12	15	directly	directly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	12	16	represent	represent	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	12	17	se-	se-	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	12	18	mantic	mantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	12	19	properties	property	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	12	20	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	12	22	type	type	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	23	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	12	24	who	who	WP
work_7jtgrdqc4zeb7lul7vifzjglr4	12	25	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	12	26	did	do	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	12	27	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	12	28	what	what	WP
work_7jtgrdqc4zeb7lul7vifzjglr4	12	29	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	12	30	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	12	31	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	12	32	whom	whom	WP
work_7jtgrdqc4zeb7lul7vifzjglr4	12	33	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	12	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	12	35	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	12	36	how	how	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	12	37	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	12	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	12	39	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	12	40	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	12	41	why	why	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	12	42	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	12	43	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	44	events	event	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	12	45	expressed	express	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	46	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	12	47	predicates	predicate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	12	48	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	12	49	typically	typically	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	12	50	verbs	verbs	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	12	51	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	12	52	nouns	noun	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	12	53	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	12	54	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	13	1	Predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	2	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	13	3	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	4	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	13	5	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	13	6	strongly	strongly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	13	7	related	related	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	13	8	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	13	10	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	13	11	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	13	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	13	14	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	15	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	13	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	13	17	ma-	ma-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	13	18	jority	jority	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	13	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	20	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	21	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	13	22	correspond	correspond	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	13	23	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	24	some	some	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	13	25	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	13	26	constituent	constituent	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	13	28	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	13	29	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	13	30	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	13	31	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	32	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	13	33	connects	connect	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	13	34	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	13	35	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	36	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	37	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	13	38	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	39	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	13	40	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	13	41	strong	strong	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	13	42	indicator	indicator	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	43	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	44	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	13	45	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	13	46	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	13	47	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	14	1	Actually	actually	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	14	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	14	3	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	14	4	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	14	5	represent	represent	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	14	6	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	14	7	abstraction	abstraction	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	14	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	14	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	14	10	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	14	11	form	form	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	14	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	14	13	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	14	14	predicative	predicative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	14	15	event	event	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	14	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	15	1	While	while	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	2	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	15	3	functions	function	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	15	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	5	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	15	6	change	change	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	15	7	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	15	9	form	form	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	15	12	event	event	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	13	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	15	14	e.g.	e.g.	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	15	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	15	16	active	active	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	15	17	vs.	vs.	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	18	passive	passive	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	15	19	forms	form	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	15	20	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	15	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	15	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	15	23	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	15	24	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	15	25	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	26	argu-	argu-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	27	ments	ment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	15	28	remain	remain	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	15	29	invariant	invariant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	15	30	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	31	their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	15	32	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	15	33	realization	realization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	15	34	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	16	1	Consequently	consequently	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	16	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	16	3	since	since	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	16	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	16	5	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	16	6	works	work	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	16	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	16	8	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	16	9	systems	system	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	16	10	have	have	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	16	11	assumed	assume	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	16	12	access	access	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	16	13	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	16	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	16	15	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	16	16	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	16	17	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	16	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	16	19	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	16	20	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	16	21	Gildea	Gildea	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	16	22	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	16	23	Jurafsky	Jurafsky	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	16	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	16	25	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	16	26	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	16	27	Carreras	Carreras	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	16	28	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	16	29	Màrquez	Màrquez	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	16	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	16	31	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	16	32	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	16	33	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	17	1	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	17	2	simple	simple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	17	3	approach	approach	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	4	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	17	5	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	17	6	obtain	obtain	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	17	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	17	8	parse	parse	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	9	trees	tree	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	17	10	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	11	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	17	12	pre	pre	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	13	-	-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	14	process	process	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	15	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	17	17	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	17	18	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	17	20	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	17	21	allows	allow	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	17	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	17	23	use	use	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	24	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	25	unrestricted	unrestricted	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	17	26	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	17	27	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	28	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	17	29	syntax	syntax	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	17	30	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	18	1	However	however	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	18	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	18	3	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	4	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	5	other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	18	6	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	7	approaches	approach	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	18	8	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	9	NLP	NLP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	18	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	18	11	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	18	12	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	18	13	been	be	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	14	shown	show	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	15	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	18	17	errors	error	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	18	18	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	19	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	18	20	syn-	syn-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	18	21	tactic	tactic	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	22	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	23	severely	severely	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	18	24	degrade	degrade	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	18	25	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	18	26	predictions	prediction	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	18	27	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	28	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	18	29	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	18	30	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	18	31	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	18	32	Gildea	Gildea	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	18	33	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	18	34	Palmer	Palmer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	18	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	18	36	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	18	37	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	18	38	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	19	1	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	19	2	common	common	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	19	3	approach	approach	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	4	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	19	5	alleviate	alleviate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	19	6	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	19	7	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	8	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	19	9	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	19	10	work	work	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	19	11	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	12	multiple	multiple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	19	13	alternative	alternative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	19	14	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	19	15	trees	tree	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	19	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	19	17	let	let	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	19	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	19	19	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	19	20	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	21	optimize	optimize	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	22	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	23	any	any	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	19	24	input	input	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	25	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	26	or	or	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	19	27	part	part	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	28	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	19	29	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	19	30	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	19	31	Toutanova	Toutanova	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	19	32	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	19	33	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	19	34	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	19	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	19	36	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	19	37	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	19	38	Punyakanok	Punyakanok	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	19	39	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	19	40	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	19	41	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	19	42	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	19	43	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	19	44	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	19	45	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	20	1	As	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	2	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	20	3	step	step	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	4	further	far	RBR
work_7jtgrdqc4zeb7lul7vifzjglr4	20	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	20	6	more	more	RBR
work_7jtgrdqc4zeb7lul7vifzjglr4	20	7	recent	recent	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	20	8	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	9	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	20	10	proposed	propose	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	11	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	20	12	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	20	13	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	20	14	predict	predict	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	15	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	20	16	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	17	aug-	aug-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	18	mented	mente	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	19	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	20	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	20	21	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	22	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	20	23	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	24	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	20	25	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	20	26	Surdeanu	Surdeanu	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	27	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	20	28	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	29	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	20	31	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	20	32	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	20	33	Hajič	hajič	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	20	34	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	35	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	36	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	20	38	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	20	39	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	20	40	Johansson	Johansson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	41	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	20	42	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	20	43	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	20	44	Titov	Titov	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	45	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	20	46	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	47	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	48	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	20	49	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	20	50	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	20	51	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	52	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	20	53	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	54	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	20	55	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	20	56	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	20	57	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	20	58	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	20	59	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	20	60	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	20	61	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	20	62	focus	focus	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	63	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	64	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	20	65	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	20	66	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	21	1	These	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	21	2	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	21	3	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	21	4	should	should	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	21	5	favor	favor	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	21	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	21	7	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	21	8	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	21	9	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	21	10	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	21	11	most	most	RBS
work_7jtgrdqc4zeb7lul7vifzjglr4	21	12	consistent	consistent	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	21	13	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	21	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	21	15	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	21	16	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	21	17	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	21	18	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	21	19	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	21	20	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	21	21	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	21	22	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	21	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	22	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	22	2	principle	principle	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	22	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	22	4	these	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	22	5	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	22	6	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	22	7	exploit	exploit	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	22	8	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	22	9	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	22	10	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	22	11	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	22	12	simultaneously	simultaneously	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	22	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	22	14	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	22	15	could	could	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	22	16	potentially	potentially	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	22	17	improve	improve	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	22	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	22	19	accuracy	accuracy	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	22	20	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	22	21	both	both	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	22	22	syn-	syn-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	22	23	tactic	tactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	22	24	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	22	25	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	22	26	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	22	27	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	23	1	One	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	23	2	difficulty	difficulty	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	23	3	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	23	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	23	5	design	design	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	23	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	23	7	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	23	8	syntactic-	syntactic-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	23	9	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	23	10	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	23	11	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	23	12	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	23	13	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	23	14	there	there	EX
work_7jtgrdqc4zeb7lul7vifzjglr4	23	15	exist	exist	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	23	16	impor-	impor-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	23	17	tant	tant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	23	18	structural	structural	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	23	19	divergences	divergence	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	23	20	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	23	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	23	22	two	two	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	23	23	layers	layer	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	23	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	24	1	219	219	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	24	2	Mary	Mary	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	24	3	loves	love	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	24	4	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	24	5	play	play	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	24	6	guitar	guitar	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	24	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	25	1	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	2	SBJ	SBJ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	3	OPRD	OPRD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	4	IM	im	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	25	5	OBJ	OBJ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	6	P	P	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	7	ARG0	ARG0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	8	ARG1	ARG1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	9	ARG0	ARG0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	10	ARG1	ARG1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	11	Figure	Figure	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	25	12	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	25	13	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	25	14	An	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	25	15	example	example	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	25	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	26	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	27	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	28	1	Das	Das	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	28	2	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	28	3	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	28	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	29	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	29	2	2012	2012	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	29	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	29	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	30	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	31	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	32	1	Riedel	Riedel	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	32	2	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	32	3	McCallum	McCallum	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	32	4	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	32	5	2011	2011	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	32	6	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	32	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	33	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	34	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	35	1	3	3	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	35	2	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	35	3	Syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	35	4	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	35	5	Semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	35	6	Dependency	Dependency	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	35	7	Model	Model	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	35	8	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	35	9	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	35	10	describe	describe	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	35	11	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	35	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	35	13	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	35	14	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	35	15	seman-	seman-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	35	16	tic	tic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	35	17	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	35	18	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	35	19	vectors	vector	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	35	20	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	35	21	binary	binary	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	35	22	variables	variable	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	35	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	36	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	2	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	36	3	denote	denote	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	36	4	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	5	yh	yh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	36	7	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	36	9	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	10	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	36	11	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	12	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	13	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	14	head	head	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	15	token	token	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	16	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	17	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	18	dependant	dependant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	19	token	token	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	20	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	21	labeled	label	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	22	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	23	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	24	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	25	l.	l.	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	26	Similarly	similarly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	36	27	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	28	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	36	29	denote	denote	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	36	30	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	31	zp	zp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	36	33	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	36	35	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	36	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	36	37	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	38	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	39	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	40	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	41	token	token	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	42	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	43	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	36	44	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	45	token	token	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	36	46	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	36	47	labeled	labeled	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	48	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	49	seman-	seman-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	50	tic	tic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	51	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	52	r.	r.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	53	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	54	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	36	55	use	use	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	36	56	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	57	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	36	58	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	36	59	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	36	60	denote	denote	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	36	61	vectors	vector	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	36	62	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	63	binary	binary	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	64	variables	variable	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	36	65	indexed	index	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	66	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	36	67	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	68	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	36	69	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	36	70	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	36	71	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	36	72	respectively	respectively	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	36	73	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	37	1	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	37	2	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	37	3	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	37	4	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	37	5	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	37	6	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	37	7	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	37	8	depen-	depen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	37	9	dency	dency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	37	10	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	37	11	could	could	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	37	12	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	37	13	defined	define	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	37	14	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	37	15	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	37	16	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	17	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	37	19	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	20	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	21	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	37	23	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	24	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	37	25	+	+	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	26	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	27	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	37	29	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	37	31	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	37	32	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	37	33	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	38	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	38	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	38	3	equation	equation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	38	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	38	5	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	38	6	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	38	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	38	8	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	38	9	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	38	10	gives	give	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	38	11	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	38	12	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	38	13	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	38	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	38	15	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	38	16	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	38	17	y.	y.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	39	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	39	3	literature	literature	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	39	5	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	6	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	39	7	standard	standard	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	39	8	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	39	9	use	use	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	39	10	arc	arc	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	11	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	39	12	factored	factored	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	39	13	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	39	14	defined	define	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	15	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	16	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	17	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	39	19	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	20	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	39	21	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	22	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	23	yh	yh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	39	25	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	39	27	l=1	l=1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	28	s	s	POS
work_7jtgrdqc4zeb7lul7vifzjglr4	39	29	syn(x	syn(x	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	39	31	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	39	33	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	39	35	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	36	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	39	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	39	38	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	39	39	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	40	overload	overload	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	39	41	s	s	POS
work_7jtgrdqc4zeb7lul7vifzjglr4	39	42	syn	syn	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	43	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	39	44	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	39	45	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	39	46	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	47	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	39	48	computes	compute	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	39	49	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	39	50	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	51	individual	individual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	39	52	labeled	label	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	39	53	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	39	54	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	39	55	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	40	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	2	discriminative	discriminative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	40	3	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	40	4	one	one	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	5	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	40	6	s	s	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	7	syn(x	syn(x	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	40	9	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	40	11	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	40	13	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	14	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	40	15	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	16	wsyn	wsyn	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	17	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	18	fsyn(x	fsyn(x	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	40	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	40	20	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	40	22	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	40	24	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	25	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	40	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	40	27	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	40	28	fsyn	fsyn	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	29	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	40	30	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	40	31	feature	feature	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	32	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	33	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	34	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	40	35	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	40	36	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	37	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	40	38	wsyn	wsyn	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	39	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	40	40	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	40	41	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	42	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	40	43	parameters	parameter	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	40	44	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	40	45	McDonald	McDonald	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	46	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	47	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	48	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	40	49	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	40	50	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	40	51	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	40	52	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	41	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	41	2	other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	41	3	term	term	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	41	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	41	5	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	41	6	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	41	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	41	8	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	41	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	41	10	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	41	11	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	41	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	41	13	gives	give	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	41	14	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	41	15	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	41	16	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	41	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	41	18	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	41	19	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	41	20	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	41	21	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	41	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	41	23	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	41	24	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	41	25	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	41	26	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	41	27	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	41	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	42	1	Previous	previous	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	42	2	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	42	3	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	42	4	empiri-	empiri-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	42	5	cally	cally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	42	6	proved	prove	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	42	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	42	8	importance	importance	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	42	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	42	10	exploiting	exploit	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	42	11	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	42	12	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	42	13	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	42	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	42	15	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	42	16	component	component	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	42	17	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	42	18	Gildea	Gildea	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	42	19	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	42	20	Ju-	Ju-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	42	21	rafsky	rafsky	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	42	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	42	23	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	42	24	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	42	25	Xue	Xue	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	42	26	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	42	27	Palmer	Palmer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	42	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	42	29	2004	2004	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	42	30	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	42	31	Punyakanok	Punyakanok	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	42	32	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	42	33	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	42	34	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	42	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	42	36	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	42	37	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	42	38	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	43	1	However	however	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	43	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	43	3	without	without	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	43	4	further	further	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	43	5	assumptions	assumption	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	43	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	43	7	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	43	8	property	property	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	43	9	makes	make	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	43	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	43	11	optimization	optimization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	43	12	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	43	13	com-	com-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	43	14	putationally	putationally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	43	15	hard	hard	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	43	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	44	1	One	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	44	2	simple	simple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	44	3	approximation	approximation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	44	4	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	44	5	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	44	6	use	use	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	44	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	44	8	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	44	9	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	44	10	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	44	11	first	first	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	44	12	compute	compute	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	44	13	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	44	14	optimal	optimal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	44	15	syn-	syn-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	44	16	tactic	tactic	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	44	17	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	44	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	44	19	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	44	20	then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	44	21	optimize	optimize	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	44	22	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	44	23	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	44	24	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	44	25	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	44	26	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	44	27	given	give	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	44	28	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	44	29	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	44	30	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	44	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	45	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	45	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	45	3	rest	rest	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	45	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	45	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	45	6	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	45	7	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	45	8	describe	describe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	45	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	45	10	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	45	11	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	45	12	searches	search	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	45	13	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	45	14	syn-	syn-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	45	15	tactic	tactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	45	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	45	17	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	45	18	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	45	19	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	45	20	jointly	jointly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	45	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	46	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	46	2	first	first	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	46	3	impose	impose	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	46	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	46	5	assumption	assumption	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	6	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	7	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	46	8	fea-	fea-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	46	9	tures	ture	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	46	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	46	12	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	46	13	component	component	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	14	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	46	15	restricted	restrict	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	16	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	46	18	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	46	19	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	20	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	21	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	46	22	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	23	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	46	24	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	46	25	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	46	27	following	follow	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	46	28	previous	previous	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	46	29	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	46	30	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	46	31	Johansson	Johansson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	46	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	46	33	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	46	34	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	46	35	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	47	1	For-	for-	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	2	mally	mally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	47	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	4	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	47	6	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	7	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	9	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	10	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	47	12	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	13	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	47	14	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	47	15	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	47	16	define	define	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	47	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	47	18	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	20	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	21	indicators	indicator	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	47	22	πp	πp	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	47	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	24	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	47	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	26	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	27	similar	similar	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	47	28	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	29	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	47	30	ones	one	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	47	31	above	above	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	47	32	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	47	33	πp	πp	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	47	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	35	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	47	36	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	37	rh	rh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	47	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	39	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	47	40	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	41	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	42	indicates	indicate	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	47	43	if	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	44	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	47	45	de-	de-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	46	pendency	pendency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	47	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	47	48	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	49	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	50	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	51	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	47	52	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	47	53	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	47	54	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	47	55	part	part	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	56	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	57	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	47	58	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	47	59	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	60	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	47	61	links	link	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	47	62	predicate	predicate	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	47	63	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	64	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	47	65	token	token	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	47	66	a.	a.	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	48	1	Figure	figure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	48	2	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	48	3	gives	give	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	48	4	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	48	5	ex-	ex-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	48	6	ample	ample	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	48	7	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	48	8	one	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	48	9	such	such	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	48	10	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	48	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	49	1	Given	give	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	49	2	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	49	3	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	49	4	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	49	5	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	49	6	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	49	7	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	49	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	49	9	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	49	10	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	49	11	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	49	12	trivial	trivial	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	49	13	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	49	14	construct	construct	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	49	15	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	49	16	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	49	17	π	π	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	49	18	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	49	19	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	49	20	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	49	21	concatenation	concatenation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	49	22	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	49	23	vectors	vector	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	49	24	πp	πp	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	49	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	49	26	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	49	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	49	28	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	49	29	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	49	30	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	49	31	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	49	32	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	49	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	49	34	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	49	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	49	36	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	49	37	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	49	38	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	49	39	z.	z.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	2	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	50	3	now	now	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	50	4	define	define	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	50	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	50	6	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	50	7	seman-	seman-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	8	tic	tic	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	9	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	10	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	11	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	12	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	14	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	16	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	50	17	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	50	18	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	50	19	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	20	zp	zp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	22	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	24	r=1	r=1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	25	s	s	POS
work_7jtgrdqc4zeb7lul7vifzjglr4	50	26	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	28	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	30	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	32	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	34	πp	πp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	36	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	50	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	38	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	39	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	50	40	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	41	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	50	42	1	1	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	50	43	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	50	44	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	50	45	s	s	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	50	46	srl	srl	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	47	computes	compute	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	50	48	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	50	49	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	50	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	51	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	50	52	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	50	53	de-	de-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	54	pendency	pendency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	55	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	56	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	57	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	58	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	59	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	60	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	61	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	62	together	together	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	50	63	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	64	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	50	65	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	50	66	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	67	πp	πp	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	50	68	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	69	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	50	70	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	50	71	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	50	72	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	51	1	As	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	2	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	51	4	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	51	5	component	component	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	51	7	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	51	8	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	9	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	51	10	typically	typically	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	51	11	defined	define	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	12	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	13	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	51	14	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	51	15	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	16	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	51	18	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	20	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	51	21	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	51	23	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	51	24	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	25	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	51	26	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	51	27	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	51	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	52	1	With	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	2	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	3	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	52	4	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	7	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	8	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	9	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	52	10	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	52	11	formulated	formulate	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	12	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	13	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	14	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	15	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	17	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	19	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	20	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	21	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	23	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	24	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	52	25	+	+	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	26	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	27	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	29	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	31	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	52	32	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	52	33	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	52	34	2	2	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	52	35	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	52	36	subject	subject	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	52	37	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	38	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	39	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	40	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	41	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	52	42	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	43	valid	valid	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	52	44	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	45	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	46	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	47	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	48	∀p	∀p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	49	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	50	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	51	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	52	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	53	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	54	zp	zp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	55	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	56	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	57	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	58	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	59	≤	≤	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	60	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	52	61	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	62	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	63	∀p	∀p	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	52	64	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	65	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	66	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	67	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	68	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	69	zp	zp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	70	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	71	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	72	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	73	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	74	≤	≤	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	75	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	52	76	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	77	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	78	∀p	∀p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	79	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	80	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	81	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	82	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	83	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	84	if	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	85	zp	zp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	86	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	87	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	88	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	89	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	90	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	52	91	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	52	92	then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	52	93	πp	πp	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	52	94	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	95	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	96	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	97	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	98	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	52	99	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	100	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	101	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	102	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	103	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	104	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	105	otherwise	otherwise	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	52	106	πp	πp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	107	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	108	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	109	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	110	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	111	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	52	112	0	0	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	113	cSubtree	cSubtree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	114	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	115	∀p	∀p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	116	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	117	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	118	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	119	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	120	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	121	πp	πp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	122	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	123	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	124	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	52	125	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	126	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	52	127	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	128	subtree	subtree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	129	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	130	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	131	Figure	Figure	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	52	132	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	52	133	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	52	134	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	135	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	136	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	137	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	52	138	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	52	139	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	52	140	top	top	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	141	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	52	142	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	52	143	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	52	144	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	52	145	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	146	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	52	147	predicates	predicate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	52	148	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	52	149	loves	love	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	52	150	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	52	151	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	52	152	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	52	153	play	play	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	52	154	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	52	155	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	52	156	bottom	bottom	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	52	157	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	52	158	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	53	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	53	2	thick	thick	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	53	3	arcs	arc	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	53	4	illustrate	illustrate	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	53	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	53	6	structural	structural	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	53	7	di-	di-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	53	8	vergence	vergence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	53	9	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	53	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	53	11	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	53	12	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	53	13	Mary	Mary	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	53	14	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	53	15	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	53	16	linked	link	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	53	17	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	53	18	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	53	19	play	play	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	53	20	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	53	21	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	53	22	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	53	23	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	53	24	involving	involve	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	53	25	three	three	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	53	26	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	53	27	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	53	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	54	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	54	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	54	3	clearly	clearly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	54	4	seen	see	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	5	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	6	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	7	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	54	8	based	base	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	9	representa-	representa-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	54	10	tions	tion	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	54	11	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	12	syntax	syntax	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	13	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	54	14	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	54	15	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	54	16	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	54	17	Surdeanu	Surdeanu	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	54	18	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	54	19	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	54	20	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	54	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	54	22	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	54	23	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	54	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	54	25	such	such	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	54	26	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	27	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	28	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	54	29	example	example	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	30	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	31	Figure	Figure	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	54	32	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	54	33	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	54	34	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	54	35	con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	36	struct	struct	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	37	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	54	38	loves	love	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	54	39	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	54	40	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	54	41	causes	cause	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	54	42	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	54	43	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	44	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	54	45	Mary	Mary	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	54	46	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	54	47	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	54	48	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	54	49	syntactically	syntactically	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	54	50	distant	distant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	54	51	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	52	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	54	53	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	54	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	54	55	play	play	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	54	56	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	54	57	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	55	1	Lin-	Lin-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	55	2	guistic	guistic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	55	3	phenomena	phenomenon	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	55	4	such	such	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	55	5	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	6	auxiliary	auxiliary	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	55	7	verbs	verb	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	55	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	55	9	control	control	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	55	11	raising	raising	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	55	13	typically	typically	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	55	14	result	result	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	55	15	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	16	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	55	17	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	55	18	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	55	19	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	55	20	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	55	21	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	55	22	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	55	23	among	among	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	24	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	55	25	direct	direct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	55	26	dependants	dependant	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	55	27	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	28	their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	55	29	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	30	—	—	:
work_7jtgrdqc4zeb7lul7vifzjglr4	55	31	e.g.	e.g.	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	55	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	55	33	about	about	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	55	34	25	25	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	55	35	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	36	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	37	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	55	38	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	55	39	distant	distant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	55	40	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	41	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	55	42	English	english	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	55	43	development	development	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	44	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	45	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	46	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	55	47	CoNLL-2009	conll-2009	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	48	shared	share	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	55	49	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	55	50	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	56	1	Besides	besides	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	56	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	56	3	standard	standard	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	56	4	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	56	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	56	6	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	56	7	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	56	8	crucially	crucially	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	56	9	depend	depend	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	56	10	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	56	11	arc	arc	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	56	12	factorizations	factorization	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	56	13	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	56	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	56	15	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	56	16	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	56	17	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	56	18	Mc-	Mc-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	56	19	Donald	Donald	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	56	20	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	56	21	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	56	22	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	56	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	56	24	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	56	25	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	56	26	Nivre	Nivre	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	56	27	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	56	28	Nilsson	Nilsson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	56	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	56	30	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	56	31	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	56	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	56	33	other-	other-	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	56	34	wise	wise	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	56	35	their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	56	36	computational	computational	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	56	37	properties	property	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	56	38	break	break	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	56	39	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	57	1	Hence	hence	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	57	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	57	3	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	57	4	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	57	5	challenging	challenge	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	57	6	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	57	7	define	define	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	57	8	efficient	efficient	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	57	9	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	57	10	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	57	11	syntac-	syntac-	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	57	12	tic	tic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	57	13	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	57	14	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	57	15	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	57	16	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	57	17	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	57	18	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	57	19	exploit	exploit	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	57	20	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	57	21	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	57	22	both	both	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	57	23	layers	layer	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	57	24	simultaneously	simultaneously	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	57	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	58	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	58	2	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	58	3	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	58	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	58	5	propose	propose	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	58	6	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	58	7	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	58	8	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	58	9	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	58	10	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	58	11	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	58	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	59	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	2	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	59	3	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	59	5	define	define	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	59	6	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	7	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	59	8	centric	centric	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	9	seman-	seman-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	10	tic	tic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	59	11	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	59	12	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	59	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	59	14	rather	rather	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	59	15	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	16	predicting	predict	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	59	17	just	just	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	59	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	59	19	ar-	ar-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	20	gument	gument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	21	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	59	22	realizes	realize	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	59	23	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	59	24	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	59	25	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	59	27	they	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	59	28	pre-	pre-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	59	29	dict	dict	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	59	30	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	59	31	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	59	32	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	59	33	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	34	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	59	35	connects	connect	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	59	36	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	59	37	predi-	predi-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	38	cate	cate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	39	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	40	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	59	41	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	59	42	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	60	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	60	2	show	show	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	60	3	how	how	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	60	4	efficient	efficient	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	60	5	pre-	pre-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	60	6	dictions	diction	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	60	7	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	60	8	these	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	60	9	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	60	10	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	60	11	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	60	12	made	make	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	60	13	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	60	14	as-	as-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	60	15	signment	signment	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	60	16	algorithms	algorithm	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	60	17	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	60	18	bipartite	bipartite	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	60	19	graphs	graph	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	60	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	61	1	Simulta-	Simulta-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	61	2	neously	neously	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	61	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	61	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	61	5	use	use	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	61	6	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	61	7	standard	standard	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	61	8	arc	arc	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	61	9	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	61	10	factored	factored	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	61	11	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	61	12	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	61	13	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	61	14	predicts	predict	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	61	15	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	61	16	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	61	17	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	61	18	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	61	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	61	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	61	21	sen-	sen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	61	22	tence	tence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	61	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	62	1	Finally	finally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	62	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	62	3	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	4	employ	employ	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	5	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	62	6	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	62	7	tech-	tech-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	62	8	niques	nique	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	62	9	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	62	10	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	11	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	12	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	13	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	62	15	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	62	16	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	62	17	Rush	Rush	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	18	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	62	19	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	20	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	62	22	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	62	23	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	62	24	Sontag	Sontag	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	25	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	26	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	27	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	62	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	62	29	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	62	30	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	62	31	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	62	32	find	find	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	62	33	agreement	agreement	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	62	34	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	62	35	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	62	36	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	62	37	de-	de-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	62	38	pendency	pendency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	62	39	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	62	40	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	62	41	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	62	42	partial	partial	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	62	43	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	62	44	trees	tree	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	62	45	linking	link	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	62	46	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	62	47	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	62	48	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	62	49	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	62	50	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	62	51	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	63	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	63	2	summary	summary	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	63	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	63	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	63	5	main	main	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	63	6	contributions	contribution	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	63	7	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	63	8	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	63	9	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	63	10	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	63	11	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	63	12	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	63	13	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	63	14	frame	frame	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	63	15	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	63	16	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	63	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	63	18	weighted	weighted	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	63	19	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	63	20	prob-	prob-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	63	21	lem	lem	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	63	22	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	63	23	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	63	24	bipartite	bipartite	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	63	25	graph	graph	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	63	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	64	1	Under	under	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	64	2	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	64	3	framework	framework	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	64	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	64	5	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	64	6	control	control	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	64	7	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	64	8	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	64	9	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	64	10	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	64	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	64	12	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	64	13	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	65	1	Key	key	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	65	2	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	65	3	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	65	4	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	65	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	65	6	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	65	7	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	65	8	efficiently	efficiently	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	65	9	search	search	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	65	10	over	over	RP
work_7jtgrdqc4zeb7lul7vifzjglr4	65	11	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	65	12	large	large	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	65	13	space	space	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	65	14	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	65	15	syn-	syn-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	65	16	tactic	tactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	65	17	realizations	realization	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	65	18	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	65	19	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	65	20	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	65	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	66	1	•	•	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	66	2	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	66	3	solve	solve	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	66	4	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	66	5	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	66	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	66	7	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	66	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	66	9	se-	se-	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	66	10	mantic	mantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	66	11	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	66	12	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	66	13	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	66	14	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	66	15	decomposi-	decomposi-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	66	16	tion	tion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	66	17	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	66	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	66	19	similar	similar	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	66	20	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	66	21	that	that	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	66	22	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	66	23	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	66	24	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	66	25	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	66	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	67	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	67	2	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	67	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	67	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	68	1	Our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	68	2	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	68	3	produces	produce	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	68	4	consistent	consistent	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	68	5	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	68	6	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	68	7	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	68	8	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	68	9	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	68	10	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	68	11	while	while	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	68	12	searching	search	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	68	13	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	68	14	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	68	15	large	large	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	68	16	space	space	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	68	17	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	68	18	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	68	19	configurations	configuration	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	68	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	69	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	69	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	69	3	experimental	experimental	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	69	4	section	section	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	69	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	69	6	compare	compare	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	69	7	joint	joint	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	69	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	69	9	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	69	10	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	69	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	70	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	70	2	final	final	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	70	3	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	70	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	5	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	70	6	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	70	7	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	70	8	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	70	9	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	70	10	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	11	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	70	12	competitive	competitive	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	70	13	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	70	15	state	state	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	16	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	70	17	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	18	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	70	19	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	70	20	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	70	21	art	art	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	22	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	70	23	improve	improve	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	70	24	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	25	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	70	26	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	70	27	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	70	28	published	publish	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	29	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	30	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	70	31	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	70	32	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	33	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	34	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	70	35	CoNLL-2009	conll-2009	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	70	36	English	English	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	70	37	dataset	dataset	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	70	38	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	71	1	2	2	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	71	2	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	71	3	Syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	71	4	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	71	5	Semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	71	6	Dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	71	7	Model	Model	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	71	8	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	71	9	first	first	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	71	10	describe	describe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	71	11	how	how	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	71	12	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	71	13	represent	represent	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	71	14	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	71	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	71	16	syn-	syn-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	71	17	tactic	tactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	71	18	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	71	19	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	71	20	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	71	21	like	like	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	71	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	71	23	one	one	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	71	24	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	71	25	Fig-	Fig-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	71	26	ure	ure	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	71	27	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	71	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	72	1	Throughout	throughout	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	72	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	72	3	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	72	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	72	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	72	6	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	72	7	assume	assume	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	72	8	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	72	9	fixed	fix	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	72	10	input	input	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	72	11	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	72	12	x	x	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	72	13	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	72	14	n	n	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	72	15	tokens	token	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	72	16	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	72	17	lexical	lexical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	72	18	predi-	predi-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	72	19	cates	cate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	72	20	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	72	21	marked	mark	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	72	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	73	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	73	2	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	73	3	also	also	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	73	4	assume	assume	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	73	5	fixed	fix	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	73	6	sets	set	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	73	7	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	73	8	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	73	9	functions	function	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	73	10	Rsyn	rsyn	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	73	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	73	12	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	73	13	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	73	14	Rsem	Rsem	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	73	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	74	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	74	2	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	74	3	represent	represent	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	74	4	depencency	depencency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	74	5	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	74	6	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	74	7	vec-	vec-	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	74	8	tors	tor	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	74	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	74	10	binary	binary	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	74	11	variables	variable	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	74	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	75	1	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	75	2	variable	variable	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	75	3	yh	yh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	75	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	75	5	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	75	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	75	7	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	75	8	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	75	9	in-	in-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	75	10	dicate	dicate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	75	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	75	12	presence	presence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	13	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	14	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	75	15	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	75	16	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	17	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	18	head	head	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	19	token	token	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	75	20	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	21	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	22	dependant	dependant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	75	23	token	token	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	24	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	25	labeled	label	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	26	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	27	syn-	syn-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	75	28	tactic	tactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	75	29	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	30	l.	l.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	75	31	Then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	75	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	75	33	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	75	34	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	75	35	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	36	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	75	37	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	75	38	de-	de-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	75	39	noted	note	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	40	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	41	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	75	42	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	43	y	y	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	44	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	45	variables	variable	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	75	46	indexed	index	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	47	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	75	48	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	75	49	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	75	50	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	76	1	Similarly	similarly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	76	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	76	3	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	76	4	variable	variable	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	76	5	zp	zp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	76	7	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	76	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	76	9	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	76	10	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	76	11	indi-	indi-	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	76	12	cate	cate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	76	13	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	76	14	presence	presence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	16	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	76	17	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	76	18	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	19	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	20	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	21	token	token	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	76	22	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	23	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	76	24	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	25	token	token	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	76	26	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	76	27	labeled	label	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	28	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	29	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	76	30	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	31	r.	r.	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	32	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	76	33	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	76	34	represent	represent	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	76	35	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	76	36	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	76	37	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	38	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	39	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	40	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	76	41	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	42	z	z	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	43	indexed	index	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	44	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	76	45	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	76	46	depen-	depen-	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	76	47	dencies	dencie	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	76	48	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	77	1	Whenever	whenever	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	77	2	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	3	enumerate	enumerate	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	4	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	77	5	depen-	depen-	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	77	6	dencies	dencie	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	77	7	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	8	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	77	10	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	77	12	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	13	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	14	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	15	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	77	16	assume	assume	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	77	17	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	77	18	they	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	19	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	77	20	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	77	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	77	22	valid	valid	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	77	23	range	range	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	77	24	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	77	25	x	x	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	77	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	77	27	i.e.	i.e.	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	78	1	0	0	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	78	2	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	3	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	4	≤	≤	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	78	5	n	n	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	78	7	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	78	8	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	9	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	10	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	11	n	n	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	78	13	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	14	6=	6=	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	78	15	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	78	17	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	18	∈	∈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	19	Rsyn	Rsyn	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	78	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	78	21	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	78	22	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	78	23	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	78	24	0	0	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	78	25	stands	stand	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	78	26	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	78	27	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	78	28	special	special	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	78	29	root	root	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	78	30	token	token	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	78	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	79	1	Similarly	similarly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	79	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	79	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	79	4	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	79	5	depen-	depen-	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	79	6	dencies	dencie	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	79	7	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	79	8	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	79	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	79	10	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	79	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	79	12	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	79	13	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	79	14	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	79	15	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	79	16	assume	assume	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	79	17	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	79	18	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	79	19	points	point	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	79	20	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	79	21	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	79	22	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	79	23	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	79	24	x	x	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	79	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	79	26	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	79	27	≤	≤	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	79	28	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	79	29	≤	≤	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	79	30	n	n	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	79	31	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	79	32	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	79	33	∈Rsem	∈Rsem	:
work_7jtgrdqc4zeb7lul7vifzjglr4	79	34	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	80	1	220	220	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	80	2	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	80	3	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	80	4	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	80	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	80	6	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	80	7	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	80	8	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	80	9	depen-	depen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	80	10	dency	dency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	80	11	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	80	12	could	could	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	80	13	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	80	14	defined	define	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	80	15	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	80	16	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	80	17	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	18	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	80	20	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	21	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	22	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	80	24	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	25	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	80	26	+	+	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	27	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	28	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	80	30	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	80	32	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	80	33	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	80	34	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	81	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	81	2	1	1	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	81	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	81	4	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	81	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	81	6	equation	equation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	81	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	81	8	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	81	9	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	81	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	81	11	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	81	12	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	81	13	gives	give	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	81	14	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	81	15	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	81	16	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	81	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	81	18	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	81	19	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	81	20	y.	y.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	82	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	82	3	literature	literature	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	82	5	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	6	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	82	7	standard	standard	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	82	8	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	82	9	use	use	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	82	10	arc	arc	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	11	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	82	12	factored	factored	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	82	13	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	82	14	defined	define	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	15	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	16	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	17	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	82	19	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	20	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	82	21	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	22	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	82	23	yh	yh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	82	25	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	82	27	l=1	l=1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	28	s	s	POS
work_7jtgrdqc4zeb7lul7vifzjglr4	82	29	syn(x	syn(x	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	82	31	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	82	33	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	82	35	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	36	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	82	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	82	38	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	82	39	2	2	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	82	40	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	82	41	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	82	42	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	43	overload	overload	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	82	44	s	s	POS
work_7jtgrdqc4zeb7lul7vifzjglr4	82	45	syn	syn	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	46	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	82	47	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	82	48	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	82	49	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	50	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	82	51	computes	compute	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	82	52	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	82	53	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	82	54	individual	individual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	82	55	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	82	56	depen-	depen-	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	82	57	dencies	dencie	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	82	58	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	83	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	2	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	83	3	discriminative	discriminative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	83	4	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	83	5	one	one	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	6	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	83	7	s	s	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	8	syn(x	syn(x	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	83	10	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	83	12	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	83	14	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	15	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	83	16	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	17	wsyn	wsyn	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	18	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	19	fsyn(x	fsyn(x	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	83	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	83	21	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	83	23	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	83	25	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	26	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	83	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	83	28	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	83	29	fsyn	fsyn	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	30	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	83	31	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	83	32	feature	feature	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	33	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	34	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	35	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	83	36	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	83	37	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	38	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	83	39	wsyn	wsyn	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	40	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	83	41	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	83	42	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	43	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	83	44	parameters	parameter	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	83	45	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	83	46	McDonald	McDonald	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	47	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	48	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	49	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	83	50	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	83	51	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	83	52	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	83	53	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	84	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	84	2	Section	section	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	84	3	6	6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	84	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	84	5	describe	describe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	84	6	how	how	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	84	7	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	84	8	trained	train	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	84	9	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	84	10	functions	function	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	84	11	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	84	12	discriminative	discriminative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	84	13	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	84	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	85	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	85	2	other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	85	3	term	term	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	85	4	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	85	5	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	85	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	86	1	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	86	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	86	3	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	86	4	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	86	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	86	6	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	86	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	86	8	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	86	9	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	86	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	86	11	gives	give	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	86	12	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	86	13	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	86	14	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	86	15	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	86	16	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	86	17	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	86	18	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	86	19	z	z	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	86	20	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	86	21	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	86	22	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	86	23	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	86	24	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	86	25	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	86	26	y.	y.	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	87	1	Previous	previous	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	87	2	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	87	3	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	87	4	empirically	empirically	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	87	5	proved	prove	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	87	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	87	7	importance	importance	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	87	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	87	9	exploit-	exploit-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	10	ing	ing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	11	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	87	12	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	87	13	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	87	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	87	15	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	87	16	component	component	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	87	17	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	87	18	Gildea	Gildea	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	19	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	87	20	Jurafsky	Jurafsky	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	87	22	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	87	23	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	87	24	Xue	Xue	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	25	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	87	26	Palmer	Palmer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	87	28	2004	2004	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	87	29	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	87	30	Punyakanok	Punyakanok	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	31	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	32	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	33	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	87	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	87	35	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	87	36	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	87	37	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	88	1	However	however	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	88	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	88	3	without	without	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	88	4	further	further	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	88	5	assumptions	assumption	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	88	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	88	7	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	88	8	property	property	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	88	9	makes	make	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	88	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	88	11	optimization	optimization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	88	12	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	88	13	computationally	computationally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	88	14	hard	hard	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	88	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	89	1	One	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	89	2	simple	simple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	89	3	approx-	approx-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	89	4	imation	imation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	89	5	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	89	6	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	89	7	use	use	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	89	8	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	89	9	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	89	10	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	89	11	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	89	12	first	first	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	89	13	compute	compute	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	89	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	89	15	optimal	optimal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	89	16	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	89	17	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	89	18	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	89	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	89	20	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	89	21	then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	89	22	optimize	optimize	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	89	23	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	89	24	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	89	25	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	89	26	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	89	27	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	89	28	z	z	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	89	29	given	give	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	89	30	y.	y.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	90	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	90	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	90	3	rest	rest	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	90	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	90	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	90	6	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	90	7	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	90	8	describe	describe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	90	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	90	10	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	90	11	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	90	12	searches	search	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	90	13	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	90	14	syn-	syn-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	90	15	tactic	tactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	90	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	90	17	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	90	18	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	90	19	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	90	20	jointly	jointly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	90	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	91	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	91	2	first	first	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	91	3	note	note	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	91	4	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	91	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	91	6	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	91	7	fixed	fix	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	91	8	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	91	9	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	91	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	91	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	91	12	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	91	13	component	component	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	91	14	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	91	15	typically	typically	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	91	16	restrict	restrict	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	91	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	91	18	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	91	19	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	91	20	representing	represent	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	91	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	91	22	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	91	23	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	91	24	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	91	25	specific	specific	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	91	26	subtree	subtree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	91	27	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	91	28	y.	y.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	92	1	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	2	example	example	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	92	4	previous	previous	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	92	5	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	6	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	92	7	restricted	restrict	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	8	such	such	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	92	9	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	92	10	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	92	12	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	92	13	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	14	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	92	15	links	link	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	92	16	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	92	17	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	18	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	19	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	92	20	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	21	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	92	22	Moschitti	Moschitti	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	92	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	92	24	2004	2004	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	92	25	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	92	26	Johansson	Johansson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	92	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	92	28	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	92	29	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	92	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	92	31	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	92	32	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	33	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	92	34	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	35	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	92	36	employ	employ	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	92	37	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	92	38	restriction	restriction	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	92	39	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	93	1	Figure	figure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	2	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	93	3	gives	give	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	93	4	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	93	5	example	example	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	93	8	sub-	sub-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	93	9	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	93	11	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	93	12	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	93	13	highlight	highlight	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	93	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	93	15	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	93	16	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	17	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	93	18	con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	19	nects	nect	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	93	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	93	21	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	93	22	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	23	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	24	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	93	25	play	play	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	26	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	93	27	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	93	28	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	93	29	Mary	Mary	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	93	30	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	93	31	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	32	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	93	33	ARG0	ARG0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	93	34	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	94	1	Formally	formally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	94	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	4	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	5	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	6	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	8	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	9	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	94	11	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	12	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	13	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	14	define	define	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	15	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	16	local	local	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	17	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	18	subtree	subtree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	19	πp	πp	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	94	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	21	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	23	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	24	repre-	repre-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	25	sented	sente	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	94	26	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	27	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	28	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	29	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	94	30	πp	πp	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	94	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	32	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	34	rh	rh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	36	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	38	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	39	indicates	indicate	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	40	if	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	41	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	42	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	43	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	44	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	45	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	46	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	47	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	48	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	49	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	50	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	51	part	part	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	52	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	53	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	54	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	55	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	56	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	57	links	link	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	94	58	pred-	pred-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	59	icate	icate	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	60	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	61	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	62	token	token	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	63	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	64	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	94	65	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	66	r.1	r.1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	67	Given	give	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	68	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	69	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	70	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	94	71	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	72	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	94	73	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	74	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	94	75	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	76	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	77	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	78	trivial	trivial	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	79	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	80	con-	con-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	81	struct	struct	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	94	82	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	83	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	84	π	π	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	85	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	86	concatenates	concatenate	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	94	87	vectors	vector	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	94	88	πp	πp	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	89	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	90	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	91	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	92	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	93	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	94	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	94	95	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	96	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	97	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	98	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	99	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	94	100	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	101	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	94	102	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	94	103	z.	z.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	95	2	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	95	3	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	4	becomes	become	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	95	5	s	s	POS
work_7jtgrdqc4zeb7lul7vifzjglr4	95	6	srl(x	srl(x	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	8	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	10	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	95	11	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	95	12	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	13	∑	∑	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	95	14	zp	zp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	16	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	18	r=1	r=1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	19	s	s	POS
work_7jtgrdqc4zeb7lul7vifzjglr4	95	20	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	22	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	24	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	26	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	28	πp	πp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	30	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	95	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	32	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	33	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	95	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	35	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	95	36	3	3	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	95	37	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	95	38	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	95	39	s	s	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	95	40	srl	srl	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	41	computes	compute	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	95	42	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	95	43	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	44	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	45	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	95	46	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	95	47	de-	de-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	48	pendency	pendency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	49	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	50	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	51	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	52	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	53	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	54	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	55	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	56	together	together	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	95	57	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	58	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	95	59	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	95	60	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	61	πp	πp	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	95	62	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	63	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	95	64	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	95	65	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	95	66	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	96	1	As	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	2	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	96	4	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	96	5	component	component	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	96	7	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	96	8	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	9	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	96	10	typically	typically	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	96	11	defined	define	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	12	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	13	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	96	14	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	96	15	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	16	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	96	18	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	20	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	96	21	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	96	23	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	96	24	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	25	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	96	26	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	96	27	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	96	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	97	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	2	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	3	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	5	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	97	6	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	7	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	8	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	9	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	10	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	11	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	13	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	15	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	16	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	17	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	19	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	20	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	97	21	+	+	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	22	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	23	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	25	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	27	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	97	28	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	97	29	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	97	30	4	4	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	97	31	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	97	32	subject	subject	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	33	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	34	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	35	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	36	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	37	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	38	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	39	valid	valid	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	40	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	41	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	42	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	43	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	44	∀p	∀p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	45	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	46	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	47	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	48	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	97	49	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	50	zp	zp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	51	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	52	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	53	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	54	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	55	≤	≤	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	56	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	97	57	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	58	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	59	∀p	∀p	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	97	60	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	61	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	62	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	63	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	97	64	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	65	zp	zp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	66	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	67	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	68	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	69	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	70	≤	≤	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	71	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	97	72	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	73	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	74	∀p	∀p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	75	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	76	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	77	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	78	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	79	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	80	if	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	81	zp	zp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	82	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	83	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	84	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	85	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	86	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	97	87	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	97	88	then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	97	89	πp	πp	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	97	90	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	91	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	92	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	93	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	94	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	95	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	96	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	97	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	98	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	99	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	100	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	101	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	102	otherwise	otherwise	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	97	103	πp	πp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	104	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	105	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	106	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	107	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	108	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	97	109	0	0	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	110	cSubtree	cSubtree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	111	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	112	∀p	∀p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	113	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	114	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	115	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	116	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	117	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	118	πp	πp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	119	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	120	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	121	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	122	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	123	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	124	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	125	subtree	subtree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	126	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	127	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	128	Constraint	Constraint	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	129	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	130	dictates	dictate	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	131	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	132	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	133	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	134	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	135	valid	valid	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	136	depen-	depen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	137	dency	dency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	138	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	139	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	97	140	see	see	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	97	141	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	97	142	Martins	Martins	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	143	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	144	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	145	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	97	146	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	97	147	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	97	148	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	97	149	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	150	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	97	151	detailed	detailed	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	97	152	specification	specification	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	97	153	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	98	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	98	2	next	next	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	98	3	two	two	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	98	4	sets	set	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	98	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	98	6	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	98	7	con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	98	8	cern	cern	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	98	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	98	10	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	98	11	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	98	12	only	only	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	98	13	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	99	1	cRole	crole	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	2	imposes	impose	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	99	3	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	4	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	99	5	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	99	6	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	7	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	99	8	realized	realize	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	9	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	10	most	most	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	99	11	once.2	once.2	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	99	12	Con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	13	versely	versely	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	99	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	99	15	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	99	16	dictates	dictate	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	99	17	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	18	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	99	19	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	20	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	99	21	realize	realize	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	99	22	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	23	most	most	RBS
work_7jtgrdqc4zeb7lul7vifzjglr4	99	24	one	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	99	25	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	99	26	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	27	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	28	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	99	29	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	99	30	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	100	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	100	2	final	final	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	100	3	two	two	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	100	4	sets	set	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	100	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	100	6	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	100	7	model	model	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	100	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	100	9	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	100	10	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	100	11	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	100	12	interdependencies	interdependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	100	13	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	101	1	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	101	2	imposes	impose	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	101	3	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	101	4	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	101	5	πp	πp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	101	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	101	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	101	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	101	9	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	101	10	represents	represent	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	101	11	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	101	12	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	101	13	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	101	14	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	101	15	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	101	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	101	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	101	18	when-	when-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	101	19	ever	ever	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	101	20	there	there	EX
work_7jtgrdqc4zeb7lul7vifzjglr4	101	21	exists	exist	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	101	22	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	101	23	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	101	24	relation	relation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	101	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	102	1	Finally	finally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	102	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	102	3	cSub-	csub-	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	102	4	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	102	5	imposes	impose	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	102	6	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	102	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	102	8	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	102	9	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	102	10	π	π	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	102	11	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	102	12	consistent	consistent	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	102	13	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	102	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	102	15	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	102	16	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	102	17	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	102	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	102	19	i.e.	i.e.	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	103	1	they	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	103	2	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	103	3	subtrees	subtree	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	103	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	104	1	1In	1in	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	104	2	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	104	3	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	104	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	104	5	say	say	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	104	6	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	104	7	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	104	8	πp	πp	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	104	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	104	10	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	104	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	104	12	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	104	13	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	104	14	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	104	15	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	104	16	predicates	predicate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	104	17	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	104	18	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	104	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	104	20	but	but	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	104	21	they	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	104	22	could	could	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	104	23	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	104	24	more	more	RBR
work_7jtgrdqc4zeb7lul7vifzjglr4	104	25	general	general	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	104	26	sub-	sub-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	104	27	trees	tree	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	104	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	105	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	105	2	condition	condition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	105	3	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	105	4	build	build	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	105	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	105	6	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	105	7	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	105	8	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	105	9	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	105	10	these	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	105	11	subtrees	subtree	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	105	12	must	must	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	105	13	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	105	14	parseable	parseable	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	105	15	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	105	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	105	17	way	way	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	105	18	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	105	19	describe	describe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	105	20	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	105	21	Section	section	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	105	22	3.1	3.1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	105	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	106	1	2In	2in	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	106	2	general	general	PDT
work_7jtgrdqc4zeb7lul7vifzjglr4	106	3	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	106	4	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	106	5	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	106	6	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	106	7	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	106	8	realized	realize	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	106	9	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	106	10	more	more	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	106	11	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	106	12	one	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	106	13	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	106	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	106	15	though	though	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	106	16	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	106	17	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	106	18	rare	rare	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	106	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	107	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	107	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	107	3	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	107	4	hard	hard	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	107	5	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	107	6	modify	modify	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	107	7	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	107	8	framework	framework	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	107	9	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	107	10	allow	allow	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	107	11	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	107	12	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	107	13	maximum	maximum	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	107	14	number	number	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	107	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	107	16	occurrences	occurrence	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	107	17	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	107	18	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	107	19	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	107	20	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	107	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	108	1	221	221	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	108	2	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	108	3	Section	section	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	108	4	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	108	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	108	6	define	define	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	108	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	108	8	process	process	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	108	9	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	108	10	optimizes	optimize	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	108	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	108	12	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	108	13	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	108	14	ignoring	ignoring	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	108	15	constraint	constraint	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	108	16	cSubtree	cSubtree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	108	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	109	1	Then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	109	2	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	109	3	Section	section	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	109	4	4	4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	109	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	109	6	describe	describe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	109	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	109	8	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	109	9	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	109	10	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	109	11	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	109	12	uses	use	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	109	13	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	109	14	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	109	15	process	process	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	109	16	repeatedly	repeatedly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	109	17	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	109	18	solve	solve	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	109	19	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	109	20	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	109	21	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	109	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	110	1	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	110	2	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	110	3	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	110	4	Assignment	Assignment	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	110	5	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	110	6	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	110	7	section	section	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	110	8	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	110	9	frame	frame	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	110	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	110	11	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	110	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	110	13	finding	find	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	110	14	se-	se-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	110	15	mantic	mantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	110	16	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	110	17	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	110	18	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	110	19	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	110	20	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	110	21	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	110	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	111	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	111	2	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	111	3	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	4	optimize	optimize	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	5	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	111	6	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	111	7	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	8	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	111	10	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	11	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	12	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	111	14	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	111	16	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	111	17	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	111	18	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	111	19	5	5	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	111	20	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	111	21	subject	subject	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	111	22	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	111	23	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	111	25	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	111	27	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	28	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	111	29	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	111	30	case	case	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	111	31	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	32	dropped	drop	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	111	33	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	111	34	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	111	35	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	111	36	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	111	37	y	y	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	111	38	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	111	39	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	111	40	optimization	optimization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	111	41	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	111	42	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	111	43	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	112	1	4	4	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	112	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	112	3	as	as	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	112	4	well	well	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	112	5	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	112	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	112	7	corresponding	correspond	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	112	8	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	112	9	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	112	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	112	11	cSubtree	cSubtree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	112	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	113	1	As	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	2	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	113	3	consequence	consequence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	113	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	113	6	note	note	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	113	7	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	113	9	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	113	10	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	113	11	π	π	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	113	12	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	113	13	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	113	14	tied	tie	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	15	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	16	any	any	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	113	17	consistency	consistency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	18	constraint	constraint	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	19	other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	113	20	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	21	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	113	22	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	23	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	113	24	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	113	25	being	be	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	113	26	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	113	27	well	well	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	113	28	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	113	29	formed	form	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	30	sequence	sequence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	31	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	32	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	113	33	linking	link	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	113	34	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	113	35	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	36	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	37	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	113	38	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	113	39	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	114	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	2	other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	114	3	words	word	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	114	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	114	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	114	6	optimal	optimal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	114	7	solution	solution	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	8	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	9	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	114	10	case	case	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	11	does	do	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	114	12	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	114	13	guarantee	guarantee	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	114	14	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	15	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	114	16	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	17	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	18	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	114	19	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	20	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	114	21	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	22	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	23	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	114	24	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	25	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	114	26	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	114	27	satisfies	satisfie	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	114	28	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	114	29	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	114	30	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	115	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	115	2	first	first	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	115	3	describe	describe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	115	4	how	how	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	115	5	these	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	115	6	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	115	7	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	115	8	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	115	9	optimized	optimize	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	115	10	locally	locally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	115	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	116	1	Then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	116	2	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	116	3	show	show	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	116	4	how	how	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	116	5	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	116	6	find	find	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	116	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	116	8	solution	solution	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	116	9	z	z	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	116	10	satisfying	satisfy	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	116	11	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	116	12	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	116	13	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	116	14	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	116	15	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	116	16	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	116	17	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	116	18	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	117	1	3.1	3.1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	117	2	Local	local	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	117	3	Optimization	Optimization	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	117	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	117	5	Syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	117	6	Paths	Paths	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	117	7	Let	let	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	117	8	ẑ	ẑ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	117	9	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	117	10	π̂	π̂	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	117	11	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	117	12	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	117	13	optimal	optimal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	117	14	values	value	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	117	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	117	16	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	117	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	118	1	5	5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	118	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	119	1	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	119	2	any	any	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	119	3	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	4	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	6	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	8	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	9	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	11	let	let	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	119	12	π̃p	π̃p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	119	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	14	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	119	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	16	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	17	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	119	18	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	19	πp	πp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	21	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	119	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	23	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	24	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	25	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	27	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	29	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	31	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	119	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	33	πp	πp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	119	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	35	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	119	36	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	119	37	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	119	38	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	119	39	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	120	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	120	2	6	6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	120	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	120	4	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	120	5	any	any	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	120	6	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	120	7	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	120	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	120	9	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	120	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	120	11	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	120	12	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	120	13	such	such	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	120	14	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	120	15	ẑp	ẑp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	120	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	120	17	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	120	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	120	19	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	120	20	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	120	21	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	120	22	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	120	23	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	120	24	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	120	25	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	120	26	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	120	27	π̂p	π̂p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	120	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	120	29	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	120	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	120	31	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	120	32	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	120	33	π̃p	π̃p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	120	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	120	35	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	120	36	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	120	37	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	120	38	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	121	1	If	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	121	2	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	121	3	was	be	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	121	4	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	121	5	true	true	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	121	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	121	7	replacing	replace	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	121	8	π̂p	π̂p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	121	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	121	10	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	121	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	121	12	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	121	13	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	121	14	π̃p	π̃p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	121	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	121	16	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	121	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	121	18	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	121	19	would	would	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	121	20	improve	improve	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	121	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	121	22	objective	objective	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	121	23	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	121	24	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	121	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	122	1	5	5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	122	2	without	without	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	122	3	violating	violate	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	122	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	122	5	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	122	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	122	7	thus	thus	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	122	8	contradicting	contradict	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	122	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	122	10	hypothesis	hypothesis	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	122	11	about	about	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	122	12	optimality	optimality	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	122	13	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	122	14	π̂.	π̂.	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	123	1	Therefore	therefore	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	123	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	123	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	123	4	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	123	5	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	123	6	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	123	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	123	8	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	123	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	123	10	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	123	11	〉	〉	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	123	12	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	123	13	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	123	14	optimize	optimize	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	123	15	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	123	16	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	123	17	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	123	18	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	123	19	locally	locally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	123	20	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	123	21	defined	define	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	123	22	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	123	23	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	123	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	124	1	6	6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	124	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	125	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	2	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	125	3	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	125	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	125	6	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	125	7	assume	assume	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	125	8	access	access	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	9	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	10	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	125	11	list	list	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	13	likely	likely	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	125	14	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	125	15	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	125	16	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	17	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	125	18	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	19	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	20	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	125	21	argu-	argu-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	125	22	ment	ment	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	125	23	candidate	candidate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	24	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	125	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	125	26	such	such	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	125	27	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	28	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	125	29	optimization	optimization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	30	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	125	31	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	125	32	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	126	1	6	6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	126	2	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	126	3	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	126	4	solved	solve	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	126	5	explicitly	explicitly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	126	6	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	126	7	looping	loop	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	126	8	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	126	9	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	126	10	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	126	11	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	126	12	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	126	13	list	list	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	126	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	127	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	127	2	main	main	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	127	3	advantage	advantage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	5	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	127	6	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	7	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	127	8	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	127	10	since	since	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	11	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	127	12	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	127	13	precomputed	precompute	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	127	15	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	127	16	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	17	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	127	18	make	make	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	127	19	unrestricted	unrestricted	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	127	20	use	use	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	21	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	22	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	127	23	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	127	24	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	127	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	128	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	2	1	1	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	128	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	4	Mary	Mary	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	5	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	6	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	7	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	8	plays	play	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	128	9	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	10	3	3	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	128	11	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	12	guitar	guitar	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	13	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	14	4	4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	15	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	16	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	17	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	18	5	5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	19	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	20	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	21	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	22	6	6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	23	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	24	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	25	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	26	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	27	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	28	ARG0	ARG0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	29	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	30	2	2	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	128	31	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	32	ARG1	ARG1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	33	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	34	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	35	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	36	ARG2	ARG2	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	128	37	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	38	4	4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	39	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	40	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	41	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	42	5	5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	43	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	44	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	45	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	46	6	6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	47	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	48	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	49	W1,1	W1,1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	50	W4,2W2,3	W4,2W2,3	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	51	W3,4	W3,4	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	52	W5,5	w5,5	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	53	W6,6W1,1	W6,6W1,1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	54	W4,2W2,3	w4,2w2,3	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	55	W3,4	W3,4	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	56	W5,5	w5,5	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	57	W6,6W1,1	W6,6W1,1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	58	W4,2W2,3	w4,2w2,3	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	59	W3,4	W3,4	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	60	W5,5	w5,5	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	61	W6,6W1,1	W6,6W1,1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	62	W4,2W2,3	w4,2w2,3	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	63	W3,4	W3,4	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	64	W5,5	w5,5	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	65	W6,6W1,1	W6,6W1,1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	66	W4,2W2,3	w4,2w2,3	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	67	W3,4	W3,4	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	68	W5,5	w5,5	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	69	W6,6W1,1	W6,6W1,1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	70	W4,2W2,3	W4,2W2,3	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	71	W3,4	W3,4	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	72	W5,5	W5,5	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	73	W6,6	W6,6	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	74	Figure	Figure	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	75	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	76	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	128	77	Illustration	illustration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	78	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	79	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	128	80	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	81	graph	graph	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	82	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	83	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	128	84	sen-	sen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	85	tence	tence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	86	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	128	87	Mary	Mary	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	88	plays	play	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	128	89	guitar	guitar	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	90	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	128	91	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	128	92	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	128	93	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	128	94	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	95	“	"	``
work_7jtgrdqc4zeb7lul7vifzjglr4	128	96	plays	play	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	128	97	”	"	''
work_7jtgrdqc4zeb7lul7vifzjglr4	128	98	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	99	have	have	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	128	100	up	up	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	101	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	128	102	three	three	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	128	103	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	128	104	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	128	105	ARG0	ARG0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	106	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	107	agent	agent	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	108	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	109	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	128	110	ARG1	ARG1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	128	111	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	112	theme	theme	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	113	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	114	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	128	115	ARG2	ARG2	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	128	116	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	117	benefactor	benefactor	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	128	118	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	128	119	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	129	1	Nodes	node	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	129	2	labeled	label	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	129	3	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	129	4	represent	represent	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	129	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	129	6	null	null	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	129	7	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	129	8	or	or	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	129	9	token	token	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	129	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	130	1	Highlighted	highlight	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	130	2	edges	edge	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	130	3	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	130	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	130	5	correct	correct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	130	6	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	130	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	131	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	131	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	131	3	simple	simple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	131	4	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	131	5	employ	employ	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	131	6	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	131	7	probabilistic	probabilistic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	131	8	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	131	9	de-	de-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	131	10	pendency	pendency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	131	11	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	131	12	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	131	13	create	create	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	131	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	131	15	list	list	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	131	16	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	131	17	likely	likely	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	131	18	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	131	19	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	131	20	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	131	21	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	131	22	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	131	23	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	131	24	pair	pair	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	131	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	132	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	132	3	experiments	experiment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	132	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	132	5	explore	explore	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	132	6	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	132	7	approach	approach	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	132	9	show	show	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	132	10	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	11	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	12	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	132	13	average	average	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	14	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	15	44	44	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	132	16	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	132	17	per	per	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	18	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	19	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	132	20	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	132	21	recover	recover	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	132	22	86.2	86.2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	132	23	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	24	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	132	25	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	132	26	correct	correct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	132	27	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	132	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	133	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	133	2	leave	leave	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	133	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	133	4	future	future	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	133	5	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	133	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	133	7	development	development	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	133	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	133	9	ef-	ef-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	133	10	ficient	ficient	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	133	11	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	133	12	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	133	13	recover	recover	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	133	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	133	15	most	most	RBS
work_7jtgrdqc4zeb7lul7vifzjglr4	133	16	likely	likely	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	133	17	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	133	18	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	133	19	linking	link	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	133	20	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	133	21	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	133	22	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	133	23	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	133	24	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	133	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	134	1	3.2	3.2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	134	2	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	134	3	Assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	134	4	Algorithm	Algorithm	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	134	5	Coming	come	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	134	6	back	back	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	134	7	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	134	8	solving	solve	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	134	9	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	134	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	135	1	5	5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	135	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	135	3	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	135	4	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	135	5	easy	easy	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	135	6	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	135	7	see	see	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	135	8	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	135	9	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	135	10	optimal	optimal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	135	11	solution	solution	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	135	12	satisfying	satisfy	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	135	13	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	135	14	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	135	15	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	135	16	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	135	17	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	135	18	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	135	19	found	find	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	135	20	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	135	21	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	135	22	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	135	23	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	135	24	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	135	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	136	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	136	2	process	process	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	136	3	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	136	4	describe	describe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	136	5	determines	determine	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	136	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	136	7	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	136	8	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	136	9	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	136	10	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	136	11	separately	separately	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	136	12	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	136	13	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	136	14	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	136	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	137	1	Assume	assume	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	137	2	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	137	3	bipartite	bipartite	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	137	4	graph	graph	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	137	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	137	6	size	size	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	137	7	N	N	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	137	8	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	137	9	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	137	10	nodes	node	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	137	11	r1	r1	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	137	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	138	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	139	1	.rN	.rN	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	139	2	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	139	3	one	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	139	4	side	side	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	139	5	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	139	6	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	139	7	nodes	node	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	139	8	a1	a1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	139	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	140	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	141	1	.aN	.an	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	141	2	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	141	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	141	4	other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	141	5	side	side	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	141	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	142	1	Assume	assume	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	142	2	also	also	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	142	3	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	142	4	matrix	matrix	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	142	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	142	6	non	non	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	142	7	-	-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	142	8	negative	negative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	142	9	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	142	10	Wi	Wi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	142	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	142	12	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	142	13	corresponding	correspond	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	142	14	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	142	15	assigning	assigning	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	142	16	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	142	17	aj	aj	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	142	18	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	142	19	role	role	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	142	20	ri	ri	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	142	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	143	1	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	143	2	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	143	3	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	143	4	algo-	algo-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	143	5	rithm	rithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	143	6	finds	find	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	143	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	143	8	bijection	bijection	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	143	9	f	f	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	143	10	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	143	11	i	i	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	143	12	→	→	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	143	13	j	j	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	143	14	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	143	15	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	143	16	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	143	17	argu-	argu-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	143	18	ments	ment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	143	19	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	143	20	maximizes	maximizes	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	143	21	∑N	∑N	,
work_7jtgrdqc4zeb7lul7vifzjglr4	143	22	i=1	i=1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	143	23	Wi	Wi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	143	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	143	25	f(i	f(i	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	143	26	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	143	27	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	144	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	144	2	Hungarian	hungarian	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	144	3	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	144	4	finds	find	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	144	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	144	6	exact	exact	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	144	7	solution	solution	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	144	8	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	144	9	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	144	10	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	144	11	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	144	12	O(N3	O(N3	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	144	13	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	144	14	time	time	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	144	15	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	144	16	Kuhn	Kuhn	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	144	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	144	18	1955	1955	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	144	19	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	144	20	Burkard	Burkard	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	144	21	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	144	22	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	144	23	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	144	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	144	25	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	144	26	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	144	27	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	145	1	All	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	145	2	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	145	3	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	145	4	left	leave	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	5	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	145	6	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	145	7	construct	construct	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	145	8	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	145	9	bipartite	bipartite	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	145	10	graph	graph	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	11	rep-	rep-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	12	resenting	resent	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	145	13	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	14	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	145	15	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	145	16	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	17	tokens	token	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	145	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	145	19	such	such	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	145	20	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	21	some	some	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	145	22	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	145	23	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	145	24	tokens	token	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	145	25	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	145	26	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	145	27	left	leave	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	28	unassigned	unassigned	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	145	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	145	30	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	145	31	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	145	32	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	145	33	common	common	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	145	34	setting	setting	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	35	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	36	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	145	37	tasks	task	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	145	38	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	146	1	Al-	al-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	146	2	gorithm	gorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	146	3	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	146	4	describes	describe	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	146	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	146	6	procedure	procedure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	146	7	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	146	8	constructing	construct	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	146	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	146	10	weighted	weighted	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	146	11	bipartite	bipartite	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	146	12	graph	graph	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	146	13	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	146	14	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	146	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	146	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	146	17	Figure	figure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	146	18	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	146	19	il-	il-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	146	20	lustrates	lustrate	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	146	21	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	146	22	example	example	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	146	23	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	146	24	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	146	25	bipartite	bipartite	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	146	26	graph	graph	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	146	27	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	147	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	147	2	then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	147	3	222	222	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	147	4	Algorithm	Algorithm	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	147	5	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	147	6	Construction	construction	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	147	7	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	147	8	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	147	9	Assignment	Assignment	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	147	10	Graph	Graph	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	147	11	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	147	12	Semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	147	13	Role	Role	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	147	14	Labeling	Labeling	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	147	15	Let	let	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	147	16	p	p	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	147	17	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	147	18	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	147	19	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	147	20	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	147	21	k	k	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	147	22	possible	possible	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	147	23	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	147	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	148	1	Let	let	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	148	2	n	n	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	148	3	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	148	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	148	5	number	number	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	148	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	148	7	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	148	8	candidates	candidate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	148	9	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	148	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	148	11	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	148	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	149	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	149	2	al-	al-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	149	3	gorithm	gorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	149	4	creates	create	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	149	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	149	6	bipartite	bipartite	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	149	7	graph	graph	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	149	8	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	149	9	N	n	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	149	10	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	149	11	n+k	n+k	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	149	12	vertices	vertex	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	149	13	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	149	14	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	149	15	side	side	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	149	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	150	1	1	1	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	150	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	151	1	Create	create	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	151	2	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	151	3	vertices	vertex	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	151	4	ri	ri	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	151	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	151	6	i	i	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	151	7	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	151	8	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	151	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	152	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	153	1	.N	.N	:
work_7jtgrdqc4zeb7lul7vifzjglr4	153	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	153	3	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	153	4	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	153	6	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	153	7	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	8	i	i	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	9	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	10	k	k	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	153	12	ri	ri	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	13	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	153	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	153	15	i	i	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	16	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	153	17	th	th	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	153	18	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	153	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	153	20	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	21	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	153	22	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	153	23	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	24	i	i	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	25	≤	≤	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	26	n	n	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	153	28	rk+i	rk+i	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	29	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	153	30	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	153	31	special	special	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	153	32	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	153	33	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	153	34	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	154	1	2	2	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	154	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	155	1	Create	create	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	155	2	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	155	3	vertices	vertice	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	155	4	aj	aj	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	155	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	155	6	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	155	7	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	155	8	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	155	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	156	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	157	1	.N	.N	:
work_7jtgrdqc4zeb7lul7vifzjglr4	157	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	157	3	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	157	4	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	157	6	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	157	7	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	8	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	9	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	10	n	n	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	157	12	aj	aj	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	13	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	157	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	157	15	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	16	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	157	17	th	th	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	18	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	157	19	candidate	candidate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	157	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	157	21	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	22	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	157	23	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	157	24	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	25	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	26	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	27	k	k	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	157	29	an+j	an+j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	30	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	157	31	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	157	32	special	special	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	157	33	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	157	34	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	157	35	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	158	1	3	3	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	158	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	159	1	Define	define	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	159	2	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	159	3	matrix	matrix	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	5	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	6	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	159	7	S	S	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	8	∈	∈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	9	R(k+1)×n	R(k+1)×n	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	159	10	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	159	11	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	159	12	a	a	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	159	13	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	159	14	Optimization	optimization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	16	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	159	17	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	159	18	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	159	19	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	20	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	159	21	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	22	i	i	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	23	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	24	k	k	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	26	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	159	27	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	28	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	29	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	30	n	n	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	31	Si	Si	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	33	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	34	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	159	35	max	max	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	36	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	37	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	39	aj	aj	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	40	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	41	ri	ri	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	42	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	43	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	44	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	45	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	46	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	47	aj	aj	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	48	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	49	ri	ri	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	50	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	51	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	52	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	53	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	54	aj	aj	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	55	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	159	56	ri	ri	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	159	57	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	159	58	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	159	59	b	b	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	60	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	159	61	Scores	Scores	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	62	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	63	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	64	assignments3	assignments3	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	65	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	159	66	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	67	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	159	68	≤	≤	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	159	69	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	70	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	71	n	n	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	72	Sk+1,j	Sk+1,j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	159	73	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	159	74	0	0	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	159	75	4	4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	159	76	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	160	1	Let	let	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	160	2	S0	s0	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	3	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	160	4	mini	mini	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	6	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	7	Si	Si	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	9	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	160	12	minimum	minimum	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	13	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	14	any	any	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	160	15	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	16	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	17	S.	S.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	18	Define	Define	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	19	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	160	20	matrix	matrix	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	21	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	22	non	non	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	23	-	-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	24	negative	negative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	25	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	160	26	W	W	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	27	∈	∈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	28	RN×N	RN×N	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	160	29	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	30	follows	follow	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	31	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	160	32	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	160	33	a	a	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	160	34	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	160	35	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	36	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	160	37	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	38	i	i	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	39	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	40	k	k	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	41	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	42	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	160	43	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	44	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	45	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	46	n	n	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	160	47	Wi	Wi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	48	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	49	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	50	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	160	51	Si	Si	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	52	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	53	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	54	−S0	−s0	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	160	55	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	160	56	b	b	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	57	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	160	58	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	59	k	k	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	60	<	<	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	160	61	i	i	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	62	≤	≤	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	63	N	n	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	160	64	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	65	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	160	66	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	67	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	68	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	69	n	n	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	160	70	Wi	Wi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	71	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	72	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	73	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	160	74	Sk+1,j	sk+1,j	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	160	75	−S0	−s0	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	160	76	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	160	77	c	c	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	78	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	160	79	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	80	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	160	81	<	<	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	160	82	i	i	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	83	≤	≤	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	84	N	N	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	85	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	86	n	n	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	87	<	<	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	160	88	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	89	≤	≤	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	90	N	N	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	91	Wi	Wi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	92	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	93	j	j	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	94	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	160	95	0	0	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	160	96	run	run	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	97	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	160	98	Hungarian	hungarian	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	99	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	100	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	101	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	160	102	weighted	weighted	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	103	graph	graph	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	104	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	160	105	obtain	obtain	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	160	106	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	160	107	bijection	bijection	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	108	f	f	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	160	109	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	160	110	ri	ri	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	111	→	→	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	160	112	aj	aj	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	113	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	160	114	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	115	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	160	116	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	117	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	118	trivial	trivial	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	119	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	160	120	recover	recover	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	160	121	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	160	122	optimal	optimal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	160	123	solution	solution	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	124	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	160	125	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	160	126	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	161	1	5	5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	161	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	162	1	Finally	finally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	162	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	162	3	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	162	4	note	note	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	162	5	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	6	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	162	7	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	162	8	simple	simple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	162	9	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	162	10	allow	allow	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	162	11	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	12	multi-	multi-	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	162	13	ple	ple	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	14	instances	instance	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	162	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	16	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	162	17	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	162	18	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	19	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	20	adding	add	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	162	21	more	more	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	162	22	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	23	nodes	node	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	162	24	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	25	step	step	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	26	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	162	27	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	162	28	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	162	29	would	would	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	162	30	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	162	31	straightforward	straightforward	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	162	32	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	162	33	add	add	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	162	34	penalties	penalty	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	162	35	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	36	step	step	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	37	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	162	38	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	39	multiple	multiple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	162	40	instances	instance	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	162	41	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	162	42	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	162	43	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	163	1	4	4	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	163	2	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	163	3	Dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	163	4	Decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	163	5	Algorithm	Algorithm	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	163	6	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	163	7	now	now	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	163	8	present	present	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	163	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	163	10	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	163	11	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	163	12	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	163	13	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	163	14	op-	op-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	163	15	timize	timize	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	163	16	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	163	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	164	1	4	4	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	164	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	164	3	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	164	4	uses	use	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	164	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	164	6	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	164	7	algorithm	algorithm	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	164	8	pre-	pre-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	164	9	sented	sente	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	164	10	above	above	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	164	11	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	164	12	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	164	13	subroutine	subroutine	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	164	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	165	1	Our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	165	2	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	165	3	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	165	4	sim-	sim-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	165	5	ilar	ilar	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	165	6	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	165	7	that	that	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	165	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	165	9	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	165	10	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	165	11	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	165	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	166	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	166	2	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	166	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	166	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	166	5	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	166	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	166	7	sense	sense	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	166	8	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	166	9	3In	3in	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	166	10	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	166	11	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	166	12	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	166	13	fix	fix	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	166	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	166	15	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	166	16	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	166	17	null	null	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	166	18	assignments	assignment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	166	19	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	166	20	0	0	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	166	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	167	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	167	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	167	3	straightforward	straightforward	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	167	4	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	167	5	compute	compute	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	167	6	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	167	7	discriminative	discriminative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	167	8	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	167	9	instead	instead	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	167	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	168	1	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	168	2	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	168	3	optimization	optimization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	168	4	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	168	5	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	168	6	decomposed	decompose	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	168	7	into	into	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	168	8	two	two	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	168	9	sub	sub	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	168	10	-	-	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	168	11	problems	problem	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	168	12	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	168	13	need	need	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	168	14	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	168	15	agree	agree	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	168	16	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	168	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	168	18	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	168	19	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	168	20	they	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	168	21	predict	predict	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	168	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	169	1	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	169	2	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	169	3	detailed	detailed	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	169	4	descrip-	descrip-	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	169	5	tion	tion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	169	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	169	7	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	169	8	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	169	9	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	169	10	applied	apply	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	169	11	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	169	12	NLP	NLP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	169	13	see	see	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	169	14	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	169	15	Sontag	Sontag	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	169	16	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	169	17	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	169	18	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	169	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	169	20	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	169	21	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	169	22	Rush	Rush	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	169	23	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	169	24	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	169	25	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	169	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	169	27	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	169	28	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	169	29	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	170	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	170	2	note	note	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	170	3	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	170	4	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	170	5	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	170	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	171	1	4	4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	171	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	3	constraint	constraint	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	4	cSubtree	cSubtree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	171	5	ties	tie	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	171	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	7	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	171	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	171	9	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	171	10	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	171	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	171	12	imposing	impose	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	171	13	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	14	any	any	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	15	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	16	πp	πp	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	171	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	171	18	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	171	20	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	21	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	22	links	link	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	171	23	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	24	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	25	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	26	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	27	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	28	argu-	argu-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	171	29	ment	ment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	30	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	31	must	must	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	171	32	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	171	33	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	34	subtree	subtree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	35	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	171	36	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	171	37	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	171	38	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	171	39	struc-	struc-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	171	40	ture	ture	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	171	41	y.	y.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	1	Formally	formally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	172	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	172	3	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	172	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	172	5	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	172	6	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	172	7	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	172	8	yh	yh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	10	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	12	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	172	13	≥	≥	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	172	14	πp	πp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	172	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	16	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	172	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	18	rh	rh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	20	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	22	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	23	∀	∀	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	24	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	26	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	28	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	172	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	30	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	172	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	32	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	172	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	172	34	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	172	35	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	173	1	These	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	173	2	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	173	3	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	173	4	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	173	5	compactly	compactly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	173	6	written	write	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	7	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	8	c	c	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	9	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	10	yh	yh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	12	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	14	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	15	≥	≥	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	16	∑	∑	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	17	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	19	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	21	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	22	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	23	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	25	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	27	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	28	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	30	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	32	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	33	∀	∀	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	34	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	36	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	38	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	39	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	40	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	173	41	c	c	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	42	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	173	43	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	173	44	constant	constant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	173	45	equal	equal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	173	46	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	47	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	173	48	number	number	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	49	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	173	50	dis-	dis-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	173	51	tinct	tinct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	173	52	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	173	53	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	173	54	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	55	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	56	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	57	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	58	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	173	59	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	60	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	173	61	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	174	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	2	addition	addition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	5	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	174	6	introduce	introduce	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	174	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	174	8	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	9	non	non	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	174	10	-	-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	174	11	negative	negative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	174	12	slack	slack	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	13	vari-	vari-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	14	ables	ables	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	15	ξ	ξ	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	174	16	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	174	18	component	component	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	19	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	20	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	174	21	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	174	22	depen-	depen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	23	dency	dency	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	24	ξh	ξh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	26	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	28	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	30	turning	turn	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	174	31	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	174	32	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	174	33	into	into	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	34	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	174	35	c	c	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	174	36	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	37	yh	yh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	39	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	40	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	41	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	42	−	−	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	43	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	174	44	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	45	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	46	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	47	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	48	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	49	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	50	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	51	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	52	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	53	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	54	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	55	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	56	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	57	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	58	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	59	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	60	−	−	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	61	ξh	ξh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	62	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	63	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	64	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	65	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	66	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	174	67	0	0	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	174	68	∀	∀	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	69	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	70	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	71	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	72	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	174	73	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	174	74	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	75	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	174	76	now	now	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	174	77	rewrite	rewrite	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	174	78	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	174	79	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	175	1	4	4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	175	2	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	3	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	175	4	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	5	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	7	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	9	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	11	ξ≥0	ξ≥0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	12	s	s	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	175	13	syn(x	syn(x	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	15	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	16	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	175	17	+	+	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	18	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	19	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	21	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	23	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	175	24	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	175	25	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	175	26	7	7	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	175	27	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	175	28	subject	subject	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	29	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	30	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	32	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	34	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	36	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	37	∀h	∀h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	39	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	40	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	41	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	42	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	175	43	c	c	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	44	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	45	yh	yh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	46	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	47	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	48	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	49	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	50	−	−	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	51	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	175	52	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	53	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	54	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	55	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	56	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	57	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	58	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	59	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	60	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	61	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	62	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	63	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	64	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	65	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	66	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	67	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	68	−	−	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	69	ξh	ξh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	70	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	71	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	72	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	175	73	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	74	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	175	75	0	0	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	76	As	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	77	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	175	78	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	79	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	80	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	175	81	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	176	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	176	2	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	176	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	176	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	176	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	176	6	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	176	7	relax	relax	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	176	8	subtree	subtree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	176	9	cons-	cons-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	176	10	traints	traint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	176	11	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	176	12	introducing	introduce	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	176	13	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	176	14	vector	vector	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	176	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	176	16	Lagrange	Lagrange	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	176	17	multipli-	multipli-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	176	18	ers	ers	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	176	19	λ	λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	176	20	indexed	index	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	176	21	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	176	22	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	176	23	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	176	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	176	25	i.e.	i.e.	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	177	1	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	177	2	coordinate	coordinate	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	177	3	λh	λh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	177	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	177	5	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	177	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	177	7	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	177	8	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	177	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	177	10	Lagrange	Lagrange	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	177	11	multiplier	multiplier	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	177	12	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	177	13	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	177	14	constraint	constraint	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	177	15	associated	associate	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	177	16	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	177	17	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	177	18	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	177	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	177	20	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	177	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	177	22	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	177	23	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	177	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	178	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	178	2	Lagrangian	Lagrangian	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	178	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	178	5	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	178	6	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	178	7	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	178	8	L(y	L(y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	10	z	z	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	178	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	12	π	π	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	178	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	14	ξ	ξ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	16	λ)=	λ)=	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	17	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	18	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	20	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	21	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	178	22	+	+	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	23	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	24	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	26	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	28	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	178	29	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	178	30	+	+	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	178	31	λ	λ	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	178	32	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	33	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	178	34	c	c	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	178	35	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	36	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	37	−	−	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	38	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	178	39	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	178	40	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	41	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	42	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	43	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	44	πp	πp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	178	45	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	46	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	178	47	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	178	48	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	49	−ξ	−ξ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	50	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	178	51	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	178	52	8)	8)	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	178	53	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	54	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	178	55	now	now	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	178	56	formulate	formulate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	178	57	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	178	58	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	179	1	7	7	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	179	2	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	179	3	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	179	4	max	max	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	179	5	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	179	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	179	7	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	179	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	179	9	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	179	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	179	11	ξ≥0	ξ≥0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	179	12	s.t	s.t	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	179	13	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	180	1	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	3	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	5	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	7	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	8	c·y−	c·y−	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	180	9	∑	∑	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	10	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	12	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	14	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	15	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	16	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	18	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	180	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	20	r−ξ=0	r−ξ=0	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	180	21	L(y	L(y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	23	z	z	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	25	π	π	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	27	ξ	ξ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	180	29	λ	λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	30	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	180	31	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	180	32	9	9	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	180	33	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	180	34	223	223	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	180	35	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	180	36	optimization	optimization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	37	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	38	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	180	39	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	180	40	property	property	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	41	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	180	42	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	180	43	optimum	optimum	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	180	44	value	value	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	45	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	180	46	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	180	47	same	same	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	180	48	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	49	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	180	50	optimum	optimum	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	51	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	180	52	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	180	53	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	181	1	7	7	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	181	2	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	181	3	any	any	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	181	4	value	value	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	181	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	181	6	λ	λ	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	181	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	182	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	182	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	182	3	because	because	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	182	4	whenever	whenever	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	182	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	182	6	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	182	7	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	182	8	satisfied	satisfied	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	182	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	182	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	182	11	terms	term	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	182	12	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	182	13	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	182	14	Lagrangian	Lagrangian	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	182	15	involving	involve	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	182	16	λ	λ	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	182	17	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	182	18	zero	zero	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	182	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	183	1	If	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	183	2	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	183	3	remove	remove	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	183	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	183	5	subtree	subtree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	183	6	con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	183	7	straints	straint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	183	8	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	183	9	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	183	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	184	1	9	9	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	184	2	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	184	3	obtain	obtain	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	184	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	184	5	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	184	6	objective	objective	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	184	7	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	184	8	D(λ	d(λ	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	184	9	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	184	10	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	184	11	max	max	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	184	12	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	184	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	184	14	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	184	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	184	16	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	184	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	184	18	ξ≥0	ξ≥0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	184	19	s.t	s.t	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	184	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	185	1	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	185	3	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	185	5	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	185	7	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	8	L(y	L(y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	185	10	z	z	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	185	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	185	12	π	π	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	185	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	185	14	ξ	ξ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	185	16	λ	λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	17	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	185	18	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	185	19	10	10	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	185	20	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	185	21	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	22	max	max	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	23	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	24	s.t	s.t	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	185	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	186	1	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	2	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	186	3	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	4	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	186	6	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	7	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	186	8	+	+	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	186	9	c	c	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	10	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	11	y	y	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	186	12	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	13	λ	λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	14	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	186	15	+	+	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	186	16	max	max	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	17	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	186	19	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	20	s.t	s.t	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	186	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	187	1	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	187	3	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	187	5	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	6	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	187	7	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	8	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	187	10	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	187	12	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	187	13	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	187	14	−λ	−λ	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	187	15	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	16	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	187	17	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	187	19	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	187	21	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	22	πp	πp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	187	24	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	187	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	187	26	r	r	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	187	27	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	187	28	+	+	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	187	29	max	max	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	30	ξ≥0	ξ≥0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	31	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	187	32	−λ	−λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	33	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	34	ξ	ξ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	35	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	187	36	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	187	37	11	11	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	187	38	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	187	39	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	187	40	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	187	41	objective	objective	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	42	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	187	43	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	187	44	upper	upper	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	187	45	bound	bound	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	46	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	47	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	187	48	opti-	opti-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	49	mal	mal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	187	50	value	value	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	51	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	52	primal	primal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	187	53	objective	objective	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	54	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	187	55	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	187	56	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	188	1	7	7	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	188	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	189	1	Thus	thus	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	189	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	189	3	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	189	4	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	189	5	interested	interested	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	189	6	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	189	7	finding	find	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	189	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	189	9	minimum	minimum	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	189	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	189	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	189	12	dual	dual	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	189	13	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	189	14	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	189	15	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	189	16	tighten	tighten	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	189	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	189	18	upper	upper	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	189	19	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	189	20	bound	bind	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	189	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	190	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	190	2	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	190	3	solve	solve	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	190	4	min	min	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	190	5	λ	λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	190	6	D(λ	d(λ	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	190	7	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	190	8	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	190	9	12	12	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	190	10	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	190	11	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	190	12	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	190	13	subgradient	subgradient	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	190	14	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	190	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	191	1	Algorithm	Algorithm	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	191	2	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	191	3	presents	present	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	191	4	pseudo	pseudo	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	191	5	-	-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	191	6	code	code	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	191	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	192	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	192	2	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	192	3	takes	take	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	192	4	advantage	advantage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	192	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	192	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	192	7	decomposed	decompose	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	192	8	form	form	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	192	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	192	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	192	11	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	192	12	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	192	13	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	192	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	193	1	11	11	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	193	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	193	3	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	193	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	193	5	have	have	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	193	6	rewritten	rewrite	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	193	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	193	8	Lagrangian	Lagrangian	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	193	9	such	such	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	193	10	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	193	11	syntac-	syntac-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	193	12	tic	tic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	193	13	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	193	14	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	193	15	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	193	16	appear	appear	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	193	17	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	193	18	separate	separate	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	193	19	terms	term	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	193	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	194	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	194	2	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	194	3	allow	allow	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	194	4	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	194	5	compute	compute	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	194	6	subgradients	subgradient	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	194	7	efficiently	efficiently	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	194	8	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	195	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	2	particular	particular	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	195	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	195	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	195	5	subgradient	subgradient	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	7	D	d	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	8	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	195	10	point	point	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	11	λ	λ	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	12	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	195	13	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	195	14	∆(λ	∆(λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	15	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	195	16	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	17	c	c	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	18	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	19	ŷ	ŷ	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	195	20	−	−	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	21	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	195	22	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	195	24	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	195	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	195	26	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	27	π̂p	π̂p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	195	29	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	195	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	195	31	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	32	−	−	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	33	ξ̂	ξ̂	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	34	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	195	35	13	13	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	195	36	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	195	37	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	195	38	ŷ	ŷ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	39	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	195	40	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	41	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	42	s.t	s.t	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	195	43	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	196	1	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	2	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	196	3	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	4	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	196	6	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	7	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	196	8	+	+	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	196	9	c	c	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	10	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	11	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	12	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	13	λ	λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	14	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	196	15	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	196	16	14	14	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	196	17	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	196	18	ẑ,π̂	ẑ,π̂	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	19	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	196	20	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	21	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	196	23	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	24	s.t	s.t	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	196	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	197	1	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	197	3	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	197	5	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	6	s	s	POS
work_7jtgrdqc4zeb7lul7vifzjglr4	197	7	srl(x	srl(x	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	197	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	197	9	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	197	11	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	197	12	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	197	13	−λ	−λ	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	197	14	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	15	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	197	16	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	197	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	197	18	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	197	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	197	20	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	21	πp	πp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	197	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	197	23	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	197	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	197	25	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	197	26	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	197	27	15	15	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	197	28	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	197	29	ξ̂	ξ̂	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	197	30	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	197	31	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	32	ξ≥0	ξ≥0	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	33	−λ	−λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	34	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	35	ξ	ξ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	36	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	197	37	16	16	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	197	38	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	197	39	Whenever	whenever	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	197	40	π̂	π̂	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	41	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	197	42	consistent	consistent	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	197	43	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	197	44	ŷ	ŷ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	197	45	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	197	46	subgradient	subgradient	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	197	47	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	197	48	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	197	49	zero	zero	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	197	50	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	197	51	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	197	52	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	197	53	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	197	54	converge	converge	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	197	55	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	198	1	When	when	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	198	2	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	198	3	π̂	π̂	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	4	contain	contain	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	198	6	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	7	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	8	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	10	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	12	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	13	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	14	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	198	15	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	198	16	in-	in-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	198	17	consistent	consistent	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	198	18	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	19	ŷ	ŷ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	198	22	associated	associate	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	23	dual	dual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	24	λh	λh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	26	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	28	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	29	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	198	30	in-	in-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	198	31	crease	crease	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	198	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	33	hence	hence	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	198	34	lowering	lower	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	198	35	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	198	36	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	37	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	38	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	198	39	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	198	40	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	198	41	use	use	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	42	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	43	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	44	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	45	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	46	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	47	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	48	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	49	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	50	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	198	51	next	next	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	198	52	iteration	iteration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	53	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	198	54	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	55	same	same	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	198	56	time	time	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	57	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	58	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	198	59	to-	to-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	198	60	tal	tal	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	198	61	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	62	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	63	that	that	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	198	64	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	65	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	198	66	increase	increase	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	198	67	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	198	68	favoring	favor	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	198	69	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	198	70	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	71	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	198	72	alternative	alternative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	198	73	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	198	74	ŷ.	ŷ.	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	199	1	As	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	2	Algorithm	Algorithm	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	3	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	199	4	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	199	5	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	199	6	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	199	7	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	8	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	9	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	10	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	199	11	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	199	12	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	199	13	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	14	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	199	15	Input	input	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	16	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	199	17	x	x	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	199	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	199	19	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	199	20	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	21	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	199	22	T	t	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	199	24	number	number	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	25	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	26	iterations	iteration	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	199	27	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	199	28	Output	output	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	29	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	199	30	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	199	31	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	199	32	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	199	33	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	199	34	ŷ	ŷ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	35	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	199	36	ẑ	ẑ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	37	Notation	Notation	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	38	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	199	39	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	40	use	use	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	41	cSem=	cSem=	.
work_7jtgrdqc4zeb7lul7vifzjglr4	199	42	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	43	∧	∧	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	44	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	45	∧	∧	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	46	cPath	cPath	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	47	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	199	48	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	199	49	λ1	λ1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	50	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	199	51	0	0	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	52	#	#	$
work_7jtgrdqc4zeb7lul7vifzjglr4	199	53	initialize	initialize	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	54	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	199	55	variables	variable	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	199	56	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	199	57	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	199	58	c	c	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	59	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	199	60	number	number	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	61	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	62	distinct	distinct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	199	63	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	64	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	65	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	199	66	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	67	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	199	68	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	69	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	70	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	71	x	x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	72	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	199	73	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	199	74	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	199	75	t	t	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	199	76	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	199	77	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	199	78	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	200	1	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	201	1	.T	.T	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	201	2	do	do	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	201	3	4	4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	201	4	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	201	5	ŷ	ŷ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	201	6	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	201	7	argmaxy	argmaxy	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	201	8	s.t	s.t	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	201	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	202	1	cTree	cTree	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	2	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	202	3	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	4	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	202	6	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	7	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	202	8	+	+	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	202	9	c	c	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	10	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	11	λt	λt	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	12	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	13	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	14	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	202	15	5	5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	202	16	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	202	17	ẑ,π̂	ẑ,π̂	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	18	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	202	19	argmax	argmax	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	20	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	202	22	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	23	s.t	s.t	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	202	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	203	1	cSem	cSem	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	2	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	203	3	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	4	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	6	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	8	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	203	9	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	203	10	−λt	−λt	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	203	11	·	·	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	12	∑	∑	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	13	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	15	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	17	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	18	π	π	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	19	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	21	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	23	r	r	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	24	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	203	25	6	6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	203	26	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	203	27	λt+1	λt+1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	28	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	203	29	λt	λt	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	30	#	#	$
work_7jtgrdqc4zeb7lul7vifzjglr4	203	31	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	203	32	variables	variable	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	203	33	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	34	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	203	35	next	next	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	203	36	iteration	iteration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	37	7	7	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	203	38	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	203	39	Set	Set	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	40	αt	αt	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	203	41	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	42	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	203	43	step	step	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	44	size	size	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	45	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	46	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	203	47	current	current	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	203	48	iteration	iteration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	49	8	8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	203	50	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	203	51	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	52	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	203	53	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	54	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	55	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	56	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	57	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	58	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	59	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	60	do	do	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	203	61	9	9	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	203	62	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	203	63	q	q	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	64	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	203	65	∑	∑	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	66	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	67	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	68	a	a	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	69	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	70	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	71	π̂	π̂	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	72	p	p	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	73	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	74	a	a	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	75	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	76	r	r	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	203	77	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	78	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	79	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	80	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	203	81	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	82	#	#	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	83	num	num	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	203	84	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	204	1	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	204	2	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	204	3	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	4	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	6	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	8	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	9	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	10	10	10	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	204	11	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	204	12	if	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	13	q	q	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	14	>	>	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	15	0	0	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	204	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	204	17	ŷh	ŷh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	19	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	21	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	22	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	204	23	0	0	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	24	then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	204	25	11	11	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	204	26	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	204	27	λt+1h	λt+1h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	29	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	31	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	32	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	204	33	λ	λ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	34	t+1	t+1	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	35	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	36	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	37	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	39	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	40	+	+	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	204	41	αtq	αtq	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	204	42	12	12	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	204	43	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	204	44	break	break	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	204	45	if	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	46	λt+1	λt+1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	47	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	204	48	λt	λt	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	49	#	#	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	50	convergence	convergence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	51	13	13	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	204	52	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	204	53	return	return	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	54	ŷ	ŷ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	55	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	56	ẑ	ẑ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	57	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	58	previous	previous	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	204	59	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	60	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	204	61	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	62	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	204	63	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	64	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	204	65	parameter	parameter	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	66	αt	αt	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	204	67	controls	control	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	204	68	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	204	69	size	size	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	70	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	71	subgradient	subgradient	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	204	72	steps	step	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	204	73	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	74	iteration	iteration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	75	t.	t.	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	76	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	204	77	key	key	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	204	78	point	point	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	79	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	80	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	204	81	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	82	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	204	83	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	84	solutions	solution	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	204	85	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	204	86	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	204	87	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	205	1	14	14	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	205	2	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	205	3	15	15	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	205	4	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	205	5	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	205	6	computed	compute	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	205	7	efficiently	efficiently	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	205	8	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	205	9	sep-	sep-	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	205	10	arate	arate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	205	11	processes	process	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	205	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	206	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	206	2	particular	particular	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	206	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	206	4	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	206	5	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	207	1	14	14	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	207	2	corresponds	correspond	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	207	3	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	4	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	207	5	standard	standard	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	207	6	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	7	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	207	8	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	207	10	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	207	11	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	12	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	207	13	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	14	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	15	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	207	17	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	207	19	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	20	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	21	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	22	have	have	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	23	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	207	24	additional	additional	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	207	25	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	26	term	term	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	27	c·λh	c·λh	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	207	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	207	29	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	207	31	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	32	—	—	:
work_7jtgrdqc4zeb7lul7vifzjglr4	207	33	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	34	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	207	35	experiments	experiment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	207	36	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	37	use	use	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	207	38	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	207	39	projected	projected	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	207	40	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	41	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	207	42	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	43	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	44	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	207	45	Eisner	eisner	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	207	46	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	207	47	2000	2000	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	207	48	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	207	49	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	208	1	To	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	208	2	calculate	calculate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	208	3	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	208	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	209	1	15	15	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	209	2	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	209	3	use	use	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	209	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	209	5	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	6	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	7	described	describe	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	8	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	9	Section	section	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	10	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	209	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	209	12	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	209	13	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	209	14	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	209	15	straight-	straight-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	16	forward	forward	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	209	17	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	209	18	introduce	introduce	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	209	19	additional	additional	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	209	20	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	21	terms	term	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	209	22	−λh	−λh	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	209	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	209	24	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	209	26	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	27	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	28	every	every	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	209	29	factor	factor	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	30	πp	πp	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	209	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	209	32	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	209	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	209	34	rh	rh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	209	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	209	36	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	209	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	209	38	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	209	39	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	210	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	210	2	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	210	3	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	210	4	shown	show	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	210	5	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	210	6	whenever	whenever	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	210	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	210	8	subgradient	subgradient	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	210	9	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	210	10	converges	converge	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	210	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	210	12	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	210	13	solutions	solution	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	210	14	ŷ	ŷ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	210	15	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	210	16	ẑ	ẑ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	210	17	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	210	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	210	19	optimal	optimal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	210	20	solutions	solution	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	210	21	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	210	22	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	210	23	original	original	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	210	24	prob-	prob-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	210	25	lem	lem	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	210	26	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	210	27	Eq	Eq	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	210	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	211	1	4	4	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	211	2	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	211	3	see	see	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	211	4	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	211	5	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	211	6	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	211	7	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	211	8	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	211	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	211	10	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	211	11	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	211	12	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	211	13	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	211	14	justifi-	justifi-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	211	15	cation	cation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	211	16	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	211	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	212	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	212	2	practice	practice	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	212	3	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	212	4	run	run	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	212	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	212	6	subgradient	subgradient	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	212	7	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	212	8	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	212	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	212	10	maximum	maximum	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	212	11	number	number	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	212	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	212	13	iterations	iteration	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	212	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	212	15	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	212	16	return	return	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	212	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	212	18	solutions	solution	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	212	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	212	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	212	21	last	last	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	212	22	iteration	iteration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	212	23	if	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	212	24	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	212	25	does	do	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	212	26	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	212	27	converge	converge	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	212	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	213	1	5	5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	213	2	Related	related	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	213	3	Work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	4	Recently	recently	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	213	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	213	6	there	there	EX
work_7jtgrdqc4zeb7lul7vifzjglr4	213	7	have	have	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	213	8	been	be	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	213	10	number	number	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	11	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	12	approaches	approach	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	213	13	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	14	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	213	15	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	16	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	17	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	213	18	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	213	19	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	213	20	dependen-	dependen-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	213	21	cies	cie	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	213	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	213	23	partly	partly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	213	24	because	because	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	25	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	26	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	213	27	availability	availability	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	28	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	29	treebanks	treebank	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	213	30	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	31	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	213	32	format	format	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	33	popularized	popularize	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	34	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	35	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	213	36	CoNLL	CoNLL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	213	37	shared	share	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	213	38	tasks	task	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	213	39	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	213	40	Surdeanu	Surdeanu	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	213	41	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	213	42	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	213	43	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	213	44	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	213	45	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	213	46	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	213	47	Hajič	hajič	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	213	48	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	213	49	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	213	50	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	213	51	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	213	52	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	213	53	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	213	54	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	214	1	Like	like	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	2	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	3	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	214	4	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	214	6	Johansson	Johansson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	214	7	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	214	8	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	214	9	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	214	10	defined	define	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	214	11	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	214	12	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	13	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	214	14	exploits	exploit	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	214	15	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	214	16	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	214	18	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	214	19	depen-	depen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	20	224	224	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	214	21	dency	dency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	22	together	together	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	214	23	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	24	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	214	25	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	214	26	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	27	connecting	connect	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	214	28	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	214	29	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	30	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	214	31	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	214	32	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	214	33	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	215	1	That	that	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	215	2	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	215	3	uses	use	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	215	4	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	215	5	approximate	approximate	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	215	6	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	215	7	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	215	8	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	215	9	employs	employ	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	215	10	k	k	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	215	11	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	215	12	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	215	13	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	215	14	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	215	15	beam	beam	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	215	16	search	search	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	215	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	216	1	Similarly	similarly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	216	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	216	3	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	216	4	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	216	5	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	216	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	217	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	217	2	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	217	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	217	4	defined	define	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	217	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	217	6	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	217	7	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	8	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	217	9	forces	force	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	217	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	217	11	predi-	predi-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	12	cate	cate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	13	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	14	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	217	15	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	217	16	represented	represent	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	17	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	217	19	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	217	20	de-	de-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	217	21	pendency	pendency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	22	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	217	24	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	25	enriching	enrich	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	217	26	arcs	arc	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	217	27	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	28	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	217	29	in-	in-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	217	30	formation	formation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	217	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	218	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	218	2	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	218	3	component	component	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	218	4	uses	use	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	218	5	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	218	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	218	7	pre	pre	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	218	8	-	-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	218	9	computed	computed	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	218	10	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	218	11	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	218	12	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	218	13	may	may	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	218	14	diverge	diverge	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	218	15	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	218	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	218	17	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	218	18	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	218	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	219	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	219	2	contrast	contrast	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	219	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	219	4	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	219	5	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	219	6	pars-	pars-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	219	7	ing	ing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	219	8	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	219	9	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	219	10	exact	exact	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	219	11	whenever	whenever	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	219	12	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	219	13	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	219	14	decomposi-	decomposi-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	219	15	tion	tion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	219	16	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	219	17	converges	converge	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	219	18	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	220	1	Titov	Titov	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	220	2	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	220	3	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	220	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	221	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	221	2	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	221	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	221	4	augmented	augment	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	221	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	221	6	transition	transition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	221	7	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	221	8	based	base	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	221	9	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	221	10	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	221	11	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	221	12	operations	operation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	221	13	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	221	14	produce	produce	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	221	15	synchronous	synchronous	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	221	16	derivations	derivation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	221	17	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	221	18	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	221	19	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	221	20	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	221	21	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	221	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	222	1	Instead	instead	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	222	2	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	222	3	explicitly	explicitly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	222	4	representing	represent	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	222	5	seman-	seman-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	222	6	tic	tic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	222	7	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	222	8	together	together	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	222	9	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	222	10	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	222	11	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	222	12	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	222	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	222	14	they	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	222	15	induce	induce	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	222	16	latent	latent	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	222	17	representations	representation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	222	18	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	222	19	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	222	20	interactions	interaction	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	222	21	be-	be-	POS
work_7jtgrdqc4zeb7lul7vifzjglr4	222	22	tween	tween	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	222	23	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	222	24	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	222	25	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	222	26	layers	layer	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	222	27	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	223	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	223	2	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	223	3	works	work	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	223	4	mentioned	mention	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	223	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	223	6	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	223	7	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	223	8	no	no	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	223	9	con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	223	10	trol	trol	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	223	11	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	223	12	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	223	13	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	223	14	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	223	15	disallow	disallow	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	223	16	label-	label-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	223	17	ing	e	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	223	18	multiple	multiple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	223	19	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	223	20	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	223	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	223	22	same	same	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	223	23	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	223	24	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	223	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	224	1	Punyakanok	Punyakanok	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	224	2	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	224	3	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	224	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	225	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	225	2	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	225	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	225	4	first	first	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	225	5	introduced	introduce	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	225	6	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	225	7	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	225	8	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	225	9	explicitly	explicitly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	225	10	controls	control	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	225	11	these	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	225	12	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	225	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	225	14	as	as	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	225	15	well	well	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	225	16	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	225	17	other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	225	18	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	225	19	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	225	20	look	look	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	225	21	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	225	22	pairwise	pairwise	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	225	23	assignments	assignment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	225	24	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	225	25	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	225	26	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	225	27	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	225	28	model	model	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	225	29	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	226	1	They	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	226	2	solve	solve	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	226	3	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	226	4	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	226	5	general	general	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	226	6	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	226	7	purpose	purpose	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	226	8	Integer	Integer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	226	9	Linear	Linear	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	226	10	Programming	Programming	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	226	11	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	226	12	ILP	ILP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	226	13	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	226	14	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	226	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	227	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	227	2	similar	similar	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	227	3	spirit	spirit	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	227	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	227	5	Riedel	Riedel	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	227	6	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	227	7	McCallum	McCallum	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	227	8	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	227	9	2011	2011	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	227	10	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	227	11	presented	present	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	227	12	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	227	13	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	227	14	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	227	15	extracting	extract	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	227	16	structured	structured	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	227	17	events	event	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	227	18	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	227	19	controls	control	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	227	20	interactions	interaction	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	227	21	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	227	22	predicate-	predicate-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	227	23	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	227	24	assignments	assignment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	227	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	228	1	They	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	228	2	take	take	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	228	3	into	into	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	228	4	account	account	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	228	5	pair-	pair-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	228	6	wise	wise	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	228	7	assignments	assignment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	228	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	228	9	solve	solve	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	228	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	228	11	optimization	optimization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	228	12	prob-	prob-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	228	13	lem	lem	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	228	14	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	228	15	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	228	16	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	228	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	229	1	More	more	RBR
work_7jtgrdqc4zeb7lul7vifzjglr4	229	2	recently	recently	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	229	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	229	4	Das	Das	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	229	5	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	229	6	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	229	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	230	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	230	2	2012	2012	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	230	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	230	4	proposed	propose	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	230	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	230	6	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	230	7	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	230	8	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	230	9	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	230	10	deals	deal	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	230	11	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	230	12	several	several	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	230	13	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	230	14	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	230	15	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	230	16	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	230	17	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	230	18	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	230	19	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	230	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	231	1	Their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	231	2	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	231	3	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	231	4	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	231	5	alternative	alternative	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	231	6	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	231	7	general	general	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	231	8	ILP	ILP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	231	9	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	231	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	232	1	To	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	232	2	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	232	3	knowl-	knowl-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	232	4	edge	edge	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	232	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	232	6	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	232	7	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	232	8	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	232	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	232	10	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	232	11	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	232	12	frames	frame	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	232	13	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	232	14	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	232	15	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	232	16	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	232	17	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	232	18	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	232	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	232	20	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	232	21	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	232	22	simple	simple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	232	23	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	232	24	exact	exact	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	232	25	algo-	algo-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	232	26	rithms	rithms	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	232	27	exist	exist	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	232	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	233	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	233	2	should	should	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	233	3	note	note	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	233	4	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	5	these	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	233	6	works	work	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	233	7	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	8	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	9	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	233	10	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	11	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	233	12	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	13	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	14	con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	15	straints	straint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	233	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	233	17	but	but	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	233	18	none	none	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	20	them	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	233	21	predicts	predict	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	233	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	233	23	underlying	underlying	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	233	24	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	233	25	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	233	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	234	1	Our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	234	2	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	234	3	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	234	4	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	234	5	follows	follow	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	234	6	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	234	7	that	that	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	234	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	234	9	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	234	10	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	234	11	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	234	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	235	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	235	2	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	235	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	235	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	236	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	236	2	both	both	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	236	3	cases	case	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	236	4	two	two	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	236	5	separate	separate	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	236	6	pro-	pro-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	236	7	cesses	cesse	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	236	8	predict	predict	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	236	9	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	236	10	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	236	11	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	236	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	236	13	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	236	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	236	15	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	236	16	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	236	17	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	236	18	seeks	seek	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	236	19	agreement	agreement	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	236	20	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	236	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	236	22	level	level	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	236	23	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	236	24	individual	individual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	236	25	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	236	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	237	1	One	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	237	2	dif-	dif-	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	3	ference	ference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	4	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	237	5	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	6	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	237	7	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	237	8	process	process	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	9	predicts	predict	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	237	10	partial	partial	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	237	11	syntax	syntax	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	12	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	237	13	restricted	restrict	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	14	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	15	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	237	16	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	237	17	connecting	connect	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	237	18	pred-	pred-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	237	19	icates	icate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	237	20	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	237	21	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	237	22	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	237	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	237	24	while	while	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	25	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	26	their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	237	27	case	case	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	28	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	237	29	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	30	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	237	31	two	two	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	237	32	processes	process	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	237	33	predicts	predict	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	237	34	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	237	35	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	237	36	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	37	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	237	38	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	237	39	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	238	1	6	6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	238	2	Experiments	experiment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	238	3	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	238	4	present	present	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	238	5	experiments	experiment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	238	6	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	238	7	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	238	8	syntactic-	syntactic-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	238	9	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	238	10	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	238	11	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	238	12	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	238	13	CoNLL-2009	CoNLL-2009	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	238	14	Shared	Shared	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	238	15	Task	Task	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	238	16	English	English	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	238	17	benchmark	benchmark	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	238	18	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	238	19	Hajič	hajič	UH
work_7jtgrdqc4zeb7lul7vifzjglr4	238	20	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	238	21	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	238	22	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	238	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	238	24	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	238	25	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	238	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	239	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	2	consists	consist	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	239	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	239	5	usual	usual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	239	6	WSJ	WSJ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	7	training	training	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	8	/	/	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	239	9	development	development	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	10	/	/	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	239	11	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	12	sections	section	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	239	13	mapped	map	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	14	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	239	15	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	16	trees	tree	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	239	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	239	18	augmented	augment	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	19	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	20	se-	se-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	21	mantic	mantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	239	22	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	23	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	239	24	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	25	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	239	26	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	27	PropBank	PropBank	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	28	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	239	29	Palmer	Palmer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	30	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	31	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	32	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	239	34	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	239	35	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	239	36	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	239	37	NomBank	NomBank	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	38	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	239	39	Meyers	Meyers	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	40	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	41	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	42	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	239	43	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	239	44	2004	2004	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	239	45	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	239	46	also	also	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	239	47	represented	represent	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	48	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	239	49	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	239	50	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	240	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	240	2	also	also	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	240	3	con-	con-	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	240	4	tains	tain	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	240	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	240	6	PropBanked	PropBanked	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	240	7	portion	portion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	240	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	240	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	240	10	Brown	Brown	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	240	11	corpus	corpus	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	240	12	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	240	13	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	240	14	out	out	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	240	15	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	240	16	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	240	17	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	240	18	domain	domain	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	240	19	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	240	20	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	240	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	241	1	Our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	241	2	goal	goal	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	241	3	was	be	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	241	4	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	241	5	evaluate	evaluate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	241	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	241	7	contributions	contribution	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	241	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	241	9	pars-	pars-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	241	10	ing	ing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	241	11	algorithms	algorithm	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	241	12	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	241	13	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	241	14	following	following	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	241	15	configurations	configuration	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	241	16	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	241	17	Base	Base	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	241	18	Pipeline	Pipeline	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	241	19	Runs	run	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	241	20	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	241	21	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	241	22	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	241	23	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	241	24	then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	241	25	runs	run	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	241	26	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	241	27	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	241	28	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	241	29	constrained	constrain	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	241	30	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	241	31	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	241	32	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	241	33	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	241	34	best	well	RBS
work_7jtgrdqc4zeb7lul7vifzjglr4	241	35	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	241	36	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	241	37	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	242	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	242	3	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	242	4	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	242	5	only	only	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	242	6	enforces	enforce	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	242	7	con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	8	straint	straint	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	9	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	242	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	242	11	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	12	simply	simply	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	242	13	classifying	classify	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	242	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	242	15	candi-	candi-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	242	16	date	date	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	17	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	18	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	19	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	242	20	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	21	into	into	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	22	one	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	242	23	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	24	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	242	25	pos-	pos-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	242	26	sible	sible	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	242	27	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	242	28	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	242	29	or	or	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	242	30	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	242	31	NULL	NULL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	242	32	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	243	1	Pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	243	2	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	243	3	Assignment	Assignment	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	243	4	Runs	run	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	243	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	243	6	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	243	7	al-	al-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	243	8	gorithm	gorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	243	9	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	243	10	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	243	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	243	12	enforcing	enforce	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	243	13	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	243	14	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	243	15	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	243	16	cArg	cArg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	243	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	243	18	but	but	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	243	19	constrained	constrain	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	243	20	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	243	21	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	243	22	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	243	23	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	243	24	best	well	RBS
work_7jtgrdqc4zeb7lul7vifzjglr4	243	25	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	243	26	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	243	27	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	244	1	Forest	forest	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	244	2	Runs	run	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	244	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	244	4	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	244	5	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	244	6	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	244	7	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	244	8	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	244	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	244	10	large	large	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	244	11	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	244	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	244	13	precomputed	precomputed	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	244	14	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	244	15	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	244	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	244	17	de-	de-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	244	18	scribed	scribe	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	244	19	below	below	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	244	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	245	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	245	2	configuration	configuration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	245	3	corresponds	correspond	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	245	4	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	245	5	running	run	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	245	6	Dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	245	7	Decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	245	8	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	245	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	245	10	single	single	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	245	11	it-	it-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	245	12	eration	eration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	245	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	245	14	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	245	15	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	245	16	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	245	17	guaranteed	guarantee	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	245	18	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	245	19	predict	predict	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	245	20	consis-	consis-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	245	21	tent	tent	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	245	22	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	245	23	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	245	24	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	245	25	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	245	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	246	1	Dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	246	2	Decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	246	3	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	246	4	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	246	5	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	246	6	Runs	run	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	246	7	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	246	8	decomposi-	decomposi-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	246	9	tion	tion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	246	10	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	246	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	246	12	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	246	13	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	246	14	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	246	15	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	246	16	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	246	17	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	246	18	precomputed	precomputed	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	246	19	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	246	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	247	1	Syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	247	2	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	247	3	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	247	4	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	247	5	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	247	6	consistent	consistent	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	247	7	when	when	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	247	8	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	247	9	reaches	reach	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	247	10	con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	247	11	vergence	vergence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	247	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	248	1	All	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	248	2	four	four	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	248	3	systems	system	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	248	4	used	use	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	248	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	248	6	same	same	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	248	7	type	type	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	248	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	248	9	discrimina-	discrimina-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	248	10	tive	tive	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	248	11	scorers	scorer	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	248	12	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	248	13	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	248	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	249	1	Next	next	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	249	2	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	249	3	provide	provide	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	249	4	details	detail	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	249	5	about	about	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	249	6	these	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	249	7	systems	system	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	249	8	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	250	1	Then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	250	2	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	250	3	present	present	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	250	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	250	5	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	250	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	251	1	6.1	6.1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	251	2	Implementation	implementation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	3	Syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	251	4	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	5	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	251	6	used	use	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	251	7	two	two	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	251	8	discriminative	discriminative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	251	9	arc-	arc-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	10	factored	factor	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	11	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	251	12	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	13	labeled	labeled	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	251	14	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	15	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	16	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	251	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	251	18	225	225	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	251	19	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	251	20	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	251	21	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	22	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	251	24	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	251	25	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	251	26	second	second	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	251	27	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	251	28	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	29	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	30	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	31	grandchildren	grandchild	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	251	32	interactions	interaction	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	251	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	251	34	both	both	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	251	35	reimplementations	reimplementation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	251	36	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	37	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	251	38	parsers	parser	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	251	39	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	251	40	McDonald	McDonald	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	251	41	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	251	42	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	251	43	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	252	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	252	2	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	252	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	252	4	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	252	5	Car-	Car-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	252	6	reras	reras	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	252	7	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	252	8	2007	2007	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	252	9	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	252	10	respectively	respectively	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	252	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	253	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	253	2	both	both	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	253	3	cases	case	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	253	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	253	5	used	use	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	253	6	projective	projective	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	253	7	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	253	8	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	253	9	algorithms	algorithm	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	253	10	based	base	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	253	11	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	253	12	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	253	13	Eisner	eisner	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	253	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	253	15	2000).4	2000).4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	253	16	To	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	253	17	learn	learn	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	253	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	253	19	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	253	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	253	21	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	253	22	used	use	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	253	23	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	253	24	log	log	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	253	25	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	253	26	linear	linear	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	253	27	loss	loss	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	253	28	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	253	29	following	follow	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	253	30	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	253	31	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	253	32	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	253	33	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	254	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	254	2	2007	2007	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	254	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	254	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	254	5	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	254	6	trains	train	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	254	7	probabilistic	probabilistic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	254	8	discriminative	discriminative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	254	9	parsers	parser	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	254	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	255	1	At	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	255	2	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	255	3	time	time	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	255	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	255	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	255	6	used	use	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	255	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	255	8	probabilistic	probabilistic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	255	9	parsers	parser	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	255	10	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	255	11	compute	compute	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	255	12	marginal	marginal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	255	13	probabilities	probability	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	255	14	p(h	p(h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	255	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	255	16	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	255	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	255	18	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	255	19	|	|	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	255	20	x	x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	255	21	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	255	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	255	23	us-	us-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	255	24	ing	e	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	255	25	inside	inside	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	255	26	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	255	27	outside	outside	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	255	28	algorithms	algorithm	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	255	29	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	255	30	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	255	31	/	/	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	255	32	second	second	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	255	33	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	255	34	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	255	35	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	255	36	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	256	1	Hence	hence	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	256	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	256	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	256	4	either	either	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	256	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	256	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	256	7	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	256	8	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	256	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	256	10	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	256	11	always	always	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	256	12	obtain	obtain	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	256	13	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	256	14	table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	256	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	256	16	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	256	17	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	256	18	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	256	19	marginal	marginal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	256	20	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	256	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	256	22	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	256	23	one	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	256	24	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	256	25	per	per	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	256	26	labeled	labeled	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	256	27	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	256	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	257	1	Then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	257	2	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	257	3	run	run	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	257	4	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	257	5	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	257	6	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	257	7	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	257	8	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	257	9	these	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	257	10	marginals	marginal	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	257	11	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	257	12	ob-	ob-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	257	13	tain	tain	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	257	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	257	15	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	257	16	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	257	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	258	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	258	2	found	find	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	258	3	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	258	5	higher	high	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	258	6	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	258	7	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	8	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	9	performed	perform	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	258	10	equally	equally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	258	11	well	well	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	258	12	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	13	development	development	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	14	us-	us-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	258	15	ing	e	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	258	16	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	258	17	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	18	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	19	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	258	20	second	second	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	258	21	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	258	22	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	23	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	24	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	258	25	predict	predict	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	258	26	trees	tree	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	258	27	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	258	28	since	since	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	29	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	258	30	run	run	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	258	31	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	258	32	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	33	multiple	multiple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	258	34	times	time	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	258	35	within	within	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	36	Dual	Dual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	258	37	Decomposition	Decomposition	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	258	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	258	39	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	258	40	strategy	strategy	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	41	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	258	42	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	258	43	faster	fast	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	258	44	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	258	45	times	time	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	258	46	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	259	1	Precomputed	Precomputed	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	259	2	Paths	Paths	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	259	3	Both	both	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	259	4	Forest	Forest	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	259	5	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	259	6	Dual	Dual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	259	7	De-	De-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	259	8	composition	composition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	259	9	run	run	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	259	10	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	259	11	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	259	12	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	259	13	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	259	14	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	259	15	precomputed	precomputed	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	259	16	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	259	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	259	18	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	259	19	here	here	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	259	20	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	259	21	explain	explain	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	259	22	how	how	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	259	23	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	259	24	build	build	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	259	25	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	259	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	260	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	260	2	first	first	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	260	3	observed	observe	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	260	4	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	5	98.4	98.4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	260	6	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	7	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	260	9	correct	correct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	260	10	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	260	11	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	12	de-	de-	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	260	13	velopment	velopment	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	260	14	data	datum	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	260	15	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	260	16	either	either	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	260	17	direct	direct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	260	18	descendants	descendant	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	260	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	260	21	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	260	23	direct	direct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	260	24	descendants	descendant	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	260	25	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	26	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	260	27	ancestor	ancestor	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	28	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	29	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	260	30	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	260	32	or	or	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	260	33	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	260	34	ancestor	ancestor	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	35	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	36	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	260	37	predicate.5	predicate.5	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	260	38	All	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	260	39	meth-	meth-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	40	ods	od	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	260	41	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	260	42	test	test	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	260	43	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	260	44	restricted	restrict	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	45	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	46	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	260	47	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	260	48	scope	scope	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	260	49	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	261	1	To	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	261	2	generate	generate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	261	3	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	261	4	list	list	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	261	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	261	6	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	261	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	261	8	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	9	did	do	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	261	10	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	261	11	follows	follow	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	261	12	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	261	13	•	•	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	261	14	Calculate	calculate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	261	15	marginals	marginal	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	261	16	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	261	17	unlabeled	unlabeled	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	261	18	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	261	19	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	261	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	261	21	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	261	22	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	261	23	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	261	24	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	261	25	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	261	26	p(h	p(h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	261	28	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	29	|	|	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	30	x	x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	31	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	261	32	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	33	∑	∑	.
work_7jtgrdqc4zeb7lul7vifzjglr4	261	34	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	261	35	p(h	p(h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	36	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	261	37	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	261	39	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	40	|	|	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	41	x	x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	261	42	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	261	43	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	262	1	Note	note	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	262	2	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	262	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	262	4	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	262	5	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	262	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	262	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	262	8	probabilities	probability	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	262	9	p(h	p(h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	262	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	262	11	m|x	m|x	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	262	12	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	262	13	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	262	14	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	262	15	h	h	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	262	16	form	form	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	262	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	262	18	distri-	distri-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	262	19	bution	bution	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	262	20	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	262	21	i.e.	i.e.	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	263	1	they	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	263	2	sum	sum	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	263	3	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	263	4	one	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	263	5	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	263	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	264	1	Then	then	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	264	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	264	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	264	4	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	264	5	m	m	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	264	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	264	7	keep	keep	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	264	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	264	9	most	most	RBS
work_7jtgrdqc4zeb7lul7vifzjglr4	264	10	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	264	11	likely	likely	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	264	12	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	264	13	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	264	14	cover	cover	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	264	15	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	264	16	least	least	RBS
work_7jtgrdqc4zeb7lul7vifzjglr4	264	17	90	90	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	264	18	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	264	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	264	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	264	21	mass	mass	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	264	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	264	23	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	264	24	prune	prune	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	264	25	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	264	26	rest	rest	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	264	27	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	265	1	•	•	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	265	2	Starting	start	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	265	3	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	265	4	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	265	5	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	265	6	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	265	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	265	8	generate	generate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	265	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	265	10	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	265	11	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	265	12	taking	take	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	265	13	any	any	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	265	14	number	number	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	265	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	265	16	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	265	17	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	265	18	as-	as-	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	265	19	cend	cend	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	265	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	265	21	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	265	22	optionally	optionally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	265	23	adding	add	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	265	24	one	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	265	25	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	265	26	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	265	27	descends	descend	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	265	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	266	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	266	2	constrained	constrain	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	266	3	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	266	4	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	266	5	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	266	6	pro-	pro-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	266	7	jective	jective	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	266	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	266	9	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	266	10	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	266	11	have	have	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	266	12	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	266	13	maximum	maximum	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	266	14	number	number	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	266	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	266	16	6	6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	266	17	4Our	4our	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	266	18	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	266	19	allows	allow	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	266	20	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	266	21	use	use	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	266	22	non	non	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	266	23	-	-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	266	24	projective	projective	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	266	25	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	266	26	pars-	pars-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	266	27	ing	ing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	266	28	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	266	29	seamlessly	seamlessly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	266	30	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	267	1	5This	5this	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	267	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	267	3	specific	specific	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	267	4	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	267	5	CoNLL-2009	conll-2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	267	6	data	datum	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	267	7	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	267	8	English	English	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	267	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	268	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	268	2	gen-	gen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	268	3	eral	eral	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	268	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	268	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	268	6	other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	268	7	languages	language	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	268	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	268	9	coverage	coverage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	268	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	268	11	these	these	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	268	12	rules	rule	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	268	13	may	may	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	268	14	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	268	15	lower	low	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	268	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	269	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	269	2	leave	leave	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	269	3	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	269	4	question	question	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	269	5	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	269	6	future	future	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	269	7	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	269	8	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	270	1	ascendant	ascendant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	270	2	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	270	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	271	1	•	•	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	271	2	Label	Label	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	3	each	each	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	271	4	unlabeled	unlabeled	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	271	5	edge	edge	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	271	6	〈	〈	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	7	h	h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	271	9	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	10	〉	〉	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	11	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	271	12	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	271	13	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	271	14	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	271	15	l	l	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	271	16	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	271	17	argmaxl	argmaxl	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	271	18	p(h	p(h	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	271	20	m	m	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	271	22	l	l	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	23	|	|	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	24	x	x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	271	25	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	271	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	272	1	On	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	2	development	development	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	3	data	datum	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	272	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	272	5	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	272	6	procedure	procedure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	7	generated	generate	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	272	8	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	272	9	average	average	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	11	43.8	43.8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	272	12	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	272	13	per	per	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	14	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	15	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	16	cover	cover	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	272	17	86.2	86.2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	272	18	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	272	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	272	21	correct	correct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	272	22	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	272	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	273	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	2	contrast	contrast	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	273	4	enumerating	enumerate	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	273	5	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	273	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	273	8	single	single	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	273	9	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	273	10	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	273	11	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	12	covers	cover	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	273	13	79.4	79.4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	273	14	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	16	correct	correct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	273	17	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	273	18	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	19	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	273	20	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	273	21	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	273	22	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	23	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	273	25	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	273	26	82.2	82.2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	273	27	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	28	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	29	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	273	30	second-	second-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	31	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	32	parser.6	parser.6	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	273	33	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	273	34	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	35	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	273	36	used	use	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	273	37	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	273	38	discriminative	discriminative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	273	39	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	40	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	41	similar	similar	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	273	42	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	273	43	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	44	those	those	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	273	45	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	46	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	273	47	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	48	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	273	49	Johansson	Johansson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	273	50	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	273	51	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	273	52	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	273	53	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	274	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	274	2	addition	addition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	274	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	274	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	274	5	included	include	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	274	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	274	7	following	follow	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	274	8	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	274	9	•	•	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	274	10	Unigram	Unigram	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	274	11	/	/	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	274	12	bigram	bigram	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	274	13	/	/	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	274	14	trigram	trigram	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	274	15	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	274	16	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	274	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	275	1	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	275	2	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	275	3	n	n	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	275	4	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	275	5	grams	gram	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	275	6	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	275	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	275	8	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	275	9	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	275	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	275	11	patterns	pattern	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	275	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	275	13	words	word	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	275	14	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	275	15	POS	POS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	275	16	tags	tag	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	275	17	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	275	18	e.g.	e.g.	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	275	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	275	20	mary+loves+to	mary+loves+to	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	275	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	275	22	mary+VB+to	mary+VB+to	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	275	23	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	275	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	276	1	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	276	2	Voice	voice	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	276	3	features	feature	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	276	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	277	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	277	2	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	277	3	voice	voice	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	277	4	together	together	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	277	5	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	277	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	277	7	word	word	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	277	8	/	/	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	277	9	POS	pos	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	277	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	277	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	277	12	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	277	13	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	277	14	e.g.	e.g.	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	277	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	277	16	pas-	pas-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	277	17	sive+mary	sive+mary	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	277	18	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	277	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	278	1	•	•	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	278	2	Path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	278	3	continuity	continuity	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	278	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	279	1	Count	Count	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	279	2	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	279	3	non	non	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	279	4	-	-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	279	5	consecutive	consecutive	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	279	6	to-	to-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	279	7	kens	ken	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	279	8	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	279	9	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	279	10	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	279	11	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	279	12	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	279	13	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	279	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	280	1	To	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	280	2	train	train	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	280	3	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	280	4	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	280	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	280	6	used	use	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	280	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	280	8	averaged	averaged	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	280	9	per-	per-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	280	10	ceptron	ceptron	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	280	11	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	280	12	Collins	Collins	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	280	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	280	14	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	280	15	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	280	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	281	1	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	281	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	281	3	base	base	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	281	4	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	281	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	281	6	trained	train	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	281	7	standard	standard	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	281	8	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	281	9	classifiers	classifier	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	281	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	282	1	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	282	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	282	3	rest	rest	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	282	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	282	5	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	282	6	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	282	7	used	use	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	282	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	282	9	structured	structure	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	282	10	Perceptron	Perceptron	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	282	11	running	run	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	282	12	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	282	13	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	282	14	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	282	15	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	282	16	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	282	17	routine	routine	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	282	18	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	283	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	2	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	283	3	case	case	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	283	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	283	6	generated	generate	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	283	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	283	8	large	large	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	283	9	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	11	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	283	12	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	283	13	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	14	training	training	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	15	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	283	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	283	17	procedure	procedure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	18	described	describe	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	19	above	above	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	283	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	283	21	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	283	22	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	283	23	set	set	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	283	24	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	283	25	loss	loss	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	26	function	function	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	27	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	283	28	penalize	penalize	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	283	29	mistakes	mistake	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	283	30	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	31	predicting	predict	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	283	32	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	283	33	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	283	34	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	35	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	36	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	283	37	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	283	38	their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	283	39	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	283	40	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	283	41	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	284	1	Dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	284	2	Decomposition	Decomposition	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	3	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	4	added	add	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	284	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	284	6	parameter	parameter	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	284	7	β	β	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	284	8	weighting	weight	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	284	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	284	10	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	284	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	284	12	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	284	13	components	component	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	284	14	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	284	15	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	284	16	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	284	17	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	284	18	follows	follow	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	284	19	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	284	20	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	284	21	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	284	22	−β	−β	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	284	23	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	284	24	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	25	syn(x	syn(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	284	27	y	y	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	28	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	284	29	+	+	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	30	β	β	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	31	s	s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	32	srl(x	srl(x	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	284	34	z	z	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	284	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	284	36	π	π	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	284	37	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	284	38	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	285	1	As	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	285	2	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	285	3	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	285	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	285	5	used	use	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	285	6	normalized	normalized	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	285	7	marginal	marginal	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	285	8	probabilities	probability	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	285	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	285	10	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	285	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	285	12	either	either	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	285	13	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	285	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	285	15	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	285	16	or	or	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	285	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	285	18	higher	high	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	285	19	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	285	20	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	285	21	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	285	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	286	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	286	2	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	286	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	286	4	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	286	5	factors	factor	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	286	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	286	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	286	8	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	286	9	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	286	10	were	be	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	286	11	normalized	normalized	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	286	12	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	286	13	every	every	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	286	14	sen-	sen-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	286	15	tence	tence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	286	16	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	286	17	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	286	18	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	286	19	-1	-1	.
work_7jtgrdqc4zeb7lul7vifzjglr4	286	20	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	286	21	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	286	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	287	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	287	2	rest	rest	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	4	details	detail	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	287	5	6One	6one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	287	6	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	287	7	evaluate	evaluate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	287	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	287	9	maximum	maximum	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	10	recall	recall	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	11	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	12	correct	correct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	13	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	287	14	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	287	15	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	287	16	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	287	17	obtained	obtain	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	287	19	irrespective	irrespective	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	287	20	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	21	whether	whether	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	287	23	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	24	path	path	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	25	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	26	correct	correct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	27	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	287	28	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	29	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	287	30	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	31	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	32	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	287	33	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	287	34	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	35	98.3	98.3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	287	36	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	287	38	while	while	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	39	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	40	single	single	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	41	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	287	42	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	287	43	trees	tree	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	287	44	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	287	45	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	46	91.9	91.9	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	287	47	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	48	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	287	49	92.7	92.7	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	287	50	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	51	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	52	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	53	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	287	54	second	second	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	287	55	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	287	56	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	287	57	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	287	58	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	288	1	226	226	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	288	2	o	o	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	288	3	LAS	LAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	288	4	UAS	UAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	288	5	semp	semp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	288	6	semr	semr	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	288	7	semF1	semF1	''
work_7jtgrdqc4zeb7lul7vifzjglr4	288	8	sempp	sempp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	288	9	Pipeline	Pipeline	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	288	10	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	288	11	85.32	85.32	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	288	12	88.86	88.86	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	288	13	86.23	86.23	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	288	14	67.67	67.67	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	288	15	75.83	75.83	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	288	16	45.64	45.64	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	288	17	w.	w.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	288	18	Assig	Assig	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	288	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	289	1	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	2	85.32	85.32	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	3	88.86	88.86	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	4	84.08	84.08	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	5	71.82	71.82	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	6	77.47	77.47	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	7	51.17	51.17	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	8	Forest	Forest	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	289	9	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	289	10	-	-	:
work_7jtgrdqc4zeb7lul7vifzjglr4	289	11	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	289	12	80.67	80.67	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	13	73.60	73.60	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	14	76.97	76.97	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	15	51.33	51.33	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	16	Pipeline	Pipeline	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	289	17	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	18	87.77	87.77	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	19	90.96	90.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	20	87.07	87.07	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	21	68.65	68.65	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	22	76.77	76.77	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	23	47.07	47.07	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	289	24	w.	w.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	289	25	Assig	Assig	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	289	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	290	1	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	290	2	87.77	87.77	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	290	3	90.96	90.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	290	4	85.21	85.21	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	290	5	73.41	73.41	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	290	6	78.87	78.87	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	290	7	53.80	53.80	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	290	8	Table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	290	9	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	290	10	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	290	11	Results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	290	12	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	290	13	development	development	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	290	14	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	290	15	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	290	16	baseline	baseline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	290	17	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	290	18	as-	as-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	290	19	signment	signment	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	290	20	pipelines	pipeline	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	290	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	290	22	running	run	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	290	23	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	290	24	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	290	25	second	second	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	290	26	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	290	27	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	290	28	syn-	syn-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	290	29	tactic	tactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	290	30	parsers	parser	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	290	31	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	290	32	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	290	33	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	290	34	Forest	Forest	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	290	35	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	290	36	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	291	1	o	o	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	291	2	indicates	indicate	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	291	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	291	4	or-	or-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	291	5	der	der	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	291	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	291	7	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	291	8	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	291	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	292	1	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	292	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	292	3	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	292	4	were	be	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	292	5	implemented	implement	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	292	6	following	follow	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	292	7	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	292	8	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	292	9	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	292	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	293	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	293	2	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	293	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	293	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	293	5	including	include	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	293	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	293	7	strategy	strategy	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	293	8	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	293	9	decreasing	decrease	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	293	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	293	11	step	step	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	293	12	size	size	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	293	13	αt	αt	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	293	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	294	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	294	2	ran	run	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	294	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	294	4	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	294	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	294	6	up	up	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	294	7	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	294	8	500	500	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	294	9	it-	it-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	294	10	erations	eration	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	294	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	294	12	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	294	13	initial	initial	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	294	14	step	step	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	294	15	size	size	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	294	16	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	294	17	0.001	0.001	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	294	18	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	295	1	6.2	6.2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	295	2	Results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	295	3	To	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	295	4	evaluate	evaluate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	295	5	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	295	6	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	295	7	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	295	8	use	use	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	295	9	unla-	unla-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	295	10	beled	beled	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	295	11	attachment	attachment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	12	score	score	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	13	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	295	14	UAS	UAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	295	15	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	295	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	295	17	i.e.	i.e.	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	295	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	295	19	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	295	20	percentage	percentage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	21	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	22	words	word	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	295	23	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	24	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	295	25	correct	correct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	295	26	head	head	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	295	28	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	295	29	labeled	label	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	295	30	attach-	attach-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	295	31	ment	ment	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	295	32	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	295	33	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	295	34	LAS	LAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	295	35	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	295	36	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	295	37	i.e.	i.e.	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	295	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	295	39	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	295	40	percentage	percentage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	41	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	42	words	word	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	295	43	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	44	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	295	45	correct	correct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	295	46	head	head	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	47	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	295	48	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	295	49	label	label	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	295	50	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	296	1	Semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	296	2	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	3	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	296	4	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	5	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	296	6	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	296	7	evaluated	evaluate	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	8	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	9	pre-	pre-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	10	cision	cision	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	11	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	296	12	semp	semp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	13	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	296	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	296	15	recall	recall	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	16	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	296	17	semr	semr	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	18	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	296	19	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	296	20	F1	F1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	296	21	measure	measure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	22	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	296	23	semF1	semf1	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	296	24	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	296	25	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	26	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	296	27	level	level	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	28	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	29	labeled	label	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	296	30	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	296	31	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	296	32	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	297	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	2	ad-	ad-	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	297	3	dition	dition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	297	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	297	6	measure	measure	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	297	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	297	8	percentage	percentage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	10	perfectly	perfectly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	297	11	pre-	pre-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	297	12	dicted	dicte	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	13	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	14	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	297	15	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	297	16	sempp).7	sempp).7	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	297	17	Table	Table	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	297	18	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	297	19	shows	show	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	297	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	297	21	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	297	22	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	23	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	297	24	development	development	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	25	set	set	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	26	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	297	27	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	297	28	three	three	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	297	29	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	297	30	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	297	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	298	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	298	2	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	298	3	see	see	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	298	4	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	298	6	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	7	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	298	8	running	run	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	298	9	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	10	improve	improve	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	298	11	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	12	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	298	13	baseline	baseline	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	298	14	pipelines	pipeline	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	298	15	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	16	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	298	17	F1	f1	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	18	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	19	about	about	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	298	20	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	298	21	points	point	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	298	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	298	23	due	due	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	24	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	25	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	298	26	application	application	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	27	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	28	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	298	29	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	298	30	constraint	constraint	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	298	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	299	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	299	2	Forest	Forest	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	299	3	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	299	4	also	also	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	299	5	shows	show	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	299	6	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	299	7	improvement	improvement	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	299	8	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	299	9	recall	recall	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	299	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	299	11	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	299	12	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	299	13	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	299	14	respect	respect	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	299	15	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	299	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	299	17	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	299	18	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	299	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	300	1	Presumably	presumably	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	300	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	300	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	300	4	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	6	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	300	7	available	available	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	300	8	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	300	10	Forest	Forest	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	300	11	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	12	allows	allow	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	300	13	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	300	14	recognize	recognize	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	300	15	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	300	16	higher	high	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	300	17	num-	num-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	18	ber	ber	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	20	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	300	21	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	22	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	300	23	expense	expense	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	24	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	25	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	300	26	lower	low	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	300	27	preci-	preci-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	300	28	sion	sion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	300	29	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	301	1	Regarding	regard	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	301	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	301	3	percentage	percentage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	5	perfect	perfect	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	301	6	predicate-	predicate-	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	301	7	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	8	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	301	9	there	there	EX
work_7jtgrdqc4zeb7lul7vifzjglr4	301	10	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	301	11	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	301	12	remarkable	remarkable	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	301	13	improve-	improve-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	14	ment	ment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	15	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	301	17	systems	system	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	301	18	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	301	19	apply	apply	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	301	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	301	21	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	301	22	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	23	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	24	con-	con-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	25	7Our	7our	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	301	26	evaluation	evaluation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	27	metrics	metric	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	301	28	differ	differ	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	301	29	slightly	slightly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	301	30	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	31	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	301	32	official	official	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	33	met-	met-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	34	ric	ric	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	301	35	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	301	36	CoNLL-2009	CoNLL-2009	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	301	37	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	302	1	That	that	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	302	2	metric	metric	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	302	3	considers	consider	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	302	4	predicate	predicate	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	302	5	senses	sense	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	302	6	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	302	7	special	special	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	302	8	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	302	9	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	302	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	302	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	302	12	thus	thus	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	302	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	302	14	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	302	15	includes	include	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	302	16	them	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	302	17	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	302	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	302	19	calculation	calculation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	302	20	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	302	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	302	22	evaluation	evaluation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	302	23	metrics	metric	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	302	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	303	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	303	2	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	303	3	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	303	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	303	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	303	6	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	303	7	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	303	8	addressing	address	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	303	9	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	303	10	sense	sense	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	303	11	disambiguation	disambiguation	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	303	12	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	303	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	303	14	conse-	conse-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	303	15	quently	quently	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	303	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	303	17	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	303	18	ignore	ignore	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	303	19	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	303	20	senses	sense	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	303	21	when	when	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	303	22	presenting	present	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	303	23	evaluation	evaluation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	303	24	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	303	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	304	1	When	when	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	304	2	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	304	3	report	report	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	304	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	304	5	performance	performance	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	304	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	304	7	CoNLL	CoNLL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	304	8	systems	systems	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	304	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	304	10	their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	304	11	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	304	12	will	will	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	304	13	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	304	14	noticeably	noticeably	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	304	15	lower	low	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	304	16	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	304	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	304	18	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	304	19	reported	report	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	304	20	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	304	21	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	304	22	shared	shared	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	304	23	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	304	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	305	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	305	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	305	3	because	because	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	305	4	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	305	5	disambiguation	disambiguation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	305	6	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	305	7	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	305	8	reasonably	reasonably	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	305	9	simple	simple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	305	10	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	305	11	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	305	12	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	305	13	very	very	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	305	14	high	high	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	305	15	baseline	baseline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	305	16	around	around	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	305	17	90	90	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	305	18	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	305	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	306	1	o	o	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	306	2	β	β	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	306	3	LAS	LAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	306	4	UAS	UAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	306	5	semp	semp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	6	semr	semr	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	7	semF1	semF1	''
work_7jtgrdqc4zeb7lul7vifzjglr4	306	8	sempp	sempp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	9	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	10	conv	conv	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	11	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	12	0.1	0.1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	13	85.32	85.32	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	14	88.86	88.86	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	15	84.09	84.09	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	16	71.84	71.84	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	17	77.48	77.48	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	18	51.77	51.77	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	19	100	100	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	20	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	21	0.4	0.4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	22	85.36	85.36	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	23	88.91	88.91	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	24	84.07	84.07	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	25	71.94	71.94	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	26	77.53	77.53	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	27	51.85	51.85	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	28	100	100	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	29	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	30	0.5	0.5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	31	85.38	85.38	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	32	88.93	88.93	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	33	84.08	84.08	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	34	72.03	72.03	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	35	77.59	77.59	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	36	51.96	51.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	37	100	100	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	38	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	39	0.6	0.6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	40	85.41	85.41	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	41	88.95	88.95	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	42	84.05	84.05	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	43	72.19	72.19	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	44	77.67	77.67	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	45	52.03	52.03	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	46	99.8	99.8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	47	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	48	0.7	0.7	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	49	85.44	85.44	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	50	89.00	89.00	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	51	84.10	84.10	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	52	72.42	72.42	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	53	77.82	77.82	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	54	52.24	52.24	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	55	99.7	99.7	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	56	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	57	0.8	0.8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	58	85.48	85.48	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	59	89.02	89.02	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	60	83.99	83.99	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	61	72.69	72.69	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	62	77.94	77.94	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	63	52.57	52.57	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	64	99.5	99.5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	65	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	66	0.9	0.9	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	67	85.39	85.39	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	68	88.93	88.93	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	69	83.68	83.68	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	70	72.82	72.82	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	71	77.88	77.88	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	72	52.49	52.49	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	73	99.8	99.8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	74	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	75	0.1	0.1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	76	87.78	87.78	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	77	90.96	90.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	78	85.20	85.20	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	79	73.11	73.11	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	80	78.69	78.69	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	81	53.74	53.74	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	82	100	100	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	83	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	84	0.4	0.4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	85	87.78	87.78	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	86	90.96	90.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	87	85.21	85.21	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	88	73.12	73.12	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	89	78.70	78.70	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	90	53.74	53.74	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	91	100	100	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	92	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	93	0.5	0.5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	94	87.78	87.78	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	95	90.96	90.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	96	85.19	85.19	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	97	73.12	73.12	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	98	78.70	78.70	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	99	53.72	53.72	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	100	100	100	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	101	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	102	0.6	0.6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	103	87.78	87.78	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	104	90.96	90.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	105	85.20	85.20	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	106	73.13	73.13	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	107	78.70	78.70	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	108	53.72	53.72	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	109	99.9	99.9	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	110	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	111	0.7	0.7	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	112	87.78	87.78	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	113	90.96	90.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	114	85.19	85.19	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	115	73.13	73.13	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	116	78.70	78.70	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	117	53.72	53.72	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	118	99.8	99.8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	119	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	120	0.8	0.8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	121	87.80	87.80	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	122	90.98	90.98	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	123	85.20	85.20	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	124	73.18	73.18	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	125	78.74	78.74	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	126	53.77	53.77	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	127	99.8	99.8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	128	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	129	0.9	0.9	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	130	87.84	87.84	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	131	91.02	91.02	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	132	85.20	85.20	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	133	73.23	73.23	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	134	78.76	78.76	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	135	53.82	53.82	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	136	100	100	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	137	Table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	138	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	306	139	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	306	140	Results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	306	141	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	142	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	306	143	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	306	144	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	145	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	146	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	147	development	development	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	148	data	datum	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	306	149	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	306	150	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	151	different	different	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	306	152	values	value	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	306	153	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	154	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	306	155	β	β	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	306	156	parame-	parame-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	157	ter	ter	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	306	158	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	307	1	o	o	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	307	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	307	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	307	4	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	307	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	307	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	307	7	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	307	8	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	307	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	308	1	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	308	2	conv	conv	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	308	3	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	308	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	308	5	percentage	percentage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	308	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	308	7	examples	example	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	308	8	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	308	9	converged	converge	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	308	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	309	1	straints	straint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	309	2	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	309	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	309	4	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	309	5	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	309	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	310	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	310	2	believe	believe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	310	3	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	310	5	cRole	cRole	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	310	6	constraint	constraint	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	7	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	310	8	ensures	ensure	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	310	9	no	no	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	310	10	repeated	repeat	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	11	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	310	12	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	13	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	310	14	given	give	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	15	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	16	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	310	17	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	310	18	key	key	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	310	19	factor	factor	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	20	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	310	21	predict	predict	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	310	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	310	23	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	310	24	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	25	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	26	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	310	27	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	28	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	310	29	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	310	30	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	311	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	311	2	Forest	Forest	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	311	3	configuration	configuration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	311	4	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	311	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	311	6	starting	starting	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	311	7	point	point	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	311	8	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	311	9	run	run	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	311	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	311	11	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	311	12	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	311	13	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	311	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	312	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	312	2	ran	run	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	312	3	ex-	ex-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	312	4	periments	periment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	312	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	312	6	various	various	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	312	7	values	value	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	312	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	312	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	312	10	β	β	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	312	11	parameter	parameter	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	312	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	313	1	Ta-	Ta-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	313	2	ble	ble	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	313	3	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	313	4	shows	show	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	313	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	313	6	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	313	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	314	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	314	2	see	see	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	314	3	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	314	4	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	314	5	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	314	6	increase	increase	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	314	7	β	β	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	314	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	314	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	314	10	SRL	SRL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	314	11	component	component	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	314	12	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	314	13	more	more	RBR
work_7jtgrdqc4zeb7lul7vifzjglr4	314	14	relative	relative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	314	15	weight	weight	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	314	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	314	17	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	314	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	314	19	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	314	20	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	314	21	changes	change	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	314	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	315	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	315	2	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	315	3	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	315	4	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	315	5	always	always	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	315	6	able	able	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	315	7	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	315	8	improve	improve	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	315	9	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	315	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	315	11	Forest	Forest	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	315	12	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	315	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	315	14	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	315	15	find	find	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	315	16	convergence	convergence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	315	17	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	315	18	more	more	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	315	19	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	315	20	99.5	99.5	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	315	21	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	315	22	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	315	23	sentences	sentence	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	315	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	316	1	Compared	compare	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	2	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	316	4	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	5	running	running	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	6	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	316	8	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	316	9	improves	improve	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	316	10	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	316	11	F1	F1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	316	12	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	13	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	316	14	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	316	15	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	16	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	316	18	but	but	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	316	19	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	316	20	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	21	higher	high	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	316	22	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	316	23	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	24	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	316	26	suggesting	suggest	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	316	27	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	28	2nd	2nd	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	316	29	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	30	predictions	prediction	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	316	31	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	316	32	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	316	33	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	316	34	quite	quite	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	316	35	accurate	accurate	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	316	36	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	317	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	317	2	also	also	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	317	3	observe	observe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	317	4	slight	slight	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	317	5	benefits	benefit	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	317	6	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	317	7	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	317	8	accuracy	accuracy	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	317	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	318	1	Table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	2	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	318	3	presents	present	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	318	4	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	318	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	6	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	318	7	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	8	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	318	10	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	11	sets	set	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	318	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	318	13	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	318	14	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	318	15	run	run	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	318	16	Pipeline	Pipeline	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	318	17	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	18	Assignment	Assignment	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	318	19	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	318	20	Dual	Dual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	318	21	Decomposition	Decomposition	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	318	22	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	23	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	318	24	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	318	25	configura-	configura-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	318	26	tion	tion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	27	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	318	28	β	β	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	318	29	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	318	30	0.8/0.9	0.8/0.9	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	318	31	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	32	1st/2nd	1st/2nd	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	318	33	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	34	syntax	syntax	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	318	35	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	318	36	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	319	1	For	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	319	2	comparison	comparison	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	319	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	319	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	319	5	table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	319	6	also	also	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	319	7	reports	report	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	319	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	319	9	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	319	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	319	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	319	12	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	319	13	CoNLL–2009	conll–2009	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	319	14	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	319	15	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	319	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	319	17	Merlo09	Merlo09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	319	18	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	319	19	Ges-	Ges-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	319	20	mundo	mundo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	319	21	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	319	22	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	319	23	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	319	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	319	25	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	319	26	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	319	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	319	28	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	319	29	proved	prove	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	319	30	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	319	31	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	319	32	very	very	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	319	33	com-	com-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	319	34	petitive	petitive	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	319	35	ranking	rank	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	319	36	third	third	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	319	37	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	319	38	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	319	39	closed	closed	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	319	40	challenge	challenge	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	319	41	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	320	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	2	also	also	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	320	3	include	include	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	4	Lluı́s09	Lluı́s09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	5	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	320	6	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	7	et	et	FW
work_7jtgrdqc4zeb7lul7vifzjglr4	320	8	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	9	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	320	11	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	320	12	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	320	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	320	14	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	320	15	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	320	16	an-	an-	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	320	17	other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	320	18	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	320	19	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	320	20	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	320	21	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	320	22	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	23	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	24	CoNLL	CoNLL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	25	–	–	:
work_7jtgrdqc4zeb7lul7vifzjglr4	320	26	2009.8	2009.8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	320	27	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	28	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	320	29	WSJ	WSJ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	30	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	31	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	32	obtains	obtain	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	320	33	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	320	34	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	320	35	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	320	36	accuracies	accuracy	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	320	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	320	38	while	while	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	39	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	320	40	Pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	41	obtains	obtain	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	320	42	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	320	43	best	good	JJS
work_7jtgrdqc4zeb7lul7vifzjglr4	320	44	se-	se-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	45	8Another	8Another	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	46	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	47	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	320	48	compare	compare	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	320	49	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	50	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	320	51	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	320	52	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	320	53	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	54	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	320	55	Jo-	Jo-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	56	hansson	hansson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	320	57	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	320	58	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	320	59	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	320	60	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	321	1	Unfortunately	unfortunately	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	321	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	321	3	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	321	4	direct	direct	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	5	comparison	comparison	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	6	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	7	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	321	8	possi-	possi-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	9	ble	ble	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	10	because	because	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	11	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	12	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	13	evaluated	evaluate	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	14	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	15	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	321	16	CoNLL-2008	CoNLL-2008	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	17	datasets	dataset	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	321	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	321	19	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	321	20	227	227	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	21	WSJ	WSJ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	22	LAS	LAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	23	UAS	UAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	24	semp	semp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	25	semr	semr	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	26	semF1	semF1	''
work_7jtgrdqc4zeb7lul7vifzjglr4	321	27	sempp	sempp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	28	Lluı́s09	Lluı́s09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	29	87.48	87.48	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	30	89.91	89.91	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	31	73.87	73.87	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	32	67.40	67.40	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	33	70.49	70.49	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	34	39.68	39.68	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	35	Merlo09	Merlo09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	36	88.79	88.79	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	37	91.26	91.26	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	38	81.00	81.00	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	39	76.45	76.45	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	40	78.66	78.66	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	41	54.80	54.80	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	42	Pipe	Pipe	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	43	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	321	44	Assig	Assig	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	45	1st	1st	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	46	86.85	86.85	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	47	89.68	89.68	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	48	85.12	85.12	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	49	73.78	73.78	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	50	79.05	79.05	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	51	54.12	54.12	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	52	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	53	1st	1st	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	54	87.04	87.04	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	55	89.89	89.89	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	56	85.03	85.03	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	57	74.56	74.56	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	58	79.45	79.45	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	59	54.92	54.92	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	60	Pipe	Pipe	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	61	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	321	62	Assig	Assig	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	63	2nd	2nd	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	64	89.19	89.19	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	65	91.62	91.62	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	66	86.11	86.11	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	67	75.16	75.16	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	68	80.26	80.26	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	69	55.96	55.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	70	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	71	2nd	2nd	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	72	89.21	89.21	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	73	91.64	91.64	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	74	86.01	86.01	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	75	74.84	74.84	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	76	80.04	80.04	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	77	55.73	55.73	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	78	Brown	Brown	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	79	LAS	LAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	80	UAS	UAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	81	semp	semp	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	82	semr	semr	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	83	semF1	semF1	''
work_7jtgrdqc4zeb7lul7vifzjglr4	321	84	sempp	sempp	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	85	Lluı́s09	Lluı́s09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	86	80.92	80.92	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	87	85.96	85.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	88	62.29	62.29	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	89	59.22	59.22	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	90	60.71	60.71	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	91	29.79	29.79	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	92	Merlo09	Merlo09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	93	80.84	80.84	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	94	86.32	86.32	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	95	68.97	68.97	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	96	63.06	63.06	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	97	65.89	65.89	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	98	38.92	38.92	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	99	Pipe	Pipe	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	100	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	321	101	Assig	Assig	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	102	1st	1st	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	103	80.96	80.96	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	104	86.58	86.58	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	105	72.91	72.91	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	106	60.16	60.16	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	107	65.93	65.93	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	108	38.44	38.44	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	109	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	110	1st	1st	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	111	81.18	81.18	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	112	86.86	86.86	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	113	72.53	72.53	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	114	60.76	60.76	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	115	66.12	66.12	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	116	38.13	38.13	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	117	Pipe	Pipe	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	118	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	321	119	Assig	Assig	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	120	2nd	2nd	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	121	82.56	82.56	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	122	87.98	87.98	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	123	73.94	73.94	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	124	61.63	61.63	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	125	67.23	67.23	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	126	38.99	38.99	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	127	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	128	2nd	2nd	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	129	82.61	82.61	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	130	88.04	88.04	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	131	74.12	74.12	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	132	61.59	61.59	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	133	67.28	67.28	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	134	38.92	38.92	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	135	Table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	136	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	321	137	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	321	138	Comparative	comparative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	139	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	321	140	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	141	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	321	142	CoNLL–2009	conll–2009	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	143	En-	en-	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	321	144	glish	glish	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	145	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	146	sets	set	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	321	147	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	321	148	namely	namely	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	321	149	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	321	150	WSJ	WSJ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	151	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	152	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	321	153	top	top	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	154	table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	155	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	321	156	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	321	157	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	321	158	out	out	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	159	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	160	domain	domain	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	161	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	162	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	163	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	321	164	Brown	Brown	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	321	165	corpus	corpus	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	166	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	321	167	bottom	bottom	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	321	168	table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	321	169	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	321	170	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	322	1	mantic	mantic	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	322	2	F1	F1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	322	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	323	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	323	2	bottom	bottom	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	323	3	part	part	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	323	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	323	5	Table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	323	6	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	323	7	presents	present	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	323	8	re-	re-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	323	9	sults	sult	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	323	10	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	323	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	323	12	out	out	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	323	13	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	323	14	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	323	15	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	323	16	domain	domain	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	323	17	Brown	brown	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	323	18	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	323	19	corpus	corpus	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	323	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	324	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	324	2	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	324	3	case	case	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	324	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	324	5	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	324	6	obtains	obtain	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	324	7	slightly	slightly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	324	8	better	well	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	324	9	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	324	10	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	324	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	324	12	rest	rest	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	324	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	324	14	both	both	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	324	15	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	324	16	terms	term	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	324	17	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	324	18	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	324	19	accuracy	accuracy	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	324	20	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	324	21	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	324	22	F1	f1	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	324	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	325	1	Table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	325	2	4	4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	325	3	shows	show	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	325	4	statistical	statistical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	325	5	significance	significance	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	325	6	tests	test	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	325	7	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	325	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	325	9	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	325	10	LAS	LAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	325	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	325	12	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	325	13	F1	f1	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	325	14	scores	score	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	325	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	325	16	Table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	325	17	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	325	18	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	326	1	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	326	2	have	have	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	326	3	applied	apply	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	326	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	326	5	sign	sign	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	326	6	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	326	7	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	326	8	Wackerly	Wackerly	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	326	9	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	326	10	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	326	11	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	326	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	326	13	2007	2007	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	326	14	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	326	15	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	326	16	approximate	approximate	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	326	17	randomization	randomization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	326	18	tests	test	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	326	19	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	326	20	Yeh	Yeh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	326	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	326	22	2000	2000	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	326	23	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	326	24	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	326	25	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	326	26	pairs	pair	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	326	27	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	326	28	systems	system	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	326	29	outputs	output	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	326	30	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	327	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	327	2	differences	difference	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	327	3	be-	be-	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	327	4	tween	tween	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	327	5	systems	system	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	327	6	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	327	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	327	8	WSJ	WSJ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	327	9	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	327	10	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	327	11	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	327	12	considered	consider	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	327	13	significant	significant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	327	14	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	327	15	almost	almost	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	327	16	all	all	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	327	17	cases	case	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	327	18	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	327	19	p	p	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	327	20	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	327	21	0.05	0.05	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	327	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	328	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	328	2	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	328	3	Brown	brown	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	328	4	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	328	5	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	328	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	328	7	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	328	8	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	328	9	more	more	RBR
work_7jtgrdqc4zeb7lul7vifzjglr4	328	10	unstable	unstable	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	328	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	328	12	dif-	dif-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	328	13	ferences	ference	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	328	14	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	328	15	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	328	16	significant	significant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	328	17	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	328	18	general	general	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	328	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	328	20	probably	probably	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	328	21	be-	be-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	328	22	cause	cause	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	328	23	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	328	24	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	328	25	relatively	relatively	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	328	26	small	small	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	328	27	size	size	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	328	28	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	328	29	that	that	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	328	30	test	test	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	328	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	329	1	Regarding	regard	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	329	2	running	run	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	329	3	times	time	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	329	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	329	5	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	329	6	implementation	implementation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	7	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	329	9	baseline	baseline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	10	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	11	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	12	2nd	2nd	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	329	13	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	14	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	15	parses	parse	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	329	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	329	17	development	development	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	18	set	set	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	19	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	329	20	1,334	1,334	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	329	21	sentences	sentence	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	329	22	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	329	23	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	24	less	less	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	329	25	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	329	26	7	7	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	329	27	minutes	minute	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	329	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	330	1	Running	run	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	330	2	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	3	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	330	5	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	6	in-	in-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	330	7	creases	crease	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	330	8	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	330	9	time	time	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	10	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	11	∼8	∼8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	330	12	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	13	due	due	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	14	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	15	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	330	16	overhead	overhead	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	17	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	18	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	330	19	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	20	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	330	21	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	331	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	331	2	Forest	Forest	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	331	3	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	331	5	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	6	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	331	7	average	average	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	9	61.3	61.3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	331	10	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	331	11	per	per	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	12	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	331	14	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	331	15	∼13	∼13	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	331	16	%	%	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	17	slower	slow	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	331	18	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	19	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	331	20	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	21	due	due	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	331	22	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	23	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	331	24	exploration	exploration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	25	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	26	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	331	27	space	space	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	28	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	331	29	precomputed	precomputed	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	331	30	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	331	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	332	1	Finally	finally	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	332	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	332	3	Dual	Dual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	332	4	Decom-	Decom-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	332	5	position	position	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	332	6	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	332	7	2nd	2nd	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	332	8	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	332	9	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	332	10	converges	converge	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	332	11	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	332	12	36.6	36.6	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	332	13	iterations	iteration	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	332	14	per	per	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	332	15	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	332	16	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	332	17	average	average	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	332	18	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	333	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	333	2	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	333	3	itera-	itera-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	333	4	tion	tion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	6	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	333	7	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	333	8	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	333	9	perform	perform	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	333	10	roughly	roughly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	333	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	333	12	same	same	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	333	13	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	14	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	15	Forest	Forest	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	333	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	333	17	while	while	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	18	subsequent	subsequent	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	333	19	iterations	iteration	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	333	20	only	only	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	333	21	need	need	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	333	22	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	333	23	re	re	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	333	24	-	-	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	333	25	parse	parse	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	333	26	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	333	27	sentence	sentence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	28	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	29	respect	respect	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	30	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	31	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	333	32	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	333	33	up-	up-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	333	34	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	333	35	slightly	slightly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	333	36	different	different	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	333	37	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	334	1	However	however	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	334	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	334	3	note	note	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	334	4	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	334	5	Merlo09	Merlo09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	334	6	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	334	7	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	334	8	applica-	applica-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	334	9	tion	tion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	334	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	334	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	334	12	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	334	13	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	334	14	Titov	Titov	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	334	15	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	334	16	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	334	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	335	1	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	335	2	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	335	3	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	335	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	336	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	336	2	that	that	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	336	3	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	336	4	authors	author	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	336	5	report	report	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	336	6	results	result	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	336	7	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	336	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	336	9	CoNLL-2008	CoNLL-2008	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	336	10	datasets	dataset	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	336	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	336	12	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	336	13	they	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	336	14	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	336	15	com-	com-	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	336	16	parable	parable	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	336	17	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	336	18	Johansson	Johansson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	336	19	’s	’	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	336	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	1	WSJ	WSJ	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	2	Brown	Brown	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	3	ME	ME	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	4	PA1	PA1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	5	DD1	DD1	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	6	PA2	PA2	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	7	DD2	DD2	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	8	ME	ME	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	9	PA1	PA1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	10	DD1	DD1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	11	PA2	PA2	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	12	DD2	DD2	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	13	LL	LL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	14	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	15	•	•	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	337	16	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	17	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	18	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	19	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	20	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	21	�	�	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	337	22	•	•	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	23	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	24	�	�	,
work_7jtgrdqc4zeb7lul7vifzjglr4	337	25	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	26	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	27	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	28	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	29	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	30	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	31	�	�	,
work_7jtgrdqc4zeb7lul7vifzjglr4	337	32	�	�	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	337	33	�	�	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	34	�	�	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	35	�	�	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	36	�	�	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	37	�	�	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	38	�	�	,
work_7jtgrdqc4zeb7lul7vifzjglr4	337	39	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	40	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	41	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	42	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	43	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	44	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	45	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	46	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	47	ME	ME	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	48	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	49	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	50	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	51	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	52	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	53	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	54	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	55	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	56	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	57	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	58	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	59	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	60	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	61	�	�	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	337	62	•	•	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	63	•	•	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	64	PA1	PA1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	65	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	66	•	•	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	67	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	68	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	69	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	70	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	71	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	72	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	73	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	74	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	75	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	76	�	�	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	337	77	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	78	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	79	•	•	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	80	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	81	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	82	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	83	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	84	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	85	DD1	DD1	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	86	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	87	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	88	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	89	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	90	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	91	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	92	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	93	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	94	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	95	•	•	:
work_7jtgrdqc4zeb7lul7vifzjglr4	337	96	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	97	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	98	•	•	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	99	�	�	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	100	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	101	PA2	PA2	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	102	•	•	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	103	�	�	.
work_7jtgrdqc4zeb7lul7vifzjglr4	337	104	�	�	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	105	Table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	106	4	4	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	107	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	337	108	Statistical	statistical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	337	109	tests	test	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	337	110	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	111	significance	significance	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	112	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	113	LAS	LAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	337	114	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	337	115	semF1	semF1	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	337	116	differences	difference	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	337	117	between	between	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	118	pairs	pair	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	337	119	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	120	systems	system	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	337	121	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	122	Table	table	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	337	123	3	3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	337	124	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	338	1	◦	◦	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	338	2	/•	/•	.
work_7jtgrdqc4zeb7lul7vifzjglr4	338	3	=	=	NFP
work_7jtgrdqc4zeb7lul7vifzjglr4	338	4	LAS	LAS	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	338	5	difference	difference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	338	6	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	338	7	significant	significant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	338	8	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	338	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	338	10	sign/	sign/	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	338	11	approxi-	approxi-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	338	12	mate	mate	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	338	13	randomization	randomization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	338	14	tests	test	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	338	15	at	at	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	338	16	0.05	0.05	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	338	17	level	level	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	338	18	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	339	1	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	339	2	/	/	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	339	3	�	�	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	339	4	=	=	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	339	5	same	same	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	339	6	mean-	mean-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	339	7	ing	ing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	339	8	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	339	9	semF1	semf1	ADD
work_7jtgrdqc4zeb7lul7vifzjglr4	339	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	340	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	340	2	legend	legend	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	340	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	340	4	systems	system	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	340	5	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	340	6	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	340	7	LL	LL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	340	8	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	340	9	Lluı́s09	Lluı́s09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	340	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	340	11	ME	me	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	340	12	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	340	13	Merlo09	Merlo09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	340	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	340	15	PA1/2	pa1/2	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	340	16	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	340	17	Pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	340	18	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	340	19	Assignment	Assignment	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	340	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	340	21	1st/2nd	1st/2nd	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	340	22	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	340	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	340	24	DD1/2	dd1/2	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	340	25	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	340	26	Dual	Dual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	340	27	Decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	340	28	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	340	29	1st/2nd	1st/2nd	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	340	30	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	340	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	341	1	dates	date	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	341	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	341	3	which	which	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	341	4	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	341	5	extremely	extremely	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	341	6	sparse	sparse	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	341	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	342	1	Our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	342	2	current	current	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	342	3	im-	im-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	342	4	plementation	plementation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	5	did	do	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	342	6	not	not	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	342	7	take	take	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	342	8	advantage	advantage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	342	11	sparsity	sparsity	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	12	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	13	updates	update	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	342	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	342	15	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	342	16	overall	overall	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	342	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	342	18	DD	DD	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	342	19	was	be	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	342	20	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	21	average	average	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	342	22	13	13	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	342	23	times	time	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	342	24	slower	slow	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	342	25	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	26	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	342	27	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	28	running	run	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	342	29	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	30	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	342	31	15	15	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	342	32	times	time	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	342	33	slower	slow	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	342	34	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	35	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	342	36	baseline	baseline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	37	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	342	38	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	343	1	7	7	LS
work_7jtgrdqc4zeb7lul7vifzjglr4	343	2	Conclusion	Conclusion	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	343	3	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	343	4	have	have	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	343	5	introduced	introduce	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	343	6	efficient	efficient	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	343	7	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	343	8	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	343	9	parse	parse	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	343	10	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	343	11	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	343	12	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	343	13	augmented	augment	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	343	14	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	343	15	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	343	16	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	343	17	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	343	18	relations	relation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	343	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	343	20	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	343	21	two	two	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	343	22	key	key	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	343	23	ideas	idea	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	343	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	344	1	One	one	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	344	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	344	3	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	344	4	predict	predict	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	344	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	344	6	local	local	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	344	7	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	344	8	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	344	9	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	344	10	links	link	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	344	11	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	344	12	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	344	13	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	344	14	its	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	344	15	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	344	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	344	17	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	344	18	seek	seek	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	344	19	agree-	agree-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	344	20	ment	ment	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	344	21	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	344	22	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	344	23	full	full	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	344	24	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	344	25	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	344	26	using	use	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	344	27	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	344	28	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	344	29	techniques	technique	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	344	30	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	345	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	345	2	second	second	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	345	3	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	345	4	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	345	5	con-	con-	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	345	6	trol	trol	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	345	7	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	345	8	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	345	9	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	345	10	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	345	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	345	12	predicate-	predicate-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	345	13	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	345	14	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	345	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	346	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	346	2	experiments	experiment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	346	3	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	346	4	observe	observe	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	346	5	large	large	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	346	6	improvements	improvement	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	346	7	resulting	result	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	346	8	from	from	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	346	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	346	10	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	346	11	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	346	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	347	1	As	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	347	2	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	347	3	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	347	4	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	347	5	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	347	6	technique	technique	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	347	7	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	347	8	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	347	9	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	347	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	347	11	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	347	12	does	do	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	347	13	improve	improve	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	347	14	over	over	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	347	15	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	347	16	pipelines	pipeline	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	347	17	when	when	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	347	18	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	347	19	use	use	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	347	20	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	347	21	first	first	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	347	22	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	347	23	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	347	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	348	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	348	2	means	mean	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	348	3	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	348	4	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	348	5	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	348	6	configu-	configu-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	348	7	ration	ration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	348	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	348	9	explicit	explicit	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	348	10	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	348	11	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	348	12	help	help	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	348	13	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	348	14	find	find	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	348	15	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	348	16	solution	solution	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	348	17	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	348	18	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	348	19	better	well	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	348	20	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	348	21	both	both	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	348	22	layers	layer	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	348	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	349	1	To	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	349	2	some	some	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	349	3	ex-	ex-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	349	4	tent	tent	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	349	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	349	6	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	349	7	empirically	empirically	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	349	8	validates	validate	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	349	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	349	10	research	research	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	349	11	objec-	objec-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	349	12	tive	tive	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	349	13	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	349	14	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	349	15	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	349	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	350	1	However	however	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	350	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	350	3	when	when	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	350	4	we	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	350	5	move	move	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	350	6	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	350	7	second	second	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	350	8	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	350	9	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	350	10	parsers	parser	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	350	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	350	12	differences	difference	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	350	13	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	350	14	respect	respect	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	350	15	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	350	16	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	350	17	pipeline	pipeline	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	350	18	are	be	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	350	19	insignificant	insignificant	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	350	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	351	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	351	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	351	3	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	351	4	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	351	5	expected	expect	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	351	6	that	that	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	351	7	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	351	8	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	351	9	parsers	parser	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	351	10	improve	improve	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	351	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	351	12	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	351	13	need	need	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	351	14	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	351	15	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	351	16	methods	method	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	351	17	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	351	18	less	less	RBR
work_7jtgrdqc4zeb7lul7vifzjglr4	351	19	critical	critical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	351	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	352	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	352	2	remains	remain	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	352	3	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	352	4	open	open	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	352	5	question	question	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	352	6	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	352	7	validate	validate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	352	8	if	if	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	352	9	large	large	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	352	10	improvements	improvement	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	352	11	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	352	12	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	352	13	achieved	achieve	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	352	14	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	352	15	integrating	integrate	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	352	16	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	352	17	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	352	18	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	352	19	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	352	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	353	1	To	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	353	2	study	study	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	353	3	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	353	4	question	question	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	353	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	353	6	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	353	7	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	353	8	necessary	necessary	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	353	9	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	353	10	have	have	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	353	11	efficient	efficient	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	353	12	pars-	pars-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	353	13	ing	ing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	353	14	algorithms	algorithm	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	353	15	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	353	16	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	353	17	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	353	18	structures	structure	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	353	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	354	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	354	2	paper	paper	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	354	3	contributes	contribute	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	354	4	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	354	5	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	354	6	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	354	7	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	354	8	has	have	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	354	9	optimality	optimality	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	354	10	228	228	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	354	11	guarantees	guarantee	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	354	12	whenever	whenever	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	354	13	it	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	354	14	converges	converge	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	354	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	355	1	Our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	355	2	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	355	3	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	355	4	incorporate	incorporate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	355	5	richer	rich	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	355	6	families	family	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	355	7	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	355	8	fea-	fea-	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	355	9	tures	ture	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	355	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	356	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	356	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	356	3	straightforward	straightforward	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	356	4	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	356	5	incorporate	incorporate	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	356	6	better	well	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	356	7	se-	se-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	356	8	mantic	mantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	356	9	representations	representation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	356	10	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	356	11	predicates	predicate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	356	12	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	356	13	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	356	14	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	356	15	just	just	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	356	16	plain	plain	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	356	17	words	word	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	356	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	356	19	e.g.	e.g.	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	357	1	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	357	2	exploiting	exploit	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	357	3	WordNet	WordNet	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	357	4	or	or	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	357	5	distributional	distributional	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	357	6	representations	representation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	357	7	as	as	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	357	8	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	357	9	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	357	10	Zapirain	Zapirain	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	357	11	et	et	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	357	12	al	al	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	357	13	.	.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	357	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	357	15	2013	2013	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	357	16	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	357	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	358	1	Potentially	potentially	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	358	2	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	358	3	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	358	4	could	could	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	358	5	result	result	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	358	6	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	358	7	larger	large	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	358	8	im-	im-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	358	9	provements	provement	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	358	10	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	358	11	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	358	12	performance	performance	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	358	13	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	358	14	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	358	15	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	358	16	se-	se-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	358	17	mantic	mantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	358	18	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	358	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	359	1	It	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	359	2	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	359	3	also	also	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	359	4	necessary	necessary	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	359	5	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	359	6	experiment	experiment	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	359	7	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	359	8	differ-	differ-	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	359	9	ent	ent	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	359	10	languages	language	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	359	11	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	359	12	where	where	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	359	13	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	359	14	performance	performance	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	359	15	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	359	16	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	359	17	parsers	parser	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	359	18	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	359	19	lower	low	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	359	20	than	than	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	359	21	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	359	22	English	English	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	359	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	359	24	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	359	25	hence	hence	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	359	26	there	there	EX
work_7jtgrdqc4zeb7lul7vifzjglr4	359	27	is	be	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	359	28	potential	potential	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	359	29	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	359	30	improvement	improvement	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	359	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	360	1	Our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	360	2	treatment	treatment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	360	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	360	4	local	local	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	360	5	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	360	6	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	360	7	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	360	8	links	link	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	360	9	predicates	predicate	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	360	10	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	360	11	argu-	argu-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	360	12	ments	ment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	360	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	360	14	based	base	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	360	15	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	360	16	explicit	explicit	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	360	17	enumeration	enumeration	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	360	18	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	360	19	likely	likely	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	360	20	paths	path	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	360	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	360	22	was	be	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	360	23	simplistic	simplistic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	360	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	361	1	Future	future	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	361	2	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	3	should	should	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	361	4	explore	explore	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	361	5	meth-	meth-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	361	6	ods	od	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	361	7	that	that	WDT
work_7jtgrdqc4zeb7lul7vifzjglr4	361	8	model	model	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	361	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	361	10	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	361	11	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	12	linking	link	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	361	13	predi-	predi-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	14	cates	cat	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	361	15	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	16	arguments	argument	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	361	17	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	361	18	whenever	whenever	WRB
work_7jtgrdqc4zeb7lul7vifzjglr4	361	19	this	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	361	20	structure	structure	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	21	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	361	22	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	361	23	parsed	parse	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	24	efficiently	efficiently	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	361	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	361	26	our	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	361	27	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	361	28	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	29	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	30	can	can	MD
work_7jtgrdqc4zeb7lul7vifzjglr4	361	31	be	be	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	361	32	employed	employ	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	33	to	to	TO
work_7jtgrdqc4zeb7lul7vifzjglr4	361	34	define	define	VB
work_7jtgrdqc4zeb7lul7vifzjglr4	361	35	an	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	361	36	efficient	efficient	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	361	37	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	361	38	system	system	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	361	39	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	362	1	Acknowledgments	acknowledgment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	362	2	We	-PRON-	PRP
work_7jtgrdqc4zeb7lul7vifzjglr4	362	3	thank	thank	VBP
work_7jtgrdqc4zeb7lul7vifzjglr4	362	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	362	5	editor	editor	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	362	6	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	362	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	362	8	anonymous	anonymous	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	362	9	reviewers	reviewer	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	362	10	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	362	11	their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	362	12	valuable	valuable	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	362	13	feedback	feedback	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	362	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	363	1	This	this	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	363	2	work	work	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	3	was	be	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	363	4	financed	finance	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	5	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	363	7	European	European	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	8	Commission	Commission	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	9	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	363	11	XLike	XLike	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	12	project	project	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	13	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	363	14	FP7-	FP7-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	15	288342	288342	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	363	16	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	363	17	;	;	:
work_7jtgrdqc4zeb7lul7vifzjglr4	363	18	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	363	19	by	by	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	363	21	Spanish	spanish	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	363	22	Government	Government	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	23	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	24	project	project	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	25	OpenMT-2	OpenMT-2	,
work_7jtgrdqc4zeb7lul7vifzjglr4	363	26	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	363	27	TIN2009	tin2009	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	28	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	363	29	14675-C03	14675-C03	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	30	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	363	31	01	01	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	363	32	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	363	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	363	34	project	project	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	35	Skater	Skater	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	36	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	363	37	TIN2012	TIN2012	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	38	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	363	39	38584-C06	38584-c06	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	363	40	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	363	41	01	01	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	363	42	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	363	43	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	363	44	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	363	45	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	363	46	Ramón	Ramón	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	47	y	y	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	48	Cajal	Cajal	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	49	contract	contract	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	50	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	363	51	Xavier	Xavier	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	52	Carreras	Carreras	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	53	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	363	54	RYC-2008	RYC-2008	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	363	55	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	363	56	02223	02223	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	363	57	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	363	58	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	364	1	References	reference	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	364	2	Rainer	Rainer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	364	3	Burkard	Burkard	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	364	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	364	5	Mario	Mario	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	364	6	Dell’Amico	Dell’Amico	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	364	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	364	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	364	9	Silvano	Silvano	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	364	10	Martello	Martello	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	364	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	365	1	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	365	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	366	1	Assignment	Assignment	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	366	2	Problems	problem	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	366	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	367	1	Society	society	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	367	2	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	367	3	Industrial	Industrial	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	367	4	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	367	5	Applied	Applied	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	367	6	Mathematics	Mathematics	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	367	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	368	1	Xavier	Xavier	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	368	2	Carreras	Carreras	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	368	3	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	368	4	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	368	5	Màrquez	Màrquez	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	368	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	369	1	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	369	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	370	1	Introduc-	Introduc-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	370	2	tion	tion	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	370	3	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	370	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	370	5	CoNLL-2005	CoNLL-2005	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	370	6	shared	share	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	370	7	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	370	8	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	370	9	Semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	370	10	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	370	11	labeling	labeling	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	370	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	371	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	371	2	Proceedings	Proceedings	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	371	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	371	5	Ninth	Ninth	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	6	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	7	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	371	8	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	9	Natural	Natural	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	10	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	11	Learning	Learning	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	12	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	371	13	CoNLL-	CoNLL-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	14	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	371	15	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	371	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	371	17	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	371	18	152–164	152–164	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	371	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	371	20	Ann	Ann	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	21	Arbor	Arbor	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	371	23	Michigan	Michigan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	371	25	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	371	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	372	1	Xavier	Xavier	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	372	2	Carreras	Carreras	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	372	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	373	1	2007	2007	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	373	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	374	1	Experiments	experiment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	374	2	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	374	3	a	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	374	4	higher	high	JJR
work_7jtgrdqc4zeb7lul7vifzjglr4	374	5	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	374	6	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	374	7	projective	projective	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	374	8	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	374	9	parser	parser	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	374	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	375	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	375	2	Proceedings	Proceedings	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	375	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	375	5	CoNLL	CoNLL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	6	Shared	Shared	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	7	Task	Task	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	8	Session	Session	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	375	10	EMNLP	EMNLP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	11	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	375	12	CoNLL	CoNLL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	13	2007	2007	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	375	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	375	15	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	375	16	957–961	957–961	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	375	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	375	18	Prague	Prague	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	375	20	Czech	Czech	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	21	Republic	Republic	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	375	23	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	375	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	376	1	Michael	Michael	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	376	2	Collins	Collins	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	376	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	377	1	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	377	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	378	1	Discriminative	discriminative	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	378	2	training	training	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	378	3	meth-	meth-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	378	4	ods	od	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	378	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	378	6	Hidden	Hidden	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	378	7	Markov	Markov	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	378	8	Models	Models	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	378	9	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	378	10	Theory	theory	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	378	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	378	12	experi-	experi-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	378	13	ments	ment	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	378	14	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	378	15	Perceptron	Perceptron	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	378	16	algorithms	algorithm	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	378	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	379	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	379	2	Proceedings	Proceedings	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	379	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	379	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	379	5	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	379	6	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	379	7	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	379	8	Empirical	Empirical	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	379	9	Methods	Methods	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	379	10	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	379	11	Natural	Natural	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	379	12	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	379	13	Processing	Processing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	379	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	379	15	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	379	16	1–8	1–8	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	379	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	379	18	July	July	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	379	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	380	1	Dipanjan	Dipanjan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	380	2	Das	Das	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	380	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	380	4	André	André	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	380	5	F.	F.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	380	6	T.	T.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	380	7	Martins	Martins	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	380	8	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	380	9	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	380	10	Noah	Noah	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	380	11	A.	A.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	380	12	Smith	Smith	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	380	13	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	381	1	2012	2012	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	381	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	382	1	An	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	382	2	exact	exact	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	382	3	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	382	4	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	382	5	algorithm	algorithm	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	382	6	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	382	7	shallow	shallow	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	382	8	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	382	9	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	382	10	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	382	11	constraints	constraint	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	382	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	383	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	383	2	Pro-	Pro-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	3	ceedings	ceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	383	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	383	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	383	6	First	First	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	7	Joint	Joint	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	8	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	9	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	383	10	Lexical	Lexical	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	383	12	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	13	Semantics	Semantics	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	383	14	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	383	15	Volume	Volume	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	16	1	1	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	383	17	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	383	18	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	383	19	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	383	20	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	383	21	main	main	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	383	22	conference	conference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	383	23	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	383	24	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	383	25	shared	shared	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	383	26	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	383	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	383	28	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	383	29	Volume	volume	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	383	30	2	2	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	383	31	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	383	32	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	383	33	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	383	34	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	383	35	Sixth	Sixth	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	36	International	International	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	37	Workshop	Workshop	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	38	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	383	39	Semantic	Semantic	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	40	Evaluation	Evaluation	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	41	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	383	42	SemEval	SemEval	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	43	’	'	''
work_7jtgrdqc4zeb7lul7vifzjglr4	383	44	12	12	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	383	45	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	383	46	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	383	47	209–217	209–217	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	383	48	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	383	49	Stroudsburg	Stroudsburg	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	50	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	383	51	PA	PA	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	52	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	383	53	USA	USA	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	383	54	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	384	1	Jason	Jason	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	384	2	Eisner	Eisner	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	384	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	385	1	2000	2000	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	385	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	386	1	Bilexical	bilexical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	386	2	grammars	grammar	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	386	3	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	386	4	their	-PRON-	PRP$
work_7jtgrdqc4zeb7lul7vifzjglr4	386	5	cubic-	cubic-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	386	6	time	time	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	386	7	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	386	8	algorithms	algorithm	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	386	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	387	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	387	2	Harry	Harry	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	387	3	Bunt	Bunt	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	387	4	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	387	5	Anton	Anton	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	387	6	Nijholt	Nijholt	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	387	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	387	8	editors	editor	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	387	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	387	10	Advances	Advances	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	387	11	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	387	12	Probabilistic	Probabilistic	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	387	13	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	387	14	Other	other	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	387	15	Parsing	Parsing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	387	16	Technologies	Technologies	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	387	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	387	18	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	387	19	29–62	29–62	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	387	20	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	388	1	Kluwer	Kluwer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	388	2	Academic	Academic	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	388	3	Publishers	Publishers	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	388	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	388	5	October	October	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	388	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	389	1	Andrea	Andrea	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	389	2	Gesmundo	Gesmundo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	389	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	389	4	James	James	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	389	5	Henderson	Henderson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	389	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	389	7	Paola	Paola	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	389	8	Merlo	Merlo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	389	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	389	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	389	11	Ivan	Ivan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	389	12	Titov	Titov	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	389	13	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	390	1	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	390	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	391	1	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	391	2	latent	latent	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	391	3	variable	variable	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	391	4	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	391	5	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	391	6	syn-	syn-	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	391	7	chronous	chronous	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	391	8	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	391	9	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	391	10	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	391	11	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	391	12	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	391	13	multiple	multiple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	391	14	lan-	lan-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	391	15	guages	guage	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	391	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	392	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	392	2	Proceedings	Proceedings	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	392	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	392	5	Thirteenth	Thirteenth	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	6	Confer-	Confer-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	7	ence	ence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	392	8	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	392	9	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	10	Natural	Natural	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	11	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	12	Learning	Learning	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	13	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	392	14	CoNLL	CoNLL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	15	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	392	16	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	392	17	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	392	18	Shared	share	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	392	19	Task	Task	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	392	21	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	392	22	37–42	37–42	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	392	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	392	24	Boulder	Boulder	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	392	26	Colorado	Colorado	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	392	28	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	392	29	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	393	1	Daniel	Daniel	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	393	2	Gildea	Gildea	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	393	3	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	393	4	Daniel	Daniel	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	393	5	Jurafsky	Jurafsky	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	393	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	394	1	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	394	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	395	1	Automatic	automatic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	395	2	la-	la-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	395	3	beling	beling	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	395	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	395	5	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	395	6	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	395	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	396	1	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	396	2	Linguistics	Linguistics	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	396	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	396	4	28(3):245–288	28(3):245–288	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	396	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	396	6	September	September	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	396	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	397	1	Daniel	Daniel	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	397	2	Gildea	Gildea	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	397	3	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	397	4	Martha	Martha	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	397	5	Palmer	Palmer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	397	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	398	1	2002	2002	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	398	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	399	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	399	2	necessity	necessity	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	399	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	399	4	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	399	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	399	6	predicate	predicate	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	399	7	argument	argument	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	399	8	recognition	recognition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	399	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	400	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	400	2	Pro-	Pro-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	3	ceedings	ceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	400	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	400	5	40th	40th	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	400	6	Annual	Annual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	7	Meeting	Meeting	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	400	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	400	10	Association	Association	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	11	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	400	12	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	13	Linguistics	Linguistics	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	400	15	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	400	16	239–246	239–246	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	400	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	400	18	Philadel-	Philadel-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	19	phia	phia	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	400	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	400	21	Pennsylvania	Pennsylvania	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	400	23	USA	USA	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	400	25	July	July	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	400	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	401	1	Jan	Jan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	2	Hajič	hajič	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	401	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	4	Massimiliano	Massimiliano	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	5	Ciaramita	Ciaramita	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	7	Richard	Richard	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	8	Johans-	Johans-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	9	son	son	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	401	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	11	Daisuke	Daisuke	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	12	Kawahara	Kawahara	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	14	Maria	Maria	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	15	Antònia	antònia	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	401	16	Martı́	Martı́	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	18	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	19	Màrquez	Màrquez	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	21	Adam	Adam	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	22	Meyers	Meyers	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	24	Joakim	Joakim	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	25	Nivre	Nivre	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	27	Sebastian	Sebastian	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	28	Padó	Padó	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	30	Jan	Jan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	31	Štěpánek	štěpánek	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	401	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	33	Pavel	Pavel	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	34	Straňák	Straňák	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	35	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	36	Mihai	Mihai	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	37	Surdeanu	Surdeanu	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	39	Nianwen	Nianwen	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	40	Xue	Xue	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	41	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	401	42	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	401	43	Yi	Yi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	44	Zhang	Zhang	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	401	45	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	402	1	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	402	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	403	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	403	2	CoNLL-	conll-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	403	3	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	403	4	shared	share	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	403	5	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	403	6	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	403	7	Syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	403	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	403	9	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	403	10	dependen-	dependen-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	403	11	cies	cie	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	403	12	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	403	13	multiple	multiple	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	403	14	languages	language	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	403	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	404	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	404	2	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	404	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	404	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	404	5	13th	13th	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	404	6	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	7	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	404	8	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	9	Natural	Natural	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	10	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	11	Learning	Learning	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	12	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	404	13	CoNLL-2009	conll-2009	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	404	14	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	404	15	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	404	16	Shared	share	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	404	17	Task	Task	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	404	19	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	404	20	1–18	1–18	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	404	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	404	22	Boulder	Boulder	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	404	24	Colorado	Colorado	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	404	26	USA	USA	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	404	28	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	404	29	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	405	1	Richard	Richard	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	405	2	Johansson	Johansson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	405	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	406	1	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	406	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	407	1	Statistical	statistical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	407	2	bistratal	bistratal	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	407	3	depen-	depen-	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	407	4	dency	dency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	407	5	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	407	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	408	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	408	2	Proceedings	Proceedings	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	408	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	408	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	408	5	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	408	6	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	408	7	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	408	8	Empirical	Empirical	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	408	9	Methods	Methods	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	408	10	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	408	11	Natural	Natural	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	408	12	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	408	13	Process-	Process-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	408	14	ing	ing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	408	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	408	16	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	408	17	561–569	561–569	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	408	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	408	19	Singapore	Singapore	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	408	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	408	21	August	August	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	408	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	409	1	Terry	Terry	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	409	2	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	409	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	409	4	Amir	Amir	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	409	5	Globerson	Globerson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	409	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	409	7	Xavier	Xavier	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	409	8	Carreras	Carreras	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	409	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	409	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	409	11	Michael	Michael	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	409	12	Collins	Collins	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	409	13	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	410	1	2007	2007	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	410	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	411	1	Structured	structured	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	411	2	prediction	prediction	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	411	3	mod-	mod-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	411	4	els	els	XX
work_7jtgrdqc4zeb7lul7vifzjglr4	411	5	via	via	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	411	6	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	411	7	matrix	matrix	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	411	8	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	411	9	tree	tree	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	411	10	theorem	theorem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	411	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	412	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	412	2	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	412	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	412	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	412	5	2007	2007	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	412	6	Joint	Joint	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	7	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	8	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	412	9	Empirical	Empirical	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	10	Methods	Methods	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	412	11	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	412	12	Nat-	Nat-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	13	ural	ural	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	412	14	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	15	Processing	Processing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	412	17	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	18	Natu-	Natu-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	19	ral	ral	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	412	20	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	21	Learning	Learning	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	22	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	412	23	EMNLP	EMNLP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	24	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	412	25	CoNLL	CoNLL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	26	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	412	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	412	28	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	412	29	141	141	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	412	30	–	–	:
work_7jtgrdqc4zeb7lul7vifzjglr4	412	31	150	150	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	412	32	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	412	33	Prague	Prague	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	34	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	412	35	Czech	Czech	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	36	Republic	Republic	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	37	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	412	38	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	412	39	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	413	1	Terry	Terry	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	2	Koo	Koo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	413	4	Alexander	Alexander	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	5	M.	M.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	6	Rush	Rush	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	413	8	Michael	Michael	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	9	Collins	Collins	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	413	11	Tommi	Tommi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	12	Jaakkola	Jaakkola	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	13	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	413	14	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	413	15	David	David	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	16	Sontag	Sontag	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	413	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	414	1	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	414	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	415	1	Dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	415	2	decompo-	decompo-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	415	3	sition	sition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	415	4	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	415	5	parsing	parse	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	415	6	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	415	7	non	non	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	415	8	-	-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	415	9	projective	projective	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	415	10	head	head	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	415	11	automata	automata	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	415	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	416	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	416	2	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	416	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	416	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	416	5	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	416	6	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	416	7	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	416	8	Empiri-	Empiri-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	416	9	cal	cal	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	416	10	Methods	Methods	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	416	11	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	416	12	Natural	Natural	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	416	13	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	416	14	Processing	Processing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	416	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	416	16	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	416	17	1288–1298	1288–1298	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	416	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	416	19	Cambridge	Cambridge	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	416	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	416	21	MA	MA	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	416	22	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	416	23	October	October	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	416	24	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	417	1	229	229	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	417	2	Harold	Harold	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	417	3	W.	W.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	417	4	Kuhn	Kuhn	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	417	5	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	418	1	1955	1955	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	418	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	419	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	419	2	Hungarian	hungarian	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	419	3	method	method	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	419	4	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	419	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	419	6	assignment	assignment	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	419	7	problem	problem	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	419	8	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	420	1	Naval	Naval	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	420	2	Research	Research	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	420	3	Logistics	Logistics	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	420	4	Quar-	Quar-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	420	5	terly	terly	RB
work_7jtgrdqc4zeb7lul7vifzjglr4	420	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	420	7	2(1	2(1	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	420	8	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	420	9	2):83–97	2):83–97	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	420	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	421	1	Xavier	Xavier	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	421	2	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	421	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	421	4	Stefan	Stefan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	421	5	Bott	Bott	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	421	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	421	7	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	421	8	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	421	9	Màrquez	Màrquez	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	421	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	422	1	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	422	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	423	1	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	423	2	second	second	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	423	3	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	423	4	order	order	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	423	5	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	423	6	eisner	eisner	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	423	7	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	423	8	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	423	9	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	423	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	423	11	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	423	12	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	423	13	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	423	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	424	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	424	2	Proceedings	Proceedings	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	424	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	424	5	Thirteenth	Thirteenth	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	6	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	7	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	424	8	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	9	Natu-	Natu-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	10	ral	ral	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	424	11	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	12	Learning	Learning	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	13	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	424	14	CoNLL	CoNLL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	15	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	424	16	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	424	17	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	424	18	Shared	share	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	424	19	Task	Task	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	424	21	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	424	22	79–84	79–84	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	424	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	424	24	Boulder	Boulder	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	424	26	Colorado	Colorado	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	424	28	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	424	29	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	425	1	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	425	2	Màrquez	Màrquez	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	425	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	425	4	Xavier	Xavier	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	425	5	Carreras	Carreras	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	425	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	425	7	Kenneth	Kenneth	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	425	8	C.	C.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	425	9	Litkowski	Litkowski	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	425	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	425	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	425	12	Suzanne	Suzanne	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	425	13	Stevenson	Stevenson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	425	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	426	1	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	426	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	427	1	Semantic	Semantic	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	427	2	Role	Role	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	427	3	Label-	Label-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	427	4	ing	ing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	427	5	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	427	6	An	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	427	7	Introduction	introduction	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	427	8	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	427	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	427	10	Special	Special	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	427	11	Issue	issue	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	427	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	428	1	Computa-	Computa-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	428	2	tional	tional	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	428	3	Linguistics	Linguistics	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	428	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	428	5	34(2):145–159	34(2):145–159	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	428	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	428	7	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	428	8	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	429	1	André	André	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	429	2	Martins	Martins	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	429	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	429	4	Noah	Noah	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	429	5	Smith	Smith	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	429	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	429	7	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	429	8	Eric	Eric	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	429	9	Xing	Xing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	429	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	430	1	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	430	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	431	1	Con-	Con-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	431	2	cise	cise	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	431	3	integer	integer	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	431	4	linear	linear	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	431	5	programming	programming	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	431	6	formulations	formulation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	431	7	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	431	8	de-	de-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	431	9	pendency	pendency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	431	10	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	431	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	432	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	2	Proceedings	Proceedings	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	432	5	Joint	Joint	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	6	Con-	Con-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	7	ference	ference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	432	10	47th	47th	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	432	11	Annual	Annual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	12	Meeting	Meeting	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	13	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	14	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	432	15	ACL	ACL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	16	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	432	17	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	432	18	4th	4th	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	432	19	International	International	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	20	Joint	Joint	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	21	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	22	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	23	Natural	Natural	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	24	Lan-	Lan-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	25	guage	guage	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	26	Processing	processing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	27	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	28	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	432	29	AFNLP	AFNLP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	30	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	432	31	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	432	32	342–350	342–350	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	33	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	432	34	Sun-	Sun-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	35	tec	tec	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	432	36	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	432	37	Singapore	Singapore	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	38	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	432	39	August	August	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	432	40	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	433	1	Ryan	Ryan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	433	2	McDonald	McDonald	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	433	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	433	4	Koby	Koby	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	433	5	Crammer	Crammer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	433	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	433	7	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	433	8	Fernando	Fernando	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	433	9	Pereira	Pereira	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	433	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	434	1	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	434	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	435	1	Online	online	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	435	2	large	large	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	435	3	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	435	4	margin	margin	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	435	5	training	training	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	435	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	435	7	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	435	8	parsers	parser	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	435	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	436	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	436	2	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	436	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	436	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	436	5	43rd	43rd	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	436	6	Annual	Annual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	436	7	Meet-	Meet-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	436	8	ing	ing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	436	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	436	10	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	436	11	Association	Association	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	436	12	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	436	13	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	436	14	Linguistics	Linguistics	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	436	15	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	436	16	ACL’05	ACL’05	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	436	17	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	436	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	436	19	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	436	20	91–98	91–98	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	436	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	436	22	Ann	Ann	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	436	23	Arbor	Arbor	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	436	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	436	25	Michigan	Michigan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	436	26	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	436	27	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	436	28	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	437	1	Adam	Adam	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	2	Meyers	Meyers	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	437	4	Ruth	Ruth	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	5	Reeves	Reeves	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	437	7	Catherine	Catherine	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	8	Macleod	Macleod	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	437	10	Rachel	Rachel	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	11	Szekely	Szekely	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	437	13	Veronika	Veronika	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	14	Zielinska	Zielinska	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	15	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	437	16	Brian	Brian	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	17	Young	Young	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	437	19	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	437	20	Ralph	Ralph	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	21	Grishman	Grishman	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	437	22	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	438	1	2004	2004	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	438	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	439	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	439	2	NomBank	NomBank	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	439	3	Project	Project	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	439	4	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	439	5	An	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	439	6	interim	interim	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	439	7	report	report	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	439	8	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	440	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	440	2	A.	a.	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	440	3	Meyers	Meyers	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	440	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	440	5	editor	editor	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	440	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	440	7	HLT	HLT	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	440	8	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	440	9	NAACL	NAACL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	440	10	2004	2004	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	440	11	Work-	work-	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	440	12	shop	shop	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	440	13	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	440	14	Frontiers	frontier	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	440	15	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	440	16	Corpus	Corpus	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	440	17	Annotation	Annotation	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	440	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	440	19	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	440	20	24–31	24–31	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	440	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	440	22	Boston	Boston	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	440	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	440	24	Massachusetts	Massachusetts	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	440	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	440	26	USA	USA	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	440	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	440	28	May.	May	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	441	1	Alessandro	Alessandro	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	441	2	Moschitti	Moschitti	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	441	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	442	1	2004	2004	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	442	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	443	1	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	443	2	study	study	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	443	3	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	443	4	convolution	convolution	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	443	5	kernels	kernel	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	443	6	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	443	7	shallow	shallow	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	443	8	statistic	statistic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	443	9	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	443	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	444	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	444	2	Proceedings	Proceedings	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	444	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	444	5	42nd	42nd	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	444	6	Meeting	Meeting	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	7	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	444	8	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	444	9	Association	Association	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	10	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	444	11	Computa-	Computa-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	12	tional	tional	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	444	13	Linguistics	Linguistics	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	14	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	444	15	ACL’04	ACL’04	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	16	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	444	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	444	18	Main	Main	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	19	Volume	Volume	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	20	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	444	21	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	444	22	335	335	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	444	23	–	–	SYM
work_7jtgrdqc4zeb7lul7vifzjglr4	444	24	342	342	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	444	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	444	26	Barcelona	Barcelona	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	444	28	Spain	Spain	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	29	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	444	30	July	July	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	444	31	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	445	1	Joakim	Joakim	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	445	2	Nivre	Nivre	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	445	3	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	445	4	Jens	Jens	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	445	5	Nilsson	Nilsson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	445	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	446	1	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	446	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	447	1	Pseudo	pseudo	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	447	2	-	-	HYPH
work_7jtgrdqc4zeb7lul7vifzjglr4	447	3	projective	projective	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	447	4	dependency	dependency	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	447	5	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	447	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	448	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	448	2	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	448	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	448	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	448	5	43rd	43rd	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	448	6	Annual	Annual	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	7	Meeting	Meeting	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	448	9	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	448	10	Association	Association	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	11	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	448	12	Computa-	Computa-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	13	tional	tional	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	448	14	Linguistics	Linguistics	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	15	(	(	-LRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	448	16	ACL’05	ACL’05	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	17	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	448	18	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	448	19	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	448	20	99–106	99–106	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	448	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	448	22	Ann	Ann	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	23	Ar-	Ar-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	24	bor	bor	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	25	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	448	26	Michigan	Michigan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	27	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	448	28	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	448	29	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	449	1	Martha	Martha	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	449	2	Palmer	Palmer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	449	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	449	4	Daniel	Daniel	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	449	5	Gildea	Gildea	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	449	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	449	7	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	449	8	Paul	Paul	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	449	9	Kingsbury	Kingsbury	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	449	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	450	1	2005	2005	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	450	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	451	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	451	2	Proposition	Proposition	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	451	3	Bank	Bank	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	451	4	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	451	5	An	an	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	451	6	annotated	annotate	VBN
work_7jtgrdqc4zeb7lul7vifzjglr4	451	7	corpus	corpus	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	451	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	451	9	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	451	10	roles	role	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	451	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	452	1	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	452	2	Linguistics	Linguistics	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	452	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	452	4	31(1):71	31(1):71	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	452	5	–	–	:
work_7jtgrdqc4zeb7lul7vifzjglr4	452	6	106	106	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	452	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	452	8	March	March	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	452	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	453	1	Vasin	Vasin	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	453	2	Punyakanok	Punyakanok	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	453	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	453	4	Dan	Dan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	453	5	Roth	Roth	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	453	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	453	7	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	453	8	Wen	Wen	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	453	9	tau	tau	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	453	10	Yih	Yih	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	453	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	454	1	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	454	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	455	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	455	2	importance	importance	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	455	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	455	4	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	455	5	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	455	6	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	455	7	inference	inference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	455	8	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	455	9	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	455	10	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	455	11	labeling	labeling	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	455	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	456	1	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	456	2	Linguistics	Linguistics	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	456	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	456	4	34(3):257–287	34(3):257–287	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	456	5	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	456	6	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	456	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	457	1	Sebastian	Sebastian	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	457	2	Riedel	Riedel	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	457	3	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	457	4	Andrew	Andrew	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	457	5	McCallum	McCallum	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	457	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	458	1	2011	2011	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	458	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	459	1	Fast	fast	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	459	2	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	459	3	robust	robust	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	459	4	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	459	5	models	model	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	459	6	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	459	7	biomedical	biomedical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	459	8	event	event	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	459	9	extraction	extraction	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	459	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	460	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	460	2	Proceedings	Proceedings	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	460	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	460	5	2011	2011	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	460	6	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	7	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	460	8	Empirical	Empirical	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	9	Methods	Methods	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	460	10	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	460	11	Natural	Natural	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	12	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	13	Processing	Processing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	460	15	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	460	16	1–12	1–12	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	460	18	Edinburgh	Edinburgh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	460	20	Scotland	Scotland	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	460	22	UK	UK	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	460	24	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	460	25	July	July	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	460	26	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	461	1	Alexander	Alexander	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	461	2	M	M	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	461	3	Rush	Rush	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	461	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	461	5	David	David	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	461	6	Sontag	Sontag	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	461	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	461	8	Michael	Michael	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	461	9	Collins	Collins	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	461	10	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	461	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	461	12	Tommi	Tommi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	461	13	Jaakkola	Jaakkola	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	461	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	462	1	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	462	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	463	1	On	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	463	2	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	463	3	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	463	4	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	463	5	linear	linear	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	463	6	programming	programming	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	463	7	relaxations	relaxation	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	463	8	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	463	9	natural	natural	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	463	10	language	language	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	463	11	processing	processing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	463	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	464	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	464	2	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	464	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	464	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	464	5	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	464	6	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	464	7	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	464	8	Empirical	Empirical	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	464	9	Methods	Methods	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	464	10	in	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	464	11	Natural	Natural	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	464	12	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	464	13	Processing	Processing	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	464	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	464	15	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	464	16	1–11	1–11	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	464	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	464	18	Cambridge	Cambridge	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	464	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	464	20	MA	MA	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	464	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	464	22	October	October	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	464	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	465	1	David	David	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	465	2	Sontag	Sontag	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	465	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	465	4	Amir	Amir	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	465	5	Globerson	Globerson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	465	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	465	7	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	465	8	Tommi	Tommi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	465	9	Jaakkola	Jaakkola	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	465	10	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	466	1	2010	2010	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	466	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	467	1	Introduction	introduction	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	467	2	to	to	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	467	3	dual	dual	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	467	4	decomposition	decomposition	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	467	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	467	6	infer-	infer-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	467	7	ence	ence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	467	8	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	468	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	468	2	S.	S.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	468	3	Sra	Sra	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	468	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	468	5	S.	S.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	468	6	Nowozin	Nowozin	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	468	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	468	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	468	9	S.	S.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	468	10	J.	J.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	468	11	Wright	Wright	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	468	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	468	13	editors	editor	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	468	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	468	15	Optimization	optimization	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	468	16	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	468	17	Machine	Machine	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	468	18	Learning	Learning	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	468	19	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	469	1	MIT	MIT	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	469	2	Press	Press	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	469	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	470	1	Mihai	Mihai	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	2	Surdeanu	Surdeanu	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	470	4	Richard	Richard	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	5	Johansson	Johansson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	470	7	Adam	Adam	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	8	Meyers	Meyers	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	470	10	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	11	Màrquez	Màrquez	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	470	13	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	470	14	Joakim	Joakim	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	15	Nivre	Nivre	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	470	16	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	471	1	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	471	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	472	1	The	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	472	2	conll	conll	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	472	3	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	472	4	shared	share	VBD
work_7jtgrdqc4zeb7lul7vifzjglr4	472	5	task	task	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	472	6	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	472	7	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	472	8	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	472	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	472	10	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	472	11	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	472	12	se-	se-	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	472	13	mantic	mantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	472	14	dependencies	dependency	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	472	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	473	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	473	2	CoNLL	CoNLL	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	3	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	473	4	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	473	5	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	473	6	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	473	7	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	473	8	Twelfth	Twelfth	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	9	Conference	Conference	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	10	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	473	11	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	12	Natu-	Natu-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	13	ral	ral	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	473	14	Language	Language	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	15	Learning	Learning	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	16	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	473	17	pages	page	VBZ
work_7jtgrdqc4zeb7lul7vifzjglr4	473	18	159–177	159–177	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	473	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	473	20	Manchester	Manchester	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	473	22	England	England	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	23	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	473	24	August	August	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	473	25	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	474	1	Ivan	Ivan	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	474	2	Titov	Titov	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	474	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	474	4	James	James	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	474	5	Henderson	Henderson	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	474	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	474	7	Paola	Paola	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	474	8	Merlo	Merlo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	474	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	474	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	474	11	Gabriele	Gabriele	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	474	12	Musillo	Musillo	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	474	13	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	475	1	2009	2009	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	475	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	476	1	Online	online	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	476	2	graph	graph	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	476	3	planarisation	planarisation	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	476	4	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	476	5	syn-	syn-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	476	6	chronous	chronous	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	476	7	parsing	parsing	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	476	8	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	476	9	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	476	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	476	11	syntactic	syntactic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	476	12	dependen-	dependen-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	476	13	cies	cie	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	476	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	477	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	477	2	Proceedings	proceeding	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	477	3	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	477	4	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	477	5	21st	21st	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	477	6	international	international	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	477	7	jont	jont	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	477	8	conference	conference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	477	9	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	477	10	Artifical	artifical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	477	11	intelligence	intelligence	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	477	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	477	13	IJCAI’09	IJCAI’09	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	477	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	477	15	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	477	16	1562–1567	1562–1567	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	477	17	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	478	1	Kristina	Kristina	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	478	2	Toutanova	Toutanova	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	478	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	478	4	Aria	Aria	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	478	5	Haghighi	Haghighi	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	478	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	478	7	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	478	8	Christopher	Christopher	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	478	9	D.	D.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	478	10	Manning	Manning	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	478	11	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	479	1	2008	2008	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	479	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	480	1	A	a	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	480	2	global	global	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	480	3	joint	joint	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	480	4	model	model	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	480	5	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	480	6	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	480	7	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	480	8	labeling	labeling	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	480	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	481	1	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	481	2	Linguistics	Linguistics	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	481	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	481	4	34(2):161	34(2):161	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	481	5	–	–	:
work_7jtgrdqc4zeb7lul7vifzjglr4	481	6	191	191	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	481	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	481	8	June	June	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	481	9	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	482	1	Dennis	Dennis	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	482	2	D.	D.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	482	3	Wackerly	Wackerly	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	482	4	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	482	5	William	William	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	482	6	Mendenhall	Mendenhall	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	482	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	482	8	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	482	9	Richard	Richard	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	482	10	L.	L.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	482	11	Scheaffer	Scheaffer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	482	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	482	13	2007	2007	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	482	14	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	483	1	Mathematical	Mathematical	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	483	2	Statis-	Statis-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	483	3	tics	tic	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	483	4	with	with	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	483	5	Applications	Applications	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	483	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	483	7	chapter	chapter	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	483	8	15	15	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	483	9	:	:	:
work_7jtgrdqc4zeb7lul7vifzjglr4	483	10	Nonparametric	nonparametric	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	483	11	statistics	statistic	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	483	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	484	1	Duxbury	Duxbury	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	484	2	Press	Press	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	484	3	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	485	1	Nianwen	Nianwen	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	485	2	Xue	Xue	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	485	3	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	485	4	Martha	Martha	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	485	5	Palmer	Palmer	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	485	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	486	1	2004	2004	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	486	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	487	1	Calibrating	calibrate	VBG
work_7jtgrdqc4zeb7lul7vifzjglr4	487	2	features	feature	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	487	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	487	4	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	487	5	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	487	6	labeling	labeling	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	487	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	488	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	488	2	Dekang	Dekang	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	488	3	Lin	Lin	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	488	4	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	488	5	Dekai	Dekai	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	488	6	Wu	Wu	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	488	7	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	488	8	editors	editor	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	488	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	488	10	Proceedings	Proceedings	NNPS
work_7jtgrdqc4zeb7lul7vifzjglr4	488	11	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	488	12	EMNLP	EMNLP	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	488	13	2004	2004	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	488	14	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	488	15	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	488	16	88–94	88–94	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	488	17	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	488	18	Barcelona	Barcelona	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	488	19	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	488	20	Spain	Spain	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	488	21	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	488	22	July	July	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	488	23	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	489	1	Alexander	Alexander	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	489	2	S.	S.	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	489	3	Yeh	Yeh	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	489	4	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	490	1	2000	2000	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	490	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	491	1	More	more	RBR
work_7jtgrdqc4zeb7lul7vifzjglr4	491	2	accurate	accurate	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	491	3	tests	test	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	491	4	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	491	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	491	6	sta-	sta-	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	491	7	tistical	tistical	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	491	8	significance	significance	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	491	9	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	491	10	result	result	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	491	11	differences	difference	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	491	12	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	492	1	In	in	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	492	2	Proceed-	Proceed-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	492	3	ings	ing	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	492	4	of	of	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	492	5	the	the	DT
work_7jtgrdqc4zeb7lul7vifzjglr4	492	6	18th	18th	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	492	7	conference	conference	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	492	8	on	on	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	492	9	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	492	10	linguis-	linguis-	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	492	11	tics	tic	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	492	12	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	492	13	pages	page	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	492	14	947–953	947–953	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	492	15	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	493	1	Beñat	Beñat	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	493	2	Zapirain	Zapirain	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	493	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	493	4	Eneko	Eneko	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	493	5	Agirre	Agirre	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	493	6	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	493	7	Lluı́s	Lluı́s	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	493	8	Màrquez	Màrquez	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	493	9	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	493	10	and	and	CC
work_7jtgrdqc4zeb7lul7vifzjglr4	493	11	Mihai	Mihai	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	493	12	Surdeanu	Surdeanu	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	493	13	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	494	1	2013	2013	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	494	2	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	495	1	Selectional	Selectional	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	495	2	preferences	preference	NNS
work_7jtgrdqc4zeb7lul7vifzjglr4	495	3	for	for	IN
work_7jtgrdqc4zeb7lul7vifzjglr4	495	4	semantic	semantic	JJ
work_7jtgrdqc4zeb7lul7vifzjglr4	495	5	role	role	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	495	6	classification	classification	NN
work_7jtgrdqc4zeb7lul7vifzjglr4	495	7	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	496	1	Computational	Computational	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	496	2	Linguistics	Linguistics	NNP
work_7jtgrdqc4zeb7lul7vifzjglr4	496	3	,	,	,
work_7jtgrdqc4zeb7lul7vifzjglr4	496	4	39(3	39(3	CD
work_7jtgrdqc4zeb7lul7vifzjglr4	496	5	)	)	-RRB-
work_7jtgrdqc4zeb7lul7vifzjglr4	496	6	.	.	.
work_7jtgrdqc4zeb7lul7vifzjglr4	497	1	230	230	CD