id	sid	tid	token	lemma	pos
fcis-32489	1	1	frontiers	frontier	NOUN
fcis-32489	1	2	in	in	ADP
fcis-32489	1	3	computing	computing	NOUN
fcis-32489	1	4	and	and	CCONJ
fcis-32489	1	5	intelligent	intelligent	ADJ
fcis-32489	1	6	systems	system	NOUN
fcis-32489	1	7	issn	issn	VERB
fcis-32489	1	8	:	:	PUNCT
fcis-32489	1	9	2832	2832	NUM
fcis-32489	1	10	-	-	SYM
fcis-32489	1	11	6024	6024	NUM
fcis-32489	1	12	|	|	NOUN
fcis-32489	1	13	vol	vol	NOUN
fcis-32489	1	14	.	.	PUNCT
fcis-32489	2	1	14	14	NUM
fcis-32489	2	2	,	,	PUNCT
fcis-32489	2	3	no	no	INTJ
fcis-32489	2	4	.	.	NOUN
fcis-32489	2	5	2	2	NUM
fcis-32489	2	6	,	,	PUNCT
fcis-32489	2	7	2025	2025	NUM
fcis-32489	2	8	34	34	NUM
fcis-32489	2	9	document‐level	document‐level	NOUN
fcis-32489	2	10	relation	relation	NOUN
fcis-32489	2	11	extraction	extraction	NOUN
fcis-32489	2	12	based	base	VERB
fcis-32489	2	13	on	on	ADP
fcis-32489	2	14	graph	graph	NOUN
fcis-32489	2	15	convolutional	convolutional	ADJ
fcis-32489	2	16	neural	neural	ADJ
fcis-32489	2	17	networks	network	NOUN
fcis-32489	2	18	pengfei	pengfei	NOUN
fcis-32489	2	19	song	song	NOUN
fcis-32489	2	20	,	,	PUNCT
fcis-32489	2	21	haoyue	haoyue	PROPN
fcis-32489	2	22	lu	lu	PROPN
fcis-32489	2	23	sias	sias	PROPN
fcis-32489	2	24	university	university	PROPN
fcis-32489	2	25	,	,	PUNCT
fcis-32489	2	26	zhengzhou	zhengzhou	PROPN
fcis-32489	2	27	,	,	PUNCT
fcis-32489	2	28	henan	henan	PROPN
fcis-32489	2	29	,	,	PUNCT
fcis-32489	2	30	china	china	PROPN
fcis-32489	2	31	abstract	abstract	PROPN
fcis-32489	2	32	:	:	PUNCT
fcis-32489	2	33	the	the	DET
fcis-32489	2	34	objective	objective	NOUN
fcis-32489	2	35	of	of	ADP
fcis-32489	2	36	extracting	extract	VERB
fcis-32489	2	37	relations	relation	NOUN
fcis-32489	2	38	between	between	ADP
fcis-32489	2	39	document	document	NOUN
fcis-32489	2	40	components	component	NOUN
fcis-32489	2	41	lies	lie	VERB
fcis-32489	2	42	in	in	ADP
fcis-32489	2	43	identifying	identify	VERB
fcis-32489	2	44	the	the	PRON
fcis-32489	2	45	within	within	ADP
fcis-32489	2	46	a	a	DET
fcis-32489	2	47	single	single	ADJ
fcis-32489	2	48	document	document	NOUN
fcis-32489	2	49	,	,	PUNCT
fcis-32489	2	50	the	the	DET
fcis-32489	2	51	focus	focus	NOUN
fcis-32489	2	52	often	often	ADV
fcis-32489	2	53	lies	lie	VERB
fcis-32489	2	54	on	on	ADP
fcis-32489	2	55	understanding	understand	VERB
fcis-32489	2	56	the	the	DET
fcis-32489	2	57	connections	connection	NOUN
fcis-32489	2	58	between	between	ADP
fcis-32489	2	59	different	different	ADJ
fcis-32489	2	60	entities	entity	NOUN
fcis-32489	2	61	.	.	PUNCT
fcis-32489	3	1	this	this	DET
fcis-32489	3	2	type	type	NOUN
fcis-32489	3	3	of	of	ADP
fcis-32489	3	4	analysis	analysis	NOUN
fcis-32489	3	5	goes	go	VERB
fcis-32489	3	6	beyond	beyond	ADP
fcis-32489	3	7	individual	individual	ADJ
fcis-32489	3	8	sentences	sentence	NOUN
fcis-32489	3	9	,	,	PUNCT
fcis-32489	3	10	as	as	SCONJ
fcis-32489	3	11	it	it	PRON
fcis-32489	3	12	demands	demand	VERB
fcis-32489	3	13	an	an	DET
fcis-32489	3	14	understanding	understanding	NOUN
fcis-32489	3	15	of	of	ADP
fcis-32489	3	16	how	how	SCONJ
fcis-32489	3	17	information	information	NOUN
fcis-32489	3	18	from	from	ADP
fcis-32489	3	19	multiple	multiple	ADJ
fcis-32489	3	20	sentences	sentence	NOUN
fcis-32489	3	21	interacts	interact	VERB
fcis-32489	3	22	to	to	PART
fcis-32489	3	23	form	form	VERB
fcis-32489	3	24	these	these	DET
fcis-32489	3	25	connections	connection	NOUN
fcis-32489	3	26	.	.	PUNCT
fcis-32489	4	1	over	over	ADP
fcis-32489	4	2	recent	recent	ADJ
fcis-32489	4	3	years	year	NOUN
fcis-32489	4	4	,	,	PUNCT
fcis-32489	4	5	the	the	DET
fcis-32489	4	6	importance	importance	NOUN
fcis-32489	4	7	of	of	ADP
fcis-32489	4	8	exploring	explore	VERB
fcis-32489	4	9	relationships	relationship	NOUN
fcis-32489	4	10	involving	involve	VERB
fcis-32489	4	11	several	several	ADJ
fcis-32489	4	12	entities	entity	NOUN
fcis-32489	4	13	simultaneously	simultaneously	ADV
fcis-32489	4	14	has	have	AUX
fcis-32489	4	15	grown	grow	VERB
fcis-32489	4	16	significantly	significantly	ADV
fcis-32489	4	17	.	.	PUNCT
fcis-32489	5	1	to	to	PART
fcis-32489	5	2	advance	advance	VERB
fcis-32489	5	3	the	the	DET
fcis-32489	5	4	field	field	NOUN
fcis-32489	5	5	of	of	ADP
fcis-32489	5	6	studying	study	VERB
fcis-32489	5	7	such	such	ADJ
fcis-32489	5	8	connections	connection	NOUN
fcis-32489	5	9	across	across	ADP
fcis-32489	5	10	entire	entire	ADJ
fcis-32489	5	11	documents	document	NOUN
fcis-32489	5	12	,	,	PUNCT
fcis-32489	5	13	a	a	DET
fcis-32489	5	14	novel	novel	ADJ
fcis-32489	5	15	collection	collection	NOUN
fcis-32489	5	16	of	of	ADP
fcis-32489	5	17	data	datum	NOUN
fcis-32489	5	18	points	point	NOUN
fcis-32489	5	19	,	,	PUNCT
fcis-32489	5	20	known	know	VERB
fcis-32489	5	21	as	as	SCONJ
fcis-32489	5	22	docred	docre	VERB
fcis-32489	5	23	,	,	PUNCT
fcis-32489	5	24	has	have	AUX
fcis-32489	5	25	been	be	AUX
fcis-32489	5	26	introduced	introduce	VERB
fcis-32489	5	27	.	.	PUNCT
fcis-32489	6	1	currently	currently	ADV
fcis-32489	6	2	,	,	PUNCT
fcis-32489	6	3	the	the	DET
fcis-32489	6	4	standard	standard	ADJ
fcis-32489	6	5	approach	approach	NOUN
fcis-32489	6	6	for	for	ADP
fcis-32489	6	7	this	this	DET
fcis-32489	6	8	task	task	NOUN
fcis-32489	6	9	involves	involve	VERB
fcis-32489	6	10	using	use	VERB
fcis-32489	6	11	bilstm	bilstm	NOUN
fcis-32489	6	12	networks	network	NOUN
fcis-32489	6	13	to	to	PART
fcis-32489	6	14	process	process	VERB
fcis-32489	6	15	the	the	DET
fcis-32489	6	16	entire	entire	ADJ
fcis-32489	6	17	document	document	NOUN
fcis-32489	6	18	as	as	ADP
fcis-32489	6	19	a	a	DET
fcis-32489	6	20	whole	whole	NOUN
fcis-32489	6	21	.	.	PUNCT
fcis-32489	7	1	however	however	ADV
fcis-32489	7	2	,	,	PUNCT
fcis-32489	7	3	this	this	DET
fcis-32489	7	4	method	method	NOUN
fcis-32489	7	5	struggles	struggle	VERB
fcis-32489	7	6	to	to	PART
fcis-32489	7	7	effectively	effectively	ADV
fcis-32489	7	8	capture	capture	VERB
fcis-32489	7	9	the	the	DET
fcis-32489	7	10	intricate	intricate	ADJ
fcis-32489	7	11	relationships	relationship	NOUN
fcis-32489	7	12	that	that	PRON
fcis-32489	7	13	exist	exist	VERB
fcis-32489	7	14	among	among	ADP
fcis-32489	7	15	various	various	ADJ
fcis-32489	7	16	entities	entity	NOUN
fcis-32489	7	17	.	.	PUNCT
fcis-32489	8	1	to	to	PART
fcis-32489	8	2	overcome	overcome	VERB
fcis-32489	8	3	this	this	DET
fcis-32489	8	4	limitation	limitation	NOUN
fcis-32489	8	5	,	,	PUNCT
fcis-32489	8	6	a	a	DET
fcis-32489	8	7	new	new	ADJ
fcis-32489	8	8	model	model	NOUN
fcis-32489	8	9	designed	design	VERB
fcis-32489	8	10	for	for	ADP
fcis-32489	8	11	document	document	NOUN
fcis-32489	8	12	-	-	PUNCT
fcis-32489	8	13	level	level	NOUN
fcis-32489	8	14	relationship	relationship	NOUN
fcis-32489	8	15	identification	identification	NOUN
fcis-32489	8	16	has	have	AUX
fcis-32489	8	17	been	be	AUX
fcis-32489	8	18	developed	develop	VERB
fcis-32489	8	19	,	,	PUNCT
fcis-32489	8	20	which	which	PRON
fcis-32489	8	21	leverages	leverage	VERB
fcis-32489	8	22	graph	graph	NOUN
fcis-32489	8	23	convolutional	convolutional	ADJ
fcis-32489	8	24	networks	network	NOUN
fcis-32489	8	25	(	(	PUNCT
fcis-32489	8	26	gcn	gcn	NOUN
fcis-32489	8	27	)	)	PUNCT
fcis-32489	8	28	.	.	PUNCT
fcis-32489	9	1	gcns	gcns	PROPN
fcis-32489	9	2	are	be	AUX
fcis-32489	9	3	particularly	particularly	ADV
fcis-32489	9	4	useful	useful	ADJ
fcis-32489	9	5	here	here	ADV
fcis-32489	9	6	because	because	SCONJ
fcis-32489	9	7	they	they	PRON
fcis-32489	9	8	can	can	AUX
fcis-32489	9	9	gather	gather	VERB
fcis-32489	9	10	information	information	NOUN
fcis-32489	9	11	from	from	ADP
fcis-32489	9	12	surrounding	surround	VERB
fcis-32489	9	13	entities	entity	NOUN
fcis-32489	9	14	,	,	PUNCT
fcis-32489	9	15	allowing	allow	VERB
fcis-32489	9	16	for	for	ADP
fcis-32489	9	17	a	a	DET
fcis-32489	9	18	more	more	ADV
fcis-32489	9	19	detailed	detailed	ADJ
fcis-32489	9	20	modeling	modeling	NOUN
fcis-32489	9	21	of	of	ADP
fcis-32489	9	22	their	their	PRON
fcis-32489	9	23	interactions	interaction	NOUN
fcis-32489	9	24	.	.	PUNCT
fcis-32489	10	1	the	the	DET
fcis-32489	10	2	proposed	propose	VERB
fcis-32489	10	3	approach	approach	NOUN
fcis-32489	10	4	starts	start	VERB
fcis-32489	10	5	by	by	ADP
fcis-32489	10	6	identifying	identify	VERB
fcis-32489	10	7	coreferential	coreferential	ADJ
fcis-32489	10	8	links	link	NOUN
fcis-32489	10	9	to	to	PART
fcis-32489	10	10	gather	gather	VERB
fcis-32489	10	11	features	feature	NOUN
fcis-32489	10	12	that	that	PRON
fcis-32489	10	13	represent	represent	VERB
fcis-32489	10	14	the	the	DET
fcis-32489	10	15	relationships	relationship	NOUN
fcis-32489	10	16	between	between	ADP
fcis-32489	10	17	pairs	pair	NOUN
fcis-32489	10	18	of	of	ADP
fcis-32489	10	19	entities	entity	NOUN
fcis-32489	10	20	.	.	PUNCT
fcis-32489	11	1	these	these	DET
fcis-32489	11	2	features	feature	NOUN
fcis-32489	11	3	are	be	AUX
fcis-32489	11	4	then	then	ADV
fcis-32489	11	5	analyzed	analyze	VERB
fcis-32489	11	6	using	use	VERB
fcis-32489	11	7	gcn	gcn	NOUN
fcis-32489	11	8	to	to	PART
fcis-32489	11	9	construct	construct	VERB
fcis-32489	11	10	a	a	DET
fcis-32489	11	11	graph	graph	NOUN
fcis-32489	11	12	structure	structure	NOUN
fcis-32489	11	13	that	that	PRON
fcis-32489	11	14	represents	represent	VERB
fcis-32489	11	15	the	the	DET
fcis-32489	11	16	entire	entire	ADJ
fcis-32489	11	17	document	document	NOUN
fcis-32489	11	18	,	,	PUNCT
fcis-32489	11	19	ultimately	ultimately	ADV
fcis-32489	11	20	revealing	reveal	VERB
fcis-32489	11	21	the	the	DET
fcis-32489	11	22	complex	complex	ADJ
fcis-32489	11	23	interactions	interaction	NOUN
fcis-32489	11	24	between	between	ADP
fcis-32489	11	25	different	different	ADJ
fcis-32489	11	26	entities	entity	NOUN
fcis-32489	11	27	.	.	PUNCT
fcis-32489	12	1	testing	test	VERB
fcis-32489	12	2	this	this	DET
fcis-32489	12	3	model	model	NOUN
fcis-32489	12	4	on	on	ADP
fcis-32489	12	5	the	the	DET
fcis-32489	12	6	large	large	ADJ
fcis-32489	12	7	-	-	PUNCT
fcis-32489	12	8	scale	scale	NOUN
fcis-32489	12	9	docred	docre	VERB
fcis-32489	12	10	dataset	dataset	NOUN
fcis-32489	12	11	from	from	ADP
fcis-32489	12	12	tsinghua	tsinghua	PROPN
fcis-32489	12	13	university	university	PROPN
fcis-32489	12	14	has	have	AUX
fcis-32489	12	15	demonstrated	demonstrate	VERB
fcis-32489	12	16	its	its	PRON
fcis-32489	12	17	strong	strong	ADJ
fcis-32489	12	18	performance	performance	NOUN
fcis-32489	12	19	in	in	ADP
fcis-32489	12	20	this	this	DET
fcis-32489	12	21	challenging	challenging	ADJ
fcis-32489	12	22	task	task	NOUN
fcis-32489	12	23	.	.	PUNCT
fcis-32489	13	1	keywords	keyword	NOUN
fcis-32489	13	2	:	:	PUNCT
fcis-32489	13	3	document	document	NOUN
fcis-32489	13	4	-	-	PUNCT
fcis-32489	13	5	level	level	NOUN
fcis-32489	13	6	relation	relation	NOUN
fcis-32489	13	7	extraction	extraction	NOUN
fcis-32489	13	8	;	;	PUNCT
fcis-32489	13	9	graph	graph	VERB
fcis-32489	13	10	convolutional	convolutional	ADJ
fcis-32489	13	11	network	network	NOUN
fcis-32489	13	12	;	;	PUNCT
fcis-32489	13	13	docred	docre	VERB
fcis-32489	13	14	.	.	PUNCT
fcis-32489	14	1	1	1	X
fcis-32489	14	2	.	.	X
fcis-32489	14	3	introduction	introduction	NOUN
fcis-32489	14	4	at	at	ADP
fcis-32489	14	5	its	its	PRON
fcis-32489	14	6	core	core	NOUN
fcis-32489	14	7	,	,	PUNCT
fcis-32489	14	8	relation	relation	NOUN
fcis-32489	14	9	extraction	extraction	NOUN
fcis-32489	14	10	(	(	PUNCT
fcis-32489	14	11	re	re	NOUN
fcis-32489	14	12	)	)	PUNCT
fcis-32489	14	13	focuses	focus	VERB
fcis-32489	14	14	on	on	ADP
fcis-32489	14	15	discerning	discern	VERB
fcis-32489	14	16	meaningful	meaningful	ADJ
fcis-32489	14	17	connections	connection	NOUN
fcis-32489	14	18	between	between	ADP
fcis-32489	14	19	entities	entity	NOUN
fcis-32489	14	20	within	within	ADP
fcis-32489	14	21	unstructured	unstructured	ADJ
fcis-32489	14	22	text	text	NOUN
fcis-32489	15	1	[	[	X
fcis-32489	15	2	1	1	NUM
fcis-32489	15	3	]	]	PUNCT
fcis-32489	15	4	and	and	CCONJ
fcis-32489	15	5	presenting	present	VERB
fcis-32489	15	6	these	these	DET
fcis-32489	15	7	connections	connection	NOUN
fcis-32489	15	8	in	in	ADP
fcis-32489	15	9	a	a	DET
fcis-32489	15	10	structured	structured	ADJ
fcis-32489	15	11	format	format	NOUN
fcis-32489	15	12	.	.	PUNCT
fcis-32489	16	1	by	by	ADP
fcis-32489	16	2	converting	convert	VERB
fcis-32489	16	3	fragmented	fragmented	ADJ
fcis-32489	16	4	information	information	NOUN
fcis-32489	16	5	into	into	ADP
fcis-32489	16	6	organized	organized	ADJ
fcis-32489	16	7	knowledge	knowledge	NOUN
fcis-32489	16	8	that	that	PRON
fcis-32489	16	9	is	be	AUX
fcis-32489	16	10	readily	readily	ADV
fcis-32489	16	11	interpretable	interpretable	ADJ
fcis-32489	16	12	,	,	PUNCT
fcis-32489	16	13	re	re	VERB
fcis-32489	16	14	serves	serve	VERB
fcis-32489	16	15	as	as	ADP
fcis-32489	16	16	a	a	DET
fcis-32489	16	17	critical	critical	ADJ
fcis-32489	16	18	component	component	NOUN
fcis-32489	16	19	in	in	ADP
fcis-32489	16	20	various	various	ADJ
fcis-32489	16	21	natural	natural	ADJ
fcis-32489	16	22	language	language	NOUN
fcis-32489	16	23	processing	processing	NOUN
fcis-32489	16	24	(	(	PUNCT
fcis-32489	16	25	nlp	nlp	ADJ
fcis-32489	16	26	)	)	PUNCT
fcis-32489	16	27	applications	application	NOUN
fcis-32489	16	28	,	,	PUNCT
fcis-32489	16	29	including	include	VERB
fcis-32489	16	30	knowledge	knowledge	NOUN
fcis-32489	16	31	graph	graph	NOUN
fcis-32489	16	32	construction	construction	NOUN
fcis-32489	16	33	[	[	X
fcis-32489	16	34	2	2	NUM
fcis-32489	16	35	]	]	PUNCT
fcis-32489	16	36	,	,	PUNCT
fcis-32489	16	37	data	data	NOUN
fcis-32489	16	38	retrieval	retrieval	NOUN
fcis-32489	16	39	,	,	PUNCT
fcis-32489	16	40	question	question	NOUN
fcis-32489	16	41	-	-	PUNCT
fcis-32489	16	42	answering	answer	VERB
fcis-32489	16	43	systems	system	NOUN
fcis-32489	16	44	[	[	X
fcis-32489	16	45	3	3	NUM
fcis-32489	16	46	]	]	PUNCT
fcis-32489	16	47	,	,	PUNCT
fcis-32489	16	48	and	and	CCONJ
fcis-32489	16	49	conversational	conversational	ADJ
fcis-32489	16	50	interfaces	interface	NOUN
fcis-32489	16	51	[	[	X
fcis-32489	16	52	4	4	NUM
fcis-32489	16	53	]	]	PUNCT
fcis-32489	16	54	nowadays	nowadays	ADV
fcis-32489	16	55	,	,	PUNCT
fcis-32489	16	56	the	the	DET
fcis-32489	16	57	task	task	NOUN
fcis-32489	16	58	of	of	ADP
fcis-32489	16	59	extracting	extract	VERB
fcis-32489	16	60	relationships	relationship	NOUN
fcis-32489	16	61	between	between	ADP
fcis-32489	16	62	entities	entity	NOUN
fcis-32489	16	63	is	be	AUX
fcis-32489	16	64	primarily	primarily	ADV
fcis-32489	16	65	carried	carry	VERB
fcis-32489	16	66	out	out	ADP
fcis-32489	16	67	at	at	ADP
fcis-32489	16	68	the	the	DET
fcis-32489	16	69	sentence	sentence	NOUN
fcis-32489	16	70	level	level	NOUN
fcis-32489	16	71	,	,	PUNCT
fcis-32489	16	72	meaning	mean	VERB
fcis-32489	16	73	identifying	identify	VERB
fcis-32489	16	74	connections	connection	NOUN
fcis-32489	16	75	within	within	ADP
fcis-32489	16	76	a	a	DET
fcis-32489	16	77	single	single	ADJ
fcis-32489	16	78	sentence	sentence	NOUN
fcis-32489	16	79	.	.	PUNCT
fcis-32489	17	1	however	however	ADV
fcis-32489	17	2	,	,	PUNCT
fcis-32489	17	3	such	such	ADJ
fcis-32489	17	4	sentence	sentence	NOUN
fcis-32489	17	5	-	-	PUNCT
fcis-32489	17	6	based	base	VERB
fcis-32489	17	7	models	model	NOUN
fcis-32489	17	8	have	have	VERB
fcis-32489	17	9	an	an	DET
fcis-32489	17	10	inherent	inherent	ADJ
fcis-32489	17	11	drawback	drawback	NOUN
fcis-32489	17	12	:	:	PUNCT
fcis-32489	17	13	they	they	PRON
fcis-32489	17	14	fail	fail	VERB
fcis-32489	17	15	to	to	PART
fcis-32489	17	16	grasp	grasp	VERB
fcis-32489	17	17	the	the	DET
fcis-32489	17	18	relationships	relationship	NOUN
fcis-32489	17	19	existing	exist	VERB
fcis-32489	17	20	between	between	ADP
fcis-32489	17	21	entities	entity	NOUN
fcis-32489	17	22	across	across	ADP
fcis-32489	17	23	multiple	multiple	ADJ
fcis-32489	17	24	sentences	sentence	NOUN
fcis-32489	17	25	.	.	PUNCT
fcis-32489	18	1	as	as	ADP
fcis-32489	18	2	a	a	DET
fcis-32489	18	3	result	result	NOUN
fcis-32489	18	4	,	,	PUNCT
fcis-32489	18	5	to	to	PART
fcis-32489	18	6	fully	fully	ADV
fcis-32489	18	7	comprehend	comprehend	VERB
fcis-32489	18	8	the	the	DET
fcis-32489	18	9	information	information	NOUN
fcis-32489	18	10	contained	contain	VERB
fcis-32489	18	11	in	in	ADP
fcis-32489	18	12	a	a	DET
fcis-32489	18	13	text	text	NOUN
fcis-32489	18	14	,	,	PUNCT
fcis-32489	18	15	extracting	extract	VERB
fcis-32489	18	16	relations	relation	NOUN
fcis-32489	18	17	from	from	ADP
fcis-32489	18	18	an	an	DET
fcis-32489	18	19	entire	entire	ADJ
fcis-32489	18	20	document	document	NOUN
fcis-32489	18	21	becomes	become	VERB
fcis-32489	18	22	essential	essential	ADJ
fcis-32489	18	23	.	.	PUNCT
fcis-32489	19	1	in	in	ADP
fcis-32489	19	2	recent	recent	ADJ
fcis-32489	19	3	times	time	NOUN
fcis-32489	19	4	,	,	PUNCT
fcis-32489	19	5	yao	yao	PROPN
fcis-32489	19	6	et	et	PROPN
fcis-32489	19	7	al	al	PROPN
fcis-32489	19	8	.	.	PUNCT
fcis-32489	20	1	[	[	X
fcis-32489	20	2	5	5	NUM
fcis-32489	20	3	]	]	PUNCT
fcis-32489	20	4	introduced	introduce	VERB
fcis-32489	20	5	a	a	DET
fcis-32489	20	6	large	large	ADJ
fcis-32489	20	7	-	-	PUNCT
fcis-32489	20	8	scale	scale	NOUN
fcis-32489	20	9	dataset	dataset	NOUN
fcis-32489	20	10	that	that	PRON
fcis-32489	20	11	is	be	AUX
fcis-32489	20	12	manually	manually	ADV
fcis-32489	20	13	annotated	annotate	VERB
fcis-32489	20	14	,	,	PUNCT
fcis-32489	20	15	which	which	PRON
fcis-32489	20	16	expands	expand	VERB
fcis-32489	20	17	the	the	DET
fcis-32489	20	18	scope	scope	NOUN
fcis-32489	20	19	of	of	ADP
fcis-32489	20	20	sentence	sentence	NOUN
fcis-32489	20	21	-	-	PUNCT
fcis-32489	20	22	level	level	NOUN
fcis-32489	20	23	relationships	relationship	NOUN
fcis-32489	20	24	to	to	ADP
fcis-32489	20	25	the	the	DET
fcis-32489	20	26	document	document	NOUN
fcis-32489	20	27	context	context	NOUN
fcis-32489	20	28	.	.	PUNCT
fcis-32489	21	1	this	this	DET
fcis-32489	21	2	dataset	dataset	NOUN
fcis-32489	21	3	includes	include	VERB
fcis-32489	21	4	numerous	numerous	ADJ
fcis-32489	21	5	relational	relational	ADJ
fcis-32489	21	6	facts	fact	NOUN
fcis-32489	21	7	and	and	CCONJ
fcis-32489	21	8	demands	demand	NOUN
fcis-32489	21	9	that	that	SCONJ
fcis-32489	21	10	models	model	NOUN
fcis-32489	21	11	predict	predict	VERB
fcis-32489	21	12	the	the	DET
fcis-32489	21	13	connections	connection	NOUN
fcis-32489	21	14	between	between	ADP
fcis-32489	21	15	every	every	DET
fcis-32489	21	16	pair	pair	NOUN
fcis-32489	21	17	of	of	ADP
fcis-32489	21	18	entities	entity	NOUN
fcis-32489	21	19	present	present	ADJ
fcis-32489	21	20	in	in	ADP
fcis-32489	21	21	the	the	DET
fcis-32489	21	22	document	document	NOUN
fcis-32489	21	23	.	.	PUNCT
fcis-32489	22	1	this	this	DET
fcis-32489	22	2	new	new	ADJ
fcis-32489	22	3	setup	setup	NOUN
fcis-32489	22	4	presents	present	VERB
fcis-32489	22	5	greater	great	ADJ
fcis-32489	22	6	difficulties	difficulty	NOUN
fcis-32489	22	7	due	due	ADP
fcis-32489	22	8	to	to	ADP
fcis-32489	22	9	the	the	DET
fcis-32489	22	10	fact	fact	NOUN
fcis-32489	22	11	that	that	SCONJ
fcis-32489	22	12	many	many	ADJ
fcis-32489	22	13	relational	relational	ADJ
fcis-32489	22	14	facts	fact	NOUN
fcis-32489	22	15	are	be	AUX
fcis-32489	22	16	spread	spread	VERB
fcis-32489	22	17	across	across	ADP
fcis-32489	22	18	several	several	ADJ
fcis-32489	22	19	sentences	sentence	NOUN
fcis-32489	22	20	and	and	CCONJ
fcis-32489	22	21	complex	complex	ADJ
fcis-32489	22	22	interactions	interaction	NOUN
fcis-32489	22	23	among	among	ADP
fcis-32489	22	24	entities	entity	NOUN
fcis-32489	22	25	need	need	VERB
fcis-32489	22	26	to	to	PART
fcis-32489	22	27	be	be	AUX
fcis-32489	22	28	considered	consider	VERB
fcis-32489	22	29	and	and	CCONJ
fcis-32489	22	30	modeled	model	VERB
fcis-32489	22	31	.	.	PUNCT
fcis-32489	23	1	but	but	CCONJ
fcis-32489	23	2	document	document	NOUN
fcis-32489	23	3	-	-	PUNCT
fcis-32489	23	4	level	level	NOUN
fcis-32489	23	5	relation	relation	NOUN
fcis-32489	23	6	extraction	extraction	NOUN
fcis-32489	23	7	also	also	ADV
fcis-32489	23	8	faces	face	VERB
fcis-32489	23	9	two	two	NUM
fcis-32489	23	10	challenges	challenge	NOUN
fcis-32489	23	11	.	.	PUNCT
fcis-32489	24	1	first	first	ADV
fcis-32489	24	2	,	,	PUNCT
fcis-32489	24	3	the	the	DET
fcis-32489	24	4	relationship	relationship	NOUN
fcis-32489	24	5	between	between	ADP
fcis-32489	24	6	two	two	NUM
fcis-32489	24	7	entities	entity	NOUN
fcis-32489	24	8	may	may	AUX
fcis-32489	24	9	involve	involve	VERB
fcis-32489	24	10	multiple	multiple	ADJ
fcis-32489	24	11	different	different	ADJ
fcis-32489	24	12	sentences	sentence	NOUN
fcis-32489	24	13	,	,	PUNCT
fcis-32489	24	14	and	and	CCONJ
fcis-32489	24	15	the	the	DET
fcis-32489	24	16	relationship	relationship	NOUN
fcis-32489	24	17	between	between	ADP
fcis-32489	24	18	entities	entity	NOUN
fcis-32489	24	19	can	can	AUX
fcis-32489	24	20	not	not	PART
fcis-32489	24	21	be	be	AUX
fcis-32489	24	22	obtained	obtain	VERB
fcis-32489	24	23	based	base	VERB
fcis-32489	24	24	on	on	ADP
fcis-32489	24	25	just	just	ADV
fcis-32489	24	26	one	one	NUM
fcis-32489	24	27	sentence	sentence	NOUN
fcis-32489	24	28	.	.	PUNCT
fcis-32489	25	1	apparently	apparently	ADV
fcis-32489	25	2	,	,	PUNCT
fcis-32489	25	3	sentence	sentence	NOUN
fcis-32489	25	4	-	-	PUNCT
fcis-32489	25	5	level	level	NOUN
fcis-32489	25	6	relation	relation	NOUN
fcis-32489	25	7	extraction	extraction	NOUN
fcis-32489	25	8	methods	method	NOUN
fcis-32489	25	9	do	do	AUX
fcis-32489	25	10	not	not	PART
fcis-32489	25	11	apply	apply	VERB
fcis-32489	25	12	to	to	ADP
fcis-32489	25	13	document	document	NOUN
fcis-32489	25	14	-	-	PUNCT
fcis-32489	25	15	level	level	NOUN
fcis-32489	25	16	relation	relation	NOUN
fcis-32489	25	17	extraction	extraction	NOUN
fcis-32489	25	18	.	.	PUNCT
fcis-32489	26	1	secondly	secondly	ADV
fcis-32489	26	2	,	,	PUNCT
fcis-32489	26	3	the	the	DET
fcis-32489	26	4	same	same	ADJ
fcis-32489	26	5	entity	entity	NOUN
fcis-32489	26	6	may	may	AUX
fcis-32489	26	7	have	have	VERB
fcis-32489	26	8	different	different	ADJ
fcis-32489	26	9	names	name	NOUN
fcis-32489	26	10	in	in	ADP
fcis-32489	26	11	a	a	DET
fcis-32489	26	12	sentence	sentence	NOUN
fcis-32489	26	13	,	,	PUNCT
fcis-32489	26	14	that	that	ADV
fcis-32489	26	15	is	is	ADV
fcis-32489	26	16	,	,	PUNCT
fcis-32489	26	17	the	the	DET
fcis-32489	26	18	same	same	ADJ
fcis-32489	26	19	entity	entity	NOUN
fcis-32489	26	20	may	may	AUX
fcis-32489	26	21	be	be	AUX
fcis-32489	26	22	mentioned	mention	VERB
fcis-32489	26	23	in	in	ADP
fcis-32489	26	24	multiple	multiple	ADJ
fcis-32489	26	25	different	different	ADJ
fcis-32489	26	26	sentences	sentence	NOUN
fcis-32489	26	27	,	,	PUNCT
fcis-32489	26	28	so	so	ADV
fcis-32489	26	29	document	document	NOUN
fcis-32489	26	30	-	-	PUNCT
fcis-32489	26	31	level	level	NOUN
fcis-32489	26	32	relation	relation	NOUN
fcis-32489	26	33	extraction	extraction	NOUN
fcis-32489	26	34	has	have	VERB
fcis-32489	26	35	a	a	DET
fcis-32489	26	36	deeper	deep	ADJ
fcis-32489	26	37	level	level	NOUN
fcis-32489	26	38	of	of	ADP
fcis-32489	26	39	feature	feature	NOUN
fcis-32489	26	40	extraction	extraction	NOUN
fcis-32489	26	41	to	to	PART
fcis-32489	26	42	aggregate	aggregate	VERB
fcis-32489	26	43	the	the	DET
fcis-32489	26	44	context	context	NOUN
fcis-32489	26	45	information	information	NOUN
fcis-32489	26	46	of	of	ADP
fcis-32489	26	47	the	the	DET
fcis-32489	26	48	entity	entity	NOUN
fcis-32489	26	49	.	.	PUNCT
fcis-32489	27	1	based	base	VERB
fcis-32489	27	2	on	on	ADP
fcis-32489	27	3	these	these	DET
fcis-32489	27	4	two	two	NUM
fcis-32489	27	5	challenges	challenge	NOUN
fcis-32489	27	6	,	,	PUNCT
fcis-32489	27	7	this	this	DET
fcis-32489	27	8	paper	paper	NOUN
fcis-32489	27	9	proposes	propose	VERB
fcis-32489	27	10	a	a	DET
fcis-32489	27	11	document	document	NOUN
fcis-32489	27	12	-	-	PUNCT
fcis-32489	27	13	level	level	NOUN
fcis-32489	27	14	relation	relation	NOUN
fcis-32489	27	15	extraction	extraction	NOUN
fcis-32489	27	16	method	method	NOUN
fcis-32489	27	17	based	base	VERB
fcis-32489	27	18	on	on	ADP
fcis-32489	27	19	graph	graph	NOUN
fcis-32489	27	20	convolutional	convolutional	ADJ
fcis-32489	27	21	neural	neural	ADJ
fcis-32489	27	22	networks	network	NOUN
fcis-32489	27	23	.	.	PUNCT
fcis-32489	28	1	the	the	DET
fcis-32489	28	2	model	model	NOUN
fcis-32489	28	3	uses	use	VERB
fcis-32489	28	4	coreference	coreference	NOUN
fcis-32489	28	5	relations	relation	NOUN
fcis-32489	28	6	at	at	ADP
fcis-32489	28	7	the	the	DET
fcis-32489	28	8	input	input	NOUN
fcis-32489	28	9	layer	layer	NOUN
fcis-32489	28	10	to	to	PART
fcis-32489	28	11	address	address	VERB
fcis-32489	28	12	the	the	DET
fcis-32489	28	13	issue	issue	NOUN
fcis-32489	28	14	of	of	ADP
fcis-32489	28	15	the	the	DET
fcis-32489	28	16	same	same	ADJ
fcis-32489	28	17	entity	entity	NOUN
fcis-32489	28	18	being	be	AUX
fcis-32489	28	19	mentioned	mention	VERB
fcis-32489	28	20	multiple	multiple	ADJ
fcis-32489	28	21	times	time	NOUN
fcis-32489	28	22	in	in	ADP
fcis-32489	28	23	a	a	DET
fcis-32489	28	24	sentence	sentence	NOUN
fcis-32489	28	25	,	,	PUNCT
fcis-32489	28	26	further	far	ADV
fcis-32489	28	27	extracting	extract	VERB
fcis-32489	28	28	the	the	DET
fcis-32489	28	29	deep	deep	ADJ
fcis-32489	28	30	feature	feature	NOUN
fcis-32489	28	31	information	information	NOUN
fcis-32489	28	32	of	of	ADP
fcis-32489	28	33	the	the	DET
fcis-32489	28	34	entity	entity	NOUN
fcis-32489	28	35	.	.	PUNCT
fcis-32489	29	1	the	the	DET
fcis-32489	29	2	graph	graph	NOUN
fcis-32489	29	3	convolution	convolution	NOUN
fcis-32489	29	4	model	model	NOUN
fcis-32489	29	5	overcomes	overcome	VERB
fcis-32489	29	6	the	the	DET
fcis-32489	29	7	problem	problem	NOUN
fcis-32489	29	8	that	that	SCONJ
fcis-32489	29	9	traditional	traditional	ADJ
fcis-32489	29	10	deep	deep	ADJ
fcis-32489	29	11	learning	learning	NOUN
fcis-32489	29	12	models	model	NOUN
fcis-32489	29	13	can	can	AUX
fcis-32489	29	14	not	not	PART
fcis-32489	29	15	aggregate	aggregate	VERB
fcis-32489	29	16	entity	entity	NOUN
fcis-32489	29	17	context	context	NOUN
fcis-32489	29	18	information	information	NOUN
fcis-32489	29	19	,	,	PUNCT
fcis-32489	29	20	effectively	effectively	ADV
fcis-32489	29	21	controlling	control	VERB
fcis-32489	29	22	the	the	DET
fcis-32489	29	23	impact	impact	NOUN
fcis-32489	29	24	of	of	ADP
fcis-32489	29	25	redundant	redundant	ADJ
fcis-32489	29	26	data	datum	NOUN
fcis-32489	29	27	on	on	ADP
fcis-32489	29	28	the	the	DET
fcis-32489	29	29	experimental	experimental	ADJ
fcis-32489	29	30	results	result	NOUN
fcis-32489	29	31	.	.	PUNCT
fcis-32489	30	1	finally	finally	ADV
fcis-32489	30	2	,	,	PUNCT
fcis-32489	30	3	the	the	DET
fcis-32489	30	4	model	model	NOUN
fcis-32489	30	5	was	be	AUX
fcis-32489	30	6	evaluated	evaluate	VERB
fcis-32489	30	7	on	on	ADP
fcis-32489	30	8	the	the	DET
fcis-32489	30	9	docred	docre	VERB
fcis-32489	30	10	dataset	dataset	NOUN
fcis-32489	30	11	,	,	PUNCT
fcis-32489	30	12	and	and	CCONJ
fcis-32489	30	13	the	the	DET
fcis-32489	30	14	results	result	NOUN
fcis-32489	30	15	showed	show	VERB
fcis-32489	30	16	that	that	SCONJ
fcis-32489	30	17	compared	compare	VERB
fcis-32489	30	18	with	with	ADP
fcis-32489	30	19	existing	exist	VERB
fcis-32489	30	20	methods	method	NOUN
fcis-32489	30	21	,	,	PUNCT
fcis-32489	30	22	the	the	DET
fcis-32489	30	23	model	model	NOUN
fcis-32489	30	24	made	make	VERB
fcis-32489	30	25	some	some	DET
fcis-32489	30	26	significant	significant	ADJ
fcis-32489	30	27	progress	progress	NOUN
fcis-32489	30	28	in	in	ADP
fcis-32489	30	29	relation	relation	NOUN
fcis-32489	30	30	extraction	extraction	NOUN
fcis-32489	30	31	.	.	PUNCT
fcis-32489	31	1	2	2	X
fcis-32489	31	2	.	.	X
fcis-32489	31	3	related	relate	VERB
fcis-32489	31	4	work	work	NOUN
fcis-32489	31	5	relation	relation	NOUN
fcis-32489	31	6	extraction	extraction	NOUN
fcis-32489	31	7	tasks	task	NOUN
fcis-32489	31	8	are	be	AUX
fcis-32489	31	9	often	often	ADV
fcis-32489	31	10	regarded	regard	VERB
fcis-32489	31	11	as	as	ADP
fcis-32489	31	12	multiclassification	multiclassification	NOUN
fcis-32489	31	13	problems	problem	NOUN
fcis-32489	31	14	,	,	PUNCT
fcis-32489	31	15	and	and	CCONJ
fcis-32489	31	16	the	the	DET
fcis-32489	31	17	most	most	ADV
fcis-32489	31	18	representative	representative	NOUN
fcis-32489	31	19	of	of	ADP
fcis-32489	31	20	traditional	traditional	ADJ
fcis-32489	31	21	methods	method	NOUN
fcis-32489	31	22	is	be	AUX
fcis-32489	31	23	the	the	DET
fcis-32489	31	24	eigenvector	eigenvector	NOUN
fcis-32489	31	25	-	-	PUNCT
fcis-32489	31	26	based	base	VERB
fcis-32489	31	27	approach	approach	NOUN
fcis-32489	31	28	[	[	X
fcis-32489	31	29	6	6	NUM
fcis-32489	31	30	]	]	PUNCT
fcis-32489	31	31	,	,	PUNCT
fcis-32489	31	32	which	which	PRON
fcis-32489	31	33	performs	perform	VERB
fcis-32489	31	34	relation	relation	NOUN
fcis-32489	31	35	extraction	extraction	NOUN
fcis-32489	31	36	by	by	ADP
fcis-32489	31	37	modeling	model	VERB
fcis-32489	31	38	the	the	DET
fcis-32489	31	39	eigenvectors	eigenvector	NOUN
fcis-32489	31	40	;	;	PUNCT
fcis-32489	31	41	kernel	kernel	PROPN
fcis-32489	31	42	function	function	NOUN
fcis-32489	31	43	-	-	PUNCT
fcis-32489	31	44	based	base	VERB
fcis-32489	31	45	methods	method	NOUN
fcis-32489	31	46	utilize	utilize	VERB
fcis-32489	31	47	the	the	DET
fcis-32489	31	48	structural	structural	ADJ
fcis-32489	31	49	information	information	NOUN
fcis-32489	31	50	of	of	ADP
fcis-32489	31	51	the	the	DET
fcis-32489	31	52	corpus	corpus	NOUN
fcis-32489	31	53	itself	itself	PRON
fcis-32489	31	54	to	to	PART
fcis-32489	31	55	achieve	achieve	VERB
fcis-32489	31	56	relation	relation	NOUN
fcis-32489	31	57	extraction	extraction	NOUN
fcis-32489	31	58	by	by	ADP
fcis-32489	31	59	calculating	calculate	VERB
fcis-32489	31	60	similarity	similarity	NOUN
fcis-32489	32	1	[	[	X
fcis-32489	32	2	7	7	NUM
fcis-32489	32	3	]	]	X
fcis-32489	32	4	;	;	PUNCT
fcis-32489	32	5	deep	deep	ADJ
fcis-32489	32	6	learning	learning	NOUN
fcis-32489	32	7	methods	method	NOUN
fcis-32489	32	8	based	base	VERB
fcis-32489	32	9	on	on	ADP
fcis-32489	32	10	neural	neural	ADJ
fcis-32489	32	11	network	network	NOUN
fcis-32489	32	12	models	model	NOUN
fcis-32489	32	13	[	[	X
fcis-32489	32	14	8	8	NUM
fcis-32489	32	15	]	]	PUNCT
fcis-32489	32	16	,	,	PUNCT
fcis-32489	32	17	such	such	ADJ
fcis-32489	32	18	as	as	ADP
fcis-32489	32	19	convolutional	convolutional	ADJ
fcis-32489	32	20	neural	neural	ADJ
fcis-32489	32	21	networks	network	NOUN
fcis-32489	32	22	(	(	PUNCT
fcis-32489	32	23	cnns)[9	cnns)[9	NUM
fcis-32489	32	24	]	]	PUNCT
fcis-32489	32	25	and	and	CCONJ
fcis-32489	32	26	recurrent	recurrent	ADJ
fcis-32489	32	27	neural	neural	ADJ
fcis-32489	32	28	networks	network	NOUN
fcis-32489	32	29	(	(	PUNCT
fcis-32489	32	30	rnns)[10	rnns)[10	PROPN
fcis-32489	32	31	]	]	PUNCT
fcis-32489	32	32	,	,	PUNCT
fcis-32489	32	33	can	can	AUX
fcis-32489	32	34	automatically	automatically	ADV
fcis-32489	32	35	learn	learn	VERB
fcis-32489	32	36	sentence	sentence	NOUN
fcis-32489	32	37	features	feature	NOUN
fcis-32489	32	38	,	,	PUNCT
fcis-32489	32	39	thus	thus	ADV
fcis-32489	32	40	avoiding	avoid	VERB
fcis-32489	32	41	problems	problem	NOUN
fcis-32489	32	42	such	such	ADJ
fcis-32489	32	43	as	as	ADP
fcis-32489	32	44	error	error	NOUN
fcis-32489	32	45	propagation	propagation	NOUN
fcis-32489	32	46	brought	bring	VERB
fcis-32489	32	47	by	by	ADP
fcis-32489	32	48	nlp	nlp	ADJ
fcis-32489	32	49	tools	tool	NOUN
fcis-32489	32	50	[	[	X
fcis-32489	32	51	11	11	NUM
fcis-32489	32	52	]	]	PUNCT
fcis-32489	32	53	,	,	PUNCT
fcis-32489	32	54	and	and	CCONJ
fcis-32489	32	55	also	also	ADV
fcis-32489	32	56	improving	improve	VERB
fcis-32489	32	57	the	the	DET
fcis-32489	32	58	effect	effect	NOUN
fcis-32489	32	59	of	of	ADP
fcis-32489	32	60	relation	relation	NOUN
fcis-32489	32	61	extraction	extraction	NOUN
fcis-32489	32	62	.	.	PUNCT
fcis-32489	33	1	however	however	ADV
fcis-32489	33	2	,	,	PUNCT
fcis-32489	33	3	these	these	DET
fcis-32489	33	4	methods	method	NOUN
fcis-32489	33	5	all	all	PRON
fcis-32489	33	6	require	require	VERB
fcis-32489	33	7	a	a	DET
fcis-32489	33	8	large	large	ADJ
fcis-32489	33	9	amount	amount	NOUN
fcis-32489	33	10	of	of	ADP
fcis-32489	33	11	manually	manually	ADV
fcis-32489	33	12	labeled	label	VERB
fcis-32489	33	13	corpora	corpora	PROPN
fcis-32489	33	14	,	,	PUNCT
fcis-32489	33	15	which	which	PRON
fcis-32489	33	16	limits	limit	VERB
fcis-32489	33	17	the	the	DET
fcis-32489	33	18	development	development	NOUN
fcis-32489	33	19	of	of	ADP
fcis-32489	33	20	relation	relation	NOUN
fcis-32489	33	21	extraction	extraction	NOUN
fcis-32489	33	22	.	.	PUNCT
fcis-32489	34	1	to	to	PART
fcis-32489	34	2	address	address	VERB
fcis-32489	34	3	the	the	DET
fcis-32489	34	4	severe	severe	ADJ
fcis-32489	34	5	shortage	shortage	NOUN
fcis-32489	34	6	of	of	ADP
fcis-32489	34	7	manually	manually	ADV
fcis-32489	34	8	annotated	annotate	VERB
fcis-32489	34	9	35	35	NUM
fcis-32489	34	10	corpora	corpus	NOUN
fcis-32489	34	11	,	,	PUNCT
fcis-32489	34	12	mintz	mintz	PROPN
fcis-32489	34	13	et	et	PROPN
fcis-32489	34	14	al	al	PROPN
fcis-32489	34	15	.	.	PUNCT
fcis-32489	35	1	[	[	X
fcis-32489	35	2	12	12	NUM
fcis-32489	35	3	]	]	PUNCT
fcis-32489	35	4	proposed	propose	VERB
fcis-32489	35	5	a	a	DET
fcis-32489	35	6	remote	remote	ADJ
fcis-32489	35	7	supervision	supervision	NOUN
fcis-32489	35	8	method	method	NOUN
fcis-32489	35	9	that	that	PRON
fcis-32489	35	10	does	do	AUX
fcis-32489	35	11	not	not	PART
fcis-32489	35	12	require	require	VERB
fcis-32489	35	13	manual	manual	ADJ
fcis-32489	35	14	annotation	annotation	NOUN
fcis-32489	35	15	,	,	PUNCT
fcis-32489	35	16	that	that	ADV
fcis-32489	35	17	is	is	ADV
fcis-32489	35	18	,	,	PUNCT
fcis-32489	35	19	using	use	VERB
fcis-32489	35	20	the	the	DET
fcis-32489	35	21	freebase	freebase	ADJ
fcis-32489	35	22	knowledge	knowledge	NOUN
fcis-32489	35	23	base	base	NOUN
fcis-32489	35	24	and	and	CCONJ
fcis-32489	35	25	wikipedia	wikipedia	PROPN
fcis-32489	35	26	text	text	PROPN
fcis-32489	35	27	for	for	ADP
fcis-32489	35	28	alignment	alignment	NOUN
fcis-32489	35	29	to	to	PART
fcis-32489	35	30	obtain	obtain	VERB
fcis-32489	35	31	large	large	ADJ
fcis-32489	35	32	-	-	PUNCT
fcis-32489	35	33	scale	scale	NOUN
fcis-32489	35	34	relation	relation	NOUN
fcis-32489	35	35	triples	triple	NOUN
fcis-32489	35	36	.	.	PUNCT
fcis-32489	36	1	however	however	ADV
fcis-32489	36	2	,	,	PUNCT
fcis-32489	36	3	the	the	DET
fcis-32489	36	4	method	method	NOUN
fcis-32489	36	5	is	be	AUX
fcis-32489	36	6	prone	prone	ADJ
fcis-32489	36	7	to	to	PART
fcis-32489	36	8	noise	noise	VERB
fcis-32489	36	9	annotation	annotation	NOUN
fcis-32489	36	10	problems	problem	NOUN
fcis-32489	36	11	,	,	PUNCT
fcis-32489	36	12	so	so	ADV
fcis-32489	36	13	filtering	filtering	NOUN
fcis-32489	36	14	noise	noise	NOUN
fcis-32489	36	15	annotations	annotation	NOUN
fcis-32489	36	16	has	have	AUX
fcis-32489	36	17	become	become	VERB
fcis-32489	36	18	the	the	DET
fcis-32489	36	19	focus	focus	NOUN
fcis-32489	36	20	of	of	ADP
fcis-32489	36	21	remote	remote	ADJ
fcis-32489	36	22	supervision	supervision	NOUN
fcis-32489	36	23	methods	method	NOUN
fcis-32489	36	24	.	.	PUNCT
fcis-32489	37	1	to	to	PART
fcis-32489	37	2	address	address	VERB
fcis-32489	37	3	the	the	DET
fcis-32489	37	4	challenges	challenge	NOUN
fcis-32489	37	5	posed	pose	VERB
fcis-32489	37	6	by	by	ADP
fcis-32489	37	7	imprecise	imprecise	ADJ
fcis-32489	37	8	annotations	annotation	NOUN
fcis-32489	37	9	and	and	CCONJ
fcis-32489	37	10	noisy	noisy	ADJ
fcis-32489	37	11	data	datum	NOUN
fcis-32489	37	12	in	in	ADP
fcis-32489	37	13	relation	relation	NOUN
fcis-32489	37	14	extraction	extraction	NOUN
fcis-32489	37	15	tasks	task	NOUN
fcis-32489	37	16	,	,	PUNCT
fcis-32489	37	17	researchers	researcher	NOUN
fcis-32489	37	18	have	have	AUX
fcis-32489	37	19	explored	explore	VERB
fcis-32489	37	20	various	various	ADJ
fcis-32489	37	21	approaches	approach	NOUN
fcis-32489	37	22	.	.	PUNCT
fcis-32489	38	1	earlier	early	ADJ
fcis-32489	38	2	works	work	NOUN
fcis-32489	38	3	,	,	PUNCT
fcis-32489	38	4	such	such	ADJ
fcis-32489	38	5	as	as	ADP
fcis-32489	38	6	those	those	PRON
fcis-32489	38	7	by	by	ADP
fcis-32489	38	8	hoffmann	hoffmann	PROPN
fcis-32489	38	9	et	et	PROPN
fcis-32489	38	10	al	al	PROPN
fcis-32489	38	11	.	.	PUNCT
fcis-32489	39	1	[	[	X
fcis-32489	39	2	13	13	NUM
fcis-32489	39	3	]	]	PUNCT
fcis-32489	39	4	and	and	CCONJ
fcis-32489	39	5	surdeanu	surdeanu	NOUN
fcis-32489	39	6	et	et	PROPN
fcis-32489	39	7	al	al	PROPN
fcis-32489	39	8	.	.	PUNCT
fcis-32489	40	1	[	[	X
fcis-32489	40	2	14	14	NUM
fcis-32489	40	3	]	]	PUNCT
fcis-32489	40	4	,	,	PUNCT
fcis-32489	40	5	introduced	introduce	VERB
fcis-32489	40	6	techniques	technique	NOUN
fcis-32489	40	7	like	like	ADP
fcis-32489	40	8	multi	multi	ADJ
fcis-32489	40	9	-	-	NOUN
fcis-32489	40	10	instance	instance	NOUN
fcis-32489	40	11	learning	learning	NOUN
fcis-32489	40	12	and	and	CCONJ
fcis-32489	40	13	multiinstance	multiinstance	NOUN
fcis-32489	40	14	multi	multi	ADJ
fcis-32489	40	15	-	-	ADJ
fcis-32489	40	16	label	label	ADJ
fcis-32489	40	17	methods	method	NOUN
fcis-32489	40	18	to	to	PART
fcis-32489	40	19	mitigate	mitigate	VERB
fcis-32489	40	20	the	the	DET
fcis-32489	40	21	impact	impact	NOUN
fcis-32489	40	22	of	of	ADP
fcis-32489	40	23	incorrect	incorrect	ADJ
fcis-32489	40	24	labeling	labeling	NOUN
fcis-32489	40	25	.	.	PUNCT
fcis-32489	41	1	additionally	additionally	ADV
fcis-32489	41	2	,	,	PUNCT
fcis-32489	41	3	benjamin	benjamin	PROPN
fcis-32489	41	4	et	et	PROPN
fcis-32489	41	5	al	al	PROPN
fcis-32489	41	6	.	.	PUNCT
fcis-32489	42	1	[	[	X
fcis-32489	42	2	15	15	NUM
fcis-32489	42	3	]	]	PUNCT
fcis-32489	42	4	combined	combine	VERB
fcis-32489	42	5	topic	topic	NOUN
fcis-32489	42	6	modeling	modeling	NOUN
fcis-32489	42	7	(	(	PUNCT
fcis-32489	42	8	lda	lda	PROPN
fcis-32489	42	9	)	)	PUNCT
fcis-32489	42	10	with	with	ADP
fcis-32489	42	11	decision	decision	NOUN
fcis-32489	42	12	learning	learn	VERB
fcis-32489	42	13	to	to	PART
fcis-32489	42	14	reduce	reduce	VERB
fcis-32489	42	15	noise	noise	NOUN
fcis-32489	42	16	in	in	ADP
fcis-32489	42	17	remote	remote	ADJ
fcis-32489	42	18	supervision	supervision	NOUN
fcis-32489	42	19	results	result	NOUN
fcis-32489	42	20	.	.	PUNCT
fcis-32489	43	1	in	in	ADP
fcis-32489	43	2	recent	recent	ADJ
fcis-32489	43	3	years	year	NOUN
fcis-32489	43	4	,	,	PUNCT
fcis-32489	43	5	deep	deep	ADJ
fcis-32489	43	6	learning	learning	NOUN
fcis-32489	43	7	has	have	AUX
fcis-32489	43	8	emerged	emerge	VERB
fcis-32489	43	9	as	as	ADP
fcis-32489	43	10	a	a	DET
fcis-32489	43	11	powerful	powerful	ADJ
fcis-32489	43	12	tool	tool	NOUN
fcis-32489	43	13	in	in	ADP
fcis-32489	43	14	this	this	DET
fcis-32489	43	15	field	field	NOUN
fcis-32489	43	16	,	,	PUNCT
fcis-32489	43	17	with	with	ADP
fcis-32489	43	18	studies	study	NOUN
fcis-32489	43	19	like	like	ADP
fcis-32489	43	20	zeng	zeng	PROPN
fcis-32489	43	21	et	et	PROPN
fcis-32489	43	22	al	al	PROPN
fcis-32489	43	23	.	.	PUNCT
fcis-32489	44	1	[	[	X
fcis-32489	44	2	16	16	NUM
fcis-32489	44	3	]	]	X
fcis-32489	44	4	utilizing	utilize	VERB
fcis-32489	44	5	convolutional	convolutional	ADJ
fcis-32489	44	6	networks	network	NOUN
fcis-32489	44	7	with	with	ADP
fcis-32489	44	8	segmented	segment	VERB
fcis-32489	44	9	max	max	PROPN
fcis-32489	44	10	pooling	pool	VERB
fcis-32489	44	11	to	to	PART
fcis-32489	44	12	automatically	automatically	ADV
fcis-32489	44	13	extract	extract	VERB
fcis-32489	44	14	sentence	sentence	NOUN
fcis-32489	44	15	features	feature	NOUN
fcis-32489	44	16	,	,	PUNCT
fcis-32489	44	17	followed	follow	VERB
fcis-32489	44	18	by	by	ADP
fcis-32489	44	19	integrating	integrate	VERB
fcis-32489	44	20	multiinstance	multiinstance	NOUN
fcis-32489	44	21	learning	learn	VERB
fcis-32489	44	22	for	for	ADP
fcis-32489	44	23	remote	remote	ADJ
fcis-32489	44	24	supervised	supervise	VERB
fcis-32489	44	25	relation	relation	NOUN
fcis-32489	44	26	extraction	extraction	NOUN
fcis-32489	44	27	;	;	PUNCT
fcis-32489	44	28	lei	lei	X
fcis-32489	44	29	et	et	NOUN
fcis-32489	44	30	al.[17	al.[17	PROPN
fcis-32489	44	31	]	]	PUNCT
fcis-32489	44	32	proposed	propose	VERB
fcis-32489	44	33	a	a	DET
fcis-32489	44	34	neural	neural	ADJ
fcis-32489	44	35	relation	relation	NOUN
fcis-32489	44	36	extraction	extraction	NOUN
fcis-32489	44	37	framework	framework	NOUN
fcis-32489	44	38	with	with	ADP
fcis-32489	44	39	bidirectional	bidirectional	ADJ
fcis-32489	44	40	knowledge	knowledge	NOUN
fcis-32489	44	41	distillation	distillation	NOUN
fcis-32489	44	42	in	in	ADP
fcis-32489	44	43	order	order	NOUN
fcis-32489	44	44	to	to	PART
fcis-32489	44	45	effectively	effectively	ADV
fcis-32489	44	46	simulate	simulate	VERB
fcis-32489	44	47	the	the	DET
fcis-32489	44	48	relationship	relationship	NOUN
fcis-32489	44	49	patterns	pattern	NOUN
fcis-32489	44	50	between	between	ADP
fcis-32489	44	51	text	text	NOUN
fcis-32489	44	52	corpora	corpora	NOUN
fcis-32489	44	53	and	and	CCONJ
fcis-32489	44	54	knowledge	knowledge	NOUN
fcis-32489	44	55	graph	graph	NOUN
fcis-32489	44	56	information	information	NOUN
fcis-32489	44	57	,	,	PUNCT
fcis-32489	44	58	which	which	PRON
fcis-32489	44	59	can	can	AUX
fcis-32489	44	60	use	use	VERB
fcis-32489	44	61	different	different	ADJ
fcis-32489	44	62	information	information	NOUN
fcis-32489	44	63	sources	source	NOUN
fcis-32489	44	64	in	in	ADP
fcis-32489	44	65	coordination	coordination	NOUN
fcis-32489	44	66	to	to	PART
fcis-32489	44	67	reduce	reduce	VERB
fcis-32489	44	68	the	the	DET
fcis-32489	44	69	noise	noise	NOUN
fcis-32489	44	70	label	label	NOUN
fcis-32489	44	71	problem	problem	NOUN
fcis-32489	44	72	in	in	ADP
fcis-32489	44	73	remote	remote	ADJ
fcis-32489	44	74	supervision	supervision	NOUN
fcis-32489	44	75	.	.	PUNCT
fcis-32489	45	1	but	but	CCONJ
fcis-32489	45	2	these	these	DET
fcis-32489	45	3	methods	method	NOUN
fcis-32489	45	4	did	do	AUX
fcis-32489	45	5	not	not	PART
fcis-32489	45	6	take	take	VERB
fcis-32489	45	7	into	into	ADP
fcis-32489	45	8	account	account	NOUN
fcis-32489	45	9	relationships	relationship	NOUN
fcis-32489	45	10	among	among	ADP
fcis-32489	45	11	multiple	multiple	ADJ
fcis-32489	45	12	sentences	sentence	NOUN
fcis-32489	45	13	.	.	PUNCT
fcis-32489	46	1	peng	peng	PROPN
fcis-32489	46	2	et	et	PROPN
fcis-32489	46	3	al	al	PROPN
fcis-32489	46	4	.	.	PUNCT
fcis-32489	47	1	[	[	X
fcis-32489	47	2	18	18	NUM
fcis-32489	47	3	]	]	PUNCT
fcis-32489	47	4	studied	study	VERB
fcis-32489	47	5	a	a	DET
fcis-32489	47	6	graphbased	graphbased	ADJ
fcis-32489	47	7	lstm	lstm	NOUN
fcis-32489	47	8	framework	framework	NOUN
fcis-32489	47	9	capable	capable	ADJ
fcis-32489	47	10	of	of	ADP
fcis-32489	47	11	extracting	extract	VERB
fcis-32489	47	12	cross	cross	NOUN
fcis-32489	47	13	-	-	NOUN
fcis-32489	47	14	sentence	sentence	ADJ
fcis-32489	47	15	n	n	CCONJ
fcis-32489	47	16	-	-	PUNCT
fcis-32489	47	17	element	element	NOUN
fcis-32489	47	18	relationships	relationship	NOUN
fcis-32489	47	19	.	.	PUNCT
fcis-32489	48	1	docred	docre	VERB
fcis-32489	48	2	,	,	PUNCT
fcis-32489	48	3	proposed	propose	VERB
fcis-32489	48	4	by	by	ADP
fcis-32489	48	5	yao	yao	PROPN
fcis-32489	48	6	et	et	PROPN
fcis-32489	48	7	al	al	PROPN
fcis-32489	48	8	.	.	PUNCT
fcis-32489	49	1	[	[	X
fcis-32489	49	2	5	5	NUM
fcis-32489	49	3	]	]	PUNCT
fcis-32489	49	4	,	,	PUNCT
fcis-32489	49	5	is	be	AUX
fcis-32489	49	6	based	base	VERB
fcis-32489	49	7	on	on	ADP
fcis-32489	49	8	large	large	ADJ
fcis-32489	49	9	document	document	NOUN
fcis-32489	49	10	-	-	PUNCT
fcis-32489	49	11	level	level	NOUN
fcis-32489	49	12	relation	relation	NOUN
fcis-32489	49	13	extraction	extraction	NOUN
fcis-32489	49	14	datasets	dataset	NOUN
fcis-32489	49	15	constructed	construct	VERB
fcis-32489	49	16	from	from	ADP
fcis-32489	49	17	wikipedia	wikipedia	PROPN
fcis-32489	49	18	and	and	CCONJ
fcis-32489	49	19	wikidata	wikidata	NOUN
fcis-32489	49	20	and	and	CCONJ
fcis-32489	49	21	tests	test	VERB
fcis-32489	49	22	several	several	ADJ
fcis-32489	49	23	state	state	NOUN
fcis-32489	49	24	-	-	PUNCT
fcis-32489	49	25	of	of	ADP
fcis-32489	49	26	-	-	PUNCT
fcis-32489	49	27	the	the	DET
fcis-32489	49	28	-	-	PUNCT
fcis-32489	49	29	art	art	NOUN
fcis-32489	49	30	neural	neural	ADJ
fcis-32489	49	31	network	network	NOUN
fcis-32489	49	32	models	model	NOUN
fcis-32489	49	33	on	on	ADP
fcis-32489	49	34	them	they	PRON
fcis-32489	49	35	.	.	PUNCT
fcis-32489	50	1	wang	wang	PROPN
fcis-32489	50	2	et	et	PROPN
fcis-32489	50	3	al	al	PROPN
fcis-32489	50	4	.	.	PUNCT
fcis-32489	51	1	[	[	X
fcis-32489	51	2	19	19	NUM
fcis-32489	51	3	]	]	PUNCT
fcis-32489	51	4	used	use	VERB
fcis-32489	51	5	transformer	transformer	NOUN
fcis-32489	51	6	[	[	X
fcis-32489	51	7	20	20	NUM
fcis-32489	51	8	]	]	X
fcis-32489	51	9	text	text	NOUN
fcis-32489	51	10	encoding	encoding	NOUN
fcis-32489	51	11	into	into	ADP
fcis-32489	51	12	context	context	NOUN
fcis-32489	51	13	representations	representation	NOUN
fcis-32489	51	14	and	and	CCONJ
fcis-32489	51	15	fine	fine	ADV
fcis-32489	51	16	-	-	PUNCT
fcis-32489	51	17	tuned	tune	VERB
fcis-32489	51	18	with	with	ADP
fcis-32489	51	19	bert	bert	PROPN
fcis-32489	51	20	to	to	PART
fcis-32489	51	21	adopt	adopt	VERB
fcis-32489	51	22	a	a	DET
fcis-32489	51	23	two	two	NUM
fcis-32489	51	24	-	-	PUNCT
fcis-32489	51	25	step	step	NOUN
fcis-32489	51	26	strategy	strategy	NOUN
fcis-32489	51	27	on	on	ADP
fcis-32489	51	28	docred	docre	VERB
fcis-32489	51	29	to	to	PART
fcis-32489	51	30	improve	improve	VERB
fcis-32489	51	31	the	the	DET
fcis-32489	51	32	model	model	NOUN
fcis-32489	51	33	's	's	PART
fcis-32489	51	34	performance	performance	NOUN
fcis-32489	51	35	.	.	PUNCT
fcis-32489	52	1	gcn	gcn	PROPN
fcis-32489	52	2	was	be	AUX
fcis-32489	52	3	first	first	ADV
fcis-32489	52	4	applied	apply	VERB
fcis-32489	52	5	by	by	ADP
fcis-32489	52	6	kipf	kipf	NOUN
fcis-32489	52	7	and	and	CCONJ
fcis-32489	52	8	welling[21	welling[21	PROPN
fcis-32489	52	9	]	]	PUNCT
fcis-32489	52	10	on	on	ADP
fcis-32489	52	11	citation	citation	NOUN
fcis-32489	52	12	networks	network	NOUN
fcis-32489	52	13	and	and	CCONJ
fcis-32489	52	14	knowledge	knowledge	NOUN
fcis-32489	52	15	graph	graph	NOUN
fcis-32489	52	16	datasets	dataset	NOUN
fcis-32489	52	17	.	.	PUNCT
fcis-32489	53	1	subsequently	subsequently	ADV
fcis-32489	53	2	,	,	PUNCT
fcis-32489	53	3	it	it	PRON
fcis-32489	53	4	was	be	AUX
fcis-32489	53	5	used	use	VERB
fcis-32489	53	6	in	in	ADP
fcis-32489	53	7	areas	area	NOUN
fcis-32489	53	8	such	such	ADJ
fcis-32489	53	9	as	as	ADP
fcis-32489	53	10	semantic	semantic	ADJ
fcis-32489	53	11	role	role	NOUN
fcis-32489	53	12	tagging	tag	VERB
fcis-32489	54	1	[	[	X
fcis-32489	54	2	22	22	NUM
fcis-32489	54	3	]	]	PUNCT
fcis-32489	54	4	,	,	PUNCT
fcis-32489	54	5	multidocument	multidocument	ADJ
fcis-32489	54	6	summarization	summarization	NOUN
fcis-32489	54	7	[	[	X
fcis-32489	54	8	23	23	NUM
fcis-32489	54	9	]	]	PUNCT
fcis-32489	54	10	,	,	PUNCT
fcis-32489	54	11	and	and	CCONJ
fcis-32489	54	12	temporal	temporal	ADJ
fcis-32489	54	13	relationship	relationship	NOUN
fcis-32489	54	14	externalization	externalization	NOUN
fcis-32489	55	1	[	[	X
fcis-32489	55	2	24	24	NUM
fcis-32489	55	3	]	]	PUNCT
fcis-32489	55	4	.	.	PUNCT
fcis-32489	56	1	zhang	zhang	PROPN
fcis-32489	56	2	et	et	PROPN
fcis-32489	56	3	al	al	PROPN
fcis-32489	56	4	.	.	PUNCT
fcis-32489	57	1	[	[	X
fcis-32489	57	2	25	25	NUM
fcis-32489	57	3	]	]	PUNCT
fcis-32489	57	4	used	use	VERB
fcis-32489	57	5	gcn	gcn	NOUN
fcis-32489	57	6	on	on	ADP
fcis-32489	57	7	dependency	dependency	NOUN
fcis-32489	57	8	trees	tree	NOUN
fcis-32489	57	9	and	and	CCONJ
fcis-32489	57	10	proposed	propose	VERB
fcis-32489	57	11	a	a	DET
fcis-32489	57	12	novel	novel	ADJ
fcis-32489	57	13	path	path	NOUN
fcis-32489	57	14	-	-	PUNCT
fcis-32489	57	15	centric	centric	ADJ
fcis-32489	57	16	pruning	pruning	NOUN
fcis-32489	57	17	technique	technique	NOUN
fcis-32489	57	18	that	that	PRON
fcis-32489	57	19	helps	help	VERB
fcis-32489	57	20	models	model	NOUN
fcis-32489	57	21	remove	remove	VERB
fcis-32489	57	22	irrelevant	irrelevant	ADJ
fcis-32489	57	23	information	information	NOUN
fcis-32489	57	24	to	to	ADP
fcis-32489	57	25	the	the	DET
fcis-32489	57	26	greatest	great	ADJ
fcis-32489	57	27	extent	extent	NOUN
fcis-32489	57	28	possible	possible	ADJ
fcis-32489	57	29	without	without	ADP
fcis-32489	57	30	disrupting	disrupt	VERB
fcis-32489	57	31	the	the	DET
fcis-32489	57	32	key	key	ADJ
fcis-32489	57	33	content	content	NOUN
fcis-32489	57	34	to	to	PART
fcis-32489	57	35	improve	improve	VERB
fcis-32489	57	36	relation	relation	NOUN
fcis-32489	57	37	extraction	extraction	NOUN
fcis-32489	57	38	.	.	PUNCT
fcis-32489	58	1	the	the	DET
fcis-32489	58	2	document	document	NOUN
fcis-32489	58	3	-	-	PUNCT
fcis-32489	58	4	level	level	NOUN
fcis-32489	58	5	graph	graph	NOUN
fcis-32489	58	6	convolutional	convolutional	ADJ
fcis-32489	58	7	network	network	NOUN
fcis-32489	58	8	structure	structure	NOUN
fcis-32489	58	9	proposed	propose	VERB
fcis-32489	58	10	in	in	ADP
fcis-32489	58	11	this	this	DET
fcis-32489	58	12	paper	paper	NOUN
fcis-32489	58	13	enables	enable	VERB
fcis-32489	58	14	deep	deep	ADJ
fcis-32489	58	15	training	training	NOUN
fcis-32489	58	16	of	of	ADP
fcis-32489	58	17	the	the	DET
fcis-32489	58	18	model	model	NOUN
fcis-32489	58	19	by	by	ADP
fcis-32489	58	20	aggregating	aggregate	VERB
fcis-32489	58	21	and	and	CCONJ
fcis-32489	58	22	modeling	model	VERB
fcis-32489	58	23	the	the	DET
fcis-32489	58	24	context	context	NOUN
fcis-32489	58	25	information	information	NOUN
fcis-32489	58	26	of	of	ADP
fcis-32489	58	27	entities	entity	NOUN
fcis-32489	58	28	,	,	PUNCT
fcis-32489	58	29	thereby	thereby	ADV
fcis-32489	58	30	capturing	capture	VERB
fcis-32489	58	31	rich	rich	ADJ
fcis-32489	58	32	local	local	ADJ
fcis-32489	58	33	and	and	CCONJ
fcis-32489	58	34	non	non	ADJ
fcis-32489	58	35	-	-	ADJ
fcis-32489	58	36	local	local	ADJ
fcis-32489	58	37	dependency	dependency	NOUN
fcis-32489	58	38	features	feature	NOUN
fcis-32489	58	39	in	in	ADP
fcis-32489	58	40	sentences	sentence	NOUN
fcis-32489	58	41	.	.	PUNCT
fcis-32489	59	1	this	this	DET
fcis-32489	59	2	paper	paper	NOUN
fcis-32489	59	3	experiments	experiment	VERB
fcis-32489	59	4	the	the	DET
fcis-32489	59	5	proposed	propose	VERB
fcis-32489	59	6	model	model	NOUN
fcis-32489	59	7	on	on	ADP
fcis-32489	59	8	the	the	DET
fcis-32489	59	9	docred	docre	VERB
fcis-32489	59	10	dataset	dataset	NOUN
fcis-32489	59	11	and	and	CCONJ
fcis-32489	59	12	achieves	achieve	VERB
fcis-32489	59	13	better	well	ADJ
fcis-32489	59	14	results	result	NOUN
fcis-32489	59	15	than	than	ADP
fcis-32489	59	16	existing	exist	VERB
fcis-32489	59	17	models	model	NOUN
fcis-32489	59	18	.	.	PUNCT
fcis-32489	60	1	3	3	X
fcis-32489	60	2	.	.	X
fcis-32489	60	3	methods	method	NOUN
fcis-32489	60	4	this	this	DET
fcis-32489	60	5	section	section	NOUN
fcis-32489	60	6	commences	commence	VERB
fcis-32489	60	7	with	with	ADP
fcis-32489	60	8	an	an	DET
fcis-32489	60	9	overview	overview	NOUN
fcis-32489	60	10	of	of	ADP
fcis-32489	60	11	the	the	DET
fcis-32489	60	12	gcn	gcn	NOUN
fcis-32489	60	13	framework	framework	NOUN
fcis-32489	60	14	employed	employ	VERB
fcis-32489	60	15	for	for	ADP
fcis-32489	60	16	extracting	extract	VERB
fcis-32489	60	17	relationships	relationship	NOUN
fcis-32489	60	18	at	at	ADP
fcis-32489	60	19	the	the	DET
fcis-32489	60	20	document	document	NOUN
fcis-32489	60	21	level	level	NOUN
fcis-32489	60	22	.	.	PUNCT
fcis-32489	61	1	the	the	DET
fcis-32489	61	2	framework	framework	NOUN
fcis-32489	61	3	takes	take	VERB
fcis-32489	61	4	documents	document	NOUN
fcis-32489	61	5	composed	compose	VERB
fcis-32489	61	6	of	of	ADP
fcis-32489	61	7	n	n	PRON
fcis-32489	61	8	terms	term	NOUN
fcis-32489	61	9	,	,	PUNCT
fcis-32489	61	10	converts	convert	VERB
fcis-32489	61	11	each	each	PRON
fcis-32489	61	12	into	into	ADP
fcis-32489	61	13	a	a	DET
fcis-32489	61	14	series	series	NOUN
fcis-32489	61	15	of	of	ADP
fcis-32489	61	16	hidden	hide	VERB
fcis-32489	61	17	state	state	NOUN
fcis-32489	61	18	vectors	vector	NOUN
fcis-32489	61	19	,	,	PUNCT
fcis-32489	61	20	and	and	CCONJ
fcis-32489	61	21	then	then	ADV
fcis-32489	61	22	computes	compute	VERB
fcis-32489	61	23	the	the	DET
fcis-32489	61	24	entity	entity	NOUN
fcis-32489	61	25	representations	representation	NOUN
fcis-32489	61	26	.	.	PUNCT
fcis-32489	62	1	these	these	DET
fcis-32489	62	2	entity	entity	NOUN
fcis-32489	62	3	features	feature	NOUN
fcis-32489	62	4	are	be	AUX
fcis-32489	62	5	combined	combine	VERB
fcis-32489	62	6	to	to	PART
fcis-32489	62	7	form	form	VERB
fcis-32489	62	8	vector	vector	NOUN
fcis-32489	62	9	embeddings	embedding	NOUN
fcis-32489	62	10	,	,	PUNCT
fcis-32489	62	11	which	which	PRON
fcis-32489	62	12	are	be	AUX
fcis-32489	62	13	subsequently	subsequently	ADV
fcis-32489	62	14	used	use	VERB
fcis-32489	62	15	to	to	PART
fcis-32489	62	16	construct	construct	VERB
fcis-32489	62	17	a	a	DET
fcis-32489	62	18	graph	graph	NOUN
fcis-32489	62	19	structure	structure	NOUN
fcis-32489	62	20	using	use	VERB
fcis-32489	62	21	gcn	gcn	NOUN
fcis-32489	62	22	.	.	PUNCT
fcis-32489	63	1	this	this	DET
fcis-32489	63	2	graph	graph	NOUN
fcis-32489	63	3	-	-	PUNCT
fcis-32489	63	4	based	base	VERB
fcis-32489	63	5	representation	representation	NOUN
fcis-32489	63	6	is	be	AUX
fcis-32489	63	7	then	then	ADV
fcis-32489	63	8	analyzed	analyze	VERB
fcis-32489	63	9	to	to	PART
fcis-32489	63	10	determine	determine	VERB
fcis-32489	63	11	the	the	DET
fcis-32489	63	12	final	final	ADJ
fcis-32489	63	13	relational	relational	ADJ
fcis-32489	63	14	predictions	prediction	NOUN
fcis-32489	63	15	.	.	PUNCT
fcis-32489	64	1	the	the	DET
fcis-32489	64	2	model	model	NOUN
fcis-32489	64	3	architecture	architecture	NOUN
fcis-32489	64	4	diagram	diagram	NOUN
fcis-32489	64	5	of	of	ADP
fcis-32489	64	6	this	this	DET
fcis-32489	64	7	paper	paper	NOUN
fcis-32489	64	8	is	be	AUX
fcis-32489	64	9	shown	show	VERB
fcis-32489	64	10	in	in	ADP
fcis-32489	64	11	figure	figure	NOUN
fcis-32489	64	12	1	1	NUM
fcis-32489	64	13	.	.	PUNCT
fcis-32489	64	14	figure	figure	NOUN
fcis-32489	64	15	1	1	NUM
fcis-32489	64	16	.	.	PUNCT
fcis-32489	64	17	model	model	NOUN
fcis-32489	64	18	architecture	architecture	NOUN
fcis-32489	64	19	diagram	diagram	NOUN
fcis-32489	64	20	3.1	3.1	NUM
fcis-32489	64	21	.	.	PUNCT
fcis-32489	65	1	input	input	NOUN
fcis-32489	65	2	layer	layer	NOUN
fcis-32489	65	3	the	the	DET
fcis-32489	65	4	input	input	NOUN
fcis-32489	65	5	layer	layer	NOUN
fcis-32489	65	6	is	be	AUX
fcis-32489	65	7	an	an	DET
fcis-32489	65	8	essential	essential	ADJ
fcis-32489	65	9	part	part	NOUN
fcis-32489	65	10	of	of	ADP
fcis-32489	65	11	relationship	relationship	NOUN
fcis-32489	65	12	extraction	extraction	NOUN
fcis-32489	65	13	tasks	task	NOUN
fcis-32489	65	14	,	,	PUNCT
fcis-32489	65	15	which	which	PRON
fcis-32489	65	16	can	can	AUX
fcis-32489	65	17	represent	represent	VERB
fcis-32489	65	18	sentences	sentence	NOUN
fcis-32489	65	19	in	in	ADP
fcis-32489	65	20	text	text	NOUN
fcis-32489	65	21	in	in	ADP
fcis-32489	65	22	a	a	DET
fcis-32489	65	23	vectorized	vectorize	VERB
fcis-32489	65	24	form	form	NOUN
fcis-32489	65	25	for	for	ADP
fcis-32489	65	26	further	further	ADJ
fcis-32489	65	27	processing	processing	NOUN
fcis-32489	65	28	.	.	PUNCT
fcis-32489	66	1	the	the	DET
fcis-32489	66	2	specific	specific	ADJ
fcis-32489	66	3	representation	representation	NOUN
fcis-32489	66	4	of	of	ADP
fcis-32489	66	5	the	the	DET
fcis-32489	66	6	input	input	NOUN
fcis-32489	66	7	layer	layer	NOUN
fcis-32489	66	8	is	be	AUX
fcis-32489	66	9	shown	show	VERB
fcis-32489	66	10	in	in	ADP
fcis-32489	66	11	figure	figure	NOUN
fcis-32489	66	12	2	2	NUM
fcis-32489	66	13	.	.	PUNCT
fcis-32489	66	14	figure	figure	NOUN
fcis-32489	66	15	2	2	NUM
fcis-32489	66	16	.	.	PUNCT
fcis-32489	66	17	schematic	schematic	ADJ
fcis-32489	66	18	diagram	diagram	NOUN
fcis-32489	66	19	of	of	ADP
fcis-32489	66	20	input	input	NOUN
fcis-32489	66	21	layer	layer	NOUN
fcis-32489	66	22	in	in	ADP
fcis-32489	66	23	figure	figure	NOUN
fcis-32489	66	24	2	2	NUM
fcis-32489	66	25	,	,	PUNCT
fcis-32489	66	26	each	each	DET
fcis-32489	66	27	word	word	NOUN
fcis-32489	66	28	in	in	ADP
fcis-32489	66	29	the	the	DET
fcis-32489	66	30	sentence	sentence	NOUN
fcis-32489	66	31	is	be	AUX
fcis-32489	66	32	composed	compose	VERB
fcis-32489	66	33	of	of	ADP
fcis-32489	66	34	word	word	NOUN
fcis-32489	66	35	embeddings	embedding	NOUN
fcis-32489	66	36	,	,	PUNCT
fcis-32489	66	37	entity	entity	NOUN
fcis-32489	66	38	types	type	NOUN
fcis-32489	66	39	,	,	PUNCT
fcis-32489	66	40	and	and	CCONJ
fcis-32489	66	41	co	co	VERB
fcis-32489	66	42	referential	referential	ADJ
fcis-32489	66	43	information	information	NOUN
fcis-32489	66	44	.	.	PUNCT
fcis-32489	67	1	white	white	PROPN
fcis-32489	67	2	represents	represent	VERB
fcis-32489	67	3	the	the	DET
fcis-32489	67	4	vector	vector	NOUN
fcis-32489	67	5	of	of	ADP
fcis-32489	67	6	words	word	NOUN
fcis-32489	67	7	in	in	ADP
fcis-32489	67	8	the	the	DET
fcis-32489	67	9	sentence	sentence	NOUN
fcis-32489	67	10	,	,	PUNCT
fcis-32489	67	11	while	while	SCONJ
fcis-32489	67	12	dark	dark	ADJ
fcis-32489	67	13	green	green	NOUN
fcis-32489	67	14	represents	represent	VERB
fcis-32489	67	15	the	the	DET
fcis-32489	67	16	head	head	NOUN
fcis-32489	67	17	and	and	CCONJ
fcis-32489	67	18	tail	tail	NOUN
fcis-32489	67	19	entities	entity	NOUN
fcis-32489	67	20	in	in	ADP
fcis-32489	67	21	the	the	DET
fcis-32489	67	22	sentence	sentence	NOUN
fcis-32489	67	23	.	.	PUNCT
fcis-32489	68	1	3.1.1	3.1.1	X
fcis-32489	68	2	.	.	PUNCT
fcis-32489	68	3	word	word	NOUN
fcis-32489	68	4	embeddings	embedding	NOUN
fcis-32489	68	5	word	word	NOUN
fcis-32489	68	6	embedding	embed	VERB
fcis-32489	68	7	is	be	AUX
fcis-32489	68	8	the	the	DET
fcis-32489	68	9	collective	collective	ADJ
fcis-32489	68	10	term	term	NOUN
fcis-32489	68	11	for	for	ADP
fcis-32489	68	12	language	language	NOUN
fcis-32489	68	13	models	model	NOUN
fcis-32489	68	14	and	and	CCONJ
fcis-32489	68	15	representation	representation	NOUN
fcis-32489	68	16	learning	learn	VERB
fcis-32489	68	17	techniques	technique	NOUN
fcis-32489	68	18	in	in	ADP
fcis-32489	68	19	natural	natural	ADJ
fcis-32489	68	20	language	language	NOUN
fcis-32489	68	21	processing	processing	NOUN
fcis-32489	68	22	(	(	PUNCT
fcis-32489	68	23	nlp	nlp	NOUN
fcis-32489	68	24	)	)	PUNCT
fcis-32489	68	25	.	.	PUNCT
fcis-32489	69	1	specifically	specifically	ADV
fcis-32489	69	2	,	,	PUNCT
fcis-32489	69	3	it	it	PRON
fcis-32489	69	4	refers	refer	VERB
fcis-32489	69	5	to	to	ADP
fcis-32489	69	6	embedding	embed	VERB
fcis-32489	69	7	a	a	DET
fcis-32489	69	8	highdimensional	highdimensional	ADJ
fcis-32489	69	9	space	space	NOUN
fcis-32489	69	10	of	of	ADP
fcis-32489	69	11	all	all	DET
fcis-32489	69	12	words	word	NOUN
fcis-32489	69	13	into	into	ADP
fcis-32489	69	14	a	a	DET
fcis-32489	69	15	low	low	ADJ
fcis-32489	69	16	-	-	PUNCT
fcis-32489	69	17	dimensional	dimensional	ADJ
fcis-32489	69	18	continuous	continuous	ADJ
fcis-32489	69	19	vector	vector	NOUN
fcis-32489	69	20	space	space	NOUN
fcis-32489	69	21	,	,	PUNCT
fcis-32489	69	22	where	where	SCONJ
fcis-32489	69	23	each	each	DET
fcis-32489	69	24	word	word	NOUN
fcis-32489	69	25	or	or	CCONJ
fcis-32489	69	26	phrase	phrase	NOUN
fcis-32489	69	27	is	be	AUX
fcis-32489	69	28	mapped	map	VERB
fcis-32489	69	29	to	to	ADP
fcis-32489	69	30	a	a	DET
fcis-32489	69	31	vector	vector	NOUN
fcis-32489	69	32	over	over	ADP
fcis-32489	69	33	the	the	DET
fcis-32489	69	34	real	real	ADJ
fcis-32489	69	35	number	number	NOUN
fcis-32489	69	36	field	field	NOUN
fcis-32489	69	37	.	.	PUNCT
fcis-32489	70	1	this	this	DET
fcis-32489	70	2	paper	paper	NOUN
fcis-32489	70	3	mainly	mainly	ADV
fcis-32489	70	4	uses	use	VERB
fcis-32489	70	5	glove	glove	NOUN
fcis-32489	70	6	word	word	NOUN
fcis-32489	70	7	embeddings	embedding	NOUN
fcis-32489	70	8	[	[	X
fcis-32489	70	9	26	26	NUM
fcis-32489	70	10	]	]	PUNCT
fcis-32489	70	11	,	,	PUNCT
fcis-32489	70	12	which	which	PRON
fcis-32489	70	13	overcomes	overcome	VERB
fcis-32489	70	14	the	the	DET
fcis-32489	70	15	shortcomings	shortcoming	NOUN
fcis-32489	70	16	of	of	ADP
fcis-32489	70	17	global	global	ADJ
fcis-32489	70	18	matrix	matrix	NOUN
fcis-32489	70	19	36	36	NUM
fcis-32489	70	20	factorization	factorization	NOUN
fcis-32489	70	21	and	and	CCONJ
fcis-32489	70	22	local	local	ADJ
fcis-32489	70	23	context	context	NOUN
fcis-32489	70	24	windows	window	NOUN
fcis-32489	70	25	by	by	ADP
fcis-32489	70	26	learning	learn	VERB
fcis-32489	70	27	word	word	NOUN
fcis-32489	70	28	vectors	vector	NOUN
fcis-32489	70	29	through	through	ADP
fcis-32489	70	30	the	the	DET
fcis-32489	70	31	statistical	statistical	ADJ
fcis-32489	70	32	information	information	NOUN
fcis-32489	70	33	of	of	ADP
fcis-32489	70	34	global	global	ADJ
fcis-32489	70	35	lexical	lexical	ADJ
fcis-32489	70	36	co	co	NOUN
fcis-32489	70	37	-	-	NOUN
fcis-32489	70	38	occurrence	occurrence	NOUN
fcis-32489	70	39	,	,	PUNCT
fcis-32489	70	40	thereby	thereby	ADV
fcis-32489	70	41	combining	combine	VERB
fcis-32489	70	42	the	the	DET
fcis-32489	70	43	statistical	statistical	ADJ
fcis-32489	70	44	information	information	NOUN
fcis-32489	70	45	with	with	ADP
fcis-32489	70	46	the	the	DET
fcis-32489	70	47	advantages	advantage	NOUN
fcis-32489	70	48	of	of	ADP
fcis-32489	70	49	the	the	DET
fcis-32489	70	50	local	local	ADJ
fcis-32489	70	51	context	context	NOUN
fcis-32489	70	52	window	window	NOUN
fcis-32489	70	53	method	method	NOUN
fcis-32489	70	54	to	to	PART
fcis-32489	70	55	better	well	ADV
fcis-32489	70	56	utilize	utilize	VERB
fcis-32489	70	57	the	the	DET
fcis-32489	70	58	context	context	NOUN
fcis-32489	70	59	.	.	PUNCT
fcis-32489	71	1	3.1.2	3.1.2	X
fcis-32489	71	2	.	.	PUNCT
fcis-32489	71	3	entity	entity	NOUN
fcis-32489	71	4	type	type	NOUN
fcis-32489	71	5	embeddings	embedding	NOUN
fcis-32489	71	6	entity	entity	NOUN
fcis-32489	71	7	type	type	NOUN
fcis-32489	71	8	embeddings	embedding	NOUN
fcis-32489	71	9	are	be	AUX
fcis-32489	71	10	achieved	achieve	VERB
fcis-32489	71	11	by	by	ADP
fcis-32489	71	12	mapping	map	VERB
fcis-32489	71	13	the	the	DET
fcis-32489	71	14	entity	entity	NOUN
fcis-32489	71	15	types	type	NOUN
fcis-32489	71	16	assigned	assign	VERB
fcis-32489	71	17	to	to	ADP
fcis-32489	71	18	words	word	NOUN
fcis-32489	71	19	(	(	PUNCT
fcis-32489	71	20	such	such	ADJ
fcis-32489	71	21	as	as	ADP
fcis-32489	71	22	per	per	ADP
fcis-32489	71	23	,	,	PUNCT
fcis-32489	71	24	loc	loc	PROPN
fcis-32489	71	25	,	,	PUNCT
fcis-32489	71	26	org	org	NOUN
fcis-32489	71	27	,	,	PUNCT
fcis-32489	71	28	etc	etc	X
fcis-32489	71	29	.	.	X
fcis-32489	71	30	)	)	PUNCT
fcis-32489	71	31	to	to	ADP
fcis-32489	71	32	vectors	vector	NOUN
fcis-32489	71	33	using	use	VERB
fcis-32489	71	34	the	the	DET
fcis-32489	71	35	embedding	embed	VERB
fcis-32489	71	36	matrix	matrix	NOUN
fcis-32489	71	37	.	.	PUNCT
fcis-32489	72	1	entity	entity	NOUN
fcis-32489	72	2	types	type	NOUN
fcis-32489	72	3	are	be	AUX
fcis-32489	72	4	assigned	assign	VERB
fcis-32489	72	5	manually	manually	ADV
fcis-32489	72	6	to	to	ADP
fcis-32489	72	7	annotated	annotated	ADJ
fcis-32489	72	8	data	datum	NOUN
fcis-32489	72	9	,	,	PUNCT
fcis-32489	72	10	and	and	CCONJ
fcis-32489	72	11	through	through	ADP
fcis-32489	72	12	the	the	DET
fcis-32489	72	13	bert	bert	PROPN
fcis-32489	72	14	model	model	NOUN
fcis-32489	72	15	to	to	PART
fcis-32489	72	16	remotely	remotely	ADV
fcis-32489	72	17	supervised	supervise	VERB
fcis-32489	72	18	data	datum	NOUN
fcis-32489	72	19	.	.	PUNCT
fcis-32489	73	1	3.1.3	3.1.3	NUM
fcis-32489	73	2	.	.	PUNCT
fcis-32489	73	3	coreference	coreference	PROPN
fcis-32489	73	4	relations	relation	NOUN
fcis-32489	73	5	co	co	NOUN
fcis-32489	73	6	-	-	NOUN
fcis-32489	73	7	reference	reference	ADJ
fcis-32489	73	8	information	information	NOUN
fcis-32489	73	9	refers	refer	VERB
fcis-32489	73	10	to	to	ADP
fcis-32489	73	11	the	the	DET
fcis-32489	73	12	use	use	NOUN
fcis-32489	73	13	of	of	ADP
fcis-32489	73	14	the	the	DET
fcis-32489	73	15	same	same	ADJ
fcis-32489	73	16	entity	entity	NOUN
fcis-32489	73	17	i	i	PROPN
fcis-32489	73	18	d	d	PROPN
fcis-32489	73	19	(	(	PUNCT
fcis-32489	73	20	determined	determine	VERB
fcis-32489	73	21	based	base	VERB
fcis-32489	73	22	on	on	ADP
fcis-32489	73	23	the	the	DET
fcis-32489	73	24	order	order	NOUN
fcis-32489	73	25	in	in	ADP
fcis-32489	73	26	which	which	PRON
fcis-32489	73	27	the	the	DET
fcis-32489	73	28	entity	entity	NOUN
fcis-32489	73	29	first	first	ADV
fcis-32489	73	30	appears	appear	VERB
fcis-32489	73	31	in	in	ADP
fcis-32489	73	32	the	the	DET
fcis-32489	73	33	document	document	NOUN
fcis-32489	73	34	)	)	PUNCT
fcis-32489	73	35	to	to	PART
fcis-32489	73	36	map	map	VERB
fcis-32489	73	37	co	co	ADJ
fcis-32489	73	38	-	-	NOUN
fcis-32489	73	39	reference	reference	ADJ
fcis-32489	73	40	relationships	relationship	NOUN
fcis-32489	73	41	to	to	ADP
fcis-32489	73	42	vectors	vector	NOUN
fcis-32489	73	43	when	when	SCONJ
fcis-32489	73	44	several	several	ADJ
fcis-32489	73	45	different	different	ADJ
fcis-32489	73	46	entities	entity	NOUN
fcis-32489	73	47	refer	refer	VERB
fcis-32489	73	48	to	to	ADP
fcis-32489	73	49	the	the	DET
fcis-32489	73	50	same	same	ADJ
fcis-32489	73	51	entity	entity	NOUN
fcis-32489	73	52	.	.	PUNCT
fcis-32489	74	1	for	for	ADP
fcis-32489	74	2	example	example	NOUN
fcis-32489	74	3	,	,	PUNCT
fcis-32489	74	4	england	england	PROPN
fcis-32489	74	5	and	and	CCONJ
fcis-32489	74	6	uk	uk	PROPN
fcis-32489	74	7	both	both	PRON
fcis-32489	74	8	refer	refer	VERB
fcis-32489	74	9	to	to	ADP
fcis-32489	74	10	the	the	DET
fcis-32489	74	11	united	united	ADJ
fcis-32489	74	12	kingdom	kingdom	PROPN
fcis-32489	74	13	.	.	PUNCT
fcis-32489	75	1	this	this	DET
fcis-32489	75	2	paper	paper	NOUN
fcis-32489	75	3	utilizes	utilize	VERB
fcis-32489	75	4	word	word	NOUN
fcis-32489	75	5	vectors	vector	NOUN
fcis-32489	75	6	,	,	PUNCT
fcis-32489	75	7	coreference	coreference	NOUN
fcis-32489	75	8	information	information	NOUN
fcis-32489	75	9	,	,	PUNCT
fcis-32489	75	10	and	and	CCONJ
fcis-32489	75	11	entity	entity	NOUN
fcis-32489	75	12	types	type	NOUN
fcis-32489	75	13	as	as	ADP
fcis-32489	75	14	basic	basic	ADJ
fcis-32489	75	15	features	feature	NOUN
fcis-32489	75	16	,	,	PUNCT
fcis-32489	75	17	representing	represent	VERB
fcis-32489	75	18	them	they	PRON
fcis-32489	75	19	in	in	ADP
fcis-32489	75	20	vector	vector	NOUN
fcis-32489	75	21	form	form	NOUN
fcis-32489	75	22	asww	asww	NOUN
fcis-32489	75	23	,	,	PUNCT
fcis-32489	75	24	wc	wc	NOUN
fcis-32489	75	25	,	,	PUNCT
fcis-32489	75	26	and	and	CCONJ
fcis-32489	75	27	wt	wt	INTJ
fcis-32489	75	28	,	,	PUNCT
fcis-32489	75	29	respectively	respectively	ADV
fcis-32489	75	30	.	.	PUNCT
fcis-32489	76	1	these	these	DET
fcis-32489	76	2	vectors	vector	NOUN
fcis-32489	76	3	are	be	AUX
fcis-32489	76	4	concatenated	concatenate	VERB
fcis-32489	76	5	to	to	PART
fcis-32489	76	6	form	form	VERB
fcis-32489	76	7	the	the	DET
fcis-32489	76	8	feature	feature	NOUN
fcis-32489	76	9	xi	xi	INTJ
fcis-32489	76	10	for	for	ADP
fcis-32489	76	11	the	the	DET
fcis-32489	76	12	initial	initial	ADJ
fcis-32489	76	13	word	word	NOUN
fcis-32489	76	14	i	i	PRON
fcis-32489	76	15	,	,	PUNCT
fcis-32489	76	16	i.e.	i.e.	X
fcis-32489	76	17	,	,	PUNCT
fcis-32489	76	18	xi=	xi=	PROPN
fcis-32489	76	19	wi	wi	PROPN
fcis-32489	77	1	w;wi	w;wi	PROPN
fcis-32489	77	2	c;wi	c;wi	PROPN
fcis-32489	77	3	t	t	PROPN
fcis-32489	77	4	.	.	PUNCT
fcis-32489	78	1	3.2	3.2	NUM
fcis-32489	78	2	.	.	PUNCT
fcis-32489	78	3	bilstm	bilstm	NOUN
fcis-32489	78	4	layer	layer	NOUN
fcis-32489	78	5	this	this	DET
fcis-32489	78	6	section	section	NOUN
fcis-32489	78	7	inputs	input	VERB
fcis-32489	78	8	the	the	DET
fcis-32489	78	9	feature	feature	NOUN
fcis-32489	78	10	information	information	NOUN
fcis-32489	78	11	obtained	obtain	VERB
fcis-32489	78	12	from	from	ADP
fcis-32489	78	13	the	the	DET
fcis-32489	78	14	input	input	NOUN
fcis-32489	78	15	layer	layer	NOUN
fcis-32489	78	16	into	into	ADP
fcis-32489	78	17	the	the	DET
fcis-32489	78	18	bilstm	bilstm	NOUN
fcis-32489	78	19	layer	layer	NOUN
fcis-32489	78	20	to	to	PART
fcis-32489	78	21	obtain	obtain	VERB
fcis-32489	78	22	the	the	DET
fcis-32489	78	23	feature	feature	NOUN
fcis-32489	78	24	representation	representation	NOUN
fcis-32489	78	25	of	of	ADP
fcis-32489	78	26	each	each	DET
fcis-32489	78	27	entity	entity	NOUN
fcis-32489	78	28	,	,	PUNCT
fcis-32489	78	29	while	while	SCONJ
fcis-32489	78	30	the	the	DET
fcis-32489	78	31	bilstm	bilstm	NOUN
fcis-32489	78	32	layer	layer	NOUN
fcis-32489	78	33	outputs	output	VERB
fcis-32489	78	34	a	a	DET
fcis-32489	78	35	vectorized	vectorize	VERB
fcis-32489	78	36	form	form	NOUN
fcis-32489	78	37	of	of	ADP
fcis-32489	78	38	word	word	NOUN
fcis-32489	78	39	sequence	sequence	NOUN
fcis-32489	78	40	information	information	NOUN
fcis-32489	78	41	.	.	PUNCT
fcis-32489	79	1	specifically	specifically	ADV
fcis-32489	79	2	,	,	PUNCT
fcis-32489	79	3	it	it	PRON
fcis-32489	79	4	involves	involve	VERB
fcis-32489	79	5	encoding	encode	VERB
fcis-32489	79	6	and	and	CCONJ
fcis-32489	79	7	analyzing	analyze	VERB
fcis-32489	79	8	the	the	DET
fcis-32489	79	9	word	word	NOUN
fcis-32489	79	10	features	feature	VERB
fcis-32489	79	11	xi	xi	PROPN
fcis-32489	79	12	obtained	obtain	VERB
fcis-32489	79	13	from	from	ADP
fcis-32489	79	14	the	the	DET
fcis-32489	79	15	input	input	NOUN
fcis-32489	79	16	layer	layer	NOUN
fcis-32489	79	17	,	,	PUNCT
fcis-32489	79	18	and	and	CCONJ
fcis-32489	79	19	concatenating	concatenate	VERB
fcis-32489	79	20	the	the	DET
fcis-32489	79	21	hidden	hide	VERB
fcis-32489	79	22	layer	layer	NOUN
fcis-32489	79	23	vector	vector	NOUN
fcis-32489	79	24	representations	representation	NOUN
fcis-32489	79	25	of	of	ADP
fcis-32489	79	26	the	the	DET
fcis-32489	79	27	forward	forward	ADJ
fcis-32489	79	28	lstm	lstm	PROPN
fcis-32489	79	29	and	and	CCONJ
fcis-32489	79	30	backward	backward	ADJ
fcis-32489	79	31	lstm	lstm	NOUN
fcis-32489	79	32	to	to	PART
fcis-32489	79	33	obtain	obtain	VERB
fcis-32489	79	34	the	the	DET
fcis-32489	79	35	sequence	sequence	NOUN
fcis-32489	79	36	h1,h2,	h1,h2,	NOUN
fcis-32489	79	37	…	…	PUNCT
fcis-32489	79	38	,hn	,hn	PUNCT
fcis-32489	79	39	that	that	PRON
fcis-32489	79	40	integrates	integrate	VERB
fcis-32489	79	41	contextual	contextual	ADJ
fcis-32489	79	42	information	information	NOUN
fcis-32489	79	43	.	.	PUNCT
fcis-32489	80	1	the	the	DET
fcis-32489	80	2	calculation	calculation	NOUN
fcis-32489	80	3	and	and	CCONJ
fcis-32489	80	4	update	update	VERB
fcis-32489	80	5	formulas	formula	NOUN
fcis-32489	80	6	for	for	ADP
fcis-32489	80	7	the	the	DET
fcis-32489	80	8	hidden	hide	VERB
fcis-32489	80	9	layer	layer	NOUN
fcis-32489	80	10	node	node	VERB
fcis-32489	80	11	ℎ	ℎ	NOUN
fcis-32489	80	12	at	at	ADP
fcis-32489	80	13	a	a	DET
fcis-32489	80	14	certain	certain	ADJ
fcis-32489	80	15	moment	moment	NOUN
fcis-32489	80	16	t	t	NOUN
fcis-32489	80	17	are	be	AUX
fcis-32489	80	18	shown	show	VERB
fcis-32489	80	19	in	in	ADP
fcis-32489	80	20	equations	equation	NOUN
fcis-32489	80	21	(	(	PUNCT
fcis-32489	80	22	1	1	NUM
fcis-32489	80	23	)	)	PUNCT
fcis-32489	80	24	to	to	ADP
fcis-32489	80	25	(	(	PUNCT
fcis-32489	80	26	6	6	NUM
fcis-32489	80	27	)	)	PUNCT
fcis-32489	80	28	.	.	PUNCT
fcis-32489	81	1	it	it	PRON
fcis-32489	82	1	=	=	NOUN
fcis-32489	82	2	σ	σ	X
fcis-32489	82	3	wxixt+whiht-1+wcict-1+bi	wxixt+whiht-1+wcict-1+bi	PROPN
fcis-32489	82	4	(	(	PUNCT
fcis-32489	82	5	1	1	NUM
fcis-32489	82	6	)	)	PUNCT
fcis-32489	82	7	ft	ft	PROPN
fcis-32489	82	8	=	=	SYM
fcis-32489	82	9	σ	σ	PROPN
fcis-32489	82	10	wxfxt+whfht-1+wcfct-1+bf	wxfxt+whfht-1+wcfct-1+bf	PROPN
fcis-32489	82	11	(	(	PUNCT
fcis-32489	82	12	2	2	NUM
fcis-32489	82	13	)	)	PUNCT
fcis-32489	82	14	gt	gt	PROPN
fcis-32489	82	15	=	=	SYM
fcis-32489	82	16	tanh	tanh	NOUN
fcis-32489	82	17	wxgxt+whght-1+wcgct-1+bg	wxgxt+whght-1+wcgct-1+bg	X
fcis-32489	82	18	(	(	PUNCT
fcis-32489	82	19	3	3	NUM
fcis-32489	82	20	)	)	PUNCT
fcis-32489	82	21	ct	ct	PROPN
fcis-32489	82	22	=	=	NOUN
fcis-32489	82	23	ft⨂ct-1	ft⨂ct-1	VERB
fcis-32489	82	24	-	-	PUNCT
fcis-32489	82	25	it⨂gt	it⨂gt	NOUN
fcis-32489	82	26	(	(	PUNCT
fcis-32489	82	27	4	4	NUM
fcis-32489	82	28	)	)	PUNCT
fcis-32489	82	29	ot	ot	NOUN
fcis-32489	82	30	=	=	NOUN
fcis-32489	82	31	σ	σ	NOUN
fcis-32489	82	32	wxoxt+whoht-1+wcoct+bo	wxoxt+whoht-1+wcoct+bo	NOUN
fcis-32489	82	33	(	(	PUNCT
fcis-32489	82	34	5	5	NUM
fcis-32489	82	35	)	)	PUNCT
fcis-32489	82	36	ht	ht	PROPN
fcis-32489	83	1	=	=	PROPN
fcis-32489	83	2	ot⨂tanh	ot⨂tanh	PROPN
fcis-32489	83	3	ct	ct	X
fcis-32489	83	4	(	(	PUNCT
fcis-32489	83	5	6	6	NUM
fcis-32489	83	6	)	)	PUNCT
fcis-32489	83	7	among	among	ADP
fcis-32489	83	8	them	they	PRON
fcis-32489	83	9	,	,	PUNCT
fcis-32489	83	10	σ	σ	PROPN
fcis-32489	83	11	represents	represent	VERB
fcis-32489	83	12	the	the	DET
fcis-32489	83	13	sigmoid	sigmoid	NOUN
fcis-32489	83	14	activation	activation	NOUN
fcis-32489	83	15	function	function	NOUN
fcis-32489	83	16	,	,	PUNCT
fcis-32489	83	17	⨂	⨂	PROPN
fcis-32489	83	18	represents	represent	VERB
fcis-32489	83	19	the	the	DET
fcis-32489	83	20	multiplication	multiplication	NOUN
fcis-32489	83	21	of	of	ADP
fcis-32489	83	22	vector	vector	NOUN
fcis-32489	83	23	elements	element	NOUN
fcis-32489	83	24	,	,	PUNCT
fcis-32489	83	25	and	and	CCONJ
fcis-32489	83	26	the	the	DET
fcis-32489	83	27	forget	forget	NOUN
fcis-32489	83	28	gate	gate	NOUN
fcis-32489	83	29	,	,	PUNCT
fcis-32489	83	30	input	input	NOUN
fcis-32489	83	31	gate	gate	NOUN
fcis-32489	83	32	,	,	PUNCT
fcis-32489	83	33	and	and	CCONJ
fcis-32489	83	34	output	output	NOUN
fcis-32489	83	35	gate	gate	NOUN
fcis-32489	83	36	are	be	AUX
fcis-32489	83	37	respectively	respectively	ADV
fcis-32489	83	38	represented	represent	VERB
fcis-32489	83	39	byft	byft	NOUN
fcis-32489	83	40	,	,	PUNCT
fcis-32489	83	41	it	it	PRON
fcis-32489	83	42	,	,	PUNCT
fcis-32489	83	43	andot	andot	NOUN
fcis-32489	83	44	.	.	PUNCT
fcis-32489	84	1	gtis	gtis	NOUN
fcis-32489	84	2	the	the	DET
fcis-32489	84	3	new	new	ADJ
fcis-32489	84	4	value	value	NOUN
fcis-32489	84	5	tensor	tensor	NOUN
fcis-32489	84	6	of	of	ADP
fcis-32489	84	7	the	the	DET
fcis-32489	84	8	unit	unit	NOUN
fcis-32489	84	9	,	,	PUNCT
fcis-32489	84	10	xt	xt	PROPN
fcis-32489	84	11	is	be	AUX
fcis-32489	84	12	the	the	DET
fcis-32489	84	13	input	input	NOUN
fcis-32489	84	14	vector	vector	NOUN
fcis-32489	84	15	at	at	ADP
fcis-32489	84	16	time	time	NOUN
fcis-32489	84	17	t	t	PROPN
fcis-32489	84	18	,	,	PUNCT
fcis-32489	84	19	ct	ct	PROPN
fcis-32489	84	20	is	be	AUX
fcis-32489	84	21	the	the	DET
fcis-32489	84	22	tensor	tensor	NOUN
fcis-32489	84	23	of	of	ADP
fcis-32489	84	24	the	the	DET
fcis-32489	84	25	forget	forget	NOUN
fcis-32489	84	26	gate	gate	NOUN
fcis-32489	84	27	,	,	PUNCT
fcis-32489	84	28	input	input	NOUN
fcis-32489	84	29	gate	gate	NOUN
fcis-32489	84	30	,	,	PUNCT
fcis-32489	84	31	and	and	CCONJ
fcis-32489	84	32	output	output	NOUN
fcis-32489	84	33	gate	gate	NOUN
fcis-32489	84	34	at	at	ADP
fcis-32489	84	35	time	time	NOUN
fcis-32489	84	36	t	t	PROPN
fcis-32489	84	37	,	,	PUNCT
fcis-32489	84	38	andhtis	andhtis	PRON
fcis-32489	84	39	the	the	DET
fcis-32489	84	40	output	output	NOUN
fcis-32489	84	41	of	of	ADP
fcis-32489	84	42	the	the	DET
fcis-32489	84	43	hidden	hide	VERB
fcis-32489	84	44	layer	layer	NOUN
fcis-32489	84	45	,	,	PUNCT
fcis-32489	84	46	wxi	wxi	NOUN
fcis-32489	84	47	,	,	PUNCT
fcis-32489	84	48	wxf	wxf	NOUN
fcis-32489	84	49	,	,	PUNCT
fcis-32489	84	50	wxg	wxg	NOUN
fcis-32489	84	51	,	,	PUNCT
fcis-32489	84	52	wxo	wxo	NOUN
fcis-32489	84	53	,	,	PUNCT
fcis-32489	84	54	wci	wci	PROPN
fcis-32489	84	55	,	,	PUNCT
fcis-32489	84	56	wcf	wcf	PROPN
fcis-32489	84	57	,	,	PUNCT
fcis-32489	84	58	wcg	wcg	PROPN
fcis-32489	84	59	representing	represent	VERB
fcis-32489	84	60	the	the	DET
fcis-32489	84	61	weight	weight	NOUN
fcis-32489	84	62	matrices	matrix	NOUN
fcis-32489	84	63	of	of	ADP
fcis-32489	84	64	𝑥	𝑥	NOUN
fcis-32489	84	65	running	run	VERB
fcis-32489	84	66	on	on	ADP
fcis-32489	84	67	different	different	ADJ
fcis-32489	84	68	gate	gate	NOUN
fcis-32489	84	69	mechanisms	mechanism	NOUN
fcis-32489	84	70	,	,	PUNCT
fcis-32489	84	71	whi	whi	PROPN
fcis-32489	84	72	,	,	PUNCT
fcis-32489	84	73	whf	whf	PROPN
fcis-32489	84	74	,	,	PUNCT
fcis-32489	84	75	whg	whg	PROPN
fcis-32489	84	76	,	,	PUNCT
fcis-32489	84	77	who	who	PRON
fcis-32489	84	78	,	,	PUNCT
fcis-32489	84	79	wco	wco	X
fcis-32489	84	80	represent	represent	VERB
fcis-32489	84	81	the	the	DET
fcis-32489	84	82	weight	weight	NOUN
fcis-32489	84	83	matrices	matrix	NOUN
fcis-32489	84	84	of	of	ADP
fcis-32489	84	85	ht	ht	PROPN
fcis-32489	84	86	on	on	ADP
fcis-32489	84	87	different	different	ADJ
fcis-32489	84	88	gate	gate	NOUN
fcis-32489	84	89	mechanisms	mechanism	NOUN
fcis-32489	84	90	,	,	PUNCT
fcis-32489	84	91	and	and	CCONJ
fcis-32489	84	92	b	b	X
fcis-32489	84	93	is	be	AUX
fcis-32489	84	94	the	the	DET
fcis-32489	84	95	bias	bias	NOUN
fcis-32489	84	96	term	term	NOUN
fcis-32489	84	97	.	.	PUNCT
fcis-32489	85	1	at	at	ADP
fcis-32489	85	2	time	time	NOUN
fcis-32489	85	3	t	t	PROPN
fcis-32489	85	4	,	,	PUNCT
fcis-32489	85	5	the	the	DET
fcis-32489	85	6	forward	forward	ADJ
fcis-32489	85	7	output	output	NOUN
fcis-32489	85	8	of	of	ADP
fcis-32489	85	9	bilstm	bilstm	NOUN
fcis-32489	85	10	isht⃗	isht⃗	PROPN
fcis-32489	85	11	,	,	PUNCT
fcis-32489	85	12	while	while	SCONJ
fcis-32489	85	13	the	the	DET
fcis-32489	85	14	reverse	reverse	ADJ
fcis-32489	85	15	output	output	NOUN
fcis-32489	85	16	isht⃖	isht⃖	VERB
fcis-32489	85	17	.	.	PUNCT
fcis-32489	86	1	then	then	ADV
fcis-32489	86	2	,	,	PUNCT
fcis-32489	86	3	concatenate	concatenate	VERB
fcis-32489	86	4	the	the	DET
fcis-32489	86	5	outputs	output	NOUN
fcis-32489	86	6	from	from	ADP
fcis-32489	86	7	both	both	DET
fcis-32489	86	8	directions	direction	NOUN
fcis-32489	86	9	to	to	PART
fcis-32489	86	10	obtain	obtain	VERB
fcis-32489	86	11	the	the	DET
fcis-32489	86	12	final	final	ADJ
fcis-32489	86	13	output	output	NOUN
fcis-32489	86	14	ht	ht	INTJ
fcis-32489	86	15	at	at	ADP
fcis-32489	86	16	time	time	NOUN
fcis-32489	86	17	t	t	PROPN
fcis-32489	86	18	,	,	PUNCT
fcis-32489	86	19	as	as	SCONJ
fcis-32489	86	20	shown	show	VERB
fcis-32489	86	21	in	in	ADP
fcis-32489	86	22	equation	equation	NOUN
fcis-32489	86	23	(	(	PUNCT
fcis-32489	86	24	7	7	NUM
fcis-32489	86	25	)	)	PUNCT
fcis-32489	86	26	.	.	PUNCT
fcis-32489	87	1	ht=	ht=	PROPN
fcis-32489	87	2	ht⃗	ht⃗	PROPN
fcis-32489	87	3	,	,	PUNCT
fcis-32489	87	4	ht⃖	ht⃖	X
fcis-32489	87	5	(	(	PUNCT
fcis-32489	87	6	7	7	NUM
fcis-32489	87	7	)	)	PUNCT
fcis-32489	87	8	3.3	3.3	NUM
fcis-32489	87	9	.	.	PUNCT
fcis-32489	88	1	figure	figure	VERB
fcis-32489	88	2	convolutional	convolutional	ADJ
fcis-32489	88	3	layers	layer	NOUN
fcis-32489	88	4	graph	graph	NOUN
fcis-32489	88	5	convolution	convolution	NOUN
fcis-32489	88	6	networks	network	NOUN
fcis-32489	88	7	involve	involve	VERB
fcis-32489	88	8	applying	apply	VERB
fcis-32489	88	9	convolutional	convolutional	ADJ
fcis-32489	88	10	neural	neural	ADJ
fcis-32489	88	11	network	network	NOUN
fcis-32489	88	12	techniques	technique	NOUN
fcis-32489	88	13	to	to	PART
fcis-32489	88	14	graph	graph	VERB
fcis-32489	88	15	data	datum	NOUN
fcis-32489	88	16	structures	structure	NOUN
fcis-32489	88	17	.	.	PUNCT
fcis-32489	89	1	consider	consider	VERB
fcis-32489	89	2	a	a	DET
fcis-32489	89	3	graph	graph	NOUN
fcis-32489	89	4	consisting	consist	VERB
fcis-32489	89	5	of	of	ADP
fcis-32489	89	6	n	n	DET
fcis-32489	89	7	nodes	node	NOUN
fcis-32489	89	8	;	;	PUNCT
fcis-32489	89	9	its	its	PRON
fcis-32489	89	10	structural	structural	ADJ
fcis-32489	89	11	information	information	NOUN
fcis-32489	89	12	can	can	AUX
fcis-32489	89	13	be	be	AUX
fcis-32489	89	14	captured	capture	VERB
fcis-32489	89	15	by	by	ADP
fcis-32489	89	16	an	an	DET
fcis-32489	89	17	n×n	n×n	PROPN
fcis-32489	89	18	adjacency	adjacency	NOUN
fcis-32489	89	19	matrix	matrix	NOUN
fcis-32489	89	20	a	a	PRON
fcis-32489	89	21	,	,	PUNCT
fcis-32489	89	22	where	where	SCONJ
fcis-32489	89	23	an	an	DET
fcis-32489	89	24	entry	entry	NOUN
fcis-32489	89	25	1a	1a	X
fcis-32489	89	26	ij	ij	NOUN
fcis-32489	89	27	(	(	PUNCT
fcis-32489	89	28	or	or	CCONJ
fcis-32489	89	29	a	a	DET
fcis-32489	89	30	corresponding	corresponding	ADJ
fcis-32489	89	31	weight	weight	NOUN
fcis-32489	89	32	)	)	PUNCT
fcis-32489	89	33	if	if	SCONJ
fcis-32489	89	34	there	there	PRON
fcis-32489	89	35	exists	exist	VERB
fcis-32489	89	36	a	a	DET
fcis-32489	89	37	connection	connection	NOUN
fcis-32489	89	38	between	between	ADP
fcis-32489	89	39	node	node	NOUN
fcis-32489	89	40	i	i	PROPN
fcis-32489	89	41	and	and	CCONJ
fcis-32489	89	42	node	node	VERB
fcis-32489	89	43	j.in	j.in	PROPN
fcis-32489	89	44	a	a	DET
fcis-32489	89	45	gcn	gcn	ADJ
fcis-32489	89	46	model	model	NOUN
fcis-32489	89	47	with	with	ADP
fcis-32489	89	48	l	l	PROPN
fcis-32489	89	49	layers	layer	NOUN
fcis-32489	89	50	,	,	PUNCT
fcis-32489	89	51	the	the	DET
fcis-32489	89	52	input	input	NOUN
fcis-32489	89	53	feature	feature	NOUN
fcis-32489	89	54	vector	vector	NOUN
fcis-32489	89	55	of	of	ADP
fcis-32489	89	56	a	a	DET
fcis-32489	89	57	node	node	NOUN
fcis-32489	89	58	i	i	PRON
fcis-32489	89	59	at	at	ADP
fcis-32489	89	60	layer	layer	NOUN
fcis-32489	89	61	l-1	l-1	NUM
fcis-32489	89	62	is	be	AUX
fcis-32489	89	63	typically	typically	ADV
fcis-32489	89	64	denoted	denote	VERB
fcis-32489	89	65	as	as	ADP
fcis-32489	89	66	ℎ	ℎ	PROPN
fcis-32489	89	67	.	.	PUNCT
fcis-32489	90	1	after	after	ADP
fcis-32489	90	2	applying	apply	VERB
fcis-32489	90	3	the	the	DET
fcis-32489	90	4	graph	graph	NOUN
fcis-32489	90	5	convolution	convolution	NOUN
fcis-32489	90	6	operation	operation	NOUN
fcis-32489	90	7	,	,	PUNCT
fcis-32489	90	8	the	the	DET
fcis-32489	90	9	corresponding	corresponding	ADJ
fcis-32489	90	10	output	output	NOUN
fcis-32489	90	11	feature	feature	NOUN
fcis-32489	90	12	vector	vector	NOUN
fcis-32489	90	13	at	at	ADP
fcis-32489	90	14	layer	layer	NOUN
fcis-32489	90	15	l	l	NOUN
fcis-32489	90	16	becomesℎ	becomesℎ	NOUN
fcis-32489	90	17	,	,	PUNCT
fcis-32489	90	18	and	and	CCONJ
fcis-32489	90	19	this	this	DET
fcis-32489	90	20	process	process	NOUN
fcis-32489	90	21	can	can	AUX
fcis-32489	90	22	be	be	AUX
fcis-32489	90	23	formally	formally	ADV
fcis-32489	90	24	expressed	express	VERB
fcis-32489	90	25	as	as	ADP
fcis-32489	90	26	:	:	PUNCT
fcis-32489	90	27	hi	hi	INTJ
fcis-32489	90	28	l	l	NOUN
fcis-32489	90	29	=	=	NOUN
fcis-32489	90	30	σ	σ	PRON
fcis-32489	90	31	aij	aij	PROPN
fcis-32489	90	32	n	n	CCONJ
fcis-32489	90	33	j=1	j=1	PROPN
fcis-32489	90	34	wlhj	wlhj	NOUN
fcis-32489	90	35	l-1+bl	l-1+bl	NOUN
fcis-32489	90	36	(	(	PUNCT
fcis-32489	90	37	8)	8)	NUM
fcis-32489	90	38	where𝑊	where𝑊	PROPN
fcis-32489	90	39	is	be	AUX
fcis-32489	90	40	a	a	DET
fcis-32489	90	41	linear	linear	ADJ
fcis-32489	90	42	transformation	transformation	NOUN
fcis-32489	90	43	,	,	PUNCT
fcis-32489	90	44	𝑏	𝑏	PROPN
fcis-32489	90	45	is	be	AUX
fcis-32489	90	46	a	a	DET
fcis-32489	90	47	bias	bias	NOUN
fcis-32489	90	48	term	term	NOUN
fcis-32489	90	49	,	,	PUNCT
fcis-32489	90	50			PROPN
fcis-32489	90	51	is	be	AUX
fcis-32489	90	52	a	a	DET
fcis-32489	90	53	nonlinear	nonlinear	ADJ
fcis-32489	90	54	function	function	NOUN
fcis-32489	90	55	(	(	PUNCT
fcis-32489	90	56	e.g.relu	e.g.relu	NOUN
fcis-32489	90	57	)	)	PUNCT
fcis-32489	90	58	.	.	PUNCT
fcis-32489	91	1	in	in	ADP
fcis-32489	91	2	the	the	DET
fcis-32489	91	3	convolution	convolution	NOUN
fcis-32489	91	4	process	process	NOUN
fcis-32489	91	5	of	of	ADP
fcis-32489	91	6	each	each	DET
fcis-32489	91	7	graph	graph	NOUN
fcis-32489	91	8	,	,	PUNCT
fcis-32489	91	9	each	each	DET
fcis-32489	91	10	node	node	NOUN
fcis-32489	91	11	collects	collect	VERB
fcis-32489	91	12	and	and	CCONJ
fcis-32489	91	13	aggregates	aggregate	VERB
fcis-32489	91	14	information	information	NOUN
fcis-32489	91	15	from	from	ADP
fcis-32489	91	16	adjacent	adjacent	ADJ
fcis-32489	91	17	nodes	node	NOUN
fcis-32489	91	18	in	in	ADP
fcis-32489	91	19	the	the	DET
fcis-32489	91	20	graph	graph	NOUN
fcis-32489	91	21	.	.	PUNCT
fcis-32489	92	1	if	if	SCONJ
fcis-32489	92	2	formula	formula	NOUN
fcis-32489	92	3	(	(	PUNCT
fcis-32489	92	4	8)	8)	NUM
fcis-32489	92	5	is	be	AUX
fcis-32489	92	6	directly	directly	ADV
fcis-32489	92	7	used	use	VERB
fcis-32489	92	8	,	,	PUNCT
fcis-32489	92	9	there	there	PRON
fcis-32489	92	10	will	will	AUX
fcis-32489	92	11	be	be	AUX
fcis-32489	92	12	a	a	DET
fcis-32489	92	13	significant	significant	ADJ
fcis-32489	92	14	magnitude	magnitude	NOUN
fcis-32489	92	15	difference	difference	NOUN
fcis-32489	92	16	between	between	ADP
fcis-32489	92	17	connections	connection	NOUN
fcis-32489	92	18	of	of	ADP
fcis-32489	92	19	different	different	ADJ
fcis-32489	92	20	nodes	node	NOUN
fcis-32489	92	21	,	,	PUNCT
fcis-32489	92	22	resulting	result	VERB
fcis-32489	92	23	in	in	ADP
fcis-32489	92	24	the	the	DET
fcis-32489	92	25	feature	feature	NOUN
fcis-32489	92	26	representation	representation	NOUN
fcis-32489	92	27	of	of	ADP
fcis-32489	92	28	the	the	DET
fcis-32489	92	29	sentence	sentence	NOUN
fcis-32489	92	30	not	not	PART
fcis-32489	92	31	containing	contain	VERB
fcis-32489	92	32	information	information	NOUN
fcis-32489	92	33	about	about	ADP
fcis-32489	92	34	the	the	DET
fcis-32489	92	35	nodes	node	NOUN
fcis-32489	92	36	themselves	themselves	PRON
fcis-32489	92	37	,	,	PUNCT
fcis-32489	92	38	but	but	CCONJ
fcis-32489	92	39	simply	simply	ADV
fcis-32489	92	40	favoring	favor	VERB
fcis-32489	92	41	higher	high	ADJ
fcis-32489	92	42	-	-	PUNCT
fcis-32489	92	43	order	order	NOUN
fcis-32489	92	44	nodes	node	NOUN
fcis-32489	92	45	.	.	PUNCT
fcis-32489	93	1	therefore	therefore	ADV
fcis-32489	93	2	,	,	PUNCT
fcis-32489	93	3	in	in	ADP
fcis-32489	93	4	practical	practical	ADJ
fcis-32489	93	5	applications	application	NOUN
fcis-32489	93	6	,	,	PUNCT
fcis-32489	93	7	normalization	normalization	NOUN
fcis-32489	93	8	is	be	AUX
fcis-32489	93	9	required	require	VERB
fcis-32489	93	10	for	for	ADP
fcis-32489	93	11	the	the	DET
fcis-32489	93	12	adjacency	adjacency	NOUN
fcis-32489	93	13	matrix𝐴	matrix𝐴	INTJ
fcis-32489	93	14	;	;	PUNCT
fcis-32489	93	15	in	in	ADP
fcis-32489	93	16	addition	addition	NOUN
fcis-32489	93	17	,	,	PUNCT
fcis-32489	93	18	nodes	node	NOUN
fcis-32489	93	19	in	in	ADP
fcis-32489	93	20	formula	formula	NOUN
fcis-32489	93	21	(	(	PUNCT
fcis-32489	93	22	8)	8)	NUM
fcis-32489	93	23	are	be	AUX
fcis-32489	93	24	not	not	PART
fcis-32489	93	25	connected	connect	VERB
fcis-32489	93	26	to	to	ADP
fcis-32489	93	27	themselves	themselves	PRON
fcis-32489	93	28	,	,	PUNCT
fcis-32489	93	29	meaning	mean	VERB
fcis-32489	93	30	that	that	SCONJ
fcis-32489	93	31	no	no	DET
fcis-32489	93	32	information	information	NOUN
fcis-32489	93	33	is	be	AUX
fcis-32489	93	34	transmitted	transmit	VERB
fcis-32489	93	35	between	between	ADP
fcis-32489	93	36	ℎ	ℎ	NOUN
fcis-32489	93	37	and	and	CCONJ
fcis-32489	93	38	sℎ	sℎ	PROPN
fcis-32489	93	39	.	.	PUNCT
fcis-32489	94	1	based	base	VERB
fcis-32489	94	2	on	on	ADP
fcis-32489	94	3	this	this	PRON
fcis-32489	94	4	,	,	PUNCT
fcis-32489	94	5	this	this	DET
fcis-32489	94	6	paper	paper	NOUN
fcis-32489	94	7	adds	add	VERB
fcis-32489	94	8	a	a	DET
fcis-32489	94	9	selfloop	selfloop	NOUN
fcis-32489	94	10	mechanism	mechanism	NOUN
fcis-32489	94	11	to	to	ADP
fcis-32489	94	12	the	the	DET
fcis-32489	94	13	graph	graph	NOUN
fcis-32489	94	14	structure	structure	NOUN
fcis-32489	94	15	,	,	PUNCT
fcis-32489	94	16	and	and	CCONJ
fcis-32489	94	17	then	then	ADV
fcis-32489	94	18	sets	set	VERB
fcis-32489	94	19	the	the	DET
fcis-32489	94	20	diagonal	diagonal	ADJ
fcis-32489	94	21	elements	element	NOUN
fcis-32489	94	22	of	of	ADP
fcis-32489	94	23	𝐴	𝐴	PROPN
fcis-32489	94	24	to	to	ADP
fcis-32489	94	25	1	1	NUM
fcis-32489	94	26	,	,	PUNCT
fcis-32489	94	27	that	that	ADV
fcis-32489	94	28	is	is	ADV
fcis-32489	94	29	,	,	PUNCT
fcis-32489	94	30	𝐴	𝐴	PROPN
fcis-32489	94	31	=	=	NOUN
fcis-32489	94	32	1	1	NUM
fcis-32489	94	33	,	,	PUNCT
fcis-32489	94	34	thus	thus	ADV
fcis-32489	94	35	obtaining	obtain	VERB
fcis-32489	94	36	the	the	DET
fcis-32489	94	37	optimized	optimize	VERB
fcis-32489	94	38	adjacency	adjacency	NOUN
fcis-32489	94	39	matrix𝐴ij	matrix𝐴ij	PROPN
fcis-32489	94	40	.	.	PUNCT
fcis-32489	95	1	finally	finally	ADV
fcis-32489	95	2	,	,	PUNCT
fcis-32489	95	3	it	it	PRON
fcis-32489	95	4	is	be	AUX
fcis-32489	95	5	fed	feed	VERB
fcis-32489	95	6	back	back	ADV
fcis-32489	95	7	to	to	ADP
fcis-32489	95	8	gcn	gcn	NOUN
fcis-32489	95	9	using	use	VERB
fcis-32489	95	10	a	a	DET
fcis-32489	95	11	nonlinear	nonlinear	ADJ
fcis-32489	95	12	function	function	NOUN
fcis-32489	95	13	.	.	PUNCT
fcis-32489	96	1	this	this	DET
fcis-32489	96	2	improvement	improvement	NOUN
fcis-32489	96	3	makes	make	VERB
fcis-32489	96	4	the	the	DET
fcis-32489	96	5	main	main	ADJ
fcis-32489	96	6	features	feature	NOUN
fcis-32489	96	7	in	in	ADP
fcis-32489	96	8	the	the	DET
fcis-32489	96	9	graph	graph	NOUN
fcis-32489	96	10	still	still	ADV
fcis-32489	96	11	the	the	DET
fcis-32489	96	12	nodes	node	NOUN
fcis-32489	96	13	themselves	themselves	PRON
fcis-32489	96	14	,	,	PUNCT
fcis-32489	96	15	which	which	PRON
fcis-32489	96	16	is	be	AUX
fcis-32489	96	17	more	more	ADJ
fcis-32489	96	18	in	in	ADP
fcis-32489	96	19	line	line	NOUN
fcis-32489	96	20	with	with	ADP
fcis-32489	96	21	the	the	DET
fcis-32489	96	22	principle	principle	NOUN
fcis-32489	96	23	of	of	ADP
fcis-32489	96	24	feature	feature	NOUN
fcis-32489	96	25	extraction	extraction	NOUN
fcis-32489	96	26	.	.	PUNCT
fcis-32489	97	1	the	the	DET
fcis-32489	97	2	revised	revise	VERB
fcis-32489	97	3	calculation	calculation	NOUN
fcis-32489	97	4	formula	formula	NOUN
fcis-32489	97	5	is	be	AUX
fcis-32489	97	6	shown	show	VERB
fcis-32489	97	7	in	in	ADP
fcis-32489	97	8	formula	formula	NOUN
fcis-32489	97	9	(	(	PUNCT
fcis-32489	97	10	9	9	NUM
fcis-32489	97	11	)	)	PUNCT
fcis-32489	97	12	.	.	PUNCT
fcis-32489	98	1	hi	hi	INTJ
fcis-32489	99	1	l	l	X
fcis-32489	99	2	=	=	PROPN
fcis-32489	99	3	σ	σ	PRON
fcis-32489	99	4	aij	aij	PROPN
fcis-32489	99	5	n	n	CCONJ
fcis-32489	99	6	j=1	j=1	PROPN
fcis-32489	99	7	wlhj	wlhj	NOUN
fcis-32489	99	8	l-1+bl	l-1+bl	NOUN
fcis-32489	99	9	(	(	PUNCT
fcis-32489	99	10	9	9	NUM
fcis-32489	99	11	)	)	PUNCT
fcis-32489	99	12	where	where	SCONJ
fcis-32489	99	13	𝐴ij	𝐴ij	PROPN
fcis-32489	99	14	is	be	AUX
fcis-32489	99	15	the	the	DET
fcis-32489	99	16	improved	improved	ADJ
fcis-32489	99	17	adjacency	adjacency	NOUN
fcis-32489	99	18	matrix	matrix	NOUN
fcis-32489	99	19	.	.	PUNCT
fcis-32489	100	1	the	the	DET
fcis-32489	100	2	gcn	gcn	PROPN
fcis-32489	100	3	model	model	NOUN
fcis-32489	100	4	described	describe	VERB
fcis-32489	100	5	above	above	ADP
fcis-32489	100	6	uses	use	VERB
fcis-32489	100	7	the	the	DET
fcis-32489	100	8	same	same	ADJ
fcis-32489	100	9	parameters	parameter	NOUN
fcis-32489	100	10	for	for	ADP
fcis-32489	100	11	all	all	DET
fcis-32489	100	12	edges	edge	NOUN
fcis-32489	100	13	in	in	ADP
fcis-32489	100	14	the	the	DET
fcis-32489	100	15	graph	graph	NOUN
fcis-32489	100	16	.	.	PUNCT
fcis-32489	101	1	use	use	VERB
fcis-32489	101	2	different	different	ADJ
fcis-32489	101	3	transition	transition	NOUN
fcis-32489	101	4	matrices	matrix	NOUN
fcis-32489	101	5	w	w	ADP
fcis-32489	101	6	for	for	ADP
fcis-32489	101	7	top	top	ADJ
fcis-32489	101	8	-	-	PUNCT
fcis-32489	101	9	down	down	NOUN
fcis-32489	101	10	,	,	PUNCT
fcis-32489	101	11	bottom	bottom	NOUN
fcis-32489	101	12	-	-	PUNCT
fcis-32489	101	13	up	up	NOUN
fcis-32489	101	14	,	,	PUNCT
fcis-32489	101	15	and	and	CCONJ
fcis-32489	101	16	self	self	NOUN
fcis-32489	101	17	-	-	PUNCT
fcis-32489	101	18	cyclic	cyclic	ADJ
fcis-32489	101	19	edges	edge	NOUN
fcis-32489	101	20	;	;	PUNCT
fcis-32489	101	21	add	add	VERB
fcis-32489	101	22	specific	specific	ADJ
fcis-32489	101	23	parameters	parameter	NOUN
fcis-32489	101	24	for	for	ADP
fcis-32489	101	25	controlling	control	VERB
fcis-32489	101	26	edge	edge	NOUN
fcis-32489	101	27	connectivity	connectivity	NOUN
fcis-32489	101	28	,	,	PUNCT
fcis-32489	101	29	similar	similar	ADJ
fcis-32489	101	30	to	to	ADP
fcis-32489	101	31	the	the	DET
fcis-32489	101	32	approach	approach	NOUN
fcis-32489	101	33	proposed	propose	VERB
fcis-32489	101	34	by	by	ADP
fcis-32489	101	35	marcheggiani	marcheggiani	NOUN
fcis-32489	101	36	and	and	CCONJ
fcis-32489	101	37	titov	titov	NOUN
fcis-32489	102	1	[	[	X
fcis-32489	102	2	27	27	NUM
fcis-32489	102	3	]	]	PUNCT
fcis-32489	102	4	in	in	ADP
fcis-32489	102	5	2017	2017	NUM
fcis-32489	102	6	.	.	PUNCT
fcis-32489	103	1	it	it	PRON
fcis-32489	103	2	was	be	AUX
fcis-32489	103	3	found	find	VERB
fcis-32489	103	4	through	through	ADP
fcis-32489	103	5	experiments	experiment	NOUN
fcis-32489	103	6	that	that	SCONJ
fcis-32489	103	7	adding	add	VERB
fcis-32489	103	8	directed	direct	VERB
fcis-32489	103	9	edges	edge	NOUN
fcis-32489	103	10	to	to	ADP
fcis-32489	103	11	the	the	DET
fcis-32489	103	12	model	model	NOUN
fcis-32489	103	13	did	do	AUX
fcis-32489	103	14	not	not	PART
fcis-32489	103	15	improve	improve	VERB
fcis-32489	103	16	the	the	DET
fcis-32489	103	17	model	model	NOUN
fcis-32489	103	18	's	's	PART
fcis-32489	103	19	performance	performance	NOUN
fcis-32489	103	20	,	,	PUNCT
fcis-32489	103	21	and	and	CCONJ
fcis-32489	103	22	that	that	SCONJ
fcis-32489	103	23	adding	add	VERB
fcis-32489	103	24	control	control	NOUN
fcis-32489	103	25	over	over	ADP
fcis-32489	103	26	edge	edge	NOUN
fcis-32489	103	27	connectivity	connectivity	NOUN
fcis-32489	103	28	reduced	reduce	VERB
fcis-32489	103	29	the	the	DET
fcis-32489	103	30	model	model	NOUN
fcis-32489	103	31	's	's	PART
fcis-32489	103	32	accuracy	accuracy	NOUN
fcis-32489	103	33	.	.	PUNCT
fcis-32489	104	1	it	it	PRON
fcis-32489	104	2	is	be	AUX
fcis-32489	104	3	hypothesized	hypothesize	VERB
fcis-32489	104	4	that	that	SCONJ
fcis-32489	104	5	this	this	PRON
fcis-32489	104	6	is	be	AUX
fcis-32489	104	7	because	because	SCONJ
fcis-32489	104	8	the	the	DET
fcis-32489	104	9	proposed	propose	VERB
fcis-32489	104	10	gcn	gcn	NOUN
fcis-32489	104	11	model	model	NOUN
fcis-32489	104	12	is	be	AUX
fcis-32489	104	13	usually	usually	ADV
fcis-32489	104	14	capable	capable	ADJ
fcis-32489	104	15	of	of	ADP
fcis-32489	104	16	capturing	capture	VERB
fcis-32489	104	17	the	the	DET
fcis-32489	104	18	information	information	NOUN
fcis-32489	104	19	of	of	ADP
fcis-32489	104	20	the	the	DET
fcis-32489	104	21	edges	edge	NOUN
fcis-32489	104	22	needed	need	VERB
fcis-32489	104	23	for	for	ADP
fcis-32489	104	24	classifying	classify	VERB
fcis-32489	104	25	relations	relation	NOUN
fcis-32489	104	26	,	,	PUNCT
fcis-32489	104	27	and	and	CCONJ
fcis-32489	104	28	adding	add	VERB
fcis-32489	104	29	features	feature	NOUN
fcis-32489	104	30	for	for	ADP
fcis-32489	104	31	the	the	DET
fcis-32489	104	32	direction	direction	NOUN
fcis-32489	104	33	and	and	CCONJ
fcis-32489	104	34	connectivity	connectivity	NOUN
fcis-32489	104	35	of	of	ADP
fcis-32489	104	36	the	the	DET
fcis-32489	104	37	edges	edge	NOUN
fcis-32489	104	38	again	again	ADV
fcis-32489	104	39	does	do	AUX
fcis-32489	104	40	not	not	PART
fcis-32489	104	41	provide	provide	VERB
fcis-32489	104	42	a	a	DET
fcis-32489	104	43	stronger	strong	ADJ
fcis-32489	104	44	classification	classification	NOUN
fcis-32489	104	45	ability	ability	NOUN
fcis-32489	104	46	for	for	ADP
fcis-32489	104	47	the	the	DET
fcis-32489	104	48	model	model	NOUN
fcis-32489	104	49	,	,	PUNCT
fcis-32489	104	50	but	but	CCONJ
fcis-32489	104	51	instead	instead	ADV
fcis-32489	104	52	causes	cause	VERB
fcis-32489	104	53	overfitting	overfitte	VERB
fcis-32489	104	54	.	.	PUNCT
fcis-32489	105	1	for	for	ADP
fcis-32489	105	2	example	example	NOUN
fcis-32489	105	3	,	,	PUNCT
fcis-32489	105	4	the	the	DET
fcis-32489	105	5	relations	relation	NOUN
fcis-32489	105	6	contained	contain	VERB
fcis-32489	105	7	in	in	ADP
fcis-32489	105	8	"	"	PUNCT
fcis-32489	105	9	a	a	PRON
fcis-32489	105	10	's	's	PART
fcis-32489	105	11	son	son	NOUN
fcis-32489	105	12	,	,	PUNCT
fcis-32489	105	13	b	b	PROPN
fcis-32489	105	14	"	"	PUNCT
fcis-32489	105	15	and	and	CCONJ
fcis-32489	105	16	"	"	PUNCT
fcis-32489	105	17	b	b	PROPN
fcis-32489	105	18	's	's	PART
fcis-32489	105	19	son	son	NOUN
fcis-32489	105	20	,	,	PUNCT
fcis-32489	105	21	a	a	PRON
fcis-32489	105	22	"	"	PUNCT
fcis-32489	105	23	can	can	AUX
fcis-32489	105	24	be	be	AUX
fcis-32489	105	25	easily	easily	ADV
fcis-32489	105	26	distinguished	distinguish	VERB
fcis-32489	105	27	by	by	ADP
fcis-32489	105	28	the	the	DET
fcis-32489	105	29	's	'	NOUN
fcis-32489	105	30	on	on	ADP
fcis-32489	105	31	different	different	ADJ
fcis-32489	105	32	entities	entity	NOUN
fcis-32489	105	33	,	,	PUNCT
fcis-32489	105	34	even	even	ADV
fcis-32489	105	35	without	without	ADP
fcis-32489	105	36	considering	consider	VERB
fcis-32489	105	37	the	the	DET
fcis-32489	105	38	directionality	directionality	NOUN
fcis-32489	105	39	of	of	ADP
fcis-32489	105	40	the	the	DET
fcis-32489	105	41	edges	edge	NOUN
fcis-32489	105	42	.	.	PUNCT
fcis-32489	106	1	3.4	3.4	NUM
fcis-32489	106	2	.	.	PUNCT
fcis-32489	107	1	categorical	categorical	ADJ
fcis-32489	107	2	layers	layer	NOUN
fcis-32489	107	3	softmax	softmax	NOUN
fcis-32489	107	4	is	be	AUX
fcis-32489	107	5	a	a	DET
fcis-32489	107	6	multi	multi	ADJ
fcis-32489	107	7	classification	classification	NOUN
fcis-32489	107	8	model	model	NOUN
fcis-32489	107	9	with	with	ADP
fcis-32489	107	10	a	a	DET
fcis-32489	107	11	wide	wide	ADJ
fcis-32489	107	12	range	range	NOUN
fcis-32489	107	13	of	of	ADP
fcis-32489	107	14	applications	application	NOUN
fcis-32489	107	15	.	.	PUNCT
fcis-32489	108	1	for	for	ADP
fcis-32489	108	2	each	each	DET
fcis-32489	108	3	named	name	VERB
fcis-32489	108	4	entity	entity	NOUN
fcis-32489	108	5	mentioned	mention	VERB
fcis-32489	108	6	(	(	PUNCT
fcis-32489	108	7	which	which	PRON
fcis-32489	108	8	37	37	NUM
fcis-32489	108	9	appears	appear	VERB
fcis-32489	108	10	in	in	ADP
fcis-32489	108	11	different	different	ADJ
fcis-32489	108	12	sentences	sentence	NOUN
fcis-32489	108	13	)	)	PUNCT
fcis-32489	108	14	mk	mk	NOUN
fcis-32489	108	15	,	,	PUNCT
fcis-32489	108	16	from	from	ADP
fcis-32489	108	17	the	the	DET
fcis-32489	108	18	s	s	PROPN
fcis-32489	108	19	-	-	PUNCT
fcis-32489	108	20	th	th	VERB
fcis-32489	108	21	word	word	NOUN
fcis-32489	108	22	to	to	ADP
fcis-32489	108	23	the	the	DET
fcis-32489	108	24	t	t	PROPN
fcis-32489	108	25	-	-	PUNCT
fcis-32489	108	26	th	th	VERB
fcis-32489	108	27	word	word	NOUN
fcis-32489	108	28	,	,	PUNCT
fcis-32489	108	29	the	the	DET
fcis-32489	108	30	calculation	calculation	NOUN
fcis-32489	108	31	is	be	AUX
fcis-32489	108	32	shown	show	VERB
fcis-32489	108	33	in	in	ADP
fcis-32489	108	34	equation	equation	NOUN
fcis-32489	108	35	(	(	PUNCT
fcis-32489	108	36	10	10	NUM
fcis-32489	108	37	)	)	PUNCT
fcis-32489	108	38	.	.	PUNCT
fcis-32489	109	1	mk=	mk=	VERB
fcis-32489	109	2	1	1	NUM
fcis-32489	109	3	t	t	PROPN
fcis-32489	109	4	-	-	PUNCT
fcis-32489	109	5	s+1	s+1	PROPN
fcis-32489	109	6	hj	hj	PROPN
fcis-32489	109	7	t	t	PROPN
fcis-32489	109	8	j	j	PROPN
fcis-32489	110	1	=	=	X
fcis-32489	110	2	s	s	X
fcis-32489	110	3	(	(	PUNCT
fcis-32489	110	4	10	10	NUM
fcis-32489	110	5	)	)	PUNCT
fcis-32489	110	6	the	the	DET
fcis-32489	110	7	entity	entity	NOUN
fcis-32489	110	8	ei	ei	NOUN
fcis-32489	110	9	with	with	ADP
fcis-32489	110	10	k	k	PROPN
fcis-32489	110	11	mentions	mention	NOUN
fcis-32489	110	12	represents	represent	VERB
fcis-32489	110	13	the	the	DET
fcis-32489	110	14	average	average	NOUN
fcis-32489	110	15	of	of	ADP
fcis-32489	110	16	these	these	DET
fcis-32489	110	17	mentions	mention	NOUN
fcis-32489	110	18	,	,	PUNCT
fcis-32489	110	19	as	as	SCONJ
fcis-32489	110	20	shown	show	VERB
fcis-32489	110	21	in	in	ADP
fcis-32489	110	22	equation	equation	NOUN
fcis-32489	110	23	(	(	PUNCT
fcis-32489	110	24	11	11	NUM
fcis-32489	110	25	)	)	PUNCT
fcis-32489	110	26	.	.	PUNCT
fcis-32489	111	1	ei=	ei=	PROPN
fcis-32489	111	2	1	1	NUM
fcis-32489	111	3	k	k	PROPN
fcis-32489	111	4	mk	mk	PROPN
fcis-32489	111	5	k	k	X
fcis-32489	111	6	(	(	PUNCT
fcis-32489	111	7	11	11	NUM
fcis-32489	111	8	)	)	PUNCT
fcis-32489	111	9	in	in	ADP
fcis-32489	111	10	this	this	DET
fcis-32489	111	11	section	section	NOUN
fcis-32489	111	12	,	,	PUNCT
fcis-32489	111	13	the	the	DET
fcis-32489	111	14	obtained	obtain	VERB
fcis-32489	111	15	entity	entity	NOUN
fcis-32489	111	16	representationsei	representationsei	VERB
fcis-32489	111	17	and	and	CCONJ
fcis-32489	111	18	ej	ej	PROPN
fcis-32489	111	19	will	will	AUX
fcis-32489	111	20	be	be	AUX
fcis-32489	111	21	concatenated	concatenate	VERB
fcis-32489	111	22	with	with	ADP
fcis-32489	111	23	the	the	DET
fcis-32489	111	24	distance	distance	NOUN
fcis-32489	111	25	features	feature	VERB
fcis-32489	111	26	dij	dij	VERB
fcis-32489	111	27	and	and	CCONJ
fcis-32489	111	28	dji	dji	NOUN
fcis-32489	111	29	of	of	ADP
fcis-32489	111	30	the	the	DET
fcis-32489	111	31	two	two	NUM
fcis-32489	111	32	entity	entity	NOUN
fcis-32489	111	33	pairs	pair	NOUN
fcis-32489	111	34	and	and	CCONJ
fcis-32489	111	35	input	input	NOUN
fcis-32489	111	36	into	into	ADP
fcis-32489	111	37	the	the	DET
fcis-32489	111	38	classification	classification	NOUN
fcis-32489	111	39	layer	layer	NOUN
fcis-32489	111	40	.	.	PUNCT
fcis-32489	112	1	finally	finally	ADV
fcis-32489	112	2	,	,	PUNCT
fcis-32489	112	3	the	the	DET
fcis-32489	112	4	probability	probability	NOUN
fcis-32489	112	5	of	of	ADP
fcis-32489	112	6	each	each	DET
fcis-32489	112	7	category	category	NOUN
fcis-32489	112	8	is	be	AUX
fcis-32489	112	9	calculated	calculate	VERB
fcis-32489	112	10	using	use	VERB
fcis-32489	112	11	the	the	DET
fcis-32489	112	12	sigmoid	sigmoid	NOUN
fcis-32489	112	13	function	function	NOUN
fcis-32489	112	14	,	,	PUNCT
fcis-32489	112	15	and	and	CCONJ
fcis-32489	112	16	the	the	DET
fcis-32489	112	17	entity	entity	NOUN
fcis-32489	112	18	with	with	ADP
fcis-32489	112	19	the	the	DET
fcis-32489	112	20	highest	high	ADJ
fcis-32489	112	21	probability	probability	NOUN
fcis-32489	112	22	is	be	AUX
fcis-32489	112	23	selected	select	VERB
fcis-32489	112	24	as	as	ADP
fcis-32489	112	25	the	the	DET
fcis-32489	112	26	relationship	relationship	NOUN
fcis-32489	112	27	between	between	ADP
fcis-32489	112	28	entities	entity	NOUN
fcis-32489	112	29	.	.	PUNCT
fcis-32489	113	1	the	the	DET
fcis-32489	113	2	calculation	calculation	NOUN
fcis-32489	113	3	formula	formula	NOUN
fcis-32489	113	4	is	be	AUX
fcis-32489	113	5	shown	show	VERB
fcis-32489	113	6	in	in	ADP
fcis-32489	113	7	equation	equation	NOUN
fcis-32489	113	8	(	(	PUNCT
fcis-32489	113	9	14	14	NUM
fcis-32489	113	10	)	)	PUNCT
fcis-32489	113	11	.	.	PUNCT
fcis-32489	114	1	ei=	ei=	PROPN
fcis-32489	114	2	ei;dij	ei;dij	PROPN
fcis-32489	114	3	(	(	PUNCT
fcis-32489	114	4	12	12	NUM
fcis-32489	114	5	)	)	PUNCT
fcis-32489	114	6	�	�	PROPN
fcis-32489	114	7	̂	̂	VERB
fcis-32489	114	8	�	�	NOUN
fcis-32489	114	9	𝑒	𝑒	X
fcis-32489	114	10	;	;	PUNCT
fcis-32489	114	11	𝑑	𝑑	X
fcis-32489	114	12	(	(	PUNCT
fcis-32489	114	13	13	13	NUM
fcis-32489	114	14	)	)	PUNCT
fcis-32489	114	15	𝑃	𝑃	PROPN
fcis-32489	114	16	𝑟|	𝑟|	PROPN
fcis-32489	114	17	�	�	PROPN
fcis-32489	114	18	̂	̂	VERB
fcis-32489	114	19	�	�	PROPN
fcis-32489	114	20	,	,	PUNCT
fcis-32489	114	21	�	�	PROPN
fcis-32489	114	22	̂	̂	PROPN
fcis-32489	114	23	�	�	PROPN
fcis-32489	114	24	𝑠𝑖𝑔𝑚𝑜𝑖𝑑	𝑠𝑖𝑔𝑚𝑜𝑖𝑑	PROPN
fcis-32489	114	25	�	�	PROPN
fcis-32489	114	26	̂	̂	SYM
fcis-32489	114	27	�	�	PROPN
fcis-32489	114	28	𝑊	𝑊	PROPN
fcis-32489	114	29	�	�	PROPN
fcis-32489	114	30	̂	̂	NOUN
fcis-32489	114	31	�	�	PROPN
fcis-32489	114	32	𝑏	𝑏	PROPN
fcis-32489	114	33	(	(	PUNCT
fcis-32489	114	34	14	14	NUM
fcis-32489	114	35	)	)	PUNCT
fcis-32489	114	36	among	among	ADP
fcis-32489	114	37	them	they	PRON
fcis-32489	114	38	,	,	PUNCT
fcis-32489	114	39	;	;	PUNCT
fcis-32489	114	40	represents	represent	VERB
fcis-32489	114	41	concatenation	concatenation	NOUN
fcis-32489	114	42	,	,	PUNCT
fcis-32489	114	43	r	r	NOUN
fcis-32489	114	44	is	be	AUX
fcis-32489	114	45	the	the	DET
fcis-32489	114	46	relationship	relationship	NOUN
fcis-32489	114	47	type	type	NOUN
fcis-32489	114	48	,	,	PUNCT
fcis-32489	114	49	wr	wr	PROPN
fcis-32489	114	50	,	,	PUNCT
fcis-32489	114	51	brare	brare	VERB
fcis-32489	114	52	the	the	DET
fcis-32489	114	53	weight	weight	NOUN
fcis-32489	114	54	parameters	parameter	NOUN
fcis-32489	114	55	and	and	CCONJ
fcis-32489	114	56	bias	bias	VERB
fcis-32489	114	57	terms	term	NOUN
fcis-32489	114	58	of	of	ADP
fcis-32489	114	59	the	the	DET
fcis-32489	114	60	relationship	relationship	NOUN
fcis-32489	114	61	type	type	NOUN
fcis-32489	114	62	.	.	PUNCT
fcis-32489	115	1	4	4	X
fcis-32489	115	2	.	.	X
fcis-32489	115	3	experiment	experiment	NOUN
fcis-32489	115	4	4.1	4.1	NUM
fcis-32489	115	5	.	.	PUNCT
fcis-32489	116	1	dataset	dataset	VERB
fcis-32489	116	2	to	to	PART
fcis-32489	116	3	construct	construct	VERB
fcis-32489	116	4	the	the	DET
fcis-32489	116	5	docred	docre	VERB
fcis-32489	116	6	dataset	dataset	NOUN
fcis-32489	116	7	,	,	PUNCT
fcis-32489	116	8	researchers	researcher	NOUN
fcis-32489	116	9	employed	employ	VERB
fcis-32489	116	10	remote	remote	ADJ
fcis-32489	116	11	supervision	supervision	NOUN
fcis-32489	116	12	techniques	technique	NOUN
fcis-32489	116	13	involving	involve	VERB
fcis-32489	116	14	wikipedia	wikipedia	PROPN
fcis-32489	116	15	documents	document	NOUN
fcis-32489	116	16	and	and	CCONJ
fcis-32489	116	17	wikidata	wikidata	NOUN
fcis-32489	116	18	.	.	PUNCT
fcis-32489	117	1	this	this	DET
fcis-32489	117	2	process	process	NOUN
fcis-32489	117	3	began	begin	VERB
fcis-32489	117	4	with	with	ADP
fcis-32489	117	5	identifying	identify	VERB
fcis-32489	117	6	named	name	VERB
fcis-32489	117	7	entities	entity	NOUN
fcis-32489	117	8	within	within	ADP
fcis-32489	117	9	each	each	DET
fcis-32489	117	10	document	document	NOUN
fcis-32489	117	11	,	,	PUNCT
fcis-32489	117	12	after	after	ADP
fcis-32489	117	13	which	which	PRON
fcis-32489	117	14	these	these	DET
fcis-32489	117	15	entities	entity	NOUN
fcis-32489	117	16	were	be	AUX
fcis-32489	117	17	mapped	map	VERB
fcis-32489	117	18	to	to	ADP
fcis-32489	117	19	corresponding	corresponding	ADJ
fcis-32489	117	20	entries	entry	NOUN
fcis-32489	117	21	in	in	ADP
fcis-32489	117	22	wikidata	wikidata	NOUN
fcis-32489	117	23	.	.	PUNCT
fcis-32489	118	1	entities	entity	NOUN
fcis-32489	118	2	sharing	share	VERB
fcis-32489	118	3	the	the	DET
fcis-32489	118	4	same	same	ADJ
fcis-32489	118	5	knowledge	knowledge	NOUN
fcis-32489	118	6	base	base	NOUN
fcis-32489	118	7	identifier	identifier	NOUN
fcis-32489	118	8	were	be	AUX
fcis-32489	118	9	then	then	ADV
fcis-32489	118	10	combined	combine	VERB
fcis-32489	118	11	into	into	ADP
fcis-32489	118	12	a	a	DET
fcis-32489	118	13	single	single	ADJ
fcis-32489	118	14	entity	entity	NOUN
fcis-32489	118	15	.	.	PUNCT
fcis-32489	119	1	additionally	additionally	ADV
fcis-32489	119	2	,	,	PUNCT
fcis-32489	119	3	relationships	relationship	NOUN
fcis-32489	119	4	between	between	ADP
fcis-32489	119	5	entity	entity	NOUN
fcis-32489	119	6	pairs	pair	NOUN
fcis-32489	119	7	were	be	AUX
fcis-32489	119	8	determined	determine	VERB
fcis-32489	119	9	by	by	ADP
fcis-32489	119	10	querying	query	VERB
fcis-32489	119	11	wikidata	wikidata	NOUN
fcis-32489	119	12	,	,	PUNCT
fcis-32489	119	13	with	with	ADP
fcis-32489	119	14	supplementary	supplementary	ADJ
fcis-32489	119	15	steps	step	NOUN
fcis-32489	119	16	such	such	ADJ
fcis-32489	119	17	as	as	ADP
fcis-32489	119	18	named	name	VERB
fcis-32489	119	19	entity	entity	NOUN
fcis-32489	119	20	and	and	CCONJ
fcis-32489	119	21	coparameter	coparameter	PROPN
fcis-32489	119	22	retrieval	retrieval	NOUN
fcis-32489	119	23	,	,	PUNCT
fcis-32489	119	24	entity	entity	NOUN
fcis-32489	119	25	merging	merging	NOUN
fcis-32489	119	26	,	,	PUNCT
fcis-32489	119	27	and	and	CCONJ
fcis-32489	119	28	collection	collection	NOUN
fcis-32489	119	29	of	of	ADP
fcis-32489	119	30	relational	relational	ADJ
fcis-32489	119	31	evidence	evidence	NOUN
fcis-32489	119	32	all	all	PRON
fcis-32489	119	33	carried	carry	VERB
fcis-32489	119	34	out	out	ADP
fcis-32489	119	35	under	under	ADP
fcis-32489	119	36	the	the	DET
fcis-32489	119	37	framework	framework	NOUN
fcis-32489	119	38	of	of	ADP
fcis-32489	119	39	remote	remote	ADJ
fcis-32489	119	40	supervision	supervision	NOUN
fcis-32489	119	41	[	[	X
fcis-32489	119	42	12	12	NUM
fcis-32489	119	43	]	]	SYM
fcis-32489	119	44	table	table	NOUN
fcis-32489	119	45	1	1	NUM
fcis-32489	119	46	.	.	PUNCT
fcis-32489	119	47	dataset	dataset	PROPN
fcis-32489	119	48	statistics	statistic	NOUN
fcis-32489	119	49	settings	setting	NOUN
fcis-32489	119	50	number	number	NOUN
fcis-32489	119	51	of	of	ADP
fcis-32489	119	52	documents	document	NOUN
fcis-32489	119	53	number	number	NOUN
fcis-32489	119	54	of	of	ADP
fcis-32489	119	55	relationships	relationship	NOUN
fcis-32489	119	56	number	number	NOUN
fcis-32489	119	57	of	of	ADP
fcis-32489	119	58	relationship	relationship	NOUN
fcis-32489	119	59	instances	instance	NOUN
fcis-32489	119	60	number	number	NOUN
fcis-32489	119	61	of	of	ADP
fcis-32489	119	62	relationship	relationship	NOUN
fcis-32489	119	63	facts	fact	NOUN
fcis-32489	119	64	train	train	VERB
fcis-32489	119	65	3053	3053	NUM
fcis-32489	119	66	96	96	NUM
fcis-32489	119	67	38269	38269	NUM
fcis-32489	119	68	34715	34715	NUM
fcis-32489	119	69	dev	dev	NOUN
fcis-32489	119	70	1000	1000	NUM
fcis-32489	119	71	96	96	NUM
fcis-32489	119	72	12332	12332	NUM
fcis-32489	119	73	11790	11790	NUM
fcis-32489	119	74	test	test	NOUN
fcis-32489	119	75	1000	1000	NUM
fcis-32489	119	76	96	96	NUM
fcis-32489	119	77	12842	12842	NUM
fcis-32489	119	78	12101	12101	NUM
fcis-32489	119	79	the	the	DET
fcis-32489	119	80	docred	docre	VERB
fcis-32489	119	81	dataset	dataset	NOUN
fcis-32489	119	82	encompasses	encompass	VERB
fcis-32489	119	83	a	a	DET
fcis-32489	119	84	diverse	diverse	ADJ
fcis-32489	119	85	spectrum	spectrum	NOUN
fcis-32489	119	86	of	of	ADP
fcis-32489	119	87	subjects	subject	NOUN
fcis-32489	119	88	.	.	PUNCT
fcis-32489	120	1	its	its	PRON
fcis-32489	120	2	entity	entity	NOUN
fcis-32489	120	3	categories	category	NOUN
fcis-32489	120	4	feature	feature	VERB
fcis-32489	120	5	individuals	individual	NOUN
fcis-32489	120	6	,	,	PUNCT
fcis-32489	120	7	locations	location	NOUN
fcis-32489	120	8	,	,	PUNCT
fcis-32489	120	9	groups	group	NOUN
fcis-32489	120	10	,	,	PUNCT
fcis-32489	120	11	periods	period	NOUN
fcis-32489	120	12	,	,	PUNCT
fcis-32489	120	13	quantities	quantity	NOUN
fcis-32489	120	14	,	,	PUNCT
fcis-32489	120	15	and	and	CCONJ
fcis-32489	120	16	proper	proper	ADJ
fcis-32489	120	17	nouns	noun	NOUN
fcis-32489	120	18	,	,	PUNCT
fcis-32489	120	19	while	while	SCONJ
fcis-32489	120	20	relationship	relationship	NOUN
fcis-32489	120	21	classifications	classification	NOUN
fcis-32489	120	22	include	include	VERB
fcis-32489	120	23	scientific	scientific	ADJ
fcis-32489	120	24	,	,	PUNCT
fcis-32489	120	25	artistic	artistic	ADJ
fcis-32489	120	26	,	,	PUNCT
fcis-32489	120	27	temporal	temporal	ADJ
fcis-32489	120	28	,	,	PUNCT
fcis-32489	120	29	and	and	CCONJ
fcis-32489	120	30	interpersonal	interpersonal	ADJ
fcis-32489	120	31	domains	domain	NOUN
fcis-32489	120	32	,	,	PUNCT
fcis-32489	120	33	among	among	ADP
fcis-32489	120	34	others	other	NOUN
fcis-32489	120	35	.	.	PUNCT
fcis-32489	121	1	achieving	achieve	VERB
fcis-32489	121	2	optimal	optimal	ADJ
fcis-32489	121	3	performance	performance	NOUN
fcis-32489	121	4	on	on	ADP
fcis-32489	121	5	this	this	DET
fcis-32489	121	6	dataset	dataset	NOUN
fcis-32489	121	7	demands	demand	NOUN
fcis-32489	121	8	an	an	DET
fcis-32489	121	9	array	array	NOUN
fcis-32489	121	10	of	of	ADP
fcis-32489	121	11	reasoning	reason	VERB
fcis-32489	121	12	capabilities	capability	NOUN
fcis-32489	121	13	,	,	PUNCT
fcis-32489	121	14	such	such	ADJ
fcis-32489	121	15	as	as	ADP
fcis-32489	121	16	pattern	pattern	NOUN
fcis-32489	121	17	identification	identification	NOUN
fcis-32489	121	18	,	,	PUNCT
fcis-32489	121	19	logical	logical	ADJ
fcis-32489	121	20	deduction	deduction	NOUN
fcis-32489	121	21	,	,	PUNCT
fcis-32489	121	22	anaphora	anaphora	ADJ
fcis-32489	121	23	resolution	resolution	NOUN
fcis-32489	121	24	,	,	PUNCT
fcis-32489	121	25	and	and	CCONJ
fcis-32489	121	26	everyday	everyday	ADJ
fcis-32489	121	27	knowledge	knowledge	NOUN
fcis-32489	121	28	inference	inference	PROPN
fcis-32489	121	29	.	.	PUNCT
fcis-32489	122	1	additionally	additionally	ADV
fcis-32489	122	2	,	,	PUNCT
fcis-32489	122	3	the	the	DET
fcis-32489	122	4	dataset	dataset	NOUN
fcis-32489	122	5	offers	offer	VERB
fcis-32489	122	6	both	both	DET
fcis-32489	122	7	annotated	annotate	VERB
fcis-32489	122	8	training	training	NOUN
fcis-32489	122	9	examples	example	NOUN
fcis-32489	122	10	(	(	PUNCT
fcis-32489	122	11	drawn	draw	VERB
fcis-32489	122	12	from	from	ADP
fcis-32489	122	13	a	a	DET
fcis-32489	122	14	combination	combination	NOUN
fcis-32489	122	15	of	of	ADP
fcis-32489	122	16	remote	remote	ADJ
fcis-32489	122	17	supervision	supervision	NOUN
fcis-32489	122	18	and	and	CCONJ
fcis-32489	122	19	manual	manual	ADJ
fcis-32489	122	20	labeling	labeling	NOUN
fcis-32489	122	21	efforts	effort	NOUN
fcis-32489	122	22	)	)	PUNCT
fcis-32489	122	23	and	and	CCONJ
fcis-32489	122	24	automatically	automatically	ADV
fcis-32489	122	25	generated	generate	VERB
fcis-32489	122	26	remote	remote	ADJ
fcis-32489	122	27	supervision	supervision	NOUN
fcis-32489	122	28	data	datum	NOUN
fcis-32489	122	29	.	.	PUNCT
fcis-32489	123	1	during	during	ADP
fcis-32489	123	2	our	our	PRON
fcis-32489	123	3	experimental	experimental	ADJ
fcis-32489	123	4	process	process	NOUN
fcis-32489	123	5	,	,	PUNCT
fcis-32489	123	6	only	only	ADV
fcis-32489	123	7	the	the	DET
fcis-32489	123	8	manually	manually	ADV
fcis-32489	123	9	annotated	annotate	VERB
fcis-32489	123	10	instances	instance	NOUN
fcis-32489	123	11	were	be	AUX
fcis-32489	123	12	utilized	utilize	VERB
fcis-32489	123	13	.	.	PUNCT
fcis-32489	124	1	detailed	detailed	ADJ
fcis-32489	124	2	statistics	statistic	NOUN
fcis-32489	124	3	regarding	regard	VERB
fcis-32489	124	4	the	the	DET
fcis-32489	124	5	dataset	dataset	NOUN
fcis-32489	124	6	are	be	AUX
fcis-32489	124	7	presented	present	VERB
fcis-32489	124	8	in	in	ADP
fcis-32489	124	9	table	table	NOUN
fcis-32489	124	10	1	1	NUM
fcis-32489	124	11	.	.	PUNCT
fcis-32489	124	12	4.2	4.2	NUM
fcis-32489	124	13	.	.	PUNCT
fcis-32489	125	1	experimental	experimental	ADJ
fcis-32489	125	2	parameter	parameter	NOUN
fcis-32489	125	3	settings	setting	NOUN
fcis-32489	125	4	table	table	VERB
fcis-32489	125	5	2	2	NUM
fcis-32489	125	6	.	.	PUNCT
fcis-32489	125	7	model	model	NOUN
fcis-32489	125	8	parameter	parameter	PROPN
fcis-32489	125	9	settings	setting	NOUN
fcis-32489	125	10	model	model	PROPN
fcis-32489	125	11	parameters	parameter	NOUN
fcis-32489	125	12	value	value	NOUN
fcis-32489	125	13	learning	learn	VERB
fcis-32489	125	14	rate	rate	NOUN
fcis-32489	125	15	0.0001	0.0001	NUM
fcis-32489	125	16	number	number	NOUN
fcis-32489	125	17	of	of	ADP
fcis-32489	125	18	network	network	NOUN
fcis-32489	125	19	iterations	iteration	VERB
fcis-32489	125	20	200	200	NUM
fcis-32489	125	21	batch	batch	NOUN
fcis-32489	125	22	number	number	NOUN
fcis-32489	125	23	of	of	ADP
fcis-32489	125	24	samples	sample	NOUN
fcis-32489	125	25	20	20	NUM
fcis-32489	125	26	number	number	NOUN
fcis-32489	125	27	of	of	ADP
fcis-32489	125	28	word	word	NOUN
fcis-32489	125	29	vector	vector	NOUN
fcis-32489	125	30	dimensions	dimension	NOUN
fcis-32489	125	31	100	100	NUM
fcis-32489	125	32	coreference	coreference	NOUN
fcis-32489	125	33	vector	vector	NOUN
fcis-32489	125	34	dimensions	dimension	NOUN
fcis-32489	125	35	20	20	NUM
fcis-32489	125	36	all	all	DET
fcis-32489	125	37	experiments	experiment	NOUN
fcis-32489	125	38	in	in	ADP
fcis-32489	125	39	this	this	DET
fcis-32489	125	40	paper	paper	NOUN
fcis-32489	125	41	are	be	AUX
fcis-32489	125	42	based	base	VERB
fcis-32489	125	43	on	on	ADP
fcis-32489	125	44	the	the	DET
fcis-32489	125	45	pytorch	pytorch	NOUN
fcis-32489	125	46	deep	deep	ADJ
fcis-32489	125	47	learning	learning	NOUN
fcis-32489	125	48	framework	framework	NOUN
fcis-32489	125	49	,	,	PUNCT
fcis-32489	125	50	using	use	VERB
fcis-32489	125	51	the	the	DET
fcis-32489	125	52	sigmoid	sigmoid	NOUN
fcis-32489	125	53	activation	activation	NOUN
fcis-32489	125	54	function	function	NOUN
fcis-32489	125	55	as	as	SCONJ
fcis-32489	125	56	the	the	DET
fcis-32489	125	57	activation	activation	NOUN
fcis-32489	125	58	function	function	NOUN
fcis-32489	125	59	within	within	ADP
fcis-32489	125	60	the	the	DET
fcis-32489	125	61	model	model	NOUN
fcis-32489	125	62	,	,	PUNCT
fcis-32489	125	63	training	train	VERB
fcis-32489	125	64	the	the	DET
fcis-32489	125	65	model	model	NOUN
fcis-32489	125	66	with	with	ADP
fcis-32489	125	67	the	the	DET
fcis-32489	125	68	adam	adam	PROPN
fcis-32489	125	69	optimization	optimization	NOUN
fcis-32489	125	70	algorithm	algorithm	NOUN
fcis-32489	125	71	,	,	PUNCT
fcis-32489	125	72	and	and	CCONJ
fcis-32489	125	73	using	use	VERB
fcis-32489	125	74	the	the	DET
fcis-32489	125	75	method	method	NOUN
fcis-32489	125	76	of	of	ADP
fcis-32489	125	77	minimizing	minimize	VERB
fcis-32489	125	78	cross	cross	NOUN
fcis-32489	125	79	-	-	NOUN
fcis-32489	125	80	entropy	entropy	NOUN
fcis-32489	125	81	to	to	PART
fcis-32489	125	82	select	select	VERB
fcis-32489	125	83	the	the	DET
fcis-32489	125	84	optimal	optimal	ADJ
fcis-32489	125	85	parameters	parameter	NOUN
fcis-32489	125	86	of	of	ADP
fcis-32489	125	87	the	the	DET
fcis-32489	125	88	model	model	NOUN
fcis-32489	125	89	.	.	PUNCT
fcis-32489	126	1	for	for	ADP
fcis-32489	126	2	the	the	DET
fcis-32489	126	3	convenience	convenience	NOUN
fcis-32489	126	4	of	of	ADP
fcis-32489	126	5	model	model	NOUN
fcis-32489	126	6	comparison	comparison	NOUN
fcis-32489	126	7	,	,	PUNCT
fcis-32489	126	8	the	the	DET
fcis-32489	126	9	common	common	ADJ
fcis-32489	126	10	parameters	parameter	NOUN
fcis-32489	126	11	of	of	ADP
fcis-32489	126	12	different	different	ADJ
fcis-32489	126	13	models	model	NOUN
fcis-32489	126	14	in	in	ADP
fcis-32489	126	15	subsequent	subsequent	ADJ
fcis-32489	126	16	experiments	experiment	NOUN
fcis-32489	126	17	were	be	AUX
fcis-32489	126	18	compared	compare	VERB
fcis-32489	126	19	using	use	VERB
fcis-32489	126	20	the	the	DET
fcis-32489	126	21	optimal	optimal	ADJ
fcis-32489	126	22	parameters	parameter	NOUN
fcis-32489	126	23	,	,	PUNCT
fcis-32489	126	24	and	and	CCONJ
fcis-32489	126	25	the	the	DET
fcis-32489	126	26	optimal	optimal	ADJ
fcis-32489	126	27	parameter	parameter	NOUN
fcis-32489	126	28	settings	setting	NOUN
fcis-32489	126	29	of	of	ADP
fcis-32489	126	30	the	the	DET
fcis-32489	126	31	models	model	NOUN
fcis-32489	126	32	are	be	AUX
fcis-32489	126	33	shown	show	VERB
fcis-32489	126	34	in	in	ADP
fcis-32489	126	35	table	table	NOUN
fcis-32489	126	36	2	2	NUM
fcis-32489	126	37	.	.	X
fcis-32489	127	1	4.3	4.3	NUM
fcis-32489	127	2	.	.	PUNCT
fcis-32489	128	1	experimental	experimental	ADJ
fcis-32489	128	2	results	result	NOUN
fcis-32489	128	3	this	this	DET
fcis-32489	128	4	paper	paper	NOUN
fcis-32489	128	5	compares	compare	VERB
fcis-32489	128	6	the	the	DET
fcis-32489	128	7	current	current	ADJ
fcis-32489	128	8	mainstream	mainstream	NOUN
fcis-32489	128	9	relation	relation	NOUN
fcis-32489	128	10	extraction	extraction	NOUN
fcis-32489	128	11	model	model	NOUN
fcis-32489	128	12	with	with	ADP
fcis-32489	128	13	the	the	DET
fcis-32489	128	14	model	model	NOUN
fcis-32489	128	15	of	of	ADP
fcis-32489	128	16	this	this	DET
fcis-32489	128	17	paper	paper	NOUN
fcis-32489	128	18	on	on	ADP
fcis-32489	128	19	the	the	DET
fcis-32489	128	20	docred	docre	VERB
fcis-32489	128	21	dataset	dataset	NOUN
fcis-32489	128	22	,	,	PUNCT
fcis-32489	128	23	and	and	CCONJ
fcis-32489	128	24	the	the	DET
fcis-32489	128	25	corresponding	corresponding	ADJ
fcis-32489	128	26	model	model	NOUN
fcis-32489	128	27	is	be	AUX
fcis-32489	128	28	as	as	SCONJ
fcis-32489	128	29	follows	follow	VERB
fcis-32489	128	30	.	.	PUNCT
fcis-32489	129	1	1)cnn[28]/lstm[29]/bilstm[30	1)cnn[28]/lstm[29]/bilstm[30	NUM
fcis-32489	129	2	]	]	X
fcis-32489	129	3	:	:	PUNCT
fcis-32489	129	4	cnn	cnn	PROPN
fcis-32489	129	5	/	/	SYM
fcis-32489	129	6	lstm	lstm	PROPN
fcis-32489	129	7	/	/	SYM
fcis-32489	129	8	bilstm	bilstm	NOUN
fcis-32489	129	9	is	be	AUX
fcis-32489	129	10	used	use	VERB
fcis-32489	129	11	as	as	ADP
fcis-32489	129	12	an	an	DET
fcis-32489	129	13	encoder	encoder	NOUN
fcis-32489	129	14	to	to	PART
fcis-32489	129	15	encode	encode	VERB
fcis-32489	129	16	the	the	DET
fcis-32489	129	17	document	document	NOUN
fcis-32489	129	18	into	into	ADP
fcis-32489	129	19	a	a	DET
fcis-32489	129	20	hidden	hide	VERB
fcis-32489	129	21	sequence	sequence	NOUN
fcis-32489	129	22	of	of	ADP
fcis-32489	129	23	state	state	NOUN
fcis-32489	129	24	vectors	vector	NOUN
fcis-32489	129	25	,	,	PUNCT
fcis-32489	129	26	and	and	CCONJ
fcis-32489	129	27	then	then	ADV
fcis-32489	129	28	a	a	DET
fcis-32489	129	29	bilinear	bilinear	NOUN
fcis-32489	129	30	function	function	NOUN
fcis-32489	129	31	is	be	AUX
fcis-32489	129	32	input	input	NOUN
fcis-32489	129	33	to	to	PART
fcis-32489	129	34	predict	predict	VERB
fcis-32489	129	35	the	the	DET
fcis-32489	129	36	relationship	relationship	NOUN
fcis-32489	129	37	of	of	ADP
fcis-32489	129	38	each	each	DET
fcis-32489	129	39	entity	entity	NOUN
fcis-32489	129	40	pair	pair	NOUN
fcis-32489	129	41	.	.	PUNCT
fcis-32489	130	1	2	2	X
fcis-32489	130	2	)	)	PUNCT
fcis-32489	130	3	context	context	NOUN
fcis-32489	130	4	-	-	PUNCT
fcis-32489	130	5	aware	aware	ADJ
fcis-32489	130	6	[	[	X
fcis-32489	130	7	31	31	NUM
fcis-32489	130	8	]	]	PUNCT
fcis-32489	130	9	:	:	PUNCT
fcis-32489	130	10	by	by	ADP
fcis-32489	130	11	using	use	VERB
fcis-32489	130	12	an	an	DET
fcis-32489	130	13	lstm	lstm	NOUN
fcis-32489	130	14	-	-	PUNCT
fcis-32489	130	15	based	base	VERB
fcis-32489	130	16	encoder	encoder	NOUN
fcis-32489	130	17	,	,	PUNCT
fcis-32489	130	18	all	all	DET
fcis-32489	130	19	the	the	DET
fcis-32489	130	20	representations	representation	NOUN
fcis-32489	130	21	of	of	ADP
fcis-32489	130	22	the	the	DET
fcis-32489	130	23	relationships	relationship	NOUN
fcis-32489	130	24	in	in	ADP
fcis-32489	130	25	the	the	DET
fcis-32489	130	26	context	context	NOUN
fcis-32489	130	27	are	be	AUX
fcis-32489	130	28	jointly	jointly	ADV
fcis-32489	130	29	learned	learn	VERB
fcis-32489	130	30	,	,	PUNCT
fcis-32489	130	31	and	and	CCONJ
fcis-32489	130	32	then	then	ADV
fcis-32489	130	33	other	other	ADJ
fcis-32489	130	34	context	context	NOUN
fcis-32489	130	35	relationships	relationship	NOUN
fcis-32489	130	36	are	be	AUX
fcis-32489	130	37	combined	combine	VERB
fcis-32489	130	38	with	with	ADP
fcis-32489	130	39	the	the	DET
fcis-32489	130	40	target	target	NOUN
fcis-32489	130	41	relationship	relationship	NOUN
fcis-32489	130	42	to	to	PART
fcis-32489	130	43	make	make	VERB
fcis-32489	130	44	the	the	DET
fcis-32489	130	45	final	final	ADJ
fcis-32489	130	46	classification	classification	NOUN
fcis-32489	130	47	.	.	PUNCT
fcis-32489	131	1	3)bert	3)bert	NUM
fcis-32489	132	1	[	[	X
fcis-32489	132	2	32	32	NUM
fcis-32489	132	3	]	]	PUNCT
fcis-32489	132	4	:	:	PUNCT
fcis-32489	132	5	use	use	VERB
fcis-32489	132	6	bert	bert	PROPN
fcis-32489	132	7	to	to	PART
fcis-32489	132	8	encode	encode	VERB
fcis-32489	132	9	the	the	DET
fcis-32489	132	10	document	document	NOUN
fcis-32489	132	11	,	,	PUNCT
fcis-32489	132	12	represent	represent	VERB
fcis-32489	132	13	entities	entity	NOUN
fcis-32489	132	14	with	with	ADP
fcis-32489	132	15	average	average	ADJ
fcis-32489	132	16	word	word	NOUN
fcis-32489	132	17	embeddings	embedding	NOUN
fcis-32489	132	18	,	,	PUNCT
fcis-32489	132	19	and	and	CCONJ
fcis-32489	132	20	use	use	VERB
fcis-32489	132	21	bilinear	bilinear	NOUN
fcis-32489	132	22	layers	layer	NOUN
fcis-32489	132	23	to	to	PART
fcis-32489	132	24	predict	predict	VERB
fcis-32489	132	25	the	the	DET
fcis-32489	132	26	relationships	relationship	NOUN
fcis-32489	132	27	between	between	ADP
fcis-32489	132	28	entity	entity	NOUN
fcis-32489	132	29	pairs	pair	NOUN
fcis-32489	132	30	.	.	PUNCT
fcis-32489	133	1	table	table	NOUN
fcis-32489	133	2	3	3	NUM
fcis-32489	133	3	.	.	PUNCT
fcis-32489	134	1	performance	performance	NOUN
fcis-32489	134	2	of	of	ADP
fcis-32489	134	3	different	different	ADJ
fcis-32489	134	4	models	model	NOUN
fcis-32489	134	5	on	on	ADP
fcis-32489	134	6	the	the	DET
fcis-32489	134	7	docred	docre	VERB
fcis-32489	134	8	dataset	dataset	NOUN
fcis-32489	134	9	model	model	NOUN
fcis-32489	134	10	dev	dev	PROPN
fcis-32489	134	11	test	test	NOUN
fcis-32489	134	12	ign	ign	PROPN
fcis-32489	134	13	f1	f1	PROPN
fcis-32489	134	14	f1	f1	PROPN
fcis-32489	134	15	ign	ign	NOUN
fcis-32489	134	16	f1	f1	PROPN
fcis-32489	134	17	f1	f1	PROPN
fcis-32489	134	18	cnn	cnn	PROPN
fcis-32489	134	19	41.58	41.58	NUM
fcis-32489	134	20	43.45	43.45	NUM
fcis-32489	134	21	40.33	40.33	NUM
fcis-32489	134	22	42.26	42.26	NUM
fcis-32489	134	23	lstm	lstm	NOUN
fcis-32489	134	24	48.44	48.44	NUM
fcis-32489	134	25	50.68	50.68	NUM
fcis-32489	134	26	47.71	47.71	NUM
fcis-32489	134	27	50.07	50.07	NUM
fcis-32489	134	28	bilstm	bilstm	NOUN
fcis-32489	134	29	48.87	48.87	NUM
fcis-32489	134	30	50.94	50.94	NUM
fcis-32489	134	31	48.78	48.78	NUM
fcis-32489	134	32	51.06	51.06	NUM
fcis-32489	134	33	context	context	NOUN
fcis-32489	134	34	-	-	PUNCT
fcis-32489	134	35	aware	aware	ADJ
fcis-32489	134	36	48.94	48.94	NUM
fcis-32489	134	37	51.09	51.09	NUM
fcis-32489	134	38	48.40	48.40	NUM
fcis-32489	134	39	50.70	50.70	NUM
fcis-32489	134	40	bilstm	bilstm	NOUN
fcis-32489	134	41	-	-	PUNCT
fcis-32489	134	42	gcn	gcn	NOUN
fcis-32489	134	43	49.02	49.02	NUM
fcis-32489	134	44	51.45	51.45	NUM
fcis-32489	134	45	49.46	49.46	NUM
fcis-32489	134	46	51.52	51.52	NUM
fcis-32489	134	47	4.4	4.4	NUM
fcis-32489	134	48	.	.	PUNCT
fcis-32489	135	1	experimental	experimental	ADJ
fcis-32489	135	2	analysis	analysis	NOUN
fcis-32489	135	3	as	as	SCONJ
fcis-32489	135	4	shown	show	VERB
fcis-32489	135	5	in	in	ADP
fcis-32489	135	6	table	table	NOUN
fcis-32489	135	7	3	3	NUM
fcis-32489	135	8	,	,	PUNCT
fcis-32489	135	9	the	the	DET
fcis-32489	135	10	f1	f1	ADJ
fcis-32489	135	11	value	value	NOUN
fcis-32489	135	12	of	of	ADP
fcis-32489	135	13	the	the	DET
fcis-32489	135	14	model	model	NOUN
fcis-32489	135	15	has	have	AUX
fcis-32489	135	16	increased	increase	VERB
fcis-32489	135	17	by	by	ADP
fcis-32489	135	18	more	more	ADJ
fcis-32489	135	19	than	than	ADP
fcis-32489	135	20	8	8	NUM
fcis-32489	135	21	%	%	NOUN
fcis-32489	135	22	compared	compare	VERB
fcis-32489	135	23	to	to	ADP
fcis-32489	135	24	cnn	cnn	PROPN
fcis-32489	135	25	.	.	PUNCT
fcis-32489	136	1	this	this	PRON
fcis-32489	136	2	is	be	AUX
fcis-32489	136	3	because	because	SCONJ
fcis-32489	136	4	the	the	DET
fcis-32489	136	5	"	"	PUNCT
fcis-32489	136	6	bilstm	bilstm	NOUN
fcis-32489	136	7	"	"	PUNCT
fcis-32489	136	8	model	model	NOUN
fcis-32489	136	9	introduced	introduce	VERB
fcis-32489	136	10	in	in	ADP
fcis-32489	136	11	the	the	DET
fcis-32489	136	12	bilstm	bilstm	NOUN
fcis-32489	136	13	-	-	PUNCT
fcis-32489	136	14	gcn	gcn	NOUN
fcis-32489	136	15	model	model	NOUN
fcis-32489	136	16	can	can	AUX
fcis-32489	136	17	learn	learn	VERB
fcis-32489	136	18	the	the	DET
fcis-32489	136	19	feature	feature	NOUN
fcis-32489	136	20	information	information	NOUN
fcis-32489	136	21	of	of	ADP
fcis-32489	136	22	the	the	DET
fcis-32489	136	23	entities	entity	NOUN
fcis-32489	136	24	in	in	ADP
fcis-32489	136	25	the	the	DET
fcis-32489	136	26	dataset	dataset	NOUN
fcis-32489	136	27	,	,	PUNCT
fcis-32489	136	28	which	which	PRON
fcis-32489	136	29	greatly	greatly	ADV
fcis-32489	136	30	overcomes	overcome	VERB
fcis-32489	136	31	the	the	DET
fcis-32489	136	32	long	long	ADJ
fcis-32489	136	33	-	-	PUNCT
fcis-32489	136	34	distance	distance	NOUN
fcis-32489	136	35	dependency	dependency	NOUN
fcis-32489	136	36	problem	problem	NOUN
fcis-32489	136	37	of	of	ADP
fcis-32489	136	38	"	"	PUNCT
fcis-32489	136	39	cnn	cnn	PROPN
fcis-32489	136	40	"	"	PUNCT
fcis-32489	136	41	.	.	PUNCT
fcis-32489	137	1	compared	compare	VERB
fcis-32489	137	2	with	with	ADP
fcis-32489	137	3	the	the	DET
fcis-32489	137	4	"	"	PUNCT
fcis-32489	137	5	lstm	lstm	ADJ
fcis-32489	137	6	"	"	PUNCT
fcis-32489	137	7	model	model	NOUN
fcis-32489	137	8	,	,	PUNCT
fcis-32489	137	9	the	the	DET
fcis-32489	137	10	"	"	PUNCT
fcis-32489	137	11	f1	f1	NOUN
fcis-32489	137	12	"	"	PUNCT
fcis-32489	137	13	value	value	NOUN
fcis-32489	137	14	of	of	ADP
fcis-32489	137	15	"	"	PUNCT
fcis-32489	137	16	bilstm	bilstm	NOUN
fcis-32489	137	17	-	-	PUNCT
fcis-32489	137	18	gcn	gcn	NOUN
fcis-32489	137	19	"	"	PUNCT
fcis-32489	137	20	has	have	AUX
fcis-32489	137	21	increased	increase	VERB
fcis-32489	137	22	by	by	ADP
fcis-32489	137	23	"	"	PUNCT
fcis-32489	137	24	0.77	0.77	NUM
fcis-32489	137	25	%	%	NOUN
fcis-32489	137	26	"	"	PUNCT
fcis-32489	137	27	,	,	PUNCT
fcis-32489	137	28	because	because	SCONJ
fcis-32489	137	29	the	the	DET
fcis-32489	137	30	model	model	NOUN
fcis-32489	137	31	can	can	AUX
fcis-32489	137	32	obtain	obtain	VERB
fcis-32489	137	33	rich	rich	ADJ
fcis-32489	137	34	semantic	semantic	ADJ
fcis-32489	137	35	context	context	NOUN
fcis-32489	137	36	information	information	NOUN
fcis-32489	137	37	.	.	PUNCT
fcis-32489	138	1	compared	compare	VERB
fcis-32489	138	2	with	with	ADP
fcis-32489	138	3	the	the	DET
fcis-32489	138	4	"	"	PUNCT
fcis-32489	138	5	bilstm	bilstm	NOUN
fcis-32489	138	6	"	"	PUNCT
fcis-32489	138	7	model	model	NOUN
fcis-32489	138	8	,	,	PUNCT
fcis-32489	138	9	the	the	DET
fcis-32489	138	10	"	"	PUNCT
fcis-32489	138	11	f1	f1	NOUN
fcis-32489	138	12	"	"	PUNCT
fcis-32489	138	13	value	value	NOUN
fcis-32489	138	14	of	of	ADP
fcis-32489	138	15	this	this	DET
fcis-32489	138	16	model	model	NOUN
fcis-32489	138	17	has	have	AUX
fcis-32489	138	18	increased	increase	VERB
fcis-32489	138	19	by	by	ADP
fcis-32489	138	20	0.51	0.51	NUM
fcis-32489	138	21	%	%	NOUN
fcis-32489	138	22	,	,	PUNCT
fcis-32489	138	23	because	because	SCONJ
fcis-32489	138	24	the	the	DET
fcis-32489	138	25	bilstm	bilstm	NOUN
fcis-32489	138	26	-	-	PUNCT
fcis-32489	138	27	gcn	gcn	NOUN
fcis-32489	138	28	model	model	NOUN
fcis-32489	138	29	,	,	PUNCT
fcis-32489	138	30	based	base	VERB
fcis-32489	138	31	on	on	ADP
fcis-32489	138	32	the	the	DET
fcis-32489	138	33	bilstm	bilstm	NOUN
fcis-32489	138	34	model	model	NOUN
fcis-32489	138	35	,	,	PUNCT
fcis-32489	138	36	uses	use	VERB
fcis-32489	138	37	graph	graph	NOUN
fcis-32489	138	38	convolution	convolution	NOUN
fcis-32489	138	39	of	of	ADP
fcis-32489	138	40	context	context	NOUN
fcis-32489	138	41	information	information	NOUN
fcis-32489	138	42	to	to	PART
fcis-32489	138	43	better	well	ADV
fcis-32489	138	44	aggregate	aggregate	VERB
fcis-32489	138	45	the	the	DET
fcis-32489	138	46	semantic	semantic	ADJ
fcis-32489	138	47	feature	feature	NOUN
fcis-32489	138	48	information	information	NOUN
fcis-32489	138	49	of	of	ADP
fcis-32489	138	50	the	the	DET
fcis-32489	138	51	target	target	NOUN
fcis-32489	138	52	entity	entity	NOUN
fcis-32489	138	53	pairs	pair	NOUN
fcis-32489	138	54	,	,	PUNCT
fcis-32489	138	55	effectively	effectively	ADV
fcis-32489	138	56	improving	improve	VERB
fcis-32489	138	57	the	the	DET
fcis-32489	138	58	aggregation	aggregation	NOUN
fcis-32489	138	59	ability	ability	NOUN
fcis-32489	138	60	of	of	ADP
fcis-32489	138	61	the	the	DET
fcis-32489	138	62	model	model	NOUN
fcis-32489	138	63	compared	compare	VERB
fcis-32489	138	64	to	to	ADP
fcis-32489	138	65	the	the	DET
fcis-32489	138	66	traditional	traditional	ADJ
fcis-32489	138	67	gcn	gcn	NOUN
fcis-32489	138	68	method	method	NOUN
fcis-32489	138	69	.	.	PUNCT
fcis-32489	139	1	compared	compare	VERB
fcis-32489	139	2	with	with	ADP
fcis-32489	139	3	the	the	DET
fcis-32489	139	4	context	context	NOUN
fcis-32489	139	5	-	-	PUNCT
fcis-32489	139	6	aware	aware	ADJ
fcis-32489	139	7	model	model	NOUN
fcis-32489	139	8	,	,	PUNCT
fcis-32489	139	9	the	the	DET
fcis-32489	139	10	f1	f1	ADJ
fcis-32489	139	11	value	value	NOUN
fcis-32489	139	12	of	of	ADP
fcis-32489	139	13	bilstm	bilstm	NOUN
fcis-32489	139	14	-	-	PUNCT
fcis-32489	139	15	gcn	gcn	NOUN
fcis-32489	139	16	has	have	AUX
fcis-32489	139	17	increased	increase	VERB
fcis-32489	139	18	by	by	ADP
fcis-32489	139	19	0.36	0.36	NUM
fcis-32489	139	20	%	%	NOUN
fcis-32489	139	21	.	.	PUNCT
fcis-32489	140	1	this	this	PRON
fcis-32489	140	2	is	be	AUX
fcis-32489	140	3	because	because	SCONJ
fcis-32489	140	4	bilstm	bilstm	NOUN
fcis-32489	140	5	-	-	PUNCT
fcis-32489	140	6	gcn	gcn	NOUN
fcis-32489	140	7	retains	retain	VERB
fcis-32489	140	8	the	the	DET
fcis-32489	140	9	context	context	NOUN
fcis-32489	140	10	feature	feature	NOUN
fcis-32489	140	11	information	information	NOUN
fcis-32489	140	12	of	of	ADP
fcis-32489	140	13	the	the	DET
fcis-32489	140	14	entities	entity	NOUN
fcis-32489	140	15	to	to	ADP
fcis-32489	140	16	a	a	DET
fcis-32489	140	17	greater	great	ADJ
fcis-32489	140	18	38	38	NUM
fcis-32489	140	19	extent	extent	NOUN
fcis-32489	140	20	,	,	PUNCT
fcis-32489	140	21	while	while	SCONJ
fcis-32489	140	22	also	also	ADV
fcis-32489	140	23	enhancing	enhance	VERB
fcis-32489	140	24	the	the	DET
fcis-32489	140	25	learning	learning	NOUN
fcis-32489	140	26	ability	ability	NOUN
fcis-32489	140	27	of	of	ADP
fcis-32489	140	28	the	the	DET
fcis-32489	140	29	model	model	NOUN
fcis-32489	140	30	and	and	CCONJ
fcis-32489	140	31	improving	improve	VERB
fcis-32489	140	32	its	its	PRON
fcis-32489	140	33	accuracy	accuracy	NOUN
fcis-32489	140	34	.	.	PUNCT
fcis-32489	141	1	5	5	X
fcis-32489	141	2	.	.	X
fcis-32489	141	3	conclusion	conclusion	NOUN
fcis-32489	141	4	this	this	DET
fcis-32489	141	5	paper	paper	NOUN
fcis-32489	141	6	proposes	propose	VERB
fcis-32489	141	7	a	a	DET
fcis-32489	141	8	document	document	NOUN
fcis-32489	141	9	-	-	PUNCT
fcis-32489	141	10	level	level	NOUN
fcis-32489	141	11	relation	relation	NOUN
fcis-32489	141	12	extraction	extraction	NOUN
fcis-32489	141	13	method	method	NOUN
fcis-32489	141	14	based	base	VERB
fcis-32489	141	15	on	on	ADP
fcis-32489	141	16	graph	graph	NOUN
fcis-32489	141	17	convolutional	convolutional	ADJ
fcis-32489	141	18	neural	neural	ADJ
fcis-32489	141	19	networks	network	NOUN
fcis-32489	141	20	,	,	PUNCT
fcis-32489	141	21	which	which	PRON
fcis-32489	141	22	learns	learn	VERB
fcis-32489	141	23	the	the	DET
fcis-32489	141	24	context	context	NOUN
fcis-32489	141	25	information	information	NOUN
fcis-32489	141	26	of	of	ADP
fcis-32489	141	27	entities	entity	NOUN
fcis-32489	141	28	through	through	ADP
fcis-32489	141	29	word	word	NOUN
fcis-32489	141	30	embeddings	embedding	NOUN
fcis-32489	141	31	,	,	PUNCT
fcis-32489	141	32	aggregates	aggregate	VERB
fcis-32489	141	33	entity	entity	NOUN
fcis-32489	141	34	feature	feature	NOUN
fcis-32489	141	35	information	information	NOUN
fcis-32489	141	36	using	use	VERB
fcis-32489	141	37	graph	graph	NOUN
fcis-32489	141	38	convolutional	convolutional	ADJ
fcis-32489	141	39	neural	neural	ADJ
fcis-32489	141	40	networks	network	NOUN
fcis-32489	141	41	,	,	PUNCT
fcis-32489	141	42	and	and	CCONJ
fcis-32489	141	43	extracts	extract	VERB
fcis-32489	141	44	longdistance	longdistance	NOUN
fcis-32489	141	45	dependency	dependency	NOUN
fcis-32489	141	46	features	feature	NOUN
fcis-32489	141	47	using	use	VERB
fcis-32489	141	48	deep	deep	ADJ
fcis-32489	141	49	graph	graph	NOUN
fcis-32489	141	50	convolutional	convolutional	ADJ
fcis-32489	141	51	networks	network	NOUN
fcis-32489	141	52	.	.	PUNCT
fcis-32489	142	1	this	this	PRON
fcis-32489	142	2	can	can	AUX
fcis-32489	142	3	effectively	effectively	ADV
fcis-32489	142	4	combine	combine	VERB
fcis-32489	142	5	local	local	ADJ
fcis-32489	142	6	and	and	CCONJ
fcis-32489	142	7	non	non	ADJ
fcis-32489	142	8	-	-	ADJ
fcis-32489	142	9	local	local	ADJ
fcis-32489	142	10	dependency	dependency	NOUN
fcis-32489	142	11	features	feature	NOUN
fcis-32489	142	12	in	in	ADP
fcis-32489	142	13	a	a	DET
fcis-32489	142	14	sentence	sentence	NOUN
fcis-32489	142	15	to	to	PART
fcis-32489	142	16	obtain	obtain	VERB
fcis-32489	142	17	a	a	DET
fcis-32489	142	18	more	more	ADV
fcis-32489	142	19	accurate	accurate	ADJ
fcis-32489	142	20	sentence	sentence	NOUN
fcis-32489	142	21	representation	representation	NOUN
fcis-32489	142	22	.	.	PUNCT
fcis-32489	143	1	through	through	ADP
fcis-32489	143	2	a	a	DET
fcis-32489	143	3	series	series	NOUN
fcis-32489	143	4	of	of	ADP
fcis-32489	143	5	experimental	experimental	ADJ
fcis-32489	143	6	comparisons	comparison	NOUN
fcis-32489	143	7	,	,	PUNCT
fcis-32489	143	8	it	it	PRON
fcis-32489	143	9	is	be	AUX
fcis-32489	143	10	proved	prove	VERB
fcis-32489	143	11	that	that	SCONJ
fcis-32489	143	12	the	the	DET
fcis-32489	143	13	method	method	NOUN
fcis-32489	143	14	proposed	propose	VERB
fcis-32489	143	15	in	in	ADP
fcis-32489	143	16	this	this	DET
fcis-32489	143	17	paper	paper	NOUN
fcis-32489	143	18	can	can	AUX
fcis-32489	143	19	effectively	effectively	ADV
fcis-32489	143	20	improve	improve	VERB
fcis-32489	143	21	the	the	DET
fcis-32489	143	22	effect	effect	NOUN
fcis-32489	143	23	of	of	ADP
fcis-32489	143	24	entity	entity	NOUN
fcis-32489	143	25	relation	relation	NOUN
fcis-32489	143	26	extraction	extraction	NOUN
fcis-32489	143	27	.	.	PUNCT
fcis-32489	144	1	the	the	DET
fcis-32489	144	2	model	model	NOUN
fcis-32489	144	3	in	in	ADP
fcis-32489	144	4	this	this	DET
fcis-32489	144	5	paper	paper	NOUN
fcis-32489	144	6	is	be	AUX
fcis-32489	144	7	based	base	VERB
fcis-32489	144	8	on	on	ADP
fcis-32489	144	9	an	an	DET
fcis-32489	144	10	english	english	ADJ
fcis-32489	144	11	dataset	dataset	NOUN
fcis-32489	144	12	.	.	PUNCT
fcis-32489	145	1	in	in	ADP
fcis-32489	145	2	subsequent	subsequent	ADJ
fcis-32489	145	3	work	work	NOUN
fcis-32489	145	4	,	,	PUNCT
fcis-32489	145	5	the	the	DET
fcis-32489	145	6	model	model	NOUN
fcis-32489	145	7	will	will	AUX
fcis-32489	145	8	be	be	AUX
fcis-32489	145	9	further	far	ADV
fcis-32489	145	10	extended	extend	VERB
fcis-32489	145	11	to	to	ADP
fcis-32489	145	12	a	a	DET
fcis-32489	145	13	chinese	chinese	ADJ
fcis-32489	145	14	corpus	corpus	PROPN
fcis-32489	145	15	dataset	dataset	NOUN
fcis-32489	145	16	.	.	PUNCT
fcis-32489	146	1	references	reference	NOUN
fcis-32489	146	2	[	[	X
fcis-32489	146	3	1	1	X
fcis-32489	146	4	]	]	PUNCT
fcis-32489	146	5	chinchor	chinchor	NOUN
fcis-32489	146	6	n	n	PROPN
fcis-32489	146	7	,	,	PUNCT
fcis-32489	146	8	marsh	marsh	PROPN
fcis-32489	146	9	e	e	PROPN
fcis-32489	146	10	,	,	PUNCT
fcis-32489	146	11	muc-7	muc-7	X
fcis-32489	146	12	information	information	NOUN
fcis-32489	146	13	extraction	extraction	NOUN
fcis-32489	146	14	task	task	NOUN
fcis-32489	146	15	definition[c]//	definition[c]//	PROPN
fcis-32489	146	16	proceedings	proceeding	NOUN
fcis-32489	146	17	of	of	ADP
fcis-32489	146	18	the	the	DET
fcis-32489	146	19	7th	7th	ADJ
fcis-32489	146	20	message	message	NOUN
fcis-32489	146	21	under	under	ADP
fcis-32489	146	22	standing	stand	VERB
fcis-32489	146	23	conference(muc-7	conference(muc-7	PROPN
fcis-32489	146	24	)	)	PUNCT
fcis-32489	146	25	.	.	PUNCT
fcis-32489	147	1	stroudsburg	stroudsburg	PROPN
fcis-32489	147	2	,	,	PUNCT
fcis-32489	147	3	pa	pa	PROPN
fcis-32489	147	4	,	,	PUNCT
fcis-32489	147	5	usa	usa	PROPN
fcis-32489	147	6	:	:	PUNCT
fcis-32489	147	7	association	association	NOUN
fcis-32489	147	8	for	for	ADP
fcis-32489	147	9	computational	computational	ADJ
fcis-32489	147	10	linguistics	linguistic	NOUN
fcis-32489	147	11	,	,	PUNCT
fcis-32489	147	12	1998	1998	NUM
fcis-32489	147	13	:	:	PUNCT
fcis-32489	147	14	359367	359367	NUM
fcis-32489	147	15	.	.	PUNCT
fcis-32489	148	1	[	[	X
fcis-32489	148	2	2	2	NUM
fcis-32489	148	3	]	]	PUNCT
fcis-32489	148	4	bayu	bayu	NOUN
fcis-32489	148	5	distiawan	distiawan	PROPN
fcis-32489	148	6	trisedya	trisedya	PROPN
fcis-32489	148	7	,	,	PUNCT
fcis-32489	148	8	gerhard	gerhard	PROPN
fcis-32489	148	9	weikum	weikum	NOUN
fcis-32489	148	10	,	,	PUNCT
fcis-32489	148	11	jianzhong	jianzhong	PROPN
fcis-32489	148	12	qi	qi	PROPN
fcis-32489	148	13	,	,	PUNCT
fcis-32489	148	14	and	and	CCONJ
fcis-32489	148	15	rui	rui	PROPN
fcis-32489	148	16	zhang	zhang	PROPN
fcis-32489	148	17	.	.	PROPN
fcis-32489	148	18	2019	2019	NUM
fcis-32489	148	19	.	.	PUNCT
fcis-32489	149	1	neural	neural	ADJ
fcis-32489	149	2	relation	relation	NOUN
fcis-32489	149	3	extraction	extraction	NOUN
fcis-32489	149	4	for	for	ADP
fcis-32489	149	5	knowledge	knowledge	NOUN
fcis-32489	149	6	base	base	NOUN
fcis-32489	149	7	enrichment	enrichment	NOUN
fcis-32489	149	8	.	.	PUNCT
fcis-32489	150	1	in	in	ADP
fcis-32489	150	2	acl	acl	PROPN
fcis-32489	150	3	,	,	PUNCT
fcis-32489	150	4	pages	page	NOUN
fcis-32489	150	5	229	229	NUM
fcis-32489	150	6	-	-	SYM
fcis-32489	150	7	240	240	NUM
fcis-32489	150	8	,	,	PUNCT
fcis-32489	150	9	florence	florence	NOUN
fcis-32489	150	10	,	,	PUNCT
fcis-32489	150	11	italy	italy	PROPN
fcis-32489	150	12	.	.	PUNCT
fcis-32489	151	1	acl	acl	PROPN
fcis-32489	151	2	.	.	PUNCT
fcis-32489	152	1	[	[	X
fcis-32489	152	2	3	3	NUM
fcis-32489	152	3	]	]	X
fcis-32489	152	4	mo	mo	PROPN
fcis-32489	152	5	yu	yu	PROPN
fcis-32489	152	6	,	,	PUNCT
fcis-32489	152	7	wenpeng	wenpeng	PROPN
fcis-32489	152	8	yin	yin	PROPN
fcis-32489	152	9	,	,	PUNCT
fcis-32489	152	10	kazi	kazi	PROPN
fcis-32489	152	11	saidul	saidul	PROPN
fcis-32489	152	12	hasan	hasan	PROPN
fcis-32489	152	13	,	,	PUNCT
fcis-32489	152	14	cicero	cicero	PROPN
fcis-32489	152	15	dos	dos	PROPN
fcis-32489	152	16	santos	santo	NOUN
fcis-32489	152	17	,	,	PUNCT
fcis-32489	152	18	bing	bing	VERB
fcis-32489	152	19	xiang	xiang	PROPN
fcis-32489	152	20	,	,	PUNCT
fcis-32489	152	21	and	and	CCONJ
fcis-32489	152	22	bowen	bowen	PROPN
fcis-32489	152	23	zhou	zhou	PROPN
fcis-32489	152	24	.	.	PROPN
fcis-32489	152	25	2017	2017	NUM
fcis-32489	152	26	.	.	PUNCT
fcis-32489	152	27	improved	improve	VERB
fcis-32489	152	28	neural	neural	ADJ
fcis-32489	152	29	relation	relation	NOUN
fcis-32489	152	30	detection	detection	NOUN
fcis-32489	152	31	for	for	ADP
fcis-32489	152	32	knowledge	knowledge	NOUN
fcis-32489	152	33	base	base	NOUN
fcis-32489	152	34	question	question	NOUN
fcis-32489	152	35	answering	answering	NOUN
fcis-32489	152	36	.	.	PUNCT
fcis-32489	153	1	in	in	ADP
fcis-32489	153	2	acl	acl	PROPN
fcis-32489	153	3	,	,	PUNCT
fcis-32489	153	4	pages	page	NOUN
fcis-32489	153	5	571	571	NUM
fcis-32489	153	6	-	-	SYM
fcis-32489	153	7	581	581	NUM
fcis-32489	153	8	,	,	PUNCT
fcis-32489	153	9	van	van	NOUN
fcis-32489	153	10	-	-	PUNCT
fcis-32489	153	11	couver	couver	PROPN
fcis-32489	153	12	,	,	PUNCT
fcis-32489	153	13	canada	canada	PROPN
fcis-32489	153	14	.	.	PUNCT
fcis-32489	154	1	acl	acl	PROPN
fcis-32489	154	2	.	.	PUNCT
fcis-32489	155	1	[	[	X
fcis-32489	155	2	4	4	NUM
fcis-32489	155	3	]	]	PUNCT
fcis-32489	155	4	tom	tom	PROPN
fcis-32489	155	5	young	young	PROPN
fcis-32489	155	6	,	,	PUNCT
fcis-32489	155	7	erik	erik	PROPN
fcis-32489	155	8	cambria	cambria	PROPN
fcis-32489	155	9	cambria	cambria	PROPN
fcis-32489	155	10	,	,	PUNCT
fcis-32489	155	11	iti	iti	PROPN
fcis-32489	155	12	chaturvedi	chaturvedi	PROPN
fcis-32489	155	13	,	,	PUNCT
fcis-32489	155	14	minlie	minlie	PROPN
fcis-32489	155	15	huang	huang	PROPN
fcis-32489	155	16	,	,	PUNCT
fcis-32489	155	17	hao	hao	PROPN
fcis-32489	155	18	zhou	zhou	PROPN
fcis-32489	155	19	,	,	PUNCT
fcis-32489	155	20	and	and	CCONJ
fcis-32489	155	21	subham	subham	VERB
fcis-32489	155	22	biswas	biswas	PROPN
fcis-32489	155	23	.	.	PUNCT
fcis-32489	156	1	2018	2018	NUM
fcis-32489	156	2	.	.	PUNCT
fcis-32489	157	1	augmenting	augment	VERB
fcis-32489	157	2	end	end	NOUN
fcis-32489	157	3	-	-	PUNCT
fcis-32489	157	4	to	to	ADP
fcis-32489	157	5	-	-	PUNCT
fcis-32489	157	6	end	end	NOUN
fcis-32489	157	7	dialog	dialog	NOUN
fcis-32489	157	8	systems	system	NOUN
fcis-32489	157	9	with	with	ADP
fcis-32489	157	10	commonsense	commonsense	ADJ
fcis-32489	157	11	knowledge	knowledge	NOUN
fcis-32489	157	12	.	.	PUNCT
fcis-32489	158	1	in	in	ADP
fcis-32489	158	2	aaai	aaai	PROPN
fcis-32489	158	3	.	.	PUNCT
fcis-32489	159	1	[	[	X
fcis-32489	159	2	5	5	NUM
fcis-32489	159	3	]	]	PUNCT
fcis-32489	159	4	yuan	yuan	NOUN
fcis-32489	159	5	yao	yao	PROPN
fcis-32489	159	6	,	,	PUNCT
fcis-32489	159	7	deming	deming	PROPN
fcis-32489	159	8	ye	ye	PROPN
fcis-32489	159	9	,	,	PUNCT
fcis-32489	159	10	peng	peng	PROPN
fcis-32489	159	11	li	li	PROPN
fcis-32489	159	12	,	,	PUNCT
fcis-32489	159	13	xu	xu	PROPN
fcis-32489	159	14	han	han	PROPN
fcis-32489	159	15	,	,	PUNCT
fcis-32489	159	16	yankai	yankai	PROPN
fcis-32489	159	17	lin	lin	PROPN
fcis-32489	159	18	,	,	PUNCT
fcis-32489	159	19	zhenghao	zhenghao	PROPN
fcis-32489	159	20	liu	liu	PROPN
fcis-32489	159	21	,	,	PUNCT
fcis-32489	159	22	zhiyuan	zhiyuan	PROPN
fcis-32489	159	23	liu	liu	PROPN
fcis-32489	159	24	,	,	PUNCT
fcis-32489	159	25	lixin	lixin	PROPN
fcis-32489	159	26	huang	huang	PROPN
fcis-32489	159	27	,	,	PUNCT
fcis-32489	159	28	jie	jie	PROPN
fcis-32489	159	29	zhou	zhou	PROPN
fcis-32489	159	30	,	,	PUNCT
fcis-32489	159	31	and	and	CCONJ
fcis-32489	159	32	maosong	maosong	PROPN
fcis-32489	159	33	sun	sun	PROPN
fcis-32489	159	34	.	.	PROPN
fcis-32489	159	35	2019	2019	NUM
fcis-32489	159	36	.	.	PUNCT
fcis-32489	160	1	docred	docre	VERB
fcis-32489	160	2	:	:	PUNCT
fcis-32489	160	3	a	a	DET
fcis-32489	160	4	large	large	ADJ
fcis-32489	160	5	-	-	PUNCT
fcis-32489	160	6	scale	scale	NOUN
fcis-32489	160	7	document	document	NOUN
fcis-32489	160	8	-	-	PUNCT
fcis-32489	160	9	level	level	NOUN
fcis-32489	160	10	relation	relation	NOUN
fcis-32489	160	11	extraction	extraction	NOUN
fcis-32489	160	12	dataset	dataset	NOUN
fcis-32489	160	13	.	.	PUNCT
fcis-32489	161	1	in	in	ADP
fcis-32489	161	2	proceedings	proceeding	NOUN
fcis-32489	161	3	ofacl	ofacl	PROPN
fcis-32489	161	4	2019	2019	NUM
fcis-32489	161	5	.	.	PUNCT
fcis-32489	162	1	[	[	X
fcis-32489	162	2	6	6	NUM
fcis-32489	162	3	]	]	X
fcis-32489	162	4	gan	gan	NOUN
fcis-32489	162	5	l	l	NOUN
fcis-32489	162	6	x	x	PROPN
fcis-32489	162	7	,	,	PUNCT
fcis-32489	162	8	wan	wan	PROPN
fcis-32489	162	9	c	c	PROPN
fcis-32489	162	10	x	x	PROPN
fcis-32489	162	11	,	,	PUNCT
fcis-32489	162	12	liu	liu	PROPN
fcis-32489	162	13	d	d	PROPN
fcis-32489	162	14	x	x	PROPN
fcis-32489	162	15	,	,	PUNCT
fcis-32489	162	16	et	et	PROPN
fcis-32489	162	17	al	al	PROPN
fcis-32489	162	18	.	.	PUNCT
fcis-32489	163	1	chinese	chinese	PROPN
fcis-32489	163	2	named	name	VERB
fcis-32489	163	3	entity	entity	NOUN
fcis-32489	163	4	relation	relation	NOUN
fcis-32489	163	5	extraction	extraction	NOUN
fcis-32489	163	6	based	base	VERB
fcis-32489	163	7	on	on	ADP
fcis-32489	163	8	syntatic	syntatic	ADJ
fcis-32489	163	9	and	and	CCONJ
fcis-32489	163	10	semantic	semantic	ADJ
fcis-32489	163	11	features	feature	NOUN
fcis-32489	163	12	[	[	X
fcis-32489	163	13	j	j	X
fcis-32489	163	14	]	]	X
fcis-32489	163	15	.	.	PUNCT
fcis-32489	164	1	journal	journal	PROPN
fcis-32489	164	2	of	of	ADP
fcis-32489	164	3	computer	computer	NOUN
fcis-32489	164	4	research	research	NOUN
fcis-32489	164	5	and	and	CCONJ
fcis-32489	164	6	development	development	NOUN
fcis-32489	164	7	,	,	PUNCT
fcis-32489	164	8	2016	2016	NUM
fcis-32489	164	9	,	,	PUNCT
fcis-32489	164	10	53(2):284	53(2):284	PROPN
fcis-32489	164	11	-	-	SYM
fcis-32489	164	12	302	302	NUM
fcis-32489	164	13	.	.	PUNCT
fcis-32489	165	1	[	[	X
fcis-32489	165	2	7	7	X
fcis-32489	165	3	]	]	X
fcis-32489	165	4	choi	choi	NOUN
fcis-32489	165	5	s	s	PROPN
fcis-32489	165	6	p	p	NOUN
fcis-32489	165	7	,	,	PUNCT
fcis-32489	165	8	lee	lee	PROPN
fcis-32489	165	9	s	s	PROPN
fcis-32489	165	10	,	,	PUNCT
fcis-32489	165	11	jung	jung	PROPN
fcis-32489	165	12	h	h	PROPN
fcis-32489	165	13	et	et	PROPN
fcis-32489	165	14	al.an	al.an	PROPN
fcis-32489	165	15	intensive	intensive	ADJ
fcis-32489	165	16	case	case	NOUN
fcis-32489	165	17	study	study	NOUN
fcis-32489	165	18	on	on	ADP
fcis-32489	165	19	kernelbased	kernelbase	VERB
fcis-32489	165	20	relation	relation	NOUN
fcis-32489	165	21	extraction[j	extraction[j	PROPN
fcis-32489	165	22	]	]	PUNCT
fcis-32489	165	23	.	.	PUNCT
fcis-32489	166	1	multimedia	multimedia	NOUN
fcis-32489	166	2	tools	tool	NOUN
fcis-32489	166	3	&	&	CCONJ
fcis-32489	166	4	applications	application	NOUN
fcis-32489	166	5	,	,	PUNCT
fcis-32489	166	6	2014	2014	NUM
fcis-32489	166	7	,	,	PUNCT
fcis-32489	166	8	71(2	71(2	NUM
fcis-32489	166	9	):	):	PUNCT
fcis-32489	166	10	741	741	NUM
fcis-32489	166	11	-	-	SYM
fcis-32489	166	12	767	767	NUM
fcis-32489	166	13	.	.	PUNCT
fcis-32489	167	1	[	[	X
fcis-32489	167	2	8	8	X
fcis-32489	167	3	]	]	X
fcis-32489	167	4	zeng	zeng	PROPN
fcis-32489	167	5	d	d	PROPN
fcis-32489	167	6	,	,	PUNCT
fcis-32489	167	7	liu	liu	PROPN
fcis-32489	167	8	k	k	PROPN
fcis-32489	167	9	,	,	PUNCT
fcis-32489	167	10	lai	lai	PROPN
fcis-32489	167	11	s	s	PROPN
fcis-32489	167	12	et	et	NOUN
fcis-32489	167	13	al.relation	al.relation	NOUN
fcis-32489	167	14	classification	classification	NOUN
fcis-32489	167	15	via	via	ADP
fcis-32489	167	16	convolutional	convolutional	ADJ
fcis-32489	167	17	deep	deep	ADJ
fcis-32489	167	18	neural	neural	ADJ
fcis-32489	167	19	network[c]//proceedings	network[c]//proceeding	NOUN
fcis-32489	167	20	of	of	ADP
fcis-32489	167	21	the	the	DET
fcis-32489	167	22	25th	25th	ADJ
fcis-32489	167	23	inter	inter	PROPN
fcis-32489	167	24	national	national	ADJ
fcis-32489	167	25	conference	conference	NOUN
fcis-32489	167	26	on	on	ADP
fcis-32489	167	27	computational	computational	ADJ
fcis-32489	167	28	linguistics	linguistic	NOUN
fcis-32489	167	29	.	.	PUNCT
fcis-32489	168	1	2014:2335	2014:2335	X
fcis-32489	168	2	-	-	SYM
fcis-32489	168	3	2344	2344	NUM
fcis-32489	168	4	.	.	PUNCT
fcis-32489	169	1	[	[	X
fcis-32489	169	2	9	9	NUM
fcis-32489	169	3	]	]	X
fcis-32489	169	4	yoon	yoon	PROPN
fcis-32489	169	5	kim	kim	PROPN
fcis-32489	169	6	.	.	PROPN
fcis-32489	169	7	2014	2014	NUM
fcis-32489	169	8	.	.	PUNCT
fcis-32489	170	1	convolutional	convolutional	ADJ
fcis-32489	170	2	neural	neural	ADJ
fcis-32489	170	3	networks	network	NOUN
fcis-32489	170	4	for	for	ADP
fcis-32489	170	5	sentence	sentence	NOUN
fcis-32489	170	6	classification	classification	NOUN
fcis-32489	170	7	.	.	PUNCT
fcis-32489	171	1	in	in	ADP
fcis-32489	171	2	proceedings	proceeding	NOUN
fcis-32489	171	3	of	of	ADP
fcis-32489	171	4	conference	conference	NOUN
fcis-32489	171	5	on	on	ADP
fcis-32489	171	6	empirical	empirical	ADJ
fcis-32489	171	7	methods	method	NOUN
fcis-32489	171	8	in	in	ADP
fcis-32489	171	9	natural	natural	ADJ
fcis-32489	171	10	language	language	NOUN
fcis-32489	171	11	processing	processing	NOUN
fcis-32489	171	12	,	,	PUNCT
fcis-32489	171	13	pages	page	NOUN
fcis-32489	171	14	1746	1746	NUM
fcis-32489	171	15	-	-	SYM
fcis-32489	171	16	1751	1751	NUM
fcis-32489	171	17	.	.	PUNCT
fcis-32489	172	1	association	association	NOUN
fcis-32489	172	2	for	for	ADP
fcis-32489	172	3	computational	computational	ADJ
fcis-32489	172	4	linguistics	linguistic	NOUN
fcis-32489	172	5	.	.	PUNCT
fcis-32489	173	1	[	[	X
fcis-32489	173	2	10	10	NUM
fcis-32489	173	3	]	]	X
fcis-32489	173	4	sunil	sunil	PROPN
fcis-32489	173	5	kumar	kumar	PROPN
fcis-32489	173	6	sahu	sahu	PROPN
fcis-32489	173	7	and	and	CCONJ
fcis-32489	173	8	ashish	ashish	PROPN
fcis-32489	173	9	anand	anand	PROPN
fcis-32489	173	10	.	.	PROPN
fcis-32489	173	11	2018	2018	NUM
fcis-32489	173	12	.	.	PUNCT
fcis-32489	174	1	drugdrug	drugdrug	NOUN
fcis-32489	174	2	interaction	interaction	NOUN
fcis-32489	174	3	extraction	extraction	NOUN
fcis-32489	174	4	from	from	ADP
fcis-32489	174	5	biomedical	biomedical	ADJ
fcis-32489	174	6	texts	text	NOUN
fcis-32489	174	7	using	use	VERB
fcis-32489	174	8	long	long	ADJ
fcis-32489	174	9	shortterm	shortterm	PROPN
fcis-32489	174	10	memory	memory	NOUN
fcis-32489	174	11	network	network	NOUN
fcis-32489	174	12	.	.	PUNCT
fcis-32489	175	1	journal	journal	PROPN
fcis-32489	175	2	of	of	ADP
fcis-32489	175	3	biomedical	biomedical	ADJ
fcis-32489	175	4	informatics	informatic	NOUN
fcis-32489	175	5	,	,	PUNCT
fcis-32489	175	6	86:15	86:15	NUM
fcis-32489	175	7	-24	-24	PUNCT
fcis-32489	175	8	.	.	PUNCT
fcis-32489	176	1	[	[	X
fcis-32489	176	2	11	11	NUM
fcis-32489	176	3	]	]	X
fcis-32489	176	4	li	li	PROPN
fcis-32489	176	5	f	f	PROPN
fcis-32489	176	6	l	l	PROPN
fcis-32489	176	7	,	,	PUNCT
fcis-32489	176	8	ke	ke	PROPN
fcis-32489	176	9	j	j	PROPN
fcis-32489	176	10	.research	.research	NOUN
fcis-32489	176	11	progress	progress	NOUN
fcis-32489	176	12	of	of	ADP
fcis-32489	176	13	entity	entity	NOUN
fcis-32489	176	14	relation	relation	NOUN
fcis-32489	176	15	ex	ex	PRON
fcis-32489	176	16	traction	traction	PROPN
fcis-32489	176	17	base	base	NOUN
fcis-32489	176	18	on	on	ADP
fcis-32489	176	19	deep	deep	ADJ
fcis-32489	176	20	learning	learning	NOUN
fcis-32489	176	21	framework	framework	NOUN
fcis-32489	176	22	[	[	X
fcis-32489	176	23	j].information	j].information	NOUN
fcis-32489	176	24	science	science	NOUN
fcis-32489	176	25	,	,	PUNCT
fcis-32489	176	26	2018	2018	NUM
fcis-32489	176	27	,	,	PUNCT
fcis-32489	176	28	v36(3):169	v36(3):169	NOUN
fcis-32489	176	29	-	-	PUNCT
fcis-32489	176	30	176	176	NUM
fcis-32489	176	31	.	.	PUNCT
fcis-32489	177	1	[	[	X
fcis-32489	177	2	12	12	NUM
fcis-32489	177	3	]	]	X
fcis-32489	177	4	mintz	mintz	PROPN
fcis-32489	177	5	m	m	PROPN
fcis-32489	177	6	,	,	PUNCT
fcis-32489	177	7	steven	steven	PROPN
fcis-32489	177	8	b	b	PROPN
fcis-32489	177	9	,	,	PUNCT
fcis-32489	177	10	rion	rion	PROPN
fcis-32489	177	11	s	s	PROPN
fcis-32489	177	12	,	,	PUNCT
fcis-32489	177	13	et	et	NOUN
fcis-32489	177	14	al	al	PROPN
fcis-32489	177	15	.distant	.distant	PUNCT
fcis-32489	177	16	super	super	ADJ
fcis-32489	177	17	vision	vision	NOUN
fcis-32489	177	18	for	for	ADP
fcis-32489	177	19	relation	relation	NOUN
fcis-32489	177	20	extraction	extraction	NOUN
fcis-32489	177	21	without	without	ADP
fcis-32489	177	22	labeled	label	VERB
fcis-32489	177	23	data[c]//proceedings	data[c]//proceedings	PROPN
fcis-32489	177	24	of	of	ADP
fcis-32489	177	25	joint	joint	ADJ
fcis-32489	177	26	conference	conference	NOUN
fcis-32489	177	27	of	of	ADP
fcis-32489	177	28	the	the	DET
fcis-32489	177	29	meeting	meeting	NOUN
fcis-32489	177	30	of	of	ADP
fcis-32489	177	31	the	the	DET
fcis-32489	177	32	acl	acl	PROPN
fcis-32489	177	33	.stroudsburg	.stroudsburg	PROPN
fcis-32489	177	34	:	:	PUNCT
fcis-32489	177	35	as	as	ADP
fcis-32489	177	36	sociation	sociation	NOUN
fcis-32489	177	37	for	for	ADP
fcis-32489	177	38	computational	computational	ADJ
fcis-32489	177	39	linguistics,2009:1003	linguistics,2009:1003	PROPN
fcis-32489	177	40	-	-	NOUN
fcis-32489	177	41	1011	1011	NUM
fcis-32489	177	42	.	.	PUNCT
fcis-32489	178	1	[	[	X
fcis-32489	178	2	13	13	NUM
fcis-32489	178	3	]	]	X
fcis-32489	178	4	hoffmann	hoffmann	PROPN
fcis-32489	178	5	r	r	PROPN
fcis-32489	178	6	,	,	PUNCT
fcis-32489	178	7	zhang	zhang	PROPN
fcis-32489	178	8	c	c	X
fcis-32489	178	9	,	,	PUNCT
fcis-32489	178	10	ling	ling	NOUN
fcis-32489	178	11	x	x	X
fcis-32489	178	12	,	,	PUNCT
fcis-32489	178	13	et	et	PROPN
fcis-32489	178	14	al	al	PROPN
fcis-32489	178	15	.	.	PUNCT
fcis-32489	179	1	knowledge	knowledge	NOUN
fcis-32489	179	2	-	-	PUNCT
fcis-32489	179	3	based	base	VERB
fcis-32489	179	4	weak	weak	ADJ
fcis-32489	179	5	supervision	supervision	NOUN
fcis-32489	179	6	for	for	ADP
fcis-32489	179	7	information	information	NOUN
fcis-32489	179	8	extraction	extraction	NOUN
fcis-32489	179	9	of	of	ADP
fcis-32489	179	10	overlapping	overlap	VERB
fcis-32489	179	11	relations	relation	NOUN
fcis-32489	179	12	[	[	X
fcis-32489	179	13	c]//proceedings	c]//proceeding	NOUN
fcis-32489	179	14	of	of	ADP
fcis-32489	179	15	the	the	DET
fcis-32489	179	16	association	association	NOUN
fcis-32489	179	17	for	for	ADP
fcis-32489	179	18	computational	computational	ADJ
fcis-32489	179	19	linguistics	linguistic	NOUN
fcis-32489	179	20	:	:	PUNCT
fcis-32489	179	21	human	human	ADJ
fcis-32489	179	22	language	language	NOUN
fcis-32489	179	23	technologies	technology	NOUN
fcis-32489	179	24	.	.	PUNCT
fcis-32489	180	1	stroudsburg	stroudsburg	PROPN
fcis-32489	180	2	:	:	PUNCT
fcis-32489	180	3	association	association	NOUN
fcis-32489	180	4	for	for	ADP
fcis-32489	180	5	computational	computational	ADJ
fcis-32489	180	6	linguistics	linguistic	NOUN
fcis-32489	180	7	,	,	PUNCT
fcis-32489	180	8	2011:541	2011:541	NUM
fcis-32489	180	9	-	-	SYM
fcis-32489	180	10	550	550	NUM
fcis-32489	180	11	.	.	PUNCT
fcis-32489	181	1	[	[	X
fcis-32489	181	2	14	14	NUM
fcis-32489	181	3	]	]	X
fcis-32489	181	4	surdeanu	surdeanu	PROPN
fcis-32489	181	5	m	m	PROPN
fcis-32489	181	6	,	,	PUNCT
fcis-32489	181	7	t	t	PROPN
fcis-32489	181	8	ibshirani	ibshirani	PROPN
fcis-32489	181	9	j	j	PROPN
fcis-32489	181	10	,	,	PUNCT
fcis-32489	181	11	nallapat	nallapat	NOUN
fcis-32489	182	1	i	i	NOUN
fcis-32489	182	2	r	r	VERB
fcis-32489	182	3	,	,	PUNCT
fcis-32489	182	4	et	et	NOUN
fcis-32489	182	5	al.multi	al.multi	NOUN
fcis-32489	182	6	-	-	PUNCT
fcis-32489	182	7	instance	instance	NOUN
fcis-32489	182	8	multi	multi	ADJ
fcis-32489	182	9	-	-	ADJ
fcis-32489	182	10	label	label	ADJ
fcis-32489	182	11	learning	learning	NOUN
fcis-32489	182	12	for	for	ADP
fcis-32489	182	13	relation	relation	NOUN
fcis-32489	182	14	extraction[c]//proceedings	extraction[c]//proceeding	NOUN
fcis-32489	182	15	of	of	ADP
fcis-32489	182	16	joint	joint	ADJ
fcis-32489	182	17	conference	conference	NOUN
fcis-32489	182	18	on	on	ADP
fcis-32489	182	19	empirical	empirical	ADJ
fcis-32489	182	20	methods	method	NOUN
fcis-32489	182	21	in	in	ADP
fcis-32489	182	22	natural	natural	ADJ
fcis-32489	182	23	language	language	NOUN
fcis-32489	182	24	processing	processing	NOUN
fcis-32489	182	25	and	and	CCONJ
fcis-32489	182	26	computational	computational	ADJ
fcis-32489	182	27	natura	natura	NOUN
fcis-32489	182	28	language	language	NOUN
fcis-32489	182	29	learning	learning	NOUN
fcis-32489	182	30	.	.	PUNCT
fcis-32489	183	1	2012	2012	NUM
fcis-32489	183	2	:	:	PUNCT
fcis-32489	183	3	455	455	NUM
fcis-32489	183	4	-	-	SYM
fcis-32489	183	5	465	465	NUM
fcis-32489	183	6	.	.	PUNCT
fcis-32489	184	1	[	[	X
fcis-32489	184	2	15	15	NUM
fcis-32489	184	3	]	]	X
fcis-32489	184	4	benjamin	benjamin	PROPN
fcis-32489	184	5	r	r	PROPN
fcis-32489	184	6	,	,	PUNCT
fcis-32489	184	7	dietrich	dietrich	PROPN
fcis-32489	184	8	k	k	PROPN
fcis-32489	184	9	.combining	.combine	VERB
fcis-32489	184	10	generative	generative	ADJ
fcis-32489	184	11	and	and	CCONJ
fcis-32489	184	12	discriminative	discriminative	NOUN
fcis-32489	184	13	model	model	NOUN
fcis-32489	184	14	scores	score	NOUN
fcis-32489	184	15	for	for	ADP
fcis-32489	184	16	distant	distant	ADJ
fcis-32489	184	17	supervision[c]//	supervision[c]//	PROPN
fcis-32489	184	18	proceedings	proceeding	NOUN
fcis-32489	184	19	of	of	ADP
fcis-32489	184	20	the	the	DET
fcis-32489	184	21	2013	2013	NUM
fcis-32489	184	22	conference	conference	NOUN
fcis-32489	184	23	on	on	ADP
fcis-32489	184	24	empirical	empirical	ADJ
fcis-32489	184	25	methods	method	NOUN
fcis-32489	184	26	in	in	ADP
fcis-32489	184	27	natural	natural	ADJ
fcis-32489	184	28	language	language	NOUN
fcis-32489	184	29	processing.2013:24	processing.2013:24	NOUN
fcis-32489	184	30	-	-	PUNCT
fcis-32489	184	31	29	29	NUM
fcis-32489	184	32	.	.	PUNCT
fcis-32489	185	1	[	[	X
fcis-32489	185	2	16	16	NUM
fcis-32489	185	3	]	]	X
fcis-32489	185	4	zeng	zeng	PROPN
fcis-32489	185	5	d	d	PROPN
fcis-32489	185	6	,	,	PUNCT
fcis-32489	185	7	liu	liu	PROPN
fcis-32489	185	8	k	k	PROPN
fcis-32489	185	9	,	,	PUNCT
fcis-32489	185	10	chen	chen	PROPN
fcis-32489	185	11	y	y	PROPN
fcis-32489	185	12	,	,	PUNCT
fcis-32489	185	13	et	et	PROPN
fcis-32489	185	14	al	al	PROPN
fcis-32489	185	15	.distant	.distant	PUNCT
fcis-32489	185	16	supervision	supervision	NOUN
fcis-32489	185	17	for	for	ADP
fcis-32489	185	18	relation	relation	NOUN
fcis-32489	185	19	extraction	extraction	NOUN
fcis-32489	185	20	via	via	ADP
fcis-32489	185	21	piecewise	piecewise	NOUN
fcis-32489	185	22	convolutional	convolutional	ADJ
fcis-32489	185	23	neural	neural	ADJ
fcis-32489	185	24	networks[c]//	networks[c]//	ADJ
fcis-32489	185	25	proceedings	proceeding	NOUN
fcis-32489	185	26	of	of	ADP
fcis-32489	185	27	conference	conference	NOUN
fcis-32489	185	28	on	on	ADP
fcis-32489	185	29	empirical	empirical	ADJ
fcis-32489	185	30	methods	method	NOUN
fcis-32489	185	31	in	in	ADP
fcis-32489	185	32	natural	natural	ADJ
fcis-32489	185	33	language	language	NOUN
fcis-32489	185	34	processing.2015:1753	processing.2015:1753	NOUN
fcis-32489	185	35	-	-	NOUN
fcis-32489	185	36	1762	1762	NUM
fcis-32489	185	37	.	.	PUNCT
fcis-32489	186	1	[	[	X
fcis-32489	186	2	17	17	NUM
fcis-32489	186	3	]	]	X
fcis-32489	186	4	lei	lei	PROPN
fcis-32489	186	5	k	k	PROPN
fcis-32489	186	6	,	,	PUNCT
fcis-32489	186	7	chen	chen	PROPN
fcis-32489	186	8	d	d	PROPN
fcis-32489	186	9	,	,	PUNCT
fcis-32489	186	10	li	li	PROPN
fcis-32489	186	11	y	y	PROPN
fcis-32489	186	12	,	,	PUNCT
fcis-32489	186	13	et	et	PROPN
fcis-32489	186	14	al.cooperative	al.cooperative	ADJ
fcis-32489	186	15	denoising	denoising	NOUN
fcis-32489	186	16	for	for	ADP
fcis-32489	186	17	distantly	distantly	ADV
fcis-32489	186	18	supervised	supervise	VERB
fcis-32489	186	19	relation	relation	NOUN
fcis-32489	186	20	ex	ex	ADJ
fcis-32489	186	21	traction[c]//proceedings	traction[c]//proceeding	NOUN
fcis-32489	186	22	of	of	ADP
fcis-32489	186	23	the	the	DET
fcis-32489	186	24	27th	27th	ADJ
fcis-32489	186	25	international	international	ADJ
fcis-32489	186	26	conference	conference	NOUN
fcis-32489	186	27	on	on	ADP
fcis-32489	186	28	computational	computational	ADJ
fcis-32489	186	29	linguistics.2018	linguistics.2018	PROPN
fcis-32489	186	30	:	:	PUNCT
fcis-32489	186	31	426	426	NUM
fcis-32489	186	32	-	-	SYM
fcis-32489	186	33	436	436	NUM
fcis-32489	186	34	.	.	PUNCT
fcis-32489	187	1	[	[	X
fcis-32489	187	2	18	18	NUM
fcis-32489	187	3	]	]	PUNCT
fcis-32489	187	4	nanyun	nanyun	PROPN
fcis-32489	187	5	peng	peng	PROPN
fcis-32489	187	6	,	,	PUNCT
fcis-32489	187	7	hoifung	hoifung	ADJ
fcis-32489	187	8	poon	poon	NOUN
fcis-32489	187	9	,	,	PUNCT
fcis-32489	187	10	chris	chris	PROPN
fcis-32489	187	11	quirk	quirk	PROPN
fcis-32489	187	12	,	,	PUNCT
fcis-32489	187	13	kristina	kristina	PROPN
fcis-32489	187	14	toutanova	toutanova	PROPN
fcis-32489	187	15	,	,	PUNCT
fcis-32489	187	16	and	and	CCONJ
fcis-32489	187	17	wen	wen	PROPN
fcis-32489	187	18	tau	tau	PROPN
fcis-32489	187	19	yih	yih	PROPN
fcis-32489	187	20	.	.	PROPN
fcis-32489	187	21	2017	2017	NUM
fcis-32489	187	22	.	.	PUNCT
fcis-32489	188	1	cross	cross	NOUN
fcis-32489	188	2	-	-	NOUN
fcis-32489	188	3	sentence	sentence	ADJ
fcis-32489	188	4	n	n	CCONJ
fcis-32489	188	5	-	-	PUNCT
fcis-32489	188	6	ary	ary	PROPN
fcis-32489	188	7	relation	relation	NOUN
fcis-32489	188	8	extraction	extraction	NOUN
fcis-32489	188	9	with	with	ADP
fcis-32489	188	10	graph	graph	NOUN
fcis-32489	188	11	lstms	lstms	ADJ
fcis-32489	188	12	.	.	PUNCT
fcis-32489	189	1	transactions	transaction	NOUN
fcis-32489	189	2	of	of	ADP
fcis-32489	189	3	the	the	DET
fcis-32489	189	4	association	association	NOUN
fcis-32489	189	5	for	for	ADP
fcis-32489	189	6	computational	computational	ADJ
fcis-32489	189	7	linguistics	linguistic	NOUN
fcis-32489	189	8	,	,	PUNCT
fcis-32489	189	9	5:101	5:101	NUM
fcis-32489	189	10	-	-	PUNCT
fcis-32489	189	11	115	115	NUM
fcis-32489	189	12	.	.	PUNCT
fcis-32489	190	1	[	[	X
fcis-32489	190	2	19	19	NUM
fcis-32489	190	3	]	]	SYM
fcis-32489	190	4	h.wang	h.wang	PROPN
fcis-32489	190	5	,	,	PUNCT
fcis-32489	190	6	c.	c.	PROPN
fcis-32489	190	7	focke	focke	PROPN
fcis-32489	190	8	,	,	PUNCT
fcis-32489	190	9	r.	r.	PROPN
fcis-32489	190	10	sylvester	sylvester	PROPN
fcis-32489	190	11	,	,	PUNCT
fcis-32489	190	12	n.	n.	PROPN
fcis-32489	190	13	mishra	mishra	PROPN
fcis-32489	190	14	,	,	PUNCT
fcis-32489	190	15	andw	andw	PROPN
fcis-32489	190	16	.	.	PUNCT
fcis-32489	191	1	wang	wang	PROPN
fcis-32489	191	2	,	,	PUNCT
fcis-32489	191	3	fine	fine	ADJ
fcis-32489	191	4	-	-	PUNCT
fcis-32489	191	5	tune	tune	NOUN
fcis-32489	191	6	bert	bert	NOUN
fcis-32489	191	7	for	for	ADP
fcis-32489	191	8	docred	docre	VERB
fcis-32489	191	9	with	with	ADP
fcis-32489	191	10	two	two	NUM
fcis-32489	191	11	-	-	PUNCT
fcis-32489	191	12	step	step	NOUN
fcis-32489	191	13	process,2019	process,2019	NOUN
fcis-32489	191	14	,	,	PUNCT
fcis-32489	191	15	arxiv:1909.11898	arxiv:1909.11898	PROPN
fcis-32489	191	16	.	.	PUNCT
fcis-32489	192	1	[	[	X
fcis-32489	192	2	online	online	X
fcis-32489	192	3	]	]	X
fcis-32489	192	4	.	.	PUNCT
fcis-32489	193	1	available	available	ADJ
fcis-32489	193	2	:	:	PUNCT
fcis-32489	194	1	http://	http://	PROPN
fcis-32489	194	2	arxiv	arxiv	PROPN
fcis-32489	194	3	.	.	PUNCT
fcis-32489	195	1	org/	org/	PRON
fcis-32489	195	2	abs/	abs/	ADJ
fcis-32489	195	3	1909	1909	NUM
fcis-32489	195	4	.	.	PUNCT
fcis-32489	196	1	11898	11898	NUM
fcis-32489	196	2	.	.	PUNCT
fcis-32489	197	1	[	[	X
fcis-32489	197	2	20	20	NUM
fcis-32489	197	3	]	]	PUNCT
fcis-32489	197	4	a.	a.	NOUN
fcis-32489	197	5	vaswani	vaswani	NOUN
fcis-32489	197	6	,	,	PUNCT
fcis-32489	197	7	n.	n.	NOUN
fcis-32489	197	8	shazeer	shazeer	NOUN
fcis-32489	197	9	,	,	PUNCT
fcis-32489	197	10	n.	n.	PROPN
fcis-32489	197	11	parmar	parmar	PROPN
fcis-32489	197	12	,	,	PUNCT
fcis-32489	197	13	j.	j.	PROPN
fcis-32489	197	14	uszkoreit	uszkoreit	PROPN
fcis-32489	197	15	,	,	PUNCT
fcis-32489	197	16	l.	l.	PROPN
fcis-32489	197	17	jones	jones	PROPN
fcis-32489	197	18	,	,	PUNCT
fcis-32489	197	19	a.	a.	PROPN
fcis-32489	197	20	n.	n.	PROPN
fcis-32489	197	21	gomez	gomez	PROPN
fcis-32489	197	22	,	,	PUNCT
fcis-32489	197	23	l.	l.	PROPN
fcis-32489	197	24	kaiser	kaiser	PROPN
fcis-32489	197	25	,	,	PUNCT
fcis-32489	197	26	and	and	CCONJ
fcis-32489	197	27	i.	i.	PROPN
fcis-32489	197	28	polosukhin	polosukhin	PROPN
fcis-32489	197	29	,	,	PUNCT
fcis-32489	197	30	attention	attention	NOUN
fcis-32489	197	31	is	be	AUX
fcis-32489	197	32	all	all	PRON
fcis-32489	197	33	you	you	PRON
fcis-32489	197	34	need	need	VERB
fcis-32489	197	35	,	,	PUNCT
fcis-32489	197	36	in	in	ADP
fcis-32489	197	37	proc	proc	NOUN
fcis-32489	197	38	.	.	PUNCT
fcis-32489	198	1	adv	adv	PROPN
fcis-32489	198	2	.	.	PUNCT
fcis-32489	198	3	neural	neural	PROPN
fcis-32489	198	4	inf	inf	PROPN
fcis-32489	198	5	.	.	PUNCT
fcis-32489	198	6	process	process	NOUN
fcis-32489	198	7	.	.	PUNCT
fcis-32489	199	1	syst	syst	PROPN
fcis-32489	199	2	.	.	PROPN
fcis-32489	199	3	,	,	PUNCT
fcis-32489	199	4	2017	2017	NUM
fcis-32489	199	5	,	,	PUNCT
fcis-32489	199	6	pp	pp	ADJ
fcis-32489	199	7	.	.	PUNCT
fcis-32489	200	1	5998	5998	NUM
fcis-32489	200	2	–	–	PUNCT
fcis-32489	200	3	6008	6008	NUM
fcis-32489	200	4	.	.	PUNCT
fcis-32489	201	1	[	[	X
fcis-32489	201	2	21	21	NUM
fcis-32489	201	3	]	]	X
fcis-32489	201	4	thomas	thomas	PROPN
fcis-32489	201	5	n	n	PRON
fcis-32489	201	6	kipf	kipf	VERB
fcis-32489	201	7	and	and	CCONJ
fcis-32489	201	8	max	max	PROPN
fcis-32489	201	9	welling	well	VERB
fcis-32489	201	10	.	.	PUNCT
fcis-32489	202	1	2017	2017	NUM
fcis-32489	202	2	.	.	PUNCT
fcis-32489	203	1	semisupervised	semisupervise	VERB
fcis-32489	203	2	classification	classification	NOUN
fcis-32489	203	3	with	with	ADP
fcis-32489	203	4	graph	graph	NOUN
fcis-32489	203	5	convolutional	convolutional	ADJ
fcis-32489	203	6	networks	network	NOUN
fcis-32489	203	7	.	.	PUNCT
fcis-32489	204	1	in	in	ADP
fcis-32489	204	2	international	international	ADJ
fcis-32489	204	3	conference	conference	NOUN
fcis-32489	204	4	on	on	ADP
fcis-32489	204	5	learning	learn	VERB
fcis-32489	204	6	representations	representation	NOUN
fcis-32489	204	7	.	.	PUNCT
fcis-32489	205	1	[	[	X
fcis-32489	205	2	22	22	NUM
fcis-32489	205	3	]	]	X
fcis-32489	205	4	diego	diego	PROPN
fcis-32489	205	5	marcheggiani	marcheggiani	PROPN
fcis-32489	205	6	and	and	CCONJ
fcis-32489	205	7	ivan	ivan	PROPN
fcis-32489	205	8	titov	titov	PROPN
fcis-32489	205	9	.	.	PUNCT
fcis-32489	206	1	2017	2017	NUM
fcis-32489	206	2	.	.	PUNCT
fcis-32489	207	1	encoding	encode	VERB
fcis-32489	207	2	sentences	sentence	NOUN
fcis-32489	207	3	with	with	ADP
fcis-32489	207	4	graph	graph	NOUN
fcis-32489	207	5	convolutional	convolutional	ADJ
fcis-32489	207	6	networks	network	NOUN
fcis-32489	207	7	for	for	ADP
fcis-32489	207	8	semantic	semantic	ADJ
fcis-32489	207	9	role	role	NOUN
fcis-32489	207	10	labeling	labeling	NOUN
fcis-32489	207	11	.	.	PUNCT
fcis-32489	208	1	in	in	ADP
fcis-32489	208	2	proceedings	proceeding	NOUN
fcis-32489	208	3	of	of	ADP
fcis-32489	208	4	conference	conference	NOUN
fcis-32489	208	5	on	on	ADP
fcis-32489	208	6	empirical	empirical	ADJ
fcis-32489	208	7	methods	method	NOUN
fcis-32489	208	8	in	in	ADP
fcis-32489	208	9	natural	natural	ADJ
fcis-32489	208	10	language	language	NOUN
fcis-32489	208	11	processing	processing	NOUN
fcis-32489	208	12	,	,	PUNCT
fcis-32489	208	13	pages	page	NOUN
fcis-32489	208	14	1506–1515	1506–1515	NUM
fcis-32489	208	15	.	.	PUNCT
fcis-32489	208	16	association	association	NOUN
fcis-32489	208	17	for	for	ADP
fcis-32489	208	18	computational	computational	ADJ
fcis-32489	208	19	linguistics	linguistic	NOUN
fcis-32489	208	20	.	.	PUNCT
fcis-32489	209	1	[	[	X
fcis-32489	209	2	.	.	NOUN
fcis-32489	209	3	2017	2017	NUM
fcis-32489	209	4	.	.	PUNCT
fcis-32489	210	1	graph	graph	NOUN
fcis-32489	210	2	-	-	PUNCT
fcis-32489	210	3	based	base	VERB
fcis-32489	210	4	neural	neural	ADJ
fcis-32489	210	5	multidocument	multidocument	NOUN
fcis-32489	210	6	summarization	summarization	NOUN
fcis-32489	210	7	.	.	PUNCT
fcis-32489	211	1	in	in	ADP
fcis-32489	211	2	proceedings	proceeding	NOUN
fcis-32489	211	3	of	of	ADP
fcis-32489	211	4	the	the	DET
fcis-32489	211	5	21st	21st	ADJ
fcis-32489	211	6	conference	conference	NOUN
fcis-32489	211	7	on	on	ADP
fcis-32489	211	8	computational	computational	ADJ
fcis-32489	211	9	linguistics	linguistic	NOUN
fcis-32489	211	10	.	.	PUNCT
fcis-32489	212	1	[	[	X
fcis-32489	212	2	23	23	NUM
fcis-32489	212	3	]	]	PUNCT
fcis-32489	212	4	michihiro	michihiro	NOUN
fcis-32489	212	5	yasunaga	yasunaga	PROPN
fcis-32489	212	6	,	,	PUNCT
fcis-32489	212	7	rui	rui	PROPN
fcis-32489	212	8	zhang	zhang	PROPN
fcis-32489	212	9	,	,	PUNCT
fcis-32489	212	10	kshitijh	kshitijh	PROPN
fcis-32489	212	11	meelu	meelu	PROPN
fcis-32489	212	12	,	,	PUNCT
fcis-32489	212	13	ayush	ayush	ADJ
fcis-32489	212	14	pareek	pareek	PROPN
fcis-32489	212	15	,	,	PUNCT
fcis-32489	212	16	krishnan	krishnan	PROPN
fcis-32489	212	17	srinivasan	srinivasan	NOUN
fcis-32489	212	18	,	,	PUNCT
fcis-32489	212	19	and	and	CCONJ
fcis-32489	212	20	dragomir	dragomir	ADJ
fcis-32489	212	21	radev	radev	PROPN
fcis-32489	212	22	.	.	PROPN
fcis-32489	213	1	2017	2017	NUM
fcis-32489	213	2	.	.	PUNCT
fcis-32489	214	1	graph	graph	NOUN
fcis-32489	214	2	-	-	PUNCT
fcis-32489	214	3	based	base	VERB
fcis-32489	214	4	neural	neural	ADJ
fcis-32489	214	5	multi	multi	ADJ
fcis-32489	214	6	-	-	ADJ
fcis-32489	214	7	document	document	ADJ
fcis-32489	214	8	summarization	summarization	NOUN
fcis-32489	214	9	.	.	PUNCT
fcis-32489	215	1	in	in	ADP
fcis-32489	215	2	proceedings	proceeding	NOUN
fcis-32489	215	3	of	of	ADP
fcis-32489	215	4	the	the	DET
fcis-32489	215	5	21st	21st	ADJ
fcis-32489	215	6	conference	conference	NOUN
fcis-32489	215	7	on	on	ADP
fcis-32489	215	8	computational	computational	ADJ
fcis-32489	215	9	natural	natural	ADJ
fcis-32489	215	10	language	language	NOUN
fcis-32489	215	11	learning	learning	NOUN
fcis-32489	215	12	,	,	PUNCT
fcis-32489	215	13	pages	page	NOUN
fcis-32489	215	14	452–462	452–462	NUM
fcis-32489	215	15	.	.	PUNCT
fcis-32489	216	1	association	association	NOUN
fcis-32489	216	2	for	for	ADP
fcis-32489	216	3	computational	computational	ADJ
fcis-32489	216	4	linguistics	linguistic	NOUN
fcis-32489	216	5	.	.	PUNCT
fcis-32489	217	1	[	[	X
fcis-32489	217	2	24	24	NUM
fcis-32489	217	3	]	]	X
fcis-32489	217	4	shikhar	shikhar	PROPN
fcis-32489	217	5	vashishth	vashishth	PROPN
fcis-32489	217	6	,	,	PUNCT
fcis-32489	217	7	shib	shib	NOUN
fcis-32489	217	8	sankar	sankar	PROPN
fcis-32489	217	9	dasgupta	dasgupta	PROPN
fcis-32489	217	10	,	,	PUNCT
fcis-32489	217	11	swayambhu	swayambhu	PROPN
fcis-32489	217	12	nath	nath	PROPN
fcis-32489	217	13	ray	ray	PROPN
fcis-32489	217	14	,	,	PUNCT
fcis-32489	217	15	and	and	CCONJ
fcis-32489	217	16	partha	partha	PROPN
fcis-32489	217	17	talukdar	talukdar	NOUN
fcis-32489	217	18	.	.	PUNCT
fcis-32489	218	1	2018	2018	NUM
fcis-32489	218	2	.	.	PUNCT
fcis-32489	219	1	dating	date	VERB
fcis-32489	219	2	documents	document	NOUN
fcis-32489	219	3	using	use	VERB
fcis-32489	219	4	graph	graph	NOUN
fcis-32489	219	5	convolution	convolution	NOUN
fcis-32489	219	6	networks	network	NOUN
fcis-32489	219	7	.	.	PUNCT
fcis-32489	220	1	in	in	ADP
fcis-32489	220	2	proceedings	proceeding	NOUN
fcis-32489	220	3	of	of	ADP
fcis-32489	220	4	the	the	DET
fcis-32489	220	5	annual	annual	ADJ
fcis-32489	220	6	meeting	meeting	NOUN
fcis-32489	220	7	of	of	ADP
fcis-32489	220	8	the	the	DET
fcis-32489	220	9	association	association	NOUN
fcis-32489	220	10	for	for	ADP
fcis-32489	220	11	computational	computational	ADJ
fcis-32489	220	12	linguistics	linguistic	NOUN
fcis-32489	220	13	,	,	PUNCT
fcis-32489	220	14	pages	page	NOUN
fcis-32489	220	15	1605	1605	NUM
fcis-32489	220	16	–	–	PUNCT
fcis-32489	220	17	1615	1615	NUM
fcis-32489	220	18	.	.	PUNCT
fcis-32489	221	1	association	association	NOUN
fcis-32489	221	2	for	for	ADP
fcis-32489	221	3	computational	computational	ADJ
fcis-32489	221	4	linguistics	linguistic	NOUN
fcis-32489	221	5	.	.	PUNCT
fcis-32489	222	1	[	[	X
fcis-32489	222	2	25	25	NUM
fcis-32489	222	3	]	]	X
fcis-32489	222	4	yuhao	yuhao	PROPN
fcis-32489	222	5	zhang	zhang	PROPN
fcis-32489	222	6	,	,	PUNCT
fcis-32489	222	7	peng	peng	PROPN
fcis-32489	222	8	qi	qi	PROPN
fcis-32489	222	9	,	,	PUNCT
fcis-32489	222	10	and	and	CCONJ
fcis-32489	222	11	christopher	christopher	PROPN
fcis-32489	222	12	d	d	PROPN
fcis-32489	222	13	manning	manning	NOUN
fcis-32489	222	14	.	.	PUNCT
fcis-32489	223	1	2018	2018	NUM
fcis-32489	223	2	.	.	PUNCT
fcis-32489	224	1	graph	graph	NOUN
fcis-32489	224	2	convolution	convolution	NOUN
fcis-32489	224	3	over	over	ADP
fcis-32489	224	4	pruned	prune	VERB
fcis-32489	224	5	dependency	dependency	NOUN
fcis-32489	224	6	trees	tree	NOUN
fcis-32489	224	7	improves	improve	VERB
fcis-32489	224	8	relation	relation	NOUN
fcis-32489	224	9	extraction	extraction	NOUN
fcis-32489	224	10	.	.	PUNCT
fcis-32489	225	1	in	in	ADP
fcis-32489	225	2	proceedings	proceeding	NOUN
fcis-32489	225	3	of	of	ADP
fcis-32489	225	4	the	the	DET
fcis-32489	225	5	2018	2018	NUM
fcis-32489	225	6	conference	conference	NOUN
fcis-32489	225	7	on	on	ADP
fcis-32489	225	8	empirical	empirical	ADJ
fcis-32489	225	9	methods	method	NOUN
fcis-32489	225	10	in	in	ADP
fcis-32489	225	11	natural	natural	ADJ
fcis-32489	225	12	language	language	NOUN
fcis-32489	225	13	processing	processing	NOUN
fcis-32489	225	14	,	,	PUNCT
fcis-32489	225	15	pages	page	NOUN
fcis-32489	225	16	2205–2215	2205–2215	NUM
fcis-32489	225	17	.	.	PUNCT
fcis-32489	226	1	association	association	NOUN
fcis-32489	226	2	for	for	ADP
fcis-32489	226	3	computational	computational	ADJ
fcis-32489	226	4	linguistics	linguistic	NOUN
fcis-32489	226	5	.	.	PUNCT
fcis-32489	227	1	[	[	X
fcis-32489	227	2	26	26	NUM
fcis-32489	227	3	]	]	X
fcis-32489	227	4	jeffrey	jeffrey	PROPN
fcis-32489	227	5	pennington	pennington	PROPN
fcis-32489	227	6	,	,	PUNCT
fcis-32489	227	7	richard	richard	PROPN
fcis-32489	227	8	socher	socher	PROPN
fcis-32489	227	9	,	,	PUNCT
fcis-32489	227	10	and	and	CCONJ
fcis-32489	227	11	christopher	christopher	PROPN
fcis-32489	227	12	manning	manning	PROPN
fcis-32489	227	13	.	.	PUNCT
fcis-32489	228	1	2014	2014	NUM
fcis-32489	228	2	.	.	PUNCT
fcis-32489	229	1	glove	glove	NOUN
fcis-32489	229	2	:	:	PUNCT
fcis-32489	229	3	global	global	ADJ
fcis-32489	229	4	vectors	vector	NOUN
fcis-32489	229	5	for	for	ADP
fcis-32489	229	6	word	word	NOUN
fcis-32489	229	7	representation	representation	NOUN
fcis-32489	229	8	.	.	PUNCT
fcis-32489	230	1	in	in	ADP
fcis-32489	230	2	proceedings	proceeding	NOUN
fcis-32489	230	3	of	of	ADP
fcis-32489	230	4	emnlp	emnlp	NOUN
fcis-32489	230	5	,	,	PUNCT
fcis-32489	230	6	pages	page	NOUN
fcis-32489	230	7	1532	1532	NUM
fcis-32489	230	8	-	-	SYM
fcis-32489	230	9	1543	1543	NUM
fcis-32489	230	10	.	.	PUNCT
fcis-32489	231	1	[	[	X
fcis-32489	231	2	27	27	NUM
fcis-32489	231	3	]	]	X
fcis-32489	231	4	diego	diego	PROPN
fcis-32489	231	5	marcheggiani	marcheggiani	PROPN
fcis-32489	231	6	and	and	CCONJ
fcis-32489	231	7	ivan	ivan	PROPN
fcis-32489	231	8	titov	titov	PROPN
fcis-32489	231	9	.	.	PUNCT
fcis-32489	232	1	2017	2017	NUM
fcis-32489	232	2	.	.	PUNCT
fcis-32489	233	1	encoding	encode	VERB
fcis-32489	233	2	sentences	sentence	NOUN
fcis-32489	233	3	with	with	ADP
fcis-32489	233	4	graph	graph	NOUN
fcis-32489	233	5	convolutional	convolutional	ADJ
fcis-32489	233	6	networks	network	NOUN
fcis-32489	233	7	for	for	ADP
fcis-32489	233	8	semantic	semantic	ADJ
fcis-32489	233	9	role	role	NOUN
fcis-32489	233	10	labeling	labeling	NOUN
fcis-32489	233	11	.	.	PUNCT
fcis-32489	234	1	proceedings	proceeding	NOUN
fcis-32489	234	2	of	of	ADP
fcis-32489	234	3	the	the	DET
fcis-32489	234	4	2017	2017	NUM
fcis-32489	234	5	conference	conference	NOUN
fcis-32489	234	6	on	on	ADP
fcis-32489	234	7	empirical	empirical	ADJ
fcis-32489	234	8	methods	method	NOUN
fcis-32489	234	9	in	in	ADP
fcis-32489	234	10	natural	natural	ADJ
fcis-32489	234	11	language	language	NOUN
fcis-32489	234	12	processing	processing	NOUN
fcis-32489	234	13	(	(	PUNCT
fcis-32489	234	14	emnlp	emnlp	ADJ
fcis-32489	234	15	2017	2017	NUM
fcis-32489	234	16	)	)	PUNCT
fcis-32489	234	17	.	.	PUNCT
fcis-32489	235	1	[	[	X
fcis-32489	235	2	28	28	NUM
fcis-32489	235	3	]	]	X
fcis-32489	235	4	daojian	daojian	PROPN
fcis-32489	235	5	zeng	zeng	PROPN
fcis-32489	235	6	,	,	PUNCT
fcis-32489	235	7	kang	kang	PROPN
fcis-32489	235	8	liu	liu	PROPN
fcis-32489	235	9	,	,	PUNCT
fcis-32489	235	10	siwei	siwei	PROPN
fcis-32489	235	11	lai	lai	PROPN
fcis-32489	235	12	,	,	PUNCT
fcis-32489	235	13	guangyou	guangyou	PROPN
fcis-32489	235	14	zhou	zhou	PROPN
fcis-32489	235	15	,	,	PUNCT
fcis-32489	235	16	and	and	CCONJ
fcis-32489	235	17	jun	jun	PROPN
fcis-32489	235	18	zhao	zhao	PROPN
fcis-32489	235	19	.	.	PUNCT
fcis-32489	236	1	2014	2014	NUM
fcis-32489	236	2	.	.	PUNCT
fcis-32489	237	1	relation	relation	NOUN
fcis-32489	237	2	classification	classification	NOUN
fcis-32489	237	3	via	via	ADP
fcis-32489	237	4	convolutional	convolutional	ADJ
fcis-32489	237	5	deep	deep	ADJ
fcis-32489	237	6	neural	neural	ADJ
fcis-32489	237	7	network	network	NOUN
fcis-32489	237	8	.	.	PUNCT
fcis-32489	238	1	in	in	ADP
fcis-32489	238	2	proceedings	proceeding	NOUN
fcis-32489	238	3	of	of	ADP
fcis-32489	238	4	coling	cole	VERB
fcis-32489	238	5	,	,	PUNCT
fcis-32489	238	6	pages	page	NOUN
fcis-32489	238	7	2335–2344	2335–2344	NUM
fcis-32489	238	8	.	.	PUNCT
fcis-32489	239	1	[	[	X
fcis-32489	239	2	29	29	NUM
fcis-32489	239	3	]	]	PUNCT
fcis-32489	239	4	sepp	sepp	PROPN
fcis-32489	239	5	hochreiter	hochreiter	PROPN
fcis-32489	239	6	and	and	CCONJ
fcis-32489	239	7	jurgen	jurgen	PROPN
fcis-32489	239	8	schmidhuber.1997.long	schmidhuber.1997.long	PROPN
fcis-32489	239	9	shortterm	shortterm	PROPN
fcis-32489	239	10	memory	memory	NOUN
fcis-32489	239	11	.	.	PUNCT
fcis-32489	240	1	neural	neural	ADJ
fcis-32489	240	2	computation	computation	NOUN
fcis-32489	240	3	,	,	PUNCT
fcis-32489	240	4	9:1735–1780	9:1735–1780	NUM
fcis-32489	240	5	.	.	PUNCT
fcis-32489	241	1	[	[	X
fcis-32489	241	2	30	30	NUM
fcis-32489	241	3	]	]	X
fcis-32489	241	4	rui	rui	PROPN
fcis-32489	241	5	cai	cai	PROPN
fcis-32489	241	6	,	,	PUNCT
fcis-32489	241	7	xiaodong	xiaodong	PROPN
fcis-32489	241	8	zhang	zhang	PROPN
fcis-32489	241	9	,	,	PUNCT
fcis-32489	241	10	and	and	CCONJ
fcis-32489	241	11	houfeng	houfeng	PROPN
fcis-32489	241	12	wang	wang	PROPN
fcis-32489	241	13	.	.	PROPN
fcis-32489	241	14	2016	2016	NUM
fcis-32489	241	15	.	.	PUNCT
fcis-32489	242	1	bidirectional	bidirectional	ADJ
fcis-32489	242	2	recurrent	recurrent	ADJ
fcis-32489	242	3	convolutional	convolutional	ADJ
fcis-32489	242	4	neural	neural	ADJ
fcis-32489	242	5	network	network	NOUN
fcis-32489	242	6	for	for	ADP
fcis-32489	242	7	relation	relation	NOUN
fcis-32489	242	8	classification	classification	NOUN
fcis-32489	242	9	.	.	PUNCT
fcis-32489	243	1	in	in	ADP
fcis-32489	243	2	proceedings	proceeding	NOUN
fcis-32489	243	3	of	of	ADP
fcis-32489	243	4	acl	acl	PROPN
fcis-32489	243	5	,	,	PUNCT
fcis-32489	243	6	pages	page	NOUN
fcis-32489	243	7	756	756	NUM
fcis-32489	243	8	-	-	SYM
fcis-32489	243	9	765	765	NUM
fcis-32489	243	10	.	.	PUNCT
fcis-32489	244	1	[	[	X
fcis-32489	244	2	31	31	NUM
fcis-32489	244	3	]	]	PUNCT
fcis-32489	244	4	daniil	daniil	PROPN
fcis-32489	244	5	sorokin	sorokin	PROPN
fcis-32489	244	6	and	and	CCONJ
fcis-32489	244	7	iryna	iryna	NOUN
fcis-32489	244	8	gurevych	gurevych	NOUN
fcis-32489	244	9	.	.	PUNCT
fcis-32489	245	1	2017	2017	NUM
fcis-32489	245	2	.	.	PUNCT
fcis-32489	246	1	contextaware	contextaware	NOUN
fcis-32489	246	2	representations	representation	NOUN
fcis-32489	246	3	for	for	ADP
fcis-32489	246	4	knowledge	knowledge	NOUN
fcis-32489	246	5	base	base	NOUN
fcis-32489	246	6	relation	relation	NOUN
fcis-32489	246	7	extraction	extraction	NOUN
fcis-32489	246	8	.	.	PUNCT
fcis-32489	247	1	in	in	ADP
fcis-32489	247	2	proceedings	proceeding	NOUN
fcis-32489	247	3	of	of	ADP
fcis-32489	247	4	emnlp	emnlp	NOUN
fcis-32489	247	5	,	,	PUNCT
fcis-32489	247	6	pages	page	NOUN
fcis-32489	247	7	1784	1784	NUM
fcis-32489	247	8	-	-	SYM
fcis-32489	247	9	1789	1789	NUM
fcis-32489	247	10	.	.	PUNCT
fcis-32489	248	1	[	[	X
fcis-32489	248	2	32	32	NUM
fcis-32489	248	3	]	]	PUNCT
fcis-32489	248	4	hong	hong	PROPN
fcis-32489	248	5	wang	wang	PROPN
fcis-32489	248	6	,	,	PUNCT
fcis-32489	248	7	christfried	christfrie	VERB
fcis-32489	248	8	focke	focke	PROPN
fcis-32489	248	9	,	,	PUNCT
fcis-32489	248	10	rob	rob	PROPN
fcis-32489	248	11	sylvester	sylvester	PROPN
fcis-32489	248	12	,	,	PUNCT
fcis-32489	248	13	nilesh	nilesh	PROPN
fcis-32489	248	14	mishra	mishra	PROPN
fcis-32489	248	15	,	,	PUNCT
fcis-32489	248	16	and	and	CCONJ
fcis-32489	248	17	william	william	PROPN
fcis-32489	248	18	wang	wang	PROPN
fcis-32489	248	19	.	.	PROPN
fcis-32489	248	20	2019	2019	NUM
fcis-32489	248	21	.	.	PUNCT
fcis-32489	249	1	fine	fine	ADJ
fcis-32489	249	2	-	-	PUNCT
fcis-32489	249	3	tune	tune	NOUN
fcis-32489	249	4	bert	bert	NOUN
fcis-32489	249	5	for	for	ADP
fcis-32489	249	6	docred	docre	VERB
fcis-32489	249	7	with	with	ADP
fcis-32489	249	8	twostep	twostep	NOUN
fcis-32489	249	9	process	process	NOUN
fcis-32489	249	10	.	.	PUNCT
fcis-32489	250	1	arxiv	arxiv	PROPN
fcis-32489	250	2	preprint	preprint	PROPN
fcis-32489	250	3	arxiv:1909.11898	arxiv:1909.11898	PROPN
fcis-32489	250	4	.	.	PUNCT