id	sid	tid	token	lemma	pos
fcis-9422	1	1	frontiers	frontier	NOUN
fcis-9422	1	2	in	in	ADP
fcis-9422	1	3	computing	computing	NOUN
fcis-9422	1	4	and	and	CCONJ
fcis-9422	1	5	intelligent	intelligent	ADJ
fcis-9422	1	6	systems	system	NOUN
fcis-9422	1	7	issn	issn	VERB
fcis-9422	1	8	:	:	PUNCT
fcis-9422	1	9	2832	2832	NUM
fcis-9422	1	10	-	-	SYM
fcis-9422	1	11	6024	6024	NUM
fcis-9422	1	12	|	|	NOUN
fcis-9422	1	13	vol	vol	NOUN
fcis-9422	1	14	.	.	PROPN
fcis-9422	2	1	4	4	NUM
fcis-9422	2	2	,	,	PUNCT
fcis-9422	2	3	no	no	INTJ
fcis-9422	2	4	.	.	NOUN
fcis-9422	2	5	1	1	NUM
fcis-9422	2	6	,	,	PUNCT
fcis-9422	2	7	2023	2023	NUM
fcis-9422	2	8	46	46	NUM
fcis-9422	2	9	entity	entity	NOUN
fcis-9422	2	10	linking	link	VERB
fcis-9422	2	11	based	base	VERB
fcis-9422	2	12	on	on	ADP
fcis-9422	2	13	roformer‐sim	roformer‐sim	NOUN
fcis-9422	2	14	for	for	ADP
fcis-9422	2	15	chinese	chinese	ADJ
fcis-9422	2	16	short	short	ADJ
fcis-9422	2	17	texts	text	NOUN
fcis-9422	2	18	weiting	weite	VERB
fcis-9422	2	19	xie	xie	PROPN
fcis-9422	2	20	southwest	southwest	PROPN
fcis-9422	2	21	minzu	minzu	PROPN
fcis-9422	2	22	university	university	PROPN
fcis-9422	2	23	,	,	PUNCT
fcis-9422	2	24	chengdu	chengdu	PROPN
fcis-9422	2	25	610200	610200	NUM
fcis-9422	2	26	,	,	PUNCT
fcis-9422	2	27	china	china	PROPN
fcis-9422	2	28	abstract	abstract	NOUN
fcis-9422	2	29	:	:	PUNCT
fcis-9422	3	1	entity	entity	NOUN
fcis-9422	3	2	linking	link	VERB
fcis-9422	3	3	is	be	AUX
fcis-9422	3	4	an	an	DET
fcis-9422	3	5	important	important	ADJ
fcis-9422	3	6	means	mean	NOUN
fcis-9422	3	7	to	to	PART
fcis-9422	3	8	identify	identify	VERB
fcis-9422	3	9	named	name	VERB
fcis-9422	3	10	entities	entity	NOUN
fcis-9422	3	11	in	in	ADP
fcis-9422	3	12	text	text	NOUN
fcis-9422	3	13	and	and	CCONJ
fcis-9422	3	14	a	a	DET
fcis-9422	3	15	key	key	ADJ
fcis-9422	3	16	technology	technology	NOUN
fcis-9422	3	17	for	for	ADP
fcis-9422	3	18	constructing	construct	VERB
fcis-9422	3	19	knowledge	knowledge	NOUN
fcis-9422	3	20	graphs	graph	NOUN
fcis-9422	3	21	,	,	PUNCT
fcis-9422	3	22	playing	play	VERB
fcis-9422	3	23	an	an	DET
fcis-9422	3	24	important	important	ADJ
fcis-9422	3	25	role	role	NOUN
fcis-9422	3	26	in	in	ADP
fcis-9422	3	27	fields	field	NOUN
fcis-9422	3	28	such	such	ADJ
fcis-9422	3	29	as	as	ADP
fcis-9422	3	30	intelligent	intelligent	ADJ
fcis-9422	3	31	question	question	NOUN
fcis-9422	3	32	answering	answering	NOUN
fcis-9422	3	33	and	and	CCONJ
fcis-9422	3	34	information	information	NOUN
fcis-9422	3	35	retrieval	retrieval	NOUN
fcis-9422	3	36	.	.	PUNCT
fcis-9422	4	1	however	however	ADV
fcis-9422	4	2	,	,	PUNCT
fcis-9422	4	3	existing	exist	VERB
fcis-9422	4	4	entity	entity	NOUN
fcis-9422	4	5	linking	link	VERB
fcis-9422	4	6	methods	method	NOUN
fcis-9422	4	7	for	for	ADP
fcis-9422	4	8	short	short	ADJ
fcis-9422	4	9	texts	text	NOUN
fcis-9422	4	10	have	have	VERB
fcis-9422	4	11	low	low	ADJ
fcis-9422	4	12	accuracy	accuracy	NOUN
fcis-9422	4	13	due	due	ADP
fcis-9422	4	14	to	to	ADP
fcis-9422	4	15	the	the	DET
fcis-9422	4	16	lack	lack	NOUN
fcis-9422	4	17	of	of	ADP
fcis-9422	4	18	rich	rich	ADJ
fcis-9422	4	19	contextual	contextual	ADJ
fcis-9422	4	20	information	information	NOUN
fcis-9422	4	21	,	,	PUNCT
fcis-9422	4	22	informal	informal	ADJ
fcis-9422	4	23	expression	expression	NOUN
fcis-9422	4	24	,	,	PUNCT
fcis-9422	4	25	and	and	CCONJ
fcis-9422	4	26	incomplete	incomplete	ADJ
fcis-9422	4	27	grammar	grammar	NOUN
fcis-9422	4	28	structures	structure	NOUN
fcis-9422	4	29	.	.	PUNCT
fcis-9422	5	1	therefore	therefore	ADV
fcis-9422	5	2	,	,	PUNCT
fcis-9422	5	3	this	this	DET
fcis-9422	5	4	paper	paper	NOUN
fcis-9422	5	5	proposes	propose	VERB
fcis-9422	5	6	a	a	DET
fcis-9422	5	7	short	short	ADJ
fcis-9422	5	8	-	-	PUNCT
fcis-9422	5	9	text	text	NOUN
fcis-9422	5	10	entity	entity	NOUN
fcis-9422	5	11	linking	link	VERB
fcis-9422	5	12	model	model	NOUN
fcis-9422	5	13	based	base	VERB
fcis-9422	5	14	on	on	ADP
fcis-9422	5	15	the	the	DET
fcis-9422	5	16	roformer	roformer	NOUN
fcis-9422	5	17	-	-	PUNCT
fcis-9422	5	18	sim	sim	ADJ
fcis-9422	5	19	pre	pre	ADJ
fcis-9422	5	20	-	-	ADJ
fcis-9422	5	21	training	training	ADJ
fcis-9422	5	22	model	model	NOUN
fcis-9422	5	23	.	.	PUNCT
fcis-9422	6	1	firstly	firstly	ADV
fcis-9422	6	2	,	,	PUNCT
fcis-9422	6	3	entity	entity	NOUN
fcis-9422	6	4	context	context	NOUN
fcis-9422	6	5	features	feature	NOUN
fcis-9422	6	6	are	be	AUX
fcis-9422	6	7	extracted	extract	VERB
fcis-9422	6	8	by	by	ADP
fcis-9422	6	9	the	the	DET
fcis-9422	6	10	roformer	roformer	NOUN
fcis-9422	6	11	-	-	PUNCT
fcis-9422	6	12	sim	sim	ADJ
fcis-9422	6	13	pre	pre	ADJ
fcis-9422	6	14	-	-	ADJ
fcis-9422	6	15	training	training	ADJ
fcis-9422	6	16	model	model	NOUN
fcis-9422	6	17	,	,	PUNCT
fcis-9422	6	18	and	and	CCONJ
fcis-9422	6	19	then	then	ADV
fcis-9422	6	20	text	text	NOUN
fcis-9422	6	21	similarity	similarity	NOUN
fcis-9422	6	22	calculation	calculation	NOUN
fcis-9422	6	23	and	and	CCONJ
fcis-9422	6	24	sorting	sorting	NOUN
fcis-9422	6	25	are	be	AUX
fcis-9422	6	26	performed	perform	VERB
fcis-9422	6	27	with	with	ADP
fcis-9422	6	28	candidate	candidate	NOUN
fcis-9422	6	29	entity	entity	NOUN
fcis-9422	6	30	description	description	NOUN
fcis-9422	6	31	texts	text	NOUN
fcis-9422	6	32	to	to	PART
fcis-9422	6	33	obtain	obtain	VERB
fcis-9422	6	34	the	the	DET
fcis-9422	6	35	corresponding	corresponding	ADJ
fcis-9422	6	36	entity	entity	NOUN
fcis-9422	6	37	in	in	ADP
fcis-9422	6	38	the	the	DET
fcis-9422	6	39	knowledge	knowledge	NOUN
fcis-9422	6	40	base	base	NOUN
fcis-9422	6	41	with	with	ADP
fcis-9422	6	42	the	the	DET
fcis-9422	6	43	disambiguated	disambiguate	VERB
fcis-9422	6	44	entity	entity	NOUN
fcis-9422	6	45	.	.	PUNCT
fcis-9422	7	1	the	the	DET
fcis-9422	7	2	experimental	experimental	ADJ
fcis-9422	7	3	results	result	NOUN
fcis-9422	7	4	show	show	VERB
fcis-9422	7	5	that	that	SCONJ
fcis-9422	7	6	the	the	DET
fcis-9422	7	7	roformer	roformer	NOUN
fcis-9422	7	8	-	-	PUNCT
fcis-9422	7	9	sim	sim	NOUN
fcis-9422	7	10	model	model	NOUN
fcis-9422	7	11	can	can	AUX
fcis-9422	7	12	provide	provide	VERB
fcis-9422	7	13	prior	prior	ADJ
fcis-9422	7	14	knowledge	knowledge	NOUN
fcis-9422	7	15	for	for	ADP
fcis-9422	7	16	entity	entity	NOUN
fcis-9422	7	17	linking	linking	NOUN
fcis-9422	7	18	,	,	PUNCT
fcis-9422	7	19	and	and	CCONJ
fcis-9422	7	20	the	the	DET
fcis-9422	7	21	proposed	propose	VERB
fcis-9422	7	22	model	model	NOUN
fcis-9422	7	23	in	in	ADP
fcis-9422	7	24	this	this	DET
fcis-9422	7	25	paper	paper	NOUN
fcis-9422	7	26	has	have	VERB
fcis-9422	7	27	an	an	DET
fcis-9422	7	28	f1	f1	ADJ
fcis-9422	7	29	value	value	NOUN
fcis-9422	7	30	of	of	ADP
fcis-9422	7	31	0.8851	0.8851	NUM
fcis-9422	7	32	,	,	PUNCT
fcis-9422	7	33	which	which	PRON
fcis-9422	7	34	is	be	AUX
fcis-9422	7	35	better	well	ADJ
fcis-9422	7	36	than	than	ADP
fcis-9422	7	37	other	other	ADJ
fcis-9422	7	38	entity	entity	NOUN
fcis-9422	7	39	linking	link	VERB
fcis-9422	7	40	models	model	NOUN
fcis-9422	7	41	based	base	VERB
fcis-9422	7	42	on	on	ADP
fcis-9422	7	43	other	other	ADJ
fcis-9422	7	44	pre	pre	ADJ
fcis-9422	7	45	-	-	ADJ
fcis-9422	7	46	training	training	ADJ
fcis-9422	7	47	models	model	NOUN
fcis-9422	7	48	.	.	PUNCT
fcis-9422	8	1	keywords	keyword	NOUN
fcis-9422	8	2	:	:	PUNCT
fcis-9422	8	3	entity	entity	NOUN
fcis-9422	8	4	linking	linking	NOUN
fcis-9422	8	5	;	;	PUNCT
fcis-9422	8	6	candidate	candidate	NOUN
fcis-9422	8	7	entity	entity	NOUN
fcis-9422	8	8	;	;	PUNCT
fcis-9422	8	9	text	text	NOUN
fcis-9422	8	10	similarity	similarity	NOUN
fcis-9422	8	11	.	.	PUNCT
fcis-9422	9	1	1	1	X
fcis-9422	9	2	.	.	X
fcis-9422	9	3	introduction	introduction	NOUN
fcis-9422	9	4	entity	entity	NOUN
fcis-9422	9	5	linking	link	VERB
fcis-9422	9	6	(	(	PUNCT
fcis-9422	9	7	el	el	NOUN
fcis-9422	9	8	)	)	PUNCT
fcis-9422	9	9	is	be	AUX
fcis-9422	9	10	an	an	DET
fcis-9422	9	11	important	important	ADJ
fcis-9422	9	12	task	task	NOUN
fcis-9422	9	13	in	in	ADP
fcis-9422	9	14	the	the	DET
fcis-9422	9	15	field	field	NOUN
fcis-9422	9	16	of	of	ADP
fcis-9422	9	17	natural	natural	ADJ
fcis-9422	9	18	language	language	NOUN
fcis-9422	9	19	processing	processing	NOUN
fcis-9422	9	20	.	.	PUNCT
fcis-9422	10	1	the	the	DET
fcis-9422	10	2	task	task	NOUN
fcis-9422	10	3	typically	typically	ADV
fcis-9422	10	4	involves	involve	VERB
fcis-9422	10	5	identifying	identify	VERB
fcis-9422	10	6	entity	entity	NOUN
fcis-9422	10	7	mentions	mention	NOUN
fcis-9422	10	8	in	in	ADP
fcis-9422	10	9	a	a	DET
fcis-9422	10	10	piece	piece	NOUN
fcis-9422	10	11	of	of	ADP
fcis-9422	10	12	text	text	NOUN
fcis-9422	10	13	and	and	CCONJ
fcis-9422	10	14	linking	link	VERB
fcis-9422	10	15	them	they	PRON
fcis-9422	10	16	to	to	ADP
fcis-9422	10	17	corresponding	correspond	VERB
fcis-9422	10	18	target	target	NOUN
fcis-9422	10	19	entities	entity	NOUN
fcis-9422	10	20	in	in	ADP
fcis-9422	10	21	a	a	DET
fcis-9422	10	22	knowledge	knowledge	NOUN
fcis-9422	10	23	base	base	NOUN
fcis-9422	10	24	.	.	PUNCT
fcis-9422	11	1	entity	entity	NOUN
fcis-9422	11	2	linking	link	VERB
fcis-9422	11	3	can	can	AUX
fcis-9422	11	4	be	be	AUX
fcis-9422	11	5	seen	see	VERB
fcis-9422	11	6	as	as	ADP
fcis-9422	11	7	consisting	consist	VERB
fcis-9422	11	8	of	of	ADP
fcis-9422	11	9	two	two	NUM
fcis-9422	11	10	subtasks	subtask	NOUN
fcis-9422	11	11	:	:	PUNCT
fcis-9422	11	12	candidate	candidate	NOUN
fcis-9422	11	13	entity	entity	NOUN
fcis-9422	11	14	retrieval	retrieval	NOUN
fcis-9422	11	15	and	and	CCONJ
fcis-9422	11	16	candidate	candidate	NOUN
fcis-9422	11	17	entity	entity	NOUN
fcis-9422	11	18	ranking	ranking	NOUN
fcis-9422	11	19	.	.	PUNCT
fcis-9422	12	1	candidate	candidate	NOUN
fcis-9422	12	2	entity	entity	NOUN
fcis-9422	12	3	retrieval	retrieval	NOUN
fcis-9422	12	4	involves	involve	VERB
fcis-9422	12	5	querying	query	VERB
fcis-9422	12	6	the	the	DET
fcis-9422	12	7	knowledge	knowledge	NOUN
fcis-9422	12	8	base	base	NOUN
fcis-9422	12	9	for	for	ADP
fcis-9422	12	10	entities	entity	NOUN
fcis-9422	12	11	that	that	PRON
fcis-9422	12	12	match	match	VERB
fcis-9422	12	13	the	the	DET
fcis-9422	12	14	entity	entity	NOUN
fcis-9422	12	15	mentions	mention	VERB
fcis-9422	12	16	to	to	PART
fcis-9422	12	17	be	be	AUX
fcis-9422	12	18	linked	link	VERB
fcis-9422	12	19	,	,	PUNCT
fcis-9422	12	20	while	while	SCONJ
fcis-9422	12	21	candidate	candidate	NOUN
fcis-9422	12	22	entity	entity	NOUN
fcis-9422	12	23	ranking	ranking	NOUN
fcis-9422	12	24	is	be	AUX
fcis-9422	12	25	a	a	DET
fcis-9422	12	26	key	key	ADJ
fcis-9422	12	27	focus	focus	NOUN
fcis-9422	12	28	of	of	ADP
fcis-9422	12	29	the	the	DET
fcis-9422	12	30	entity	entity	NOUN
fcis-9422	12	31	linking	link	VERB
fcis-9422	12	32	process	process	NOUN
fcis-9422	12	33	.	.	PUNCT
fcis-9422	13	1	this	this	PRON
fcis-9422	13	2	involves	involve	VERB
fcis-9422	13	3	calculating	calculate	VERB
fcis-9422	13	4	and	and	CCONJ
fcis-9422	13	5	ranking	rank	VERB
fcis-9422	13	6	the	the	DET
fcis-9422	13	7	relevance	relevance	NOUN
fcis-9422	13	8	of	of	ADP
fcis-9422	13	9	all	all	DET
fcis-9422	13	10	candidate	candidate	NOUN
fcis-9422	13	11	entities	entity	NOUN
fcis-9422	13	12	in	in	ADP
fcis-9422	13	13	the	the	DET
fcis-9422	13	14	candidate	candidate	NOUN
fcis-9422	13	15	entity	entity	NOUN
fcis-9422	13	16	set	set	VERB
fcis-9422	13	17	to	to	ADP
fcis-9422	13	18	the	the	DET
fcis-9422	13	19	entity	entity	NOUN
fcis-9422	13	20	mention	mention	NOUN
fcis-9422	13	21	,	,	PUNCT
fcis-9422	13	22	and	and	CCONJ
fcis-9422	13	23	selecting	select	VERB
fcis-9422	13	24	the	the	DET
fcis-9422	13	25	candidate	candidate	NOUN
fcis-9422	13	26	entity	entity	NOUN
fcis-9422	13	27	with	with	ADP
fcis-9422	13	28	the	the	DET
fcis-9422	13	29	highest	high	ADJ
fcis-9422	13	30	relevance	relevance	NOUN
fcis-9422	13	31	score	score	NOUN
fcis-9422	13	32	as	as	SCONJ
fcis-9422	13	33	the	the	DET
fcis-9422	13	34	entity	entity	NOUN
fcis-9422	13	35	linked	link	VERB
fcis-9422	13	36	to	to	ADP
fcis-9422	13	37	the	the	DET
fcis-9422	13	38	entity	entity	NOUN
fcis-9422	13	39	mention	mention	NOUN
fcis-9422	13	40	[	[	X
fcis-9422	13	41	1	1	NUM
fcis-9422	13	42	]	]	PUNCT
fcis-9422	13	43	.	.	PUNCT
fcis-9422	14	1	currently	currently	ADV
fcis-9422	14	2	,	,	PUNCT
fcis-9422	14	3	in	in	ADP
fcis-9422	14	4	entity	entity	NOUN
fcis-9422	14	5	linking	link	VERB
fcis-9422	14	6	tasks	task	NOUN
fcis-9422	14	7	,	,	PUNCT
fcis-9422	14	8	linking	link	VERB
fcis-9422	14	9	entities	entity	NOUN
fcis-9422	14	10	in	in	ADP
fcis-9422	14	11	short	short	ADJ
fcis-9422	14	12	texts	text	NOUN
fcis-9422	14	13	is	be	AUX
fcis-9422	14	14	more	more	ADV
fcis-9422	14	15	challenging	challenging	ADJ
fcis-9422	14	16	than	than	ADP
fcis-9422	14	17	in	in	ADP
fcis-9422	14	18	long	long	ADJ
fcis-9422	14	19	texts	text	NOUN
fcis-9422	14	20	because	because	SCONJ
fcis-9422	14	21	long	long	ADJ
fcis-9422	14	22	texts	text	NOUN
fcis-9422	14	23	provide	provide	VERB
fcis-9422	14	24	rich	rich	ADJ
fcis-9422	14	25	semantic	semantic	ADJ
fcis-9422	14	26	information	information	NOUN
fcis-9422	14	27	that	that	PRON
fcis-9422	14	28	can	can	AUX
fcis-9422	14	29	provide	provide	VERB
fcis-9422	14	30	effective	effective	ADJ
fcis-9422	14	31	entity	entity	NOUN
fcis-9422	14	32	context	context	NOUN
fcis-9422	14	33	information	information	NOUN
fcis-9422	14	34	to	to	PART
fcis-9422	14	35	improve	improve	VERB
fcis-9422	14	36	entity	entity	NOUN
fcis-9422	14	37	linking	link	VERB
fcis-9422	14	38	accuracy	accuracy	NOUN
fcis-9422	14	39	[	[	X
fcis-9422	14	40	2	2	NUM
fcis-9422	14	41	]	]	PUNCT
fcis-9422	14	42	.	.	PUNCT
fcis-9422	15	1	for	for	ADP
fcis-9422	15	2	short	short	ADJ
fcis-9422	15	3	texts	text	NOUN
fcis-9422	15	4	,	,	PUNCT
fcis-9422	15	5	the	the	DET
fcis-9422	15	6	lack	lack	NOUN
fcis-9422	15	7	of	of	ADP
fcis-9422	15	8	rich	rich	ADJ
fcis-9422	15	9	contextual	contextual	ADJ
fcis-9422	15	10	information	information	NOUN
fcis-9422	15	11	makes	make	VERB
fcis-9422	15	12	entity	entity	NOUN
fcis-9422	15	13	disambiguation	disambiguation	NOUN
fcis-9422	15	14	difficult	difficult	ADJ
fcis-9422	15	15	.	.	PUNCT
fcis-9422	16	1	especially	especially	ADV
fcis-9422	16	2	in	in	ADP
fcis-9422	16	3	chinese	chinese	ADJ
fcis-9422	16	4	texts	text	NOUN
fcis-9422	16	5	,	,	PUNCT
fcis-9422	16	6	due	due	ADP
fcis-9422	16	7	to	to	ADP
fcis-9422	16	8	the	the	DET
fcis-9422	16	9	variability	variability	NOUN
fcis-9422	16	10	of	of	ADP
fcis-9422	16	11	grammar	grammar	NOUN
fcis-9422	16	12	and	and	CCONJ
fcis-9422	16	13	richness	richness	NOUN
fcis-9422	16	14	of	of	ADP
fcis-9422	16	15	word	word	NOUN
fcis-9422	16	16	meanings	meaning	NOUN
fcis-9422	16	17	,	,	PUNCT
fcis-9422	16	18	it	it	PRON
fcis-9422	16	19	is	be	AUX
fcis-9422	16	20	necessary	necessary	ADJ
fcis-9422	16	21	to	to	PART
fcis-9422	16	22	have	have	VERB
fcis-9422	16	23	a	a	DET
fcis-9422	16	24	precise	precise	ADJ
fcis-9422	16	25	understanding	understanding	NOUN
fcis-9422	16	26	of	of	ADP
fcis-9422	16	27	the	the	DET
fcis-9422	16	28	entity	entity	NOUN
fcis-9422	16	29	context	context	NOUN
fcis-9422	16	30	.	.	PUNCT
fcis-9422	17	1	currently	currently	ADV
fcis-9422	17	2	,	,	PUNCT
fcis-9422	17	3	there	there	PRON
fcis-9422	17	4	are	be	VERB
fcis-9422	17	5	various	various	ADJ
fcis-9422	17	6	methods	method	NOUN
fcis-9422	17	7	to	to	PART
fcis-9422	17	8	achieve	achieve	VERB
fcis-9422	17	9	entity	entity	NOUN
fcis-9422	17	10	linking	linking	NOUN
fcis-9422	17	11	,	,	PUNCT
fcis-9422	17	12	such	such	ADJ
fcis-9422	17	13	as	as	ADP
fcis-9422	17	14	rule	rule	NOUN
fcis-9422	17	15	-	-	PUNCT
fcis-9422	17	16	based	base	VERB
fcis-9422	17	17	methods	method	NOUN
fcis-9422	17	18	,	,	PUNCT
fcis-9422	17	19	machine	machine	NOUN
fcis-9422	17	20	learning	learning	NOUN
fcis-9422	17	21	-	-	PUNCT
fcis-9422	17	22	based	base	VERB
fcis-9422	17	23	methods	method	NOUN
fcis-9422	17	24	,	,	PUNCT
fcis-9422	17	25	and	and	CCONJ
fcis-9422	17	26	deep	deep	ADJ
fcis-9422	17	27	learning	learning	NOUN
fcis-9422	17	28	methods	method	NOUN
fcis-9422	17	29	.	.	PUNCT
fcis-9422	18	1	rulebased	rulebase	VERB
fcis-9422	18	2	methods	method	NOUN
fcis-9422	18	3	typically	typically	ADV
fcis-9422	18	4	rely	rely	VERB
fcis-9422	18	5	on	on	ADP
fcis-9422	18	6	manually	manually	ADV
fcis-9422	18	7	defined	define	VERB
fcis-9422	18	8	rules	rule	NOUN
fcis-9422	18	9	and	and	CCONJ
fcis-9422	18	10	rule	rule	NOUN
fcis-9422	18	11	libraries	library	NOUN
fcis-9422	18	12	to	to	PART
fcis-9422	18	13	match	match	VERB
fcis-9422	18	14	and	and	CCONJ
fcis-9422	18	15	link	link	VERB
fcis-9422	18	16	entities	entity	NOUN
fcis-9422	18	17	.	.	PUNCT
fcis-9422	19	1	machine	machine	NOUN
fcis-9422	19	2	learningbased	learningbase	VERB
fcis-9422	19	3	methods	method	NOUN
fcis-9422	19	4	learn	learn	VERB
fcis-9422	19	5	entity	entity	NOUN
fcis-9422	19	6	linking	link	VERB
fcis-9422	19	7	models	model	NOUN
fcis-9422	19	8	from	from	ADP
fcis-9422	19	9	large	large	ADJ
fcis-9422	19	10	-	-	PUNCT
fcis-9422	19	11	scale	scale	NOUN
fcis-9422	19	12	annotated	annotate	VERB
fcis-9422	19	13	data	datum	NOUN
fcis-9422	19	14	by	by	ADP
fcis-9422	19	15	constructing	construct	VERB
fcis-9422	19	16	features	feature	NOUN
fcis-9422	19	17	,	,	PUNCT
fcis-9422	19	18	selecting	select	VERB
fcis-9422	19	19	models	model	NOUN
fcis-9422	19	20	,	,	PUNCT
fcis-9422	19	21	and	and	CCONJ
fcis-9422	19	22	training	train	VERB
fcis-9422	19	23	them	they	PRON
fcis-9422	19	24	.	.	PUNCT
fcis-9422	20	1	deep	deep	ADJ
fcis-9422	20	2	learning	learning	NOUN
fcis-9422	20	3	methods	method	NOUN
fcis-9422	20	4	learn	learn	VERB
fcis-9422	20	5	features	feature	NOUN
fcis-9422	20	6	and	and	CCONJ
fcis-9422	20	7	train	train	NOUN
fcis-9422	20	8	models	model	NOUN
fcis-9422	20	9	through	through	ADP
fcis-9422	20	10	neural	neural	ADJ
fcis-9422	20	11	network	network	NOUN
fcis-9422	20	12	structures	structure	NOUN
fcis-9422	20	13	to	to	PART
fcis-9422	20	14	link	link	VERB
fcis-9422	20	15	entities	entity	NOUN
fcis-9422	20	16	.	.	PUNCT
fcis-9422	21	1	traditional	traditional	ADJ
fcis-9422	21	2	rule	rule	NOUN
fcis-9422	21	3	-	-	PUNCT
fcis-9422	21	4	based	base	VERB
fcis-9422	21	5	and	and	CCONJ
fcis-9422	21	6	machine	machine	NOUN
fcis-9422	21	7	learning	learning	NOUN
fcis-9422	21	8	-	-	PUNCT
fcis-9422	21	9	based	base	VERB
fcis-9422	21	10	methods	method	NOUN
fcis-9422	21	11	may	may	AUX
fcis-9422	21	12	not	not	PART
fcis-9422	21	13	accurately	accurately	ADV
fcis-9422	21	14	capture	capture	VERB
fcis-9422	21	15	the	the	DET
fcis-9422	21	16	semantic	semantic	ADJ
fcis-9422	21	17	information	information	NOUN
fcis-9422	21	18	of	of	ADP
fcis-9422	21	19	text	text	NOUN
fcis-9422	21	20	and	and	CCONJ
fcis-9422	21	21	entities	entity	NOUN
fcis-9422	21	22	,	,	PUNCT
fcis-9422	21	23	and	and	CCONJ
fcis-9422	21	24	their	their	PRON
fcis-9422	21	25	generalization	generalization	NOUN
fcis-9422	21	26	ability	ability	NOUN
fcis-9422	21	27	needs	need	VERB
fcis-9422	21	28	to	to	PART
fcis-9422	21	29	be	be	AUX
fcis-9422	21	30	improved	improve	VERB
fcis-9422	21	31	because	because	SCONJ
fcis-9422	21	32	they	they	PRON
fcis-9422	21	33	require	require	VERB
fcis-9422	21	34	manually	manually	ADV
fcis-9422	21	35	designed	design	VERB
fcis-9422	21	36	features	feature	NOUN
fcis-9422	21	37	or	or	CCONJ
fcis-9422	21	38	shallow	shallow	ADJ
fcis-9422	21	39	models	model	NOUN
fcis-9422	21	40	for	for	ADP
fcis-9422	21	41	feature	feature	NOUN
fcis-9422	21	42	extraction	extraction	NOUN
fcis-9422	21	43	,	,	PUNCT
fcis-9422	21	44	as	as	ADV
fcis-9422	21	45	well	well	ADV
fcis-9422	21	46	as	as	ADP
fcis-9422	21	47	handdesigned	handdesigned	ADJ
fcis-9422	21	48	rules	rule	NOUN
fcis-9422	21	49	or	or	CCONJ
fcis-9422	21	50	the	the	DET
fcis-9422	21	51	selection	selection	NOUN
fcis-9422	21	52	of	of	ADP
fcis-9422	21	53	specific	specific	ADJ
fcis-9422	21	54	algorithms	algorithm	NOUN
fcis-9422	21	55	and	and	CCONJ
fcis-9422	21	56	models	model	NOUN
fcis-9422	21	57	.	.	PUNCT
fcis-9422	22	1	deep	deep	ADJ
fcis-9422	22	2	learning	learning	NOUN
fcis-9422	22	3	methods	method	NOUN
fcis-9422	22	4	automatically	automatically	ADV
fcis-9422	22	5	learn	learn	VERB
fcis-9422	22	6	deep	deep	ADJ
fcis-9422	22	7	feature	feature	NOUN
fcis-9422	22	8	representations	representation	NOUN
fcis-9422	22	9	of	of	ADP
fcis-9422	22	10	text	text	NOUN
fcis-9422	22	11	and	and	CCONJ
fcis-9422	22	12	entities	entity	NOUN
fcis-9422	22	13	,	,	PUNCT
fcis-9422	22	14	making	make	VERB
fcis-9422	22	15	semantic	semantic	ADJ
fcis-9422	22	16	understanding	understanding	NOUN
fcis-9422	22	17	of	of	ADP
fcis-9422	22	18	text	text	NOUN
fcis-9422	22	19	and	and	CCONJ
fcis-9422	22	20	entities	entity	NOUN
fcis-9422	22	21	more	more	ADV
fcis-9422	22	22	accurate	accurate	ADJ
fcis-9422	22	23	and	and	CCONJ
fcis-9422	22	24	comprehensive	comprehensive	ADJ
fcis-9422	22	25	.	.	PUNCT
fcis-9422	23	1	they	they	PRON
fcis-9422	23	2	also	also	ADV
fcis-9422	23	3	have	have	VERB
fcis-9422	23	4	stronger	strong	ADJ
fcis-9422	23	5	generalization	generalization	NOUN
fcis-9422	23	6	abilities	ability	NOUN
fcis-9422	23	7	and	and	CCONJ
fcis-9422	23	8	can	can	AUX
fcis-9422	23	9	adapt	adapt	VERB
fcis-9422	23	10	to	to	ADP
fcis-9422	23	11	more	more	ADV
fcis-9422	23	12	complex	complex	ADJ
fcis-9422	23	13	and	and	CCONJ
fcis-9422	23	14	varied	varied	ADJ
fcis-9422	23	15	entity	entity	NOUN
fcis-9422	23	16	linking	link	VERB
fcis-9422	23	17	tasks	task	NOUN
fcis-9422	23	18	through	through	ADP
fcis-9422	23	19	large	large	ADJ
fcis-9422	23	20	-	-	PUNCT
fcis-9422	23	21	scale	scale	NOUN
fcis-9422	23	22	training	training	NOUN
fcis-9422	23	23	data	datum	NOUN
fcis-9422	23	24	and	and	CCONJ
fcis-9422	23	25	powerful	powerful	ADJ
fcis-9422	23	26	model	model	NOUN
fcis-9422	23	27	expression	expression	NOUN
fcis-9422	23	28	abilities	ability	NOUN
fcis-9422	23	29	[	[	X
fcis-9422	23	30	3	3	NUM
fcis-9422	23	31	-	-	SYM
fcis-9422	23	32	5	5	NUM
fcis-9422	23	33	]	]	PUNCT
fcis-9422	23	34	.	.	PUNCT
fcis-9422	24	1	with	with	ADP
fcis-9422	24	2	the	the	DET
fcis-9422	24	3	continuous	continuous	ADJ
fcis-9422	24	4	development	development	NOUN
fcis-9422	24	5	of	of	ADP
fcis-9422	24	6	deep	deep	ADJ
fcis-9422	24	7	learning	learning	NOUN
fcis-9422	24	8	and	and	CCONJ
fcis-9422	24	9	natural	natural	ADJ
fcis-9422	24	10	language	language	NOUN
fcis-9422	24	11	processing	processing	NOUN
fcis-9422	24	12	technologies	technology	NOUN
fcis-9422	24	13	,	,	PUNCT
fcis-9422	24	14	significant	significant	ADJ
fcis-9422	24	15	progress	progress	NOUN
fcis-9422	24	16	has	have	AUX
fcis-9422	24	17	been	be	AUX
fcis-9422	24	18	made	make	VERB
fcis-9422	24	19	in	in	ADP
fcis-9422	24	20	the	the	DET
fcis-9422	24	21	field	field	NOUN
fcis-9422	24	22	of	of	ADP
fcis-9422	24	23	chinese	chinese	ADJ
fcis-9422	24	24	entity	entity	NOUN
fcis-9422	24	25	linking	link	VERB
fcis-9422	24	26	.	.	PUNCT
fcis-9422	25	1	for	for	ADP
fcis-9422	25	2	example	example	NOUN
fcis-9422	25	3	,	,	PUNCT
fcis-9422	25	4	researchers	researcher	NOUN
fcis-9422	25	5	have	have	AUX
fcis-9422	25	6	proposed	propose	VERB
fcis-9422	25	7	models	model	NOUN
fcis-9422	25	8	based	base	VERB
fcis-9422	25	9	on	on	ADP
fcis-9422	25	10	convolutional	convolutional	ADJ
fcis-9422	25	11	neural	neural	ADJ
fcis-9422	25	12	networks	network	NOUN
fcis-9422	25	13	(	(	PUNCT
fcis-9422	25	14	cnn	cnn	PROPN
fcis-9422	25	15	)	)	PUNCT
fcis-9422	26	1	[	[	X
fcis-9422	26	2	6	6	NUM
fcis-9422	26	3	]	]	PUNCT
fcis-9422	26	4	and	and	CCONJ
fcis-9422	26	5	recurrent	recurrent	ADJ
fcis-9422	26	6	neural	neural	ADJ
fcis-9422	26	7	networks	network	NOUN
fcis-9422	26	8	(	(	PUNCT
fcis-9422	26	9	rnn	rnn	PROPN
fcis-9422	26	10	)	)	PUNCT
fcis-9422	27	1	[	[	X
fcis-9422	27	2	7	7	NUM
fcis-9422	27	3	]	]	PUNCT
fcis-9422	27	4	,	,	PUNCT
fcis-9422	27	5	as	as	ADV
fcis-9422	27	6	well	well	ADV
fcis-9422	27	7	as	as	ADP
fcis-9422	27	8	methods	method	NOUN
fcis-9422	27	9	for	for	ADP
fcis-9422	27	10	entity	entity	NOUN
fcis-9422	27	11	linking	link	VERB
fcis-9422	27	12	using	use	VERB
fcis-9422	27	13	pre	pre	ADJ
fcis-9422	27	14	-	-	ADJ
fcis-9422	27	15	trained	train	VERB
fcis-9422	27	16	language	language	NOUN
fcis-9422	27	17	models	model	NOUN
fcis-9422	27	18	such	such	ADJ
fcis-9422	27	19	as	as	ADP
fcis-9422	27	20	bert	bert	NOUN
fcis-9422	27	21	[	[	X
fcis-9422	27	22	8	8	NUM
fcis-9422	27	23	]	]	PUNCT
fcis-9422	27	24	and	and	CCONJ
fcis-9422	27	25	elmo	elmo	PROPN
fcis-9422	28	1	[	[	X
fcis-9422	28	2	9	9	NUM
fcis-9422	28	3	]	]	PUNCT
fcis-9422	28	4	,	,	PUNCT
fcis-9422	28	5	which	which	PRON
fcis-9422	28	6	have	have	AUX
fcis-9422	28	7	achieved	achieve	VERB
fcis-9422	28	8	high	high	ADJ
fcis-9422	28	9	linking	linking	NOUN
fcis-9422	28	10	accuracy	accuracy	NOUN
fcis-9422	28	11	and	and	CCONJ
fcis-9422	28	12	robustness	robustness	NOUN
fcis-9422	28	13	in	in	ADP
fcis-9422	28	14	entity	entity	NOUN
fcis-9422	28	15	linking	link	VERB
fcis-9422	28	16	tasks	task	NOUN
fcis-9422	28	17	.	.	PUNCT
fcis-9422	29	1	zhou	zhou	PROPN
fcis-9422	29	2	pengcheng	pengcheng	PROPN
fcis-9422	30	1	[	[	X
fcis-9422	30	2	10	10	NUM
fcis-9422	30	3	]	]	PUNCT
fcis-9422	30	4	and	and	CCONJ
fcis-9422	30	5	others	other	NOUN
fcis-9422	30	6	used	use	VERB
fcis-9422	30	7	n	n	CCONJ
fcis-9422	30	8	-	-	PUNCT
fcis-9422	30	9	grams	gram	NOUN
fcis-9422	30	10	combined	combine	VERB
fcis-9422	30	11	with	with	ADP
fcis-9422	30	12	part	part	NOUN
fcis-9422	30	13	-	-	PUNCT
fcis-9422	30	14	of	of	ADP
fcis-9422	30	15	-	-	PUNCT
fcis-9422	30	16	speech	speech	NOUN
fcis-9422	30	17	tagging	tagging	NOUN
fcis-9422	30	18	and	and	CCONJ
fcis-9422	30	19	multiple	multiple	ADJ
fcis-9422	30	20	mention	mention	NOUN
fcis-9422	30	21	-	-	PUNCT
fcis-9422	30	22	entity	entity	NOUN
fcis-9422	30	23	dictionaries	dictionary	NOUN
fcis-9422	30	24	to	to	PART
fcis-9422	30	25	obtain	obtain	VERB
fcis-9422	30	26	candidate	candidate	NOUN
fcis-9422	30	27	entities	entity	NOUN
fcis-9422	30	28	,	,	PUNCT
fcis-9422	30	29	calculated	calculate	VERB
fcis-9422	30	30	the	the	DET
fcis-9422	30	31	relevance	relevance	NOUN
fcis-9422	30	32	of	of	ADP
fcis-9422	30	33	entity	entity	NOUN
fcis-9422	30	34	sequences	sequence	NOUN
fcis-9422	30	35	through	through	ADP
fcis-9422	30	36	multiple	multiple	ADJ
fcis-9422	30	37	knowledge	knowledge	NOUN
fcis-9422	30	38	bases	basis	NOUN
fcis-9422	30	39	(	(	PUNCT
fcis-9422	30	40	wikipedia	wikipedia	PROPN
fcis-9422	30	41	and	and	CCONJ
fcis-9422	30	42	freebase	freebase	NOUN
fcis-9422	30	43	)	)	PUNCT
fcis-9422	30	44	and	and	CCONJ
fcis-9422	30	45	entity	entity	NOUN
fcis-9422	30	46	categories	category	NOUN
fcis-9422	30	47	,	,	PUNCT
fcis-9422	30	48	and	and	CCONJ
fcis-9422	30	49	selected	select	VERB
fcis-9422	30	50	the	the	DET
fcis-9422	30	51	entity	entity	NOUN
fcis-9422	30	52	with	with	ADP
fcis-9422	30	53	the	the	DET
fcis-9422	30	54	highest	high	ADJ
fcis-9422	30	55	relevance	relevance	NOUN
fcis-9422	30	56	as	as	ADP
fcis-9422	30	57	the	the	DET
fcis-9422	30	58	linking	linking	NOUN
fcis-9422	30	59	result	result	NOUN
fcis-9422	30	60	.	.	PUNCT
fcis-9422	31	1	zeng	zeng	PROPN
fcis-9422	31	2	et	et	PROPN
fcis-9422	31	3	al	al	PROPN
fcis-9422	31	4	.	.	PUNCT
fcis-9422	32	1	[	[	X
fcis-9422	32	2	11	11	NUM
fcis-9422	32	3	]	]	PUNCT
fcis-9422	32	4	proposed	propose	VERB
fcis-9422	32	5	a	a	DET
fcis-9422	32	6	short	short	ADJ
fcis-9422	32	7	-	-	PUNCT
fcis-9422	32	8	text	text	NOUN
fcis-9422	32	9	entity	entity	NOUN
fcis-9422	32	10	linking	link	VERB
fcis-9422	32	11	method	method	NOUN
fcis-9422	32	12	based	base	VERB
fcis-9422	32	13	on	on	ADP
fcis-9422	32	14	a	a	DET
fcis-9422	32	15	biattentional	biattentional	ADJ
fcis-9422	32	16	lstm	lstm	ADJ
fcis-9422	32	17	network	network	NOUN
fcis-9422	32	18	,	,	PUNCT
fcis-9422	32	19	but	but	CCONJ
fcis-9422	32	20	the	the	DET
fcis-9422	32	21	lstm	lstm	NOUN
fcis-9422	32	22	model	model	NOUN
fcis-9422	32	23	can	can	AUX
fcis-9422	32	24	not	not	PART
fcis-9422	32	25	capture	capture	VERB
fcis-9422	32	26	backward	backward	ADJ
fcis-9422	32	27	semantic	semantic	ADJ
fcis-9422	32	28	information	information	NOUN
fcis-9422	32	29	and	and	CCONJ
fcis-9422	32	30	has	have	VERB
fcis-9422	32	31	more	more	ADJ
fcis-9422	32	32	model	model	NOUN
fcis-9422	32	33	parameters	parameter	NOUN
fcis-9422	32	34	.	.	PUNCT
fcis-9422	33	1	hu	hu	PROPN
fcis-9422	33	2	et	et	PROPN
fcis-9422	33	3	al	al	PROPN
fcis-9422	33	4	.	.	PUNCT
fcis-9422	34	1	[	[	X
fcis-9422	34	2	12	12	NUM
fcis-9422	34	3	]	]	PUNCT
fcis-9422	34	4	then	then	ADV
fcis-9422	34	5	constructed	construct	VERB
fcis-9422	34	6	a	a	DET
fcis-9422	34	7	symmetric	symmetric	ADJ
fcis-9422	34	8	bilstm	bilstm	NOUN
fcis-9422	34	9	model	model	NOUN
fcis-9422	34	10	with	with	ADP
fcis-9422	34	11	dual	dual	ADJ
fcis-9422	34	12	attention	attention	NOUN
fcis-9422	34	13	mechanisms	mechanism	NOUN
fcis-9422	34	14	,	,	PUNCT
fcis-9422	34	15	which	which	PRON
fcis-9422	34	16	extracts	extract	VERB
fcis-9422	34	17	entity	entity	NOUN
fcis-9422	34	18	semantic	semantic	ADJ
fcis-9422	34	19	features	feature	VERB
fcis-9422	34	20	more	more	ADV
fcis-9422	34	21	comprehensively	comprehensively	ADV
fcis-9422	34	22	by	by	ADP
fcis-9422	34	23	utilizing	utilize	VERB
fcis-9422	34	24	structural	structural	ADJ
fcis-9422	34	25	information	information	NOUN
fcis-9422	34	26	and	and	CCONJ
fcis-9422	34	27	attention	attention	NOUN
fcis-9422	34	28	mechanisms	mechanism	NOUN
fcis-9422	34	29	,	,	PUNCT
fcis-9422	34	30	and	and	CCONJ
fcis-9422	34	31	bilstm	bilstm	NOUN
fcis-9422	34	32	can	can	AUX
fcis-9422	34	33	better	well	ADV
fcis-9422	34	34	capture	capture	VERB
fcis-9422	34	35	bidirectional	bidirectional	ADJ
fcis-9422	34	36	semantic	semantic	ADJ
fcis-9422	34	37	dependencies	dependency	NOUN
fcis-9422	34	38	.	.	PUNCT
fcis-9422	35	1	devlin	devlin	PROPN
fcis-9422	35	2	et	et	PROPN
fcis-9422	35	3	al	al	PROPN
fcis-9422	35	4	.	.	PUNCT
fcis-9422	36	1	[	[	X
fcis-9422	36	2	13	13	NUM
fcis-9422	36	3	]	]	PUNCT
fcis-9422	36	4	proposed	propose	VERB
fcis-9422	36	5	the	the	DET
fcis-9422	36	6	pre	pre	ADJ
fcis-9422	36	7	-	-	ADJ
fcis-9422	36	8	trained	train	VERB
fcis-9422	36	9	model	model	NOUN
fcis-9422	36	10	bert	bert	PROPN
fcis-9422	36	11	,	,	PUNCT
fcis-9422	36	12	which	which	PRON
fcis-9422	36	13	performs	perform	VERB
fcis-9422	36	14	well	well	ADV
fcis-9422	36	15	in	in	ADP
fcis-9422	36	16	various	various	ADJ
fcis-9422	36	17	natural	natural	ADJ
fcis-9422	36	18	language	language	NOUN
fcis-9422	36	19	processing	processing	NOUN
fcis-9422	36	20	tasks	task	NOUN
fcis-9422	36	21	and	and	CCONJ
fcis-9422	36	22	has	have	AUX
fcis-9422	36	23	been	be	AUX
fcis-9422	36	24	used	use	VERB
fcis-9422	36	25	by	by	ADP
fcis-9422	36	26	many	many	ADJ
fcis-9422	36	27	researchers	researcher	NOUN
fcis-9422	36	28	in	in	ADP
fcis-9422	36	29	entity	entity	NOUN
fcis-9422	36	30	linking	link	VERB
fcis-9422	36	31	tasks	task	NOUN
fcis-9422	36	32	.	.	PUNCT
fcis-9422	37	1	chen	chen	PROPN
fcis-9422	37	2	et	et	PROPN
fcis-9422	37	3	al	al	PROPN
fcis-9422	37	4	.	.	PUNCT
fcis-9422	38	1	[	[	X
fcis-9422	38	2	14	14	NUM
fcis-9422	38	3	]	]	PUNCT
fcis-9422	38	4	modeled	model	VERB
fcis-9422	38	5	the	the	DET
fcis-9422	38	6	context	context	NOUN
fcis-9422	38	7	information	information	NOUN
fcis-9422	38	8	of	of	ADP
fcis-9422	38	9	entities	entity	NOUN
fcis-9422	38	10	in	in	ADP
fcis-9422	38	11	knowledge	knowledge	NOUN
fcis-9422	38	12	bases	basis	NOUN
fcis-9422	38	13	using	use	VERB
fcis-9422	38	14	bert	bert	NOUN
fcis-9422	38	15	,	,	PUNCT
fcis-9422	38	16	and	and	CCONJ
fcis-9422	38	17	the	the	DET
fcis-9422	38	18	entity	entity	NOUN
fcis-9422	38	19	embedding	embed	VERB
fcis-9422	38	20	based	base	VERB
fcis-9422	38	21	on	on	ADP
fcis-9422	38	22	bert	bert	PROPN
fcis-9422	38	23	can	can	AUX
fcis-9422	38	24	better	well	ADV
fcis-9422	38	25	capture	capture	VERB
fcis-9422	38	26	entity	entity	NOUN
fcis-9422	38	27	type	type	NOUN
fcis-9422	38	28	information	information	NOUN
fcis-9422	38	29	.	.	PUNCT
fcis-9422	39	1	cheng	cheng	PROPN
fcis-9422	39	2	et	et	PROPN
fcis-9422	39	3	al	al	PROPN
fcis-9422	39	4	.	.	PUNCT
fcis-9422	40	1	[	[	X
fcis-9422	40	2	15	15	NUM
fcis-9422	40	3	]	]	PUNCT
fcis-9422	40	4	proposed	propose	VERB
fcis-9422	40	5	the	the	DET
fcis-9422	40	6	bert	bert	PROPN
fcis-9422	40	7	-	-	PUNCT
fcis-9422	40	8	ene	ene	NOUN
fcis-9422	40	9	model	model	NOUN
fcis-9422	40	10	based	base	VERB
fcis-9422	40	11	on	on	ADP
fcis-9422	40	12	bert	bert	PROPN
fcis-9422	40	13	and	and	CCONJ
fcis-9422	40	14	treated	treat	VERB
fcis-9422	40	15	the	the	DET
fcis-9422	40	16	shorttext	shorttext	ADJ
fcis-9422	40	17	entity	entity	NOUN
fcis-9422	40	18	linking	link	VERB
fcis-9422	40	19	problem	problem	NOUN
fcis-9422	40	20	as	as	ADP
fcis-9422	40	21	a	a	DET
fcis-9422	40	22	binary	binary	ADJ
fcis-9422	40	23	classification	classification	NOUN
fcis-9422	40	24	task	task	NOUN
fcis-9422	40	25	.	.	PUNCT
fcis-9422	41	1	zhao	zhao	PROPN
fcis-9422	41	2	et	et	PROPN
fcis-9422	41	3	al	al	PROPN
fcis-9422	41	4	.	.	PUNCT
fcis-9422	42	1	[	[	X
fcis-9422	42	2	16	16	NUM
fcis-9422	42	3	]	]	PUNCT
fcis-9422	42	4	proposed	propose	VERB
fcis-9422	42	5	a	a	DET
fcis-9422	42	6	method	method	NOUN
fcis-9422	42	7	combining	combine	VERB
fcis-9422	42	8	bert	bert	NOUN
fcis-9422	42	9	with	with	ADP
fcis-9422	42	10	graph	graph	NOUN
fcis-9422	42	11	models	model	NOUN
fcis-9422	42	12	,	,	PUNCT
fcis-9422	42	13	which	which	PRON
fcis-9422	42	14	can	can	AUX
fcis-9422	42	15	better	well	ADV
fcis-9422	42	16	capture	capture	VERB
fcis-9422	42	17	the	the	DET
fcis-9422	42	18	relevance	relevance	NOUN
fcis-9422	42	19	between	between	ADP
fcis-9422	42	20	entity	entity	NOUN
fcis-9422	42	21	mentions	mention	NOUN
fcis-9422	42	22	and	and	CCONJ
fcis-9422	42	23	candidate	candidate	NOUN
fcis-9422	42	24	entities	entity	NOUN
fcis-9422	42	25	.	.	PUNCT
fcis-9422	43	1	47	47	NUM
fcis-9422	43	2	with	with	ADP
fcis-9422	43	3	the	the	DET
fcis-9422	43	4	rapid	rapid	ADJ
fcis-9422	43	5	development	development	NOUN
fcis-9422	43	6	of	of	ADP
fcis-9422	43	7	the	the	DET
fcis-9422	43	8	internet	internet	NOUN
fcis-9422	43	9	and	and	CCONJ
fcis-9422	43	10	digital	digital	ADJ
fcis-9422	43	11	information	information	NOUN
fcis-9422	43	12	,	,	PUNCT
fcis-9422	43	13	a	a	DET
fcis-9422	43	14	large	large	ADJ
fcis-9422	43	15	amount	amount	NOUN
fcis-9422	43	16	of	of	ADP
fcis-9422	43	17	chinese	chinese	ADJ
fcis-9422	43	18	textual	textual	ADJ
fcis-9422	43	19	data	datum	NOUN
fcis-9422	43	20	has	have	AUX
fcis-9422	43	21	emerged	emerge	VERB
fcis-9422	43	22	,	,	PUNCT
fcis-9422	43	23	making	make	VERB
fcis-9422	43	24	it	it	PRON
fcis-9422	43	25	a	a	DET
fcis-9422	43	26	key	key	ADJ
fcis-9422	43	27	problem	problem	NOUN
fcis-9422	43	28	in	in	ADP
fcis-9422	43	29	applications	application	NOUN
fcis-9422	43	30	such	such	ADJ
fcis-9422	43	31	as	as	ADP
fcis-9422	43	32	information	information	NOUN
fcis-9422	43	33	extraction	extraction	NOUN
fcis-9422	43	34	,	,	PUNCT
fcis-9422	43	35	information	information	NOUN
fcis-9422	43	36	retrieval	retrieval	NOUN
fcis-9422	43	37	,	,	PUNCT
fcis-9422	43	38	and	and	CCONJ
fcis-9422	43	39	semantic	semantic	ADJ
fcis-9422	43	40	search	search	NOUN
fcis-9422	43	41	to	to	PART
fcis-9422	43	42	efficiently	efficiently	ADV
fcis-9422	43	43	identify	identify	VERB
fcis-9422	43	44	and	and	CCONJ
fcis-9422	43	45	link	link	VERB
fcis-9422	43	46	entities	entity	NOUN
fcis-9422	43	47	from	from	ADP
fcis-9422	43	48	these	these	DET
fcis-9422	43	49	texts	text	NOUN
fcis-9422	43	50	.	.	PUNCT
fcis-9422	44	1	given	give	VERB
fcis-9422	44	2	the	the	DET
fcis-9422	44	3	current	current	ADJ
fcis-9422	44	4	issues	issue	NOUN
fcis-9422	44	5	of	of	ADP
fcis-9422	44	6	colloquialization	colloquialization	NOUN
fcis-9422	44	7	and	and	CCONJ
fcis-9422	44	8	lack	lack	NOUN
fcis-9422	44	9	of	of	ADP
fcis-9422	44	10	rich	rich	ADJ
fcis-9422	44	11	context	context	NOUN
fcis-9422	44	12	in	in	ADP
fcis-9422	44	13	chinese	chinese	ADJ
fcis-9422	44	14	short	short	ADJ
fcis-9422	44	15	texts	text	NOUN
fcis-9422	44	16	,	,	PUNCT
fcis-9422	44	17	this	this	DET
fcis-9422	44	18	paper	paper	NOUN
fcis-9422	44	19	proposes	propose	VERB
fcis-9422	44	20	a	a	DET
fcis-9422	44	21	chinese	chinese	ADJ
fcis-9422	44	22	short	short	ADJ
fcis-9422	44	23	text	text	NOUN
fcis-9422	44	24	entity	entity	NOUN
fcis-9422	44	25	linking	link	VERB
fcis-9422	44	26	method	method	NOUN
fcis-9422	44	27	based	base	VERB
fcis-9422	44	28	on	on	ADP
fcis-9422	44	29	the	the	DET
fcis-9422	44	30	roformer	roformer	NOUN
fcis-9422	44	31	-	-	PUNCT
fcis-9422	44	32	sim	sim	ADJ
fcis-9422	44	33	pre	pre	ADJ
fcis-9422	44	34	-	-	ADJ
fcis-9422	44	35	training	training	ADJ
fcis-9422	44	36	model	model	NOUN
fcis-9422	44	37	,	,	PUNCT
fcis-9422	44	38	which	which	PRON
fcis-9422	44	39	improves	improve	VERB
fcis-9422	44	40	the	the	DET
fcis-9422	44	41	accuracy	accuracy	NOUN
fcis-9422	44	42	of	of	ADP
fcis-9422	44	43	entity	entity	NOUN
fcis-9422	44	44	linking	linking	NOUN
fcis-9422	44	45	.	.	PUNCT
fcis-9422	45	1	the	the	DET
fcis-9422	45	2	main	main	ADJ
fcis-9422	45	3	contributions	contribution	NOUN
fcis-9422	45	4	of	of	ADP
fcis-9422	45	5	this	this	DET
fcis-9422	45	6	paper	paper	NOUN
fcis-9422	45	7	are	be	AUX
fcis-9422	45	8	as	as	SCONJ
fcis-9422	45	9	follows	follow	VERB
fcis-9422	45	10	:	:	PUNCT
fcis-9422	45	11	(	(	PUNCT
fcis-9422	45	12	1	1	X
fcis-9422	45	13	)	)	PUNCT
fcis-9422	45	14	proposing	propose	VERB
fcis-9422	45	15	a	a	DET
fcis-9422	45	16	chinese	chinese	ADJ
fcis-9422	45	17	short	short	ADJ
fcis-9422	45	18	-	-	PUNCT
fcis-9422	45	19	text	text	NOUN
fcis-9422	45	20	entity	entity	NOUN
fcis-9422	45	21	linking	link	VERB
fcis-9422	45	22	method	method	NOUN
fcis-9422	45	23	based	base	VERB
fcis-9422	45	24	on	on	ADP
fcis-9422	45	25	the	the	DET
fcis-9422	45	26	roformer	roformer	NOUN
fcis-9422	45	27	-	-	PUNCT
fcis-9422	45	28	sim	sim	ADJ
fcis-9422	45	29	pre	pre	ADJ
fcis-9422	45	30	-	-	ADJ
fcis-9422	45	31	training	training	ADJ
fcis-9422	45	32	model	model	NOUN
fcis-9422	45	33	.	.	PUNCT
fcis-9422	46	1	the	the	DET
fcis-9422	46	2	roformer	roformer	NOUN
fcis-9422	46	3	-	-	PUNCT
fcis-9422	46	4	sim	sim	ADJ
fcis-9422	46	5	pre	pre	ADJ
fcis-9422	46	6	-	-	ADJ
fcis-9422	46	7	training	training	ADJ
fcis-9422	46	8	model	model	NOUN
fcis-9422	46	9	is	be	AUX
fcis-9422	46	10	used	use	VERB
fcis-9422	46	11	to	to	PART
fcis-9422	46	12	extract	extract	VERB
fcis-9422	46	13	contextual	contextual	ADJ
fcis-9422	46	14	features	feature	NOUN
fcis-9422	46	15	,	,	PUNCT
fcis-9422	46	16	which	which	PRON
fcis-9422	46	17	are	be	AUX
fcis-9422	46	18	then	then	ADV
fcis-9422	46	19	used	use	VERB
fcis-9422	46	20	to	to	PART
fcis-9422	46	21	calculate	calculate	VERB
fcis-9422	46	22	text	text	NOUN
fcis-9422	46	23	similarity	similarity	NOUN
fcis-9422	46	24	with	with	ADP
fcis-9422	46	25	candidate	candidate	NOUN
fcis-9422	46	26	entity	entity	NOUN
fcis-9422	46	27	description	description	NOUN
fcis-9422	46	28	text	text	NOUN
fcis-9422	46	29	.	.	PUNCT
fcis-9422	47	1	the	the	DET
fcis-9422	47	2	similarity	similarity	NOUN
fcis-9422	47	3	scores	score	NOUN
fcis-9422	47	4	are	be	AUX
fcis-9422	47	5	then	then	ADV
fcis-9422	47	6	sorted	sort	VERB
fcis-9422	47	7	,	,	PUNCT
fcis-9422	47	8	and	and	CCONJ
fcis-9422	47	9	the	the	DET
fcis-9422	47	10	entity	entity	NOUN
fcis-9422	47	11	with	with	ADP
fcis-9422	47	12	the	the	DET
fcis-9422	47	13	highest	high	ADJ
fcis-9422	47	14	score	score	NOUN
fcis-9422	47	15	is	be	AUX
fcis-9422	47	16	selected	select	VERB
fcis-9422	47	17	as	as	ADP
fcis-9422	47	18	the	the	DET
fcis-9422	47	19	correct	correct	ADJ
fcis-9422	47	20	linked	link	VERB
fcis-9422	47	21	entity	entity	NOUN
fcis-9422	47	22	.	.	PUNCT
fcis-9422	48	1	(	(	PUNCT
fcis-9422	48	2	2	2	NUM
fcis-9422	48	3	)	)	PUNCT
fcis-9422	48	4	by	by	ADP
fcis-9422	48	5	using	use	VERB
fcis-9422	48	6	the	the	DET
fcis-9422	48	7	roformer	roformer	NOUN
fcis-9422	48	8	-	-	PUNCT
fcis-9422	48	9	sim	sim	ADJ
fcis-9422	48	10	pre	pre	ADJ
fcis-9422	48	11	-	-	ADJ
fcis-9422	48	12	training	training	ADJ
fcis-9422	48	13	model	model	NOUN
fcis-9422	48	14	,	,	PUNCT
fcis-9422	48	15	prior	prior	ADJ
fcis-9422	48	16	knowledge	knowledge	NOUN
fcis-9422	48	17	is	be	AUX
fcis-9422	48	18	provided	provide	VERB
fcis-9422	48	19	for	for	ADP
fcis-9422	48	20	the	the	DET
fcis-9422	48	21	entity	entity	NOUN
fcis-9422	48	22	linking	link	VERB
fcis-9422	48	23	task	task	NOUN
fcis-9422	48	24	,	,	PUNCT
fcis-9422	48	25	achieving	achieve	VERB
fcis-9422	48	26	a	a	DET
fcis-9422	48	27	significant	significant	ADJ
fcis-9422	48	28	improvement	improvement	NOUN
fcis-9422	48	29	with	with	ADP
fcis-9422	48	30	an	an	DET
fcis-9422	48	31	f1	f1	ADJ
fcis-9422	48	32	score	score	NOUN
fcis-9422	48	33	of	of	ADP
fcis-9422	48	34	0.8851	0.8851	NUM
fcis-9422	48	35	on	on	ADP
fcis-9422	48	36	the	the	DET
fcis-9422	48	37	ccks2020	ccks2020	PROPN
fcis-9422	48	38	dataset	dataset	NOUN
fcis-9422	48	39	.	.	PUNCT
fcis-9422	49	1	2	2	X
fcis-9422	49	2	.	.	X
fcis-9422	49	3	methodology	methodology	NOUN
fcis-9422	49	4	2.1	2.1	NUM
fcis-9422	49	5	.	.	PUNCT
fcis-9422	50	1	short	short	ADJ
fcis-9422	50	2	text	text	NOUN
fcis-9422	50	3	entity	entity	NOUN
fcis-9422	50	4	linking	link	VERB
fcis-9422	50	5	process	process	NOUN
fcis-9422	50	6	entity	entity	NOUN
fcis-9422	50	7	linking	link	VERB
fcis-9422	50	8	aims	aim	VERB
fcis-9422	50	9	to	to	PART
fcis-9422	50	10	map	map	VERB
fcis-9422	50	11	the	the	DET
fcis-9422	50	12	entity	entity	NOUN
fcis-9422	50	13	mentions	mention	NOUN
fcis-9422	50	14	mentioned	mention	VERB
fcis-9422	50	15	in	in	ADP
fcis-9422	50	16	the	the	DET
fcis-9422	50	17	text	text	NOUN
fcis-9422	50	18	to	to	ADP
fcis-9422	50	19	the	the	DET
fcis-9422	50	20	corresponding	corresponding	ADJ
fcis-9422	50	21	entities	entity	NOUN
fcis-9422	50	22	in	in	ADP
fcis-9422	50	23	the	the	DET
fcis-9422	50	24	given	give	VERB
fcis-9422	50	25	knowledge	knowledge	NOUN
fcis-9422	50	26	base	base	NOUN
fcis-9422	50	27	.	.	PUNCT
fcis-9422	51	1	the	the	DET
fcis-9422	51	2	purpose	purpose	NOUN
fcis-9422	51	3	is	be	AUX
fcis-9422	51	4	to	to	PART
fcis-9422	51	5	solve	solve	VERB
fcis-9422	51	6	the	the	DET
fcis-9422	51	7	problem	problem	NOUN
fcis-9422	51	8	of	of	ADP
fcis-9422	51	9	ambiguity	ambiguity	NOUN
fcis-9422	51	10	caused	cause	VERB
fcis-9422	51	11	by	by	ADP
fcis-9422	51	12	entity	entity	NOUN
fcis-9422	51	13	mentions	mention	NOUN
fcis-9422	51	14	in	in	ADP
fcis-9422	51	15	the	the	DET
fcis-9422	51	16	text	text	NOUN
fcis-9422	51	17	corresponding	correspond	VERB
fcis-9422	51	18	to	to	ADP
fcis-9422	51	19	entities	entity	NOUN
fcis-9422	51	20	in	in	ADP
fcis-9422	51	21	different	different	ADJ
fcis-9422	51	22	contexts	contexts	NOUN
fcis-9422	51	23	.	.	PUNCT
fcis-9422	52	1	this	this	DET
fcis-9422	52	2	article	article	NOUN
fcis-9422	52	3	divides	divide	VERB
fcis-9422	52	4	the	the	DET
fcis-9422	52	5	short	short	ADJ
fcis-9422	52	6	text	text	NOUN
fcis-9422	52	7	entity	entity	NOUN
fcis-9422	52	8	linking	link	VERB
fcis-9422	52	9	task	task	NOUN
fcis-9422	52	10	into	into	ADP
fcis-9422	52	11	two	two	NUM
fcis-9422	52	12	stages	stage	NOUN
fcis-9422	52	13	:	:	PUNCT
fcis-9422	52	14	candidate	candidate	NOUN
fcis-9422	52	15	entity	entity	NOUN
fcis-9422	52	16	generation	generation	NOUN
fcis-9422	52	17	and	and	CCONJ
fcis-9422	52	18	candidate	candidate	NOUN
fcis-9422	52	19	entity	entity	NOUN
fcis-9422	52	20	ranking	ranking	NOUN
fcis-9422	52	21	,	,	PUNCT
fcis-9422	52	22	as	as	SCONJ
fcis-9422	52	23	shown	show	VERB
fcis-9422	52	24	in	in	ADP
fcis-9422	52	25	figure	figure	NOUN
fcis-9422	52	26	1	1	NUM
fcis-9422	52	27	.	.	PUNCT
fcis-9422	52	28	figure	figure	NOUN
fcis-9422	52	29	1	1	NUM
fcis-9422	52	30	.	.	PUNCT
fcis-9422	52	31	short	short	ADJ
fcis-9422	52	32	text	text	NOUN
fcis-9422	52	33	entity	entity	NOUN
fcis-9422	52	34	linking	link	VERB
fcis-9422	52	35	workflow	workflow	NOUN
fcis-9422	52	36	the	the	DET
fcis-9422	52	37	essence	essence	NOUN
fcis-9422	52	38	of	of	ADP
fcis-9422	52	39	entity	entity	NOUN
fcis-9422	52	40	linking	link	VERB
fcis-9422	52	41	is	be	AUX
fcis-9422	52	42	essentially	essentially	ADV
fcis-9422	52	43	to	to	PART
fcis-9422	52	44	calculate	calculate	VERB
fcis-9422	52	45	the	the	DET
fcis-9422	52	46	text	text	NOUN
fcis-9422	52	47	similarity	similarity	NOUN
fcis-9422	52	48	between	between	ADP
fcis-9422	52	49	the	the	DET
fcis-9422	52	50	short	short	ADJ
fcis-9422	52	51	text	text	NOUN
fcis-9422	52	52	containing	contain	VERB
fcis-9422	52	53	the	the	DET
fcis-9422	52	54	entity	entity	NOUN
fcis-9422	52	55	mention	mention	NOUN
fcis-9422	52	56	and	and	CCONJ
fcis-9422	52	57	the	the	DET
fcis-9422	52	58	candidate	candidate	NOUN
fcis-9422	52	59	entity	entity	NOUN
fcis-9422	52	60	description	description	NOUN
fcis-9422	52	61	text	text	NOUN
fcis-9422	52	62	,	,	PUNCT
fcis-9422	52	63	and	and	CCONJ
fcis-9422	52	64	to	to	PART
fcis-9422	52	65	sort	sort	VERB
fcis-9422	52	66	the	the	DET
fcis-9422	52	67	similarity	similarity	NOUN
fcis-9422	52	68	scores	score	NOUN
fcis-9422	52	69	of	of	ADP
fcis-9422	52	70	the	the	DET
fcis-9422	52	71	candidate	candidate	NOUN
fcis-9422	52	72	entities	entity	NOUN
fcis-9422	52	73	,	,	PUNCT
fcis-9422	52	74	so	so	SCONJ
fcis-9422	52	75	as	as	SCONJ
fcis-9422	52	76	to	to	PART
fcis-9422	52	77	link	link	VERB
fcis-9422	52	78	the	the	DET
fcis-9422	52	79	entities	entity	NOUN
fcis-9422	52	80	to	to	ADP
fcis-9422	52	81	the	the	DET
fcis-9422	52	82	corresponding	corresponding	ADJ
fcis-9422	52	83	entities	entity	NOUN
fcis-9422	52	84	in	in	ADP
fcis-9422	52	85	the	the	DET
fcis-9422	52	86	knowledge	knowledge	NOUN
fcis-9422	52	87	base	base	NOUN
fcis-9422	52	88	.	.	PUNCT
fcis-9422	53	1	that	that	PRON
fcis-9422	53	2	is	be	AUX
fcis-9422	53	3	,	,	PUNCT
fcis-9422	53	4	the	the	DET
fcis-9422	53	5	short	short	ADJ
fcis-9422	53	6	text	text	NOUN
fcis-9422	53	7	containing	contain	VERB
fcis-9422	53	8	the	the	DET
fcis-9422	53	9	entity	entity	NOUN
fcis-9422	53	10	mention	mention	NOUN
fcis-9422	53	11	and	and	CCONJ
fcis-9422	53	12	the	the	DET
fcis-9422	53	13	entity	entity	NOUN
fcis-9422	53	14	description	description	NOUN
fcis-9422	53	15	text	text	NOUN
fcis-9422	53	16	are	be	AUX
fcis-9422	53	17	input	input	NOUN
fcis-9422	53	18	to	to	ADP
fcis-9422	53	19	the	the	DET
fcis-9422	53	20	model	model	NOUN
fcis-9422	53	21	as	as	ADP
fcis-9422	53	22	a	a	DET
fcis-9422	53	23	text	text	NOUN
fcis-9422	53	24	pair	pair	NOUN
fcis-9422	53	25	,	,	PUNCT
fcis-9422	53	26	then	then	ADV
fcis-9422	53	27	the	the	DET
fcis-9422	53	28	scores	score	NOUN
fcis-9422	53	29	of	of	ADP
fcis-9422	53	30	the	the	DET
fcis-9422	53	31	candidate	candidate	NOUN
fcis-9422	53	32	entities	entity	NOUN
fcis-9422	53	33	are	be	AUX
fcis-9422	53	34	calculated	calculate	VERB
fcis-9422	53	35	and	and	CCONJ
fcis-9422	53	36	sorted	sort	VERB
fcis-9422	53	37	,	,	PUNCT
fcis-9422	53	38	and	and	CCONJ
fcis-9422	53	39	finally	finally	ADV
fcis-9422	53	40	the	the	DET
fcis-9422	53	41	candidate	candidate	NOUN
fcis-9422	53	42	entity	entity	NOUN
fcis-9422	53	43	with	with	ADP
fcis-9422	53	44	the	the	DET
fcis-9422	53	45	highest	high	ADJ
fcis-9422	53	46	score	score	NOUN
fcis-9422	53	47	is	be	AUX
fcis-9422	53	48	selected	select	VERB
fcis-9422	53	49	as	as	ADP
fcis-9422	53	50	the	the	DET
fcis-9422	53	51	linked	link	VERB
fcis-9422	53	52	entity	entity	NOUN
fcis-9422	53	53	.	.	PUNCT
fcis-9422	54	1	2.2	2.2	NUM
fcis-9422	54	2	.	.	PUNCT
fcis-9422	55	1	pre	pre	VERB
fcis-9422	55	2	-	-	ADJ
fcis-9422	55	3	trained	train	VERB
fcis-9422	55	4	language	language	NOUN
fcis-9422	55	5	model	model	NOUN
fcis-9422	55	6	roformersim	roformersim	ADJ
fcis-9422	55	7	roformer	roformer	NOUN
fcis-9422	55	8	-	-	PUNCT
fcis-9422	55	9	sim	sim	NOUN
fcis-9422	55	10	model	model	NOUN
fcis-9422	55	11	combines	combine	VERB
fcis-9422	55	12	retrieval	retrieval	NOUN
fcis-9422	55	13	and	and	CCONJ
fcis-9422	55	14	generation	generation	NOUN
fcis-9422	55	15	and	and	CCONJ
fcis-9422	55	16	was	be	AUX
fcis-9422	55	17	released	release	VERB
fcis-9422	55	18	by	by	ADP
fcis-9422	55	19	the	the	DET
fcis-9422	55	20	team	team	NOUN
fcis-9422	55	21	led	lead	VERB
fcis-9422	55	22	by	by	ADP
fcis-9422	55	23	jianlin	jianlin	PROPN
fcis-9422	55	24	su	su	PROPN
fcis-9422	55	25	in	in	ADP
fcis-9422	55	26	june	june	PROPN
fcis-9422	55	27	2021	2021	NUM
fcis-9422	55	28	.	.	PUNCT
fcis-9422	56	1	before	before	ADP
fcis-9422	56	2	roformer	roformer	NOUN
fcis-9422	56	3	-	-	PUNCT
fcis-9422	56	4	sim	sim	NOUN
fcis-9422	56	5	,	,	PUNCT
fcis-9422	56	6	the	the	DET
fcis-9422	56	7	authors	author	NOUN
fcis-9422	56	8	also	also	ADV
fcis-9422	56	9	proposed	propose	VERB
fcis-9422	56	10	a	a	DET
fcis-9422	56	11	similar	similar	ADJ
fcis-9422	56	12	model	model	NOUN
fcis-9422	56	13	called	call	VERB
fcis-9422	56	14	simbert	simbert	NOUN
fcis-9422	56	15	.	.	PUNCT
fcis-9422	57	1	roformer	roformer	NOUN
fcis-9422	57	2	-	-	PUNCT
fcis-9422	57	3	sim	sim	NOUN
fcis-9422	57	4	is	be	AUX
fcis-9422	57	5	a	a	DET
fcis-9422	57	6	further	further	ADJ
fcis-9422	57	7	integration	integration	NOUN
fcis-9422	57	8	and	and	CCONJ
fcis-9422	57	9	optimization	optimization	NOUN
fcis-9422	57	10	of	of	ADP
fcis-9422	57	11	simbert	simbert	NOUN
fcis-9422	57	12	-	-	PUNCT
fcis-9422	57	13	related	relate	VERB
fcis-9422	57	14	techniques	technique	NOUN
fcis-9422	57	15	and	and	CCONJ
fcis-9422	57	16	can	can	AUX
fcis-9422	57	17	be	be	AUX
fcis-9422	57	18	used	use	VERB
fcis-9422	57	19	to	to	PART
fcis-9422	57	20	automatically	automatically	ADV
fcis-9422	57	21	generate	generate	VERB
fcis-9422	57	22	similar	similar	ADJ
fcis-9422	57	23	questions	question	NOUN
fcis-9422	57	24	and	and	CCONJ
fcis-9422	57	25	as	as	ADP
fcis-9422	57	26	an	an	DET
fcis-9422	57	27	auxiliary	auxiliary	ADJ
fcis-9422	57	28	data	datum	NOUN
fcis-9422	57	29	augmentation	augmentation	NOUN
fcis-9422	57	30	tool	tool	NOUN
fcis-9422	57	31	.	.	PUNCT
fcis-9422	58	1	the	the	DET
fcis-9422	58	2	roformer	roformer	PROPN
fcis-9422	58	3	model	model	NOUN
fcis-9422	58	4	improves	improve	VERB
fcis-9422	58	5	the	the	DET
fcis-9422	58	6	position	position	NOUN
fcis-9422	58	7	encoding	encoding	NOUN
fcis-9422	58	8	of	of	ADP
fcis-9422	58	9	the	the	DET
fcis-9422	58	10	transformer	transformer	NOUN
fcis-9422	58	11	by	by	ADP
fcis-9422	58	12	using	use	VERB
fcis-9422	58	13	rotary	rotary	ADJ
fcis-9422	58	14	position	position	NOUN
fcis-9422	58	15	embedding	embed	VERB
fcis-9422	58	16	(	(	PUNCT
fcis-9422	58	17	rope	rope	NOUN
fcis-9422	58	18	)	)	PUNCT
fcis-9422	58	19	instead	instead	ADV
fcis-9422	58	20	of	of	ADP
fcis-9422	58	21	absolute	absolute	ADJ
fcis-9422	58	22	position	position	NOUN
fcis-9422	58	23	encoding	encoding	NOUN
fcis-9422	58	24	in	in	ADP
fcis-9422	58	25	the	the	DET
fcis-9422	58	26	transformer	transformer	NOUN
fcis-9422	58	27	.	.	PUNCT
fcis-9422	59	1	this	this	PRON
fcis-9422	59	2	is	be	AUX
fcis-9422	59	3	a	a	DET
fcis-9422	59	4	design	design	NOUN
fcis-9422	59	5	that	that	PRON
fcis-9422	59	6	can	can	AUX
fcis-9422	59	7	achieve	achieve	VERB
fcis-9422	59	8	relative	relative	ADJ
fcis-9422	59	9	position	position	NOUN
fcis-9422	59	10	encoding	encoding	NOUN
fcis-9422	59	11	using	use	VERB
fcis-9422	59	12	the	the	DET
fcis-9422	59	13	attention	attention	NOUN
fcis-9422	59	14	mechanism	mechanism	NOUN
fcis-9422	59	15	in	in	ADP
fcis-9422	59	16	combination	combination	NOUN
fcis-9422	59	17	with	with	ADP
fcis-9422	59	18	the	the	DET
fcis-9422	59	19	"	"	PUNCT
fcis-9422	59	20	absolute	absolute	ADJ
fcis-9422	59	21	position	position	NOUN
fcis-9422	59	22	encoding	encoding	NOUN
fcis-9422	59	23	"	"	PUNCT
fcis-9422	59	24	method	method	NOUN
fcis-9422	59	25	.	.	PUNCT
fcis-9422	60	1	generally	generally	ADV
fcis-9422	60	2	speaking	speak	VERB
fcis-9422	60	3	,	,	PUNCT
fcis-9422	60	4	absolute	absolute	ADJ
fcis-9422	60	5	position	position	NOUN
fcis-9422	60	6	encoding	encoding	NOUN
fcis-9422	60	7	has	have	VERB
fcis-9422	60	8	the	the	DET
fcis-9422	60	9	advantages	advantage	NOUN
fcis-9422	60	10	of	of	ADP
fcis-9422	60	11	simple	simple	ADJ
fcis-9422	60	12	implementation	implementation	NOUN
fcis-9422	60	13	and	and	CCONJ
fcis-9422	60	14	fast	fast	ADJ
fcis-9422	60	15	computation	computation	NOUN
fcis-9422	60	16	speed	speed	NOUN
fcis-9422	60	17	,	,	PUNCT
fcis-9422	60	18	while	while	SCONJ
fcis-9422	60	19	relative	relative	ADJ
fcis-9422	60	20	position	position	NOUN
fcis-9422	60	21	encoding	encode	VERB
fcis-9422	60	22	directly	directly	ADV
fcis-9422	60	23	reflects	reflect	VERB
fcis-9422	60	24	relative	relative	ADJ
fcis-9422	60	25	position	position	NOUN
fcis-9422	60	26	signals	signal	NOUN
fcis-9422	60	27	and	and	CCONJ
fcis-9422	60	28	often	often	ADV
fcis-9422	60	29	has	have	VERB
fcis-9422	60	30	better	well	ADJ
fcis-9422	60	31	performance	performance	NOUN
fcis-9422	60	32	in	in	ADP
fcis-9422	60	33	practice	practice	NOUN
fcis-9422	60	34	.	.	PUNCT
fcis-9422	61	1	the	the	DET
fcis-9422	61	2	main	main	ADJ
fcis-9422	61	3	framework	framework	NOUN
fcis-9422	61	4	of	of	ADP
fcis-9422	61	5	roformer	roformer	NOUN
fcis-9422	61	6	-	-	PUNCT
fcis-9422	61	7	sim	sim	NOUN
fcis-9422	61	8	is	be	AUX
fcis-9422	61	9	the	the	DET
fcis-9422	61	10	roformer	roformer	PROPN
fcis-9422	61	11	model	model	NOUN
fcis-9422	61	12	.	.	PUNCT
fcis-9422	62	1	in	in	ADP
fcis-9422	62	2	addition	addition	NOUN
fcis-9422	62	3	,	,	PUNCT
fcis-9422	62	4	roformer	roformer	NOUN
fcis-9422	62	5	-	-	PUNCT
fcis-9422	62	6	sim	sim	NOUN
fcis-9422	62	7	uses	use	VERB
fcis-9422	62	8	more	more	ADJ
fcis-9422	62	9	training	training	NOUN
fcis-9422	62	10	data	datum	NOUN
fcis-9422	62	11	and	and	CCONJ
fcis-9422	62	12	is	be	AUX
fcis-9422	62	13	not	not	PART
fcis-9422	62	14	limited	limit	VERB
fcis-9422	62	15	to	to	PART
fcis-9422	62	16	only	only	ADV
fcis-9422	62	17	question	question	VERB
fcis-9422	62	18	sentences	sentence	NOUN
fcis-9422	62	19	like	like	ADP
fcis-9422	62	20	simbert	simbert	NOUN
fcis-9422	62	21	.	.	PUNCT
fcis-9422	63	1	roformer	roformer	NOUN
fcis-9422	63	2	-	-	PUNCT
fcis-9422	63	3	sim	sim	NOUN
fcis-9422	63	4	can	can	AUX
fcis-9422	63	5	be	be	AUX
fcis-9422	63	6	used	use	VERB
fcis-9422	63	7	to	to	PART
fcis-9422	63	8	generate	generate	VERB
fcis-9422	63	9	similar	similar	ADJ
fcis-9422	63	10	sentences	sentence	NOUN
fcis-9422	63	11	for	for	ADP
fcis-9422	63	12	general	general	ADJ
fcis-9422	63	13	sentences	sentence	NOUN
fcis-9422	63	14	,	,	PUNCT
fcis-9422	63	15	making	make	VERB
fcis-9422	63	16	it	it	PRON
fcis-9422	63	17	more	more	ADV
fcis-9422	63	18	versatile	versatile	ADJ
fcis-9422	63	19	.	.	PUNCT
fcis-9422	64	1	other	other	ADJ
fcis-9422	64	2	training	training	NOUN
fcis-9422	64	3	details	detail	NOUN
fcis-9422	64	4	include	include	VERB
fcis-9422	64	5	the	the	DET
fcis-9422	64	6	use	use	NOUN
fcis-9422	64	7	of	of	ADP
fcis-9422	64	8	a	a	DET
fcis-9422	64	9	larger	large	ADJ
fcis-9422	64	10	batch	batch	NOUN
fcis-9422	64	11	size	size	NOUN
fcis-9422	64	12	and	and	CCONJ
fcis-9422	64	13	maxlen	maxlen	ADJ
fcis-9422	64	14	in	in	ADP
fcis-9422	64	15	roformer	roformer	NOUN
fcis-9422	64	16	-	-	PUNCT
fcis-9422	64	17	sim	sim	NOUN
fcis-9422	64	18	.	.	PUNCT
fcis-9422	65	1	the	the	DET
fcis-9422	65	2	key	key	NOUN
fcis-9422	65	3	to	to	ADP
fcis-9422	65	4	both	both	CCONJ
fcis-9422	65	5	simbert	simbert	NOUN
fcis-9422	65	6	and	and	CCONJ
fcis-9422	65	7	roformer	roformer	NOUN
fcis-9422	65	8	-	-	PUNCT
fcis-9422	65	9	sim	sim	NOUN
fcis-9422	65	10	lies	lie	VERB
fcis-9422	65	11	in	in	ADP
fcis-9422	65	12	the	the	DET
fcis-9422	65	13	construction	construction	NOUN
fcis-9422	65	14	of	of	ADP
fcis-9422	65	15	training	training	NOUN
fcis-9422	65	16	corpus	corpus	NOUN
fcis-9422	65	17	.	.	PUNCT
fcis-9422	66	1	the	the	DET
fcis-9422	66	2	training	training	NOUN
fcis-9422	66	3	corpus	corpus	NOUN
fcis-9422	66	4	of	of	ADP
fcis-9422	66	5	roformer	roformer	NOUN
fcis-9422	66	6	-	-	PUNCT
fcis-9422	66	7	sim	sim	NOUN
fcis-9422	66	8	includes	include	VERB
fcis-9422	66	9	two	two	NUM
fcis-9422	66	10	parts	part	NOUN
fcis-9422	66	11	:	:	PUNCT
fcis-9422	66	12	similar	similar	ADJ
fcis-9422	66	13	question	question	NOUN
fcis-9422	66	14	types	type	NOUN
fcis-9422	66	15	and	and	CCONJ
fcis-9422	66	16	general	general	ADJ
fcis-9422	66	17	similar	similar	ADJ
fcis-9422	66	18	types	type	NOUN
fcis-9422	66	19	.	.	PUNCT
fcis-9422	67	1	for	for	ADP
fcis-9422	67	2	similar	similar	ADJ
fcis-9422	67	3	question	question	NOUN
fcis-9422	67	4	types	type	NOUN
fcis-9422	67	5	,	,	PUNCT
fcis-9422	67	6	similar	similar	ADJ
fcis-9422	67	7	questions	question	NOUN
fcis-9422	67	8	are	be	AUX
fcis-9422	67	9	collected	collect	VERB
fcis-9422	67	10	from	from	ADP
fcis-9422	67	11	baidu	baidu	NOUN
fcis-9422	67	12	knows	know	VERB
fcis-9422	67	13	,	,	PUNCT
fcis-9422	67	14	and	and	CCONJ
fcis-9422	67	15	then	then	ADV
fcis-9422	67	16	further	far	ADV
fcis-9422	67	17	cleaned	clean	VERB
fcis-9422	67	18	by	by	ADP
fcis-9422	67	19	rules	rule	NOUN
fcis-9422	67	20	like	like	ADP
fcis-9422	67	21	simbert	simbert	NOUN
fcis-9422	67	22	.	.	PUNCT
fcis-9422	68	1	for	for	ADP
fcis-9422	68	2	general	general	ADJ
fcis-9422	68	3	similar	similar	ADJ
fcis-9422	68	4	types	type	NOUN
fcis-9422	68	5	,	,	PUNCT
fcis-9422	68	6	since	since	SCONJ
fcis-9422	68	7	there	there	PRON
fcis-9422	68	8	is	be	VERB
fcis-9422	68	9	no	no	DET
fcis-9422	68	10	existing	exist	VERB
fcis-9422	68	11	data	datum	NOUN
fcis-9422	68	12	to	to	PART
fcis-9422	68	13	collect	collect	VERB
fcis-9422	68	14	,	,	PUNCT
fcis-9422	68	15	two	two	NUM
fcis-9422	68	16	unsupervised	unsupervised	ADJ
fcis-9422	68	17	methods	method	NOUN
fcis-9422	68	18	are	be	AUX
fcis-9422	68	19	proposed	propose	VERB
fcis-9422	68	20	to	to	PART
fcis-9422	68	21	construct	construct	VERB
fcis-9422	68	22	similar	similar	ADJ
fcis-9422	68	23	sentence	sentence	NOUN
fcis-9422	68	24	pairs	pair	NOUN
fcis-9422	68	25	.	.	PUNCT
fcis-9422	69	1	the	the	DET
fcis-9422	69	2	first	first	ADJ
fcis-9422	69	3	method	method	NOUN
fcis-9422	69	4	is	be	AUX
fcis-9422	69	5	based	base	VERB
fcis-9422	69	6	on	on	ADP
fcis-9422	69	7	the	the	DET
fcis-9422	69	8	idea	idea	NOUN
fcis-9422	69	9	that	that	SCONJ
fcis-9422	69	10	"	"	PUNCT
fcis-9422	69	11	the	the	DET
fcis-9422	69	12	answers	answer	NOUN
fcis-9422	69	13	to	to	ADP
fcis-9422	69	14	the	the	DET
fcis-9422	69	15	same	same	ADJ
fcis-9422	69	16	question	question	NOUN
fcis-9422	69	17	are	be	AUX
fcis-9422	69	18	similar	similar	ADJ
fcis-9422	69	19	"	"	PUNCT
fcis-9422	69	20	.	.	PUNCT
fcis-9422	70	1	assuming	assume	VERB
fcis-9422	70	2	that	that	SCONJ
fcis-9422	70	3	there	there	PRON
fcis-9422	70	4	is	be	VERB
fcis-9422	70	5	a	a	DET
fcis-9422	70	6	question	question	NOUN
fcis-9422	70	7	-	-	PUNCT
fcis-9422	70	8	and	and	CCONJ
fcis-9422	70	9	-	-	PUNCT
fcis-9422	70	10	answer	answer	NOUN
fcis-9422	70	11	corpus	corpus	NOUN
fcis-9422	70	12	with	with	ADP
fcis-9422	70	13	multiple	multiple	ADJ
fcis-9422	70	14	answers	answer	NOUN
fcis-9422	70	15	to	to	ADP
fcis-9422	70	16	the	the	DET
fcis-9422	70	17	same	same	ADJ
fcis-9422	70	18	question	question	NOUN
fcis-9422	70	19	,	,	PUNCT
fcis-9422	70	20	each	each	DET
fcis-9422	70	21	answer	answer	NOUN
fcis-9422	70	22	can	can	AUX
fcis-9422	70	23	be	be	AUX
fcis-9422	70	24	divided	divide	VERB
fcis-9422	70	25	into	into	ADP
fcis-9422	70	26	sentences	sentence	NOUN
fcis-9422	70	27	,	,	PUNCT
fcis-9422	70	28	and	and	CCONJ
fcis-9422	70	29	then	then	ADV
fcis-9422	70	30	a	a	DET
fcis-9422	70	31	pre	pre	ADJ
fcis-9422	70	32	-	-	ADJ
fcis-9422	70	33	existing	existing	ADJ
fcis-9422	70	34	similarity	similarity	NOUN
fcis-9422	70	35	function	function	NOUN
fcis-9422	70	36	can	can	AUX
fcis-9422	70	37	be	be	AUX
fcis-9422	70	38	used	use	VERB
fcis-9422	70	39	to	to	PART
fcis-9422	70	40	compare	compare	VERB
fcis-9422	70	41	the	the	DET
fcis-9422	70	42	similarity	similarity	NOUN
fcis-9422	70	43	between	between	ADP
fcis-9422	70	44	answers	answer	NOUN
fcis-9422	70	45	.	.	PUNCT
fcis-9422	71	1	pairs	pair	NOUN
fcis-9422	71	2	of	of	ADP
fcis-9422	71	3	sentences	sentence	NOUN
fcis-9422	71	4	with	with	ADP
fcis-9422	71	5	similarity	similarity	NOUN
fcis-9422	71	6	scores	score	NOUN
fcis-9422	71	7	above	above	ADP
fcis-9422	71	8	a	a	DET
fcis-9422	71	9	certain	certain	ADJ
fcis-9422	71	10	threshold	threshold	NOUN
fcis-9422	71	11	are	be	AUX
fcis-9422	71	12	selected	select	VERB
fcis-9422	71	13	as	as	ADP
fcis-9422	71	14	similar	similar	ADJ
fcis-9422	71	15	sentence	sentence	NOUN
fcis-9422	71	16	pairs	pair	NOUN
fcis-9422	71	17	.	.	PUNCT
fcis-9422	72	1	the	the	DET
fcis-9422	72	2	second	second	ADJ
fcis-9422	72	3	method	method	NOUN
fcis-9422	72	4	is	be	AUX
fcis-9422	72	5	based	base	VERB
fcis-9422	72	6	on	on	ADP
fcis-9422	72	7	the	the	DET
fcis-9422	72	8	idea	idea	NOUN
fcis-9422	72	9	that	that	SCONJ
fcis-9422	72	10	"	"	PUNCT
fcis-9422	72	11	sentences	sentence	NOUN
fcis-9422	72	12	in	in	ADP
fcis-9422	72	13	the	the	DET
fcis-9422	72	14	same	same	ADJ
fcis-9422	72	15	passage	passage	NOUN
fcis-9422	72	16	are	be	AUX
fcis-9422	72	17	similar	similar	ADJ
fcis-9422	72	18	"	"	PUNCT
fcis-9422	72	19	.	.	PUNCT
fcis-9422	73	1	each	each	DET
fcis-9422	73	2	passage	passage	NOUN
fcis-9422	73	3	is	be	AUX
fcis-9422	73	4	divided	divide	VERB
fcis-9422	73	5	into	into	ADP
fcis-9422	73	6	sentences	sentence	NOUN
fcis-9422	73	7	,	,	PUNCT
fcis-9422	73	8	and	and	CCONJ
fcis-9422	73	9	a	a	DET
fcis-9422	73	10	pre	pre	ADJ
fcis-9422	73	11	-	-	ADJ
fcis-9422	73	12	existing	existing	ADJ
fcis-9422	73	13	similarity	similarity	NOUN
fcis-9422	73	14	function	function	NOUN
fcis-9422	73	15	is	be	AUX
fcis-9422	73	16	used	use	VERB
fcis-9422	73	17	to	to	PART
fcis-9422	73	18	calculate	calculate	VERB
fcis-9422	73	19	the	the	DET
fcis-9422	73	20	similarity	similarity	NOUN
fcis-9422	73	21	between	between	ADP
fcis-9422	73	22	each	each	DET
fcis-9422	73	23	pair	pair	NOUN
fcis-9422	73	24	of	of	ADP
fcis-9422	73	25	sentences	sentence	NOUN
fcis-9422	73	26	.	.	PUNCT
fcis-9422	74	1	pairs	pair	NOUN
fcis-9422	74	2	of	of	ADP
fcis-9422	74	3	sentences	sentence	NOUN
fcis-9422	74	4	with	with	ADP
fcis-9422	74	5	similarity	similarity	NOUN
fcis-9422	74	6	scores	score	NOUN
fcis-9422	74	7	above	above	ADP
fcis-9422	74	8	a	a	DET
fcis-9422	74	9	certain	certain	ADJ
fcis-9422	74	10	threshold	threshold	NOUN
fcis-9422	74	11	are	be	AUX
fcis-9422	74	12	selected	select	VERB
fcis-9422	74	13	as	as	ADP
fcis-9422	74	14	similar	similar	ADJ
fcis-9422	74	15	sentence	sentence	NOUN
fcis-9422	74	16	pairs	pair	NOUN
fcis-9422	74	17	.	.	PUNCT
fcis-9422	75	1	the	the	DET
fcis-9422	75	2	similarity	similarity	NOUN
fcis-9422	75	3	function	function	NOUN
fcis-9422	75	4	is	be	AUX
fcis-9422	75	5	a	a	DET
fcis-9422	75	6	variant	variant	NOUN
fcis-9422	75	7	of	of	ADP
fcis-9422	75	8	jaccard	jaccard	ADJ
fcis-9422	75	9	similarity	similarity	NOUN
fcis-9422	75	10	that	that	PRON
fcis-9422	75	11	directly	directly	ADV
fcis-9422	75	12	uses	use	VERB
fcis-9422	75	13	a	a	DET
fcis-9422	75	14	rule	rule	NOUN
fcis-9422	75	15	-	-	PUNCT
fcis-9422	75	16	based	base	VERB
fcis-9422	75	17	character	character	NOUN
fcis-9422	75	18	-	-	PUNCT
fcis-9422	75	19	level	level	NOUN
fcis-9422	75	20	similarity	similarity	NOUN
fcis-9422	75	21	,	,	PUNCT
fcis-9422	75	22	and	and	CCONJ
fcis-9422	75	23	semantic	semantic	ADJ
fcis-9422	75	24	relevance	relevance	NOUN
fcis-9422	75	25	is	be	AUX
fcis-9422	75	26	obtained	obtain	VERB
fcis-9422	75	27	through	through	ADP
fcis-9422	75	28	the	the	DET
fcis-9422	75	29	intra	intra	ADJ
fcis-9422	75	30	-	-	ADJ
fcis-9422	75	31	chapter	chapter	ADJ
fcis-9422	75	32	correlation	correlation	NOUN
fcis-9422	75	33	and	and	CCONJ
fcis-9422	75	34	the	the	DET
fcis-9422	75	35	generalization	generalization	NOUN
fcis-9422	75	36	ability	ability	NOUN
fcis-9422	75	37	of	of	ADP
fcis-9422	75	38	the	the	DET
fcis-9422	75	39	pre	pre	ADJ
fcis-9422	75	40	-	-	ADJ
fcis-9422	75	41	trained	train	VERB
fcis-9422	75	42	model	model	NOUN
fcis-9422	75	43	itself	itself	PRON
fcis-9422	75	44	.	.	PUNCT
fcis-9422	76	1	through	through	ADP
fcis-9422	76	2	the	the	DET
fcis-9422	76	3	first	first	ADJ
fcis-9422	76	4	approach	approach	NOUN
fcis-9422	76	5	,	,	PUNCT
fcis-9422	76	6	about	about	ADV
fcis-9422	76	7	4.5	4.5	NUM
fcis-9422	76	8	million	million	NUM
fcis-9422	76	9	(	(	PUNCT
fcis-9422	76	10	pseudo	pseudo	NOUN
fcis-9422	76	11	)	)	PUNCT
fcis-9422	76	12	similar	similar	ADJ
fcis-9422	76	13	sentence	sentence	NOUN
fcis-9422	76	14	pairs	pair	NOUN
fcis-9422	76	15	were	be	AUX
fcis-9422	76	16	constructed	construct	VERB
fcis-9422	76	17	from	from	ADP
fcis-9422	76	18	several	several	ADJ
fcis-9422	76	19	reading	reading	NOUN
fcis-9422	76	20	comprehension	comprehension	NOUN
fcis-9422	76	21	datasets	dataset	NOUN
fcis-9422	76	22	;	;	PUNCT
fcis-9422	76	23	through	through	ADP
fcis-9422	76	24	the	the	DET
fcis-9422	76	25	second	second	ADJ
fcis-9422	76	26	approach	approach	NOUN
fcis-9422	76	27	,	,	PUNCT
fcis-9422	76	28	about	about	ADV
fcis-9422	76	29	4.7	4.7	NUM
fcis-9422	76	30	million	million	NUM
fcis-9422	76	31	(	(	PUNCT
fcis-9422	76	32	pseudo	pseudo	NOUN
fcis-9422	76	33	)	)	PUNCT
fcis-9422	76	34	similar	similar	ADJ
fcis-9422	76	35	sentence	sentence	NOUN
fcis-9422	76	36	pairs	pair	NOUN
fcis-9422	76	37	were	be	AUX
fcis-9422	76	38	constructed	construct	VERB
fcis-9422	76	39	from	from	ADP
fcis-9422	76	40	over	over	ADP
fcis-9422	76	41	30	30	NUM
fcis-9422	76	42	gb	gb	NOUN
fcis-9422	76	43	of	of	ADP
fcis-9422	76	44	parallel	parallel	ADJ
fcis-9422	76	45	corpus	corpu	NOUN
fcis-9422	76	46	;	;	PUNCT
fcis-9422	76	47	and	and	CCONJ
fcis-9422	76	48	48	48	NUM
fcis-9422	76	49	about	about	ADV
fcis-9422	76	50	30	30	NUM
fcis-9422	76	51	million	million	NUM
fcis-9422	76	52	groups	group	NOUN
fcis-9422	76	53	of	of	ADP
fcis-9422	76	54	similar	similar	ADJ
fcis-9422	76	55	sentences	sentence	NOUN
fcis-9422	76	56	were	be	AUX
fcis-9422	76	57	crawled	crawl	VERB
fcis-9422	76	58	(	(	PUNCT
fcis-9422	76	59	each	each	DET
fcis-9422	76	60	group	group	NOUN
fcis-9422	76	61	can	can	AUX
fcis-9422	76	62	form	form	VERB
fcis-9422	76	63	multiple	multiple	ADJ
fcis-9422	76	64	pairs	pair	NOUN
fcis-9422	76	65	)	)	PUNCT
fcis-9422	76	66	.	.	PUNCT
fcis-9422	77	1	it	it	PRON
fcis-9422	77	2	can	can	AUX
fcis-9422	77	3	be	be	AUX
fcis-9422	77	4	seen	see	VERB
fcis-9422	77	5	that	that	SCONJ
fcis-9422	77	6	the	the	DET
fcis-9422	77	7	number	number	NOUN
fcis-9422	77	8	of	of	ADP
fcis-9422	77	9	question	question	NOUN
fcis-9422	77	10	sentences	sentence	NOUN
fcis-9422	77	11	far	far	ADV
fcis-9422	77	12	exceeds	exceed	VERB
fcis-9422	77	13	that	that	PRON
fcis-9422	77	14	of	of	ADP
fcis-9422	77	15	general	general	ADJ
fcis-9422	77	16	sentences	sentence	NOUN
fcis-9422	77	17	,	,	PUNCT
fcis-9422	77	18	so	so	SCONJ
fcis-9422	77	19	they	they	PRON
fcis-9422	77	20	were	be	AUX
fcis-9422	77	21	sampled	sample	VERB
fcis-9422	77	22	in	in	ADP
fcis-9422	77	23	a	a	DET
fcis-9422	77	24	1:1	1:1	NUM
fcis-9422	77	25	ratio	ratio	NOUN
fcis-9422	77	26	to	to	PART
fcis-9422	77	27	ensure	ensure	VERB
fcis-9422	77	28	balance	balance	NOUN
fcis-9422	77	29	between	between	ADP
fcis-9422	77	30	each	each	DET
fcis-9422	77	31	type	type	NOUN
fcis-9422	77	32	of	of	ADP
fcis-9422	77	33	sample	sample	NOUN
fcis-9422	77	34	.	.	PUNCT
fcis-9422	78	1	during	during	ADP
fcis-9422	78	2	training	training	NOUN
fcis-9422	78	3	,	,	PUNCT
fcis-9422	78	4	in	in	ADP
fcis-9422	78	5	order	order	NOUN
fcis-9422	78	6	to	to	PART
fcis-9422	78	7	enhance	enhance	VERB
fcis-9422	78	8	the	the	DET
fcis-9422	78	9	generation	generation	NOUN
fcis-9422	78	10	ability	ability	NOUN
fcis-9422	78	11	of	of	ADP
fcis-9422	78	12	the	the	DET
fcis-9422	78	13	model	model	NOUN
fcis-9422	78	14	,	,	PUNCT
fcis-9422	78	15	a	a	DET
fcis-9422	78	16	portion	portion	NOUN
fcis-9422	78	17	of	of	ADP
fcis-9422	78	18	the	the	DET
fcis-9422	78	19	input	input	NOUN
fcis-9422	78	20	tokens	token	NOUN
fcis-9422	78	21	in	in	ADP
fcis-9422	78	22	the	the	DET
fcis-9422	78	23	input	input	NOUN
fcis-9422	78	24	sentence	sentence	NOUN
fcis-9422	78	25	are	be	AUX
fcis-9422	78	26	randomly	randomly	ADV
fcis-9422	78	27	replaced	replace	VERB
fcis-9422	78	28	with	with	ADP
fcis-9422	78	29	[	[	X
fcis-9422	78	30	mask	mask	NOUN
fcis-9422	78	31	]	]	PUNCT
fcis-9422	78	32	,	,	PUNCT
fcis-9422	78	33	which	which	PRON
fcis-9422	78	34	is	be	AUX
fcis-9422	78	35	a	a	DET
fcis-9422	78	36	pretraining	pretraine	VERB
fcis-9422	78	37	method	method	NOUN
fcis-9422	78	38	first	first	ADV
fcis-9422	78	39	proposed	propose	VERB
fcis-9422	78	40	by	by	ADP
fcis-9422	78	41	bert	bert	PROPN
fcis-9422	78	42	.	.	PUNCT
fcis-9422	79	1	however	however	ADV
fcis-9422	79	2	,	,	PUNCT
fcis-9422	79	3	unlike	unlike	ADP
fcis-9422	79	4	bert	bert	PROPN
fcis-9422	79	5	which	which	PRON
fcis-9422	79	6	inputs	input	VERB
fcis-9422	79	7	noisy	noisy	ADJ
fcis-9422	79	8	sentences	sentence	NOUN
fcis-9422	79	9	and	and	CCONJ
fcis-9422	79	10	outputs	output	VERB
fcis-9422	79	11	the	the	DET
fcis-9422	79	12	original	original	ADJ
fcis-9422	79	13	sentence	sentence	NOUN
fcis-9422	79	14	,	,	PUNCT
fcis-9422	79	15	roformer	roformer	NOUN
fcis-9422	79	16	-	-	PUNCT
fcis-9422	79	17	sim	sim	NOUN
fcis-9422	79	18	inputs	input	NOUN
fcis-9422	79	19	noisy	noisy	ADJ
fcis-9422	79	20	sentences	sentence	NOUN
fcis-9422	79	21	and	and	CCONJ
fcis-9422	79	22	outputs	output	VERB
fcis-9422	79	23	a	a	DET
fcis-9422	79	24	similar	similar	ADJ
fcis-9422	79	25	sentence	sentence	NOUN
fcis-9422	79	26	to	to	ADP
fcis-9422	79	27	the	the	DET
fcis-9422	79	28	original	original	ADJ
fcis-9422	79	29	one	one	NUM
fcis-9422	79	30	.	.	PUNCT
fcis-9422	80	1	1	1	NUM
fcis-9422	80	2	0	0	NUM
fcis-9422	80	3	0	0	NUM
fcis-9422	80	4	0	0	NUM
fcis-9422	80	5	0	0	NUM
fcis-9422	80	6	1	1	NUM
fcis-9422	80	7	0	0	NUM
fcis-9422	80	8	0	0	NUM
fcis-9422	80	9	0	0	NUM
fcis-9422	80	10	0	0	NUM
fcis-9422	80	11	0	0	NUM
fcis-9422	80	12	0	0	NUM
fcis-9422	80	13	1	1	NUM
fcis-9422	80	14	0	0	NUM
fcis-9422	80	15	0	0	NUM
fcis-9422	80	16	0	0	NUM
fcis-9422	80	17	0	0	NUM
fcis-9422	80	18	1	1	NUM
fcis-9422	80	19	0	0	NUM
fcis-9422	80	20	0	0	NUM
fcis-9422	80	21	0	0	NUM
fcis-9422	80	22	0	0	NUM
fcis-9422	80	23	0	0	NUM
fcis-9422	80	24	0	0	NUM
fcis-9422	80	25	1	1	NUM
fcis-9422	80	26	0	0	NUM
fcis-9422	80	27	0	0	NUM
fcis-9422	80	28	0	0	NUM
fcis-9422	80	29	0	0	NUM
fcis-9422	80	30	1	1	NUM
fcis-9422	80	31	cls	cls	NOUN
fcis-9422	80	32	白	白	NOUN
fcis-9422	80	33	云	云	ADP
fcis-9422	80	34	山	山	X
fcis-9422	80	35	主	主	X
fcis-9422	80	36	峰	峰	X
fcis-9422	80	37	海	海	NOUN
fcis-9422	80	38	拔	拔	PROPN
fcis-9422	80	39	3	3	NUM
fcis-9422	80	40	8	8	NUM
fcis-9422	80	41	2	2	NUM
fcis-9422	80	42	米	米	PROPN
fcis-9422	80	43	sep	sep	PROPN
fcis-9422	80	44	白	白	PROPN
fcis-9422	80	45	云	云	ADP
fcis-9422	80	46	山	山	X
fcis-9422	80	47	高	高	NOUN
fcis-9422	80	48	3	3	NUM
fcis-9422	80	49	8	8	NUM
fcis-9422	80	50	2	2	NUM
fcis-9422	80	51	米	米	PROPN
fcis-9422	80	52	sep	sep	PROPN
fcis-9422	80	53	cls	cls	NOUN
fcis-9422	81	1	白	白	NOUN
fcis-9422	81	2	云	云	ADP
fcis-9422	81	3	山	山	X
fcis-9422	81	4	高	高	NOUN
fcis-9422	81	5	3	3	NUM
fcis-9422	81	6	8	8	NUM
fcis-9422	81	7	2	2	NUM
fcis-9422	81	8	米	米	PROPN
fcis-9422	81	9	sep	sep	PROPN
fcis-9422	81	10	白	白	PROPN
fcis-9422	81	11	云	云	ADP
fcis-9422	81	12	山	山	X
fcis-9422	81	13	主	主	X
fcis-9422	81	14	峰	峰	X
fcis-9422	81	15	海	海	NOUN
fcis-9422	81	16	拔	拔	PROPN
fcis-9422	81	17	3	3	NUM
fcis-9422	81	18	8	8	NUM
fcis-9422	81	19	2	2	NUM
fcis-9422	81	20	米	米	PROPN
fcis-9422	81	21	sep	sep	PROPN
fcis-9422	81	22	cls	cls	NOUN
fcis-9422	81	23	广	广	X
fcis-9422	81	24	州	州	X
fcis-9422	81	25	的	的	X
fcis-9422	81	26	简	简	X
fcis-9422	81	27	称	称	ADP
fcis-9422	81	28	是	是	PROPN
fcis-9422	81	29	什	什	PROPN
fcis-9422	81	30	么	么	PROPN
fcis-9422	81	31	呢	呢	ADV
fcis-9422	81	32	sep	sep	NOUN
fcis-9422	81	33	广	广	PROPN
fcis-9422	81	34	州	州	X
fcis-9422	81	35	又	又	X
fcis-9422	81	36	简	简	X
fcis-9422	81	37	称	称	ADP
fcis-9422	81	38	什	什	X
fcis-9422	81	39	么	么	INTJ
fcis-9422	81	40	sep	sep	PROPN
fcis-9422	81	41	cls	cls	NOUN
fcis-9422	81	42	广	广	PROPN
fcis-9422	81	43	州	州	X
fcis-9422	81	44	又	又	X
fcis-9422	81	45	简	简	X
fcis-9422	81	46	称	称	ADP
fcis-9422	81	47	什	什	PROPN
fcis-9422	82	1	么	么	INTJ
fcis-9422	82	2	sep	sep	NOUN
fcis-9422	82	3	广	广	PROPN
fcis-9422	82	4	州	州	X
fcis-9422	82	5	的	的	X
fcis-9422	82	6	简	简	X
fcis-9422	82	7	称	称	ADP
fcis-9422	82	8	是	是	PROPN
fcis-9422	82	9	什	什	PROPN
fcis-9422	82	10	么	么	PROPN
fcis-9422	82	11	呢	呢	ADP
fcis-9422	82	12	sep	sep	PROPN
fcis-9422	82	13	cls	cls	NOUN
fcis-9422	83	1	白	白	PRON
fcis-9422	83	2	切	切	X
fcis-9422	83	3	鸡	鸡	X
fcis-9422	83	4	怎	怎	X
fcis-9422	83	5	么	么	PROPN
fcis-9422	83	6	做	做	X
fcis-9422	83	7	？	？	PUNCT
fcis-9422	83	8	sep	sep	PROPN
fcis-9422	83	9	如	如	PROPN
fcis-9422	84	1	何	何	VERB
fcis-9422	84	2	做	做	NOUN
fcis-9422	84	3	一	一	NOUN
fcis-9422	84	4	只	只	X
fcis-9422	84	5	好	好	X
fcis-9422	84	6	吃	吃	X
fcis-9422	84	7	的	的	PROPN
fcis-9422	84	8	白	白	X
fcis-9422	84	9	切	切	X
fcis-9422	84	10	鸡	鸡	NOUN
fcis-9422	84	11	sep	sep	PROPN
fcis-9422	84	12	cls	cls	NOUN
fcis-9422	84	13	如	如	X
fcis-9422	84	14	何	何	NOUN
fcis-9422	84	15	做	做	NOUN
fcis-9422	84	16	一	一	NOUN
fcis-9422	85	1	只	只	X
fcis-9422	85	2	好	好	X
fcis-9422	85	3	吃	吃	X
fcis-9422	85	4	的	的	PROPN
fcis-9422	85	5	白	白	PRON
fcis-9422	85	6	切	切	X
fcis-9422	85	7	鸡	鸡	PROPN
fcis-9422	85	8	sep	sep	PROPN
fcis-9422	85	9	白	白	PRON
fcis-9422	85	10	切	切	X
fcis-9422	85	11	鸡	鸡	X
fcis-9422	85	12	怎	怎	X
fcis-9422	85	13	么	么	PROPN
fcis-9422	85	14	做	做	NOUN
fcis-9422	85	15	？	？	PUNCT
fcis-9422	85	16	sep	sep	PROPN
fcis-9422	85	17	cls	cls	PROPN
fcis-9422	85	18	cls	cls	PROPN
fcis-9422	85	19	cls	cls	PROPN
fcis-9422	85	20	cls	cls	PROPN
fcis-9422	85	21	cls	cls	PROPN
fcis-9422	85	22	cls	cls	NOUN
fcis-9422	85	23	figure	figure	NOUN
fcis-9422	85	24	2	2	NUM
fcis-9422	85	25	.	.	X
fcis-9422	85	26	diagram	diagram	NOUN
fcis-9422	85	27	of	of	ADP
fcis-9422	85	28	simbert	simbert	PROPN
fcis-9422	85	29	training	training	NOUN
fcis-9422	85	30	method	method	NOUN
fcis-9422	85	31	.	.	PUNCT
fcis-9422	86	1	after	after	ADP
fcis-9422	86	2	training	train	VERB
fcis-9422	86	3	roformer	roformer	NOUN
fcis-9422	86	4	-	-	PUNCT
fcis-9422	86	5	sim	sim	NOUN
fcis-9422	86	6	,	,	PUNCT
fcis-9422	86	7	the	the	DET
fcis-9422	86	8	retrieval	retrieval	NOUN
fcis-9422	86	9	performance	performance	NOUN
fcis-9422	86	10	of	of	ADP
fcis-9422	86	11	simbert	simbert	NOUN
fcis-9422	86	12	is	be	AUX
fcis-9422	86	13	further	far	ADV
fcis-9422	86	14	transferred	transfer	VERB
fcis-9422	86	15	to	to	ADP
fcis-9422	86	16	roformer	roformer	NOUN
fcis-9422	86	17	-	-	PUNCT
fcis-9422	86	18	sim	sim	NOUN
fcis-9422	86	19	through	through	ADP
fcis-9422	86	20	distillation	distillation	NOUN
fcis-9422	86	21	,	,	PUNCT
fcis-9422	86	22	so	so	SCONJ
fcis-9422	86	23	that	that	SCONJ
fcis-9422	86	24	the	the	DET
fcis-9422	86	25	retrieval	retrieval	NOUN
fcis-9422	86	26	performance	performance	NOUN
fcis-9422	86	27	of	of	ADP
fcis-9422	86	28	roformersim	roformersim	NOUN
fcis-9422	86	29	is	be	AUX
fcis-9422	86	30	basically	basically	ADV
fcis-9422	86	31	on	on	ADP
fcis-9422	86	32	par	par	NOUN
fcis-9422	86	33	with	with	ADP
fcis-9422	86	34	or	or	CCONJ
fcis-9422	86	35	even	even	ADV
fcis-9422	86	36	better	well	ADJ
fcis-9422	86	37	than	than	ADP
fcis-9422	86	38	that	that	PRON
fcis-9422	86	39	of	of	ADP
fcis-9422	86	40	simbert	simbert	NOUN
fcis-9422	86	41	.	.	PUNCT
fcis-9422	87	1	the	the	DET
fcis-9422	87	2	distillation	distillation	NOUN
fcis-9422	87	3	method	method	NOUN
fcis-9422	87	4	is	be	AUX
fcis-9422	87	5	simple	simple	ADJ
fcis-9422	87	6	:	:	PUNCT
fcis-9422	87	7	if	if	SCONJ
fcis-9422	87	8	the	the	DET
fcis-9422	87	9	sentence	sentence	NOUN
fcis-9422	87	10	vectors	vector	NOUN
fcis-9422	87	11	generated	generate	VERB
fcis-9422	87	12	by	by	ADP
fcis-9422	87	13	simbert	simbert	PROPN
fcis-9422	87	14	for	for	ADP
fcis-9422	87	15	a	a	DET
fcis-9422	87	16	batch	batch	NOUN
fcis-9422	87	17	of	of	ADP
fcis-9422	87	18	sentences	sentence	NOUN
fcis-9422	87	19	are	be	AUX
fcis-9422	87	20	u1	u1	NOUN
fcis-9422	87	21	,	,	PUNCT
fcis-9422	87	22	u2	u2	NOUN
fcis-9422	87	23	,	,	PUNCT
fcis-9422	87	24	...	...	PUNCT
fcis-9422	87	25	,	,	PUNCT
fcis-9422	87	26	un	un	PROPN
fcis-9422	87	27	,	,	PUNCT
fcis-9422	87	28	and	and	CCONJ
fcis-9422	87	29	the	the	DET
fcis-9422	87	30	sentence	sentence	NOUN
fcis-9422	87	31	vectors	vector	NOUN
fcis-9422	87	32	generated	generate	VERB
fcis-9422	87	33	by	by	ADP
fcis-9422	87	34	roformer	roformer	NOUN
fcis-9422	87	35	-	-	PUNCT
fcis-9422	87	36	sim	sim	NOUN
fcis-9422	87	37	are	be	AUX
fcis-9422	87	38	v1	v1	NOUN
fcis-9422	87	39	,	,	PUNCT
fcis-9422	87	40	v2	v2	PROPN
fcis-9422	87	41	,	,	PUNCT
fcis-9422	87	42	...	...	PUNCT
fcis-9422	87	43	,	,	PUNCT
fcis-9422	87	44	vn	vn	INTJ
fcis-9422	87	45	,	,	PUNCT
fcis-9422	87	46	then	then	ADV
fcis-9422	87	47	:	:	PUNCT
fcis-9422	87	48			PRON
fcis-9422	87	49			PROPN
fcis-9422	87	50			PROPN
fcis-9422	87	51			PROPN
fcis-9422	88	1	n	n	CCONJ
fcis-9422	88	2	i	i	PRON
fcis-9422	88	3	n	n	CCONJ
fcis-9422	88	4	j	j	PROPN
fcis-9422	88	5	jijin	jijin	PROPN
fcis-9422	88	6	vvuul	vvuul	NOUN
fcis-9422	88	7	1	1	NUM
fcis-9422	88	8	2	2	NUM
fcis-9422	88	9	1	1	NUM
fcis-9422	88	10	)	)	PUNCT
fcis-9422	88	11	)	)	PUNCT
fcis-9422	89	1	,	,	PUNCT
fcis-9422	89	2	cos(),(cos(2	cos(),(cos(2	NOUN
fcis-9422	89	3			X
fcis-9422	89	4	(	(	PUNCT
fcis-9422	89	5	1	1	NUM
fcis-9422	89	6	)	)	PUNCT
fcis-9422	89	7	to	to	PART
fcis-9422	89	8	prevent	prevent	VERB
fcis-9422	89	9	the	the	DET
fcis-9422	89	10	model	model	NOUN
fcis-9422	89	11	from	from	ADP
fcis-9422	89	12	"	"	PUNCT
fcis-9422	89	13	forgetting	forget	VERB
fcis-9422	89	14	"	"	PUNCT
fcis-9422	89	15	the	the	DET
fcis-9422	89	16	generation	generation	NOUN
fcis-9422	89	17	model	model	NOUN
fcis-9422	89	18	,	,	PUNCT
fcis-9422	89	19	a	a	DET
fcis-9422	89	20	generation	generation	NOUN
fcis-9422	89	21	loss	loss	NOUN
fcis-9422	89	22	is	be	AUX
fcis-9422	89	23	also	also	ADV
fcis-9422	89	24	added	add	VERB
fcis-9422	89	25	during	during	ADP
fcis-9422	89	26	distillation	distillation	NOUN
fcis-9422	89	27	,	,	PUNCT
fcis-9422	89	28	l	l	NOUN
fcis-9422	89	29	=	=	NOUN
fcis-9422	89	30	lsim+lgen	lsim+lgen	PROPN
fcis-9422	89	31	.	.	PUNCT
fcis-9422	90	1	roformer	roformer	NOUN
fcis-9422	90	2	-	-	PUNCT
fcis-9422	90	3	sim	sim	NOUN
fcis-9422	90	4	is	be	AUX
fcis-9422	90	5	based	base	VERB
fcis-9422	90	6	on	on	ADP
fcis-9422	90	7	the	the	DET
fcis-9422	90	8	roformer	roformer	NOUN
fcis-9422	90	9	architecture	architecture	NOUN
fcis-9422	90	10	and	and	CCONJ
fcis-9422	90	11	trained	train	VERB
fcis-9422	90	12	using	use	VERB
fcis-9422	90	13	a	a	DET
fcis-9422	90	14	large	large	ADJ
fcis-9422	90	15	number	number	NOUN
fcis-9422	90	16	of	of	ADP
fcis-9422	90	17	similar	similar	ADJ
fcis-9422	90	18	pairs	pair	NOUN
fcis-9422	90	19	of	of	ADP
fcis-9422	90	20	sentences	sentence	NOUN
fcis-9422	90	21	to	to	PART
fcis-9422	90	22	perform	perform	VERB
fcis-9422	90	23	sentence	sentence	NOUN
fcis-9422	90	24	expansion	expansion	NOUN
fcis-9422	90	25	.	.	PUNCT
fcis-9422	91	1	there	there	PRON
fcis-9422	91	2	are	be	VERB
fcis-9422	91	3	two	two	NUM
fcis-9422	91	4	tasks	task	NOUN
fcis-9422	91	5	during	during	ADP
fcis-9422	91	6	training	training	NOUN
fcis-9422	91	7	:	:	PUNCT
fcis-9422	91	8	generation	generation	NOUN
fcis-9422	91	9	and	and	CCONJ
fcis-9422	91	10	retrieval	retrieval	NOUN
fcis-9422	91	11	.	.	PUNCT
fcis-9422	92	1	in	in	ADP
fcis-9422	92	2	the	the	DET
fcis-9422	92	3	generation	generation	NOUN
fcis-9422	92	4	task	task	NOUN
fcis-9422	92	5	,	,	PUNCT
fcis-9422	92	6	similar	similar	ADJ
fcis-9422	92	7	to	to	ADP
fcis-9422	92	8	bert	bert	PROPN
fcis-9422	92	9	,	,	PUNCT
fcis-9422	92	10	a	a	DET
fcis-9422	92	11	noisy	noisy	ADJ
fcis-9422	92	12	input	input	NOUN
fcis-9422	92	13	sentence	sentence	NOUN
fcis-9422	92	14	is	be	AUX
fcis-9422	92	15	provided	provide	VERB
fcis-9422	92	16	,	,	PUNCT
fcis-9422	92	17	and	and	CCONJ
fcis-9422	92	18	the	the	DET
fcis-9422	92	19	model	model	NOUN
fcis-9422	92	20	outputs	output	VERB
fcis-9422	92	21	a	a	DET
fcis-9422	92	22	similar	similar	ADJ
fcis-9422	92	23	sentence	sentence	NOUN
fcis-9422	92	24	to	to	ADP
fcis-9422	92	25	the	the	DET
fcis-9422	92	26	original	original	ADJ
fcis-9422	92	27	sentence	sentence	NOUN
fcis-9422	92	28	.	.	PUNCT
fcis-9422	93	1	in	in	ADP
fcis-9422	93	2	the	the	DET
fcis-9422	93	3	retrieval	retrieval	NOUN
fcis-9422	93	4	task	task	NOUN
fcis-9422	93	5	,	,	PUNCT
fcis-9422	93	6	the	the	DET
fcis-9422	93	7	retrieval	retrieval	NOUN
fcis-9422	93	8	performance	performance	NOUN
fcis-9422	93	9	of	of	ADP
fcis-9422	93	10	simbert	simbert	PROPN
fcis-9422	93	11	is	be	AUX
fcis-9422	93	12	transferred	transfer	VERB
fcis-9422	93	13	to	to	ADP
fcis-9422	93	14	roformer	roformer	NOUN
fcis-9422	93	15	-	-	PUNCT
fcis-9422	93	16	sim	sim	NOUN
fcis-9422	93	17	through	through	ADP
fcis-9422	93	18	distillation	distillation	NOUN
fcis-9422	93	19	.	.	PUNCT
fcis-9422	94	1	2.3	2.3	NUM
fcis-9422	94	2	.	.	PUNCT
fcis-9422	95	1	candidate	candidate	NOUN
fcis-9422	95	2	entity	entity	NOUN
fcis-9422	95	3	generation	generation	NOUN
fcis-9422	95	4	the	the	DET
fcis-9422	95	5	generation	generation	NOUN
fcis-9422	95	6	of	of	ADP
fcis-9422	95	7	candidate	candidate	NOUN
fcis-9422	95	8	entities	entity	NOUN
fcis-9422	95	9	is	be	AUX
fcis-9422	95	10	a	a	DET
fcis-9422	95	11	crucial	crucial	ADJ
fcis-9422	95	12	prerequisite	prerequisite	NOUN
fcis-9422	95	13	for	for	ADP
fcis-9422	95	14	entity	entity	NOUN
fcis-9422	95	15	linking	linking	NOUN
fcis-9422	95	16	.	.	PUNCT
fcis-9422	96	1	the	the	DET
fcis-9422	96	2	main	main	ADJ
fcis-9422	96	3	goal	goal	NOUN
fcis-9422	96	4	of	of	ADP
fcis-9422	96	5	candidate	candidate	NOUN
fcis-9422	96	6	entity	entity	NOUN
fcis-9422	96	7	generation	generation	NOUN
fcis-9422	96	8	is	be	AUX
fcis-9422	96	9	to	to	PART
fcis-9422	96	10	obtain	obtain	VERB
fcis-9422	96	11	a	a	DET
fcis-9422	96	12	candidate	candidate	NOUN
fcis-9422	96	13	entity	entity	NOUN
fcis-9422	96	14	set	set	VERB
fcis-9422	96	15	for	for	ADP
fcis-9422	96	16	each	each	DET
fcis-9422	96	17	entity	entity	NOUN
fcis-9422	96	18	mention	mention	NOUN
fcis-9422	96	19	.	.	PUNCT
fcis-9422	97	1	the	the	DET
fcis-9422	97	2	selection	selection	NOUN
fcis-9422	97	3	of	of	ADP
fcis-9422	97	4	candidate	candidate	NOUN
fcis-9422	97	5	entities	entity	NOUN
fcis-9422	97	6	has	have	VERB
fcis-9422	97	7	a	a	DET
fcis-9422	97	8	great	great	ADJ
fcis-9422	97	9	impact	impact	NOUN
fcis-9422	97	10	on	on	ADP
fcis-9422	97	11	the	the	DET
fcis-9422	97	12	accuracy	accuracy	NOUN
fcis-9422	97	13	of	of	ADP
fcis-9422	97	14	the	the	DET
fcis-9422	97	15	entity	entity	NOUN
fcis-9422	97	16	linking	link	VERB
fcis-9422	97	17	model	model	NOUN
fcis-9422	97	18	.	.	PUNCT
fcis-9422	98	1	the	the	DET
fcis-9422	98	2	generation	generation	NOUN
fcis-9422	98	3	of	of	ADP
fcis-9422	98	4	candidate	candidate	NOUN
fcis-9422	98	5	entities	entity	NOUN
fcis-9422	98	6	first	first	ADV
fcis-9422	98	7	requires	require	VERB
fcis-9422	98	8	a	a	DET
fcis-9422	98	9	given	give	VERB
fcis-9422	98	10	entity	entity	NOUN
fcis-9422	98	11	mention	mention	NOUN
fcis-9422	98	12	,	,	PUNCT
fcis-9422	98	13	and	and	CCONJ
fcis-9422	98	14	then	then	ADV
fcis-9422	98	15	a	a	DET
fcis-9422	98	16	candidate	candidate	NOUN
fcis-9422	98	17	entity	entity	NOUN
fcis-9422	98	18	list	list	NOUN
fcis-9422	98	19	corresponding	correspond	VERB
fcis-9422	98	20	to	to	ADP
fcis-9422	98	21	the	the	DET
fcis-9422	98	22	entity	entity	NOUN
fcis-9422	98	23	mention	mention	NOUN
fcis-9422	98	24	is	be	AUX
fcis-9422	98	25	obtained	obtain	VERB
fcis-9422	98	26	based	base	VERB
fcis-9422	98	27	on	on	ADP
fcis-9422	98	28	knowledge	knowledge	NOUN
fcis-9422	98	29	,	,	PUNCT
fcis-9422	98	30	rules	rule	NOUN
fcis-9422	98	31	,	,	PUNCT
fcis-9422	98	32	and	and	CCONJ
fcis-9422	98	33	other	other	ADJ
fcis-9422	98	34	information	information	NOUN
fcis-9422	98	35	.	.	PUNCT
fcis-9422	99	1	the	the	DET
fcis-9422	99	2	quality	quality	NOUN
fcis-9422	99	3	of	of	ADP
fcis-9422	99	4	the	the	DET
fcis-9422	99	5	candidate	candidate	NOUN
fcis-9422	99	6	entity	entity	NOUN
fcis-9422	99	7	set	set	VERB
fcis-9422	99	8	depends	depend	VERB
fcis-9422	99	9	on	on	ADP
fcis-9422	99	10	the	the	DET
fcis-9422	99	11	size	size	NOUN
fcis-9422	99	12	of	of	ADP
fcis-9422	99	13	the	the	DET
fcis-9422	99	14	candidate	candidate	NOUN
fcis-9422	99	15	entity	entity	NOUN
fcis-9422	99	16	set	set	VERB
fcis-9422	99	17	and	and	CCONJ
fcis-9422	99	18	whether	whether	SCONJ
fcis-9422	99	19	it	it	PRON
fcis-9422	99	20	contains	contain	VERB
fcis-9422	99	21	the	the	DET
fcis-9422	99	22	target	target	NOUN
fcis-9422	99	23	entity	entity	NOUN
fcis-9422	99	24	[	[	X
fcis-9422	99	25	17	17	NUM
fcis-9422	99	26	]	]	PUNCT
fcis-9422	99	27	.	.	PUNCT
fcis-9422	100	1	the	the	DET
fcis-9422	100	2	number	number	NOUN
fcis-9422	100	3	of	of	ADP
fcis-9422	100	4	candidate	candidate	NOUN
fcis-9422	100	5	entity	entity	NOUN
fcis-9422	100	6	sets	set	NOUN
fcis-9422	100	7	can	can	AUX
fcis-9422	100	8	not	not	PART
fcis-9422	100	9	be	be	AUX
fcis-9422	100	10	too	too	ADV
fcis-9422	100	11	many	many	ADJ
fcis-9422	100	12	or	or	CCONJ
fcis-9422	100	13	too	too	ADV
fcis-9422	100	14	few	few	ADJ
fcis-9422	100	15	,	,	PUNCT
fcis-9422	100	16	as	as	SCONJ
fcis-9422	100	17	too	too	ADV
fcis-9422	100	18	many	many	ADJ
fcis-9422	100	19	can	can	AUX
fcis-9422	100	20	lead	lead	VERB
fcis-9422	100	21	to	to	ADP
fcis-9422	100	22	low	low	ADJ
fcis-9422	100	23	experimental	experimental	ADJ
fcis-9422	100	24	efficiency	efficiency	NOUN
fcis-9422	100	25	and	and	CCONJ
fcis-9422	100	26	too	too	ADV
fcis-9422	100	27	few	few	ADJ
fcis-9422	100	28	may	may	AUX
fcis-9422	100	29	result	result	VERB
fcis-9422	100	30	in	in	ADP
fcis-9422	100	31	the	the	DET
fcis-9422	100	32	target	target	NOUN
fcis-9422	100	33	entity	entity	NOUN
fcis-9422	100	34	not	not	PART
fcis-9422	100	35	being	be	AUX
fcis-9422	100	36	included	include	VERB
fcis-9422	100	37	.	.	PUNCT
fcis-9422	101	1	this	this	DET
fcis-9422	101	2	article	article	NOUN
fcis-9422	101	3	generates	generate	VERB
fcis-9422	101	4	candidate	candidate	NOUN
fcis-9422	101	5	entities	entity	NOUN
fcis-9422	101	6	by	by	ADP
fcis-9422	101	7	constructing	construct	VERB
fcis-9422	101	8	a	a	DET
fcis-9422	101	9	query	query	NOUN
fcis-9422	101	10	dictionary	dictionary	NOUN
fcis-9422	101	11	.	.	PUNCT
fcis-9422	102	1	specifically	specifically	ADV
fcis-9422	102	2	,	,	PUNCT
fcis-9422	102	3	the	the	DET
fcis-9422	102	4	main	main	ADJ
fcis-9422	102	5	process	process	NOUN
fcis-9422	102	6	of	of	ADP
fcis-9422	102	7	this	this	DET
fcis-9422	102	8	method	method	NOUN
fcis-9422	102	9	is	be	AUX
fcis-9422	102	10	as	as	SCONJ
fcis-9422	102	11	follows	follow	VERB
fcis-9422	102	12	:	:	PUNCT
fcis-9422	102	13	1.generate	1.generate	NUM
fcis-9422	102	14	query	query	NOUN
fcis-9422	102	15	dictionaries	dictionary	NOUN
fcis-9422	102	16	.	.	PUNCT
fcis-9422	103	1	preprocess	preprocess	VERB
fcis-9422	103	2	the	the	DET
fcis-9422	103	3	knowledge	knowledge	NOUN
fcis-9422	103	4	base	base	NOUN
fcis-9422	103	5	and	and	CCONJ
fcis-9422	103	6	generate	generate	VERB
fcis-9422	103	7	two	two	NUM
fcis-9422	103	8	query	query	NOUN
fcis-9422	103	9	dictionaries	dictionary	NOUN
fcis-9422	103	10	.	.	PUNCT
fcis-9422	104	1	one	one	NUM
fcis-9422	104	2	dictionary	dictionary	NOUN
fcis-9422	104	3	is	be	AUX
fcis-9422	104	4	used	use	VERB
fcis-9422	104	5	to	to	PART
fcis-9422	104	6	search	search	VERB
fcis-9422	104	7	for	for	ADP
fcis-9422	104	8	the	the	DET
fcis-9422	104	9	corresponding	corresponding	ADJ
fcis-9422	104	10	entity	entity	NOUN
fcis-9422	104	11	object	object	NOUN
fcis-9422	104	12	in	in	ADP
fcis-9422	104	13	the	the	DET
fcis-9422	104	14	knowledge	knowledge	NOUN
fcis-9422	104	15	base	base	NOUN
fcis-9422	104	16	through	through	ADP
fcis-9422	104	17	the	the	DET
fcis-9422	104	18	entity	entity	NOUN
fcis-9422	104	19	i	i	PROPN
fcis-9422	104	20	d	d	PROPN
fcis-9422	104	21	,	,	PUNCT
fcis-9422	104	22	and	and	CCONJ
fcis-9422	104	23	the	the	DET
fcis-9422	104	24	other	other	ADJ
fcis-9422	104	25	dictionary	dictionary	NOUN
fcis-9422	104	26	is	be	AUX
fcis-9422	104	27	used	use	VERB
fcis-9422	104	28	to	to	PART
fcis-9422	104	29	search	search	VERB
fcis-9422	104	30	for	for	ADP
fcis-9422	104	31	the	the	DET
fcis-9422	104	32	corresponding	correspond	VERB
fcis-9422	104	33	entity	entity	NOUN
fcis-9422	104	34	object	object	NOUN
fcis-9422	104	35	through	through	ADP
fcis-9422	104	36	the	the	DET
fcis-9422	104	37	entity	entity	NOUN
fcis-9422	104	38	mention	mention	NOUN
fcis-9422	104	39	name	name	NOUN
fcis-9422	104	40	,	,	PUNCT
fcis-9422	104	41	with	with	ADP
fcis-9422	104	42	the	the	DET
fcis-9422	104	43	key	key	NOUN
fcis-9422	104	44	of	of	ADP
fcis-9422	104	45	the	the	DET
fcis-9422	104	46	dictionary	dictionary	NOUN
fcis-9422	104	47	being	be	AUX
fcis-9422	104	48	the	the	DET
fcis-9422	104	49	entity	entity	NOUN
fcis-9422	104	50	name	name	NOUN
fcis-9422	104	51	and	and	CCONJ
fcis-9422	104	52	the	the	DET
fcis-9422	104	53	value	value	NOUN
fcis-9422	104	54	being	be	AUX
fcis-9422	104	55	the	the	DET
fcis-9422	104	56	set	set	NOUN
fcis-9422	104	57	of	of	ADP
fcis-9422	104	58	ids	id	NOUN
fcis-9422	104	59	of	of	ADP
fcis-9422	104	60	the	the	DET
fcis-9422	104	61	entities	entity	NOUN
fcis-9422	104	62	with	with	ADP
fcis-9422	104	63	the	the	DET
fcis-9422	104	64	same	same	ADJ
fcis-9422	104	65	name	name	NOUN
fcis-9422	104	66	.	.	PUNCT
fcis-9422	105	1	2.generate	2.generate	NUM
fcis-9422	105	2	candidate	candidate	NOUN
fcis-9422	105	3	entity	entity	NOUN
fcis-9422	105	4	set	set	NOUN
fcis-9422	105	5	.	.	PUNCT
fcis-9422	106	1	after	after	ADP
fcis-9422	106	2	generating	generate	VERB
fcis-9422	106	3	the	the	DET
fcis-9422	106	4	query	query	NOUN
fcis-9422	106	5	dictionary	dictionary	NOUN
fcis-9422	106	6	,	,	PUNCT
fcis-9422	106	7	it	it	PRON
fcis-9422	106	8	is	be	AUX
fcis-9422	106	9	necessary	necessary	ADJ
fcis-9422	106	10	to	to	PART
fcis-9422	106	11	recall	recall	VERB
fcis-9422	106	12	the	the	DET
fcis-9422	106	13	candidate	candidate	NOUN
fcis-9422	106	14	entity	entity	NOUN
fcis-9422	106	15	set	set	VERB
fcis-9422	106	16	to	to	PART
fcis-9422	106	17	find	find	VERB
fcis-9422	106	18	candidate	candidate	NOUN
fcis-9422	106	19	entities	entity	NOUN
fcis-9422	106	20	that	that	PRON
fcis-9422	106	21	may	may	AUX
fcis-9422	106	22	be	be	AUX
fcis-9422	106	23	related	relate	VERB
fcis-9422	106	24	to	to	ADP
fcis-9422	106	25	the	the	DET
fcis-9422	106	26	entities	entity	NOUN
fcis-9422	106	27	mentioned	mention	VERB
fcis-9422	106	28	in	in	ADP
fcis-9422	106	29	the	the	DET
fcis-9422	106	30	short	short	ADJ
fcis-9422	106	31	text	text	NOUN
fcis-9422	106	32	.	.	PUNCT
fcis-9422	107	1	for	for	SCONJ
fcis-9422	107	2	each	each	DET
fcis-9422	107	3	entity	entity	NOUN
fcis-9422	107	4	mention	mention	VERB
fcis-9422	107	5	appearing	appear	VERB
fcis-9422	107	6	in	in	ADP
fcis-9422	107	7	the	the	DET
fcis-9422	107	8	short	short	ADJ
fcis-9422	107	9	text	text	NOUN
fcis-9422	107	10	,	,	PUNCT
fcis-9422	107	11	the	the	DET
fcis-9422	107	12	i	i	PROPN
fcis-9422	107	13	d	d	PROPN
fcis-9422	107	14	set	set	NOUN
fcis-9422	107	15	of	of	ADP
fcis-9422	107	16	the	the	DET
fcis-9422	107	17	corresponding	corresponding	ADJ
fcis-9422	107	18	entity	entity	NOUN
fcis-9422	107	19	in	in	ADP
fcis-9422	107	20	the	the	DET
fcis-9422	107	21	knowledge	knowledge	NOUN
fcis-9422	107	22	base	base	NOUN
fcis-9422	107	23	can	can	AUX
fcis-9422	107	24	be	be	AUX
fcis-9422	107	25	extracted	extract	VERB
fcis-9422	107	26	through	through	ADP
fcis-9422	107	27	the	the	DET
fcis-9422	107	28	two	two	NUM
fcis-9422	107	29	query	query	NOUN
fcis-9422	107	30	dictionaries	dictionary	NOUN
fcis-9422	107	31	in	in	ADP
fcis-9422	107	32	turn	turn	NOUN
fcis-9422	107	33	,	,	PUNCT
fcis-9422	107	34	thereby	thereby	ADV
fcis-9422	107	35	generating	generate	VERB
fcis-9422	107	36	the	the	DET
fcis-9422	107	37	candidate	candidate	NOUN
fcis-9422	107	38	entity	entity	NOUN
fcis-9422	107	39	set	set	VERB
fcis-9422	107	40	.	.	PUNCT
fcis-9422	108	1	2.4	2.4	NUM
fcis-9422	108	2	.	.	PUNCT
fcis-9422	109	1	entity	entity	NOUN
fcis-9422	109	2	linking	link	VERB
fcis-9422	109	3	model	model	NOUN
fcis-9422	109	4	construction	construction	NOUN
fcis-9422	109	5	the	the	DET
fcis-9422	109	6	chinese	chinese	ADJ
fcis-9422	109	7	short	short	ADJ
fcis-9422	109	8	-	-	PUNCT
fcis-9422	109	9	text	text	NOUN
fcis-9422	109	10	entity	entity	NOUN
fcis-9422	109	11	linking	link	VERB
fcis-9422	109	12	model	model	NOUN
fcis-9422	109	13	proposed	propose	VERB
fcis-9422	109	14	in	in	ADP
fcis-9422	109	15	this	this	DET
fcis-9422	109	16	paper	paper	NOUN
fcis-9422	109	17	based	base	VERB
fcis-9422	109	18	on	on	ADP
fcis-9422	109	19	the	the	DET
fcis-9422	109	20	roformer	roformer	NOUN
fcis-9422	109	21	-	-	PUNCT
fcis-9422	109	22	sim	sim	ADJ
fcis-9422	109	23	pre	pre	ADJ
fcis-9422	109	24	-	-	ADJ
fcis-9422	109	25	training	training	ADJ
fcis-9422	109	26	model	model	NOUN
fcis-9422	109	27	is	be	AUX
fcis-9422	109	28	shown	show	VERB
fcis-9422	109	29	in	in	ADP
fcis-9422	109	30	figure	figure	NOUN
fcis-9422	109	31	3	3	NUM
fcis-9422	109	32	.	.	PUNCT
fcis-9422	110	1	the	the	DET
fcis-9422	110	2	model	model	NOUN
fcis-9422	110	3	combines	combine	VERB
fcis-9422	110	4	the	the	DET
fcis-9422	110	5	short	short	ADJ
fcis-9422	110	6	text	text	NOUN
fcis-9422	110	7	with	with	ADP
fcis-9422	110	8	added	add	VERB
fcis-9422	110	9	entity	entity	NOUN
fcis-9422	110	10	mention	mention	VERB
fcis-9422	110	11	position	position	NOUN
fcis-9422	110	12	tags	tag	NOUN
fcis-9422	110	13	and	and	CCONJ
fcis-9422	110	14	the	the	DET
fcis-9422	110	15	candidate	candidate	NOUN
fcis-9422	110	16	entity	entity	NOUN
fcis-9422	110	17	description	description	NOUN
fcis-9422	110	18	text	text	NOUN
fcis-9422	110	19	as	as	ADP
fcis-9422	110	20	input	input	NOUN
fcis-9422	110	21	and	and	CCONJ
fcis-9422	110	22	outputs	output	VERB
fcis-9422	110	23	the	the	DET
fcis-9422	110	24	similarity	similarity	NOUN
fcis-9422	110	25	score	score	NOUN
fcis-9422	110	26	between	between	ADP
fcis-9422	110	27	the	the	DET
fcis-9422	110	28	two	two	NUM
fcis-9422	110	29	.	.	PUNCT
fcis-9422	111	1	the	the	DET
fcis-9422	111	2	entity	entity	NOUN
fcis-9422	111	3	linking	link	VERB
fcis-9422	111	4	model	model	NOUN
fcis-9422	111	5	proposed	propose	VERB
fcis-9422	111	6	in	in	ADP
fcis-9422	111	7	this	this	DET
fcis-9422	111	8	paper	paper	NOUN
fcis-9422	111	9	first	first	ADV
fcis-9422	111	10	uses	use	VERB
fcis-9422	111	11	the	the	DET
fcis-9422	111	12	roformer	roformer	NOUN
fcis-9422	111	13	-	-	PUNCT
fcis-9422	111	14	sim	sim	ADJ
fcis-9422	111	15	pre	pre	ADJ
fcis-9422	111	16	-	-	ADJ
fcis-9422	111	17	training	training	ADJ
fcis-9422	111	18	model	model	NOUN
fcis-9422	111	19	to	to	PART
fcis-9422	111	20	extract	extract	VERB
fcis-9422	111	21	contextual	contextual	ADJ
fcis-9422	111	22	features	feature	NOUN
fcis-9422	111	23	and	and	CCONJ
fcis-9422	111	24	then	then	ADV
fcis-9422	111	25	uses	use	VERB
fcis-9422	111	26	a	a	DET
fcis-9422	111	27	fully	fully	ADV
fcis-9422	111	28	connected	connect	VERB
fcis-9422	111	29	layer	layer	NOUN
fcis-9422	111	30	to	to	PART
fcis-9422	111	31	obtain	obtain	VERB
fcis-9422	111	32	the	the	DET
fcis-9422	111	33	final	final	ADJ
fcis-9422	111	34	result	result	NOUN
fcis-9422	111	35	.	.	PUNCT
fcis-9422	112	1	figure	figure	VERB
fcis-9422	112	2	3	3	NUM
fcis-9422	112	3	.	.	PUNCT
fcis-9422	113	1	chinese	chinese	ADJ
fcis-9422	113	2	short	short	ADJ
fcis-9422	113	3	-	-	PUNCT
fcis-9422	113	4	text	text	NOUN
fcis-9422	113	5	entity	entity	NOUN
fcis-9422	113	6	linking	link	VERB
fcis-9422	113	7	model	model	NOUN
fcis-9422	113	8	based	base	VERB
fcis-9422	113	9	on	on	ADP
fcis-9422	113	10	roformer	roformer	NOUN
fcis-9422	113	11	-	-	PUNCT
fcis-9422	113	12	sim	sim	ADJ
fcis-9422	113	13	pre	pre	ADJ
fcis-9422	113	14	-	-	ADJ
fcis-9422	113	15	training	training	ADJ
fcis-9422	113	16	model	model	NOUN
fcis-9422	113	17	for	for	ADP
fcis-9422	113	18	the	the	DET
fcis-9422	113	19	entity	entity	NOUN
fcis-9422	113	20	linking	link	VERB
fcis-9422	113	21	task	task	NOUN
fcis-9422	113	22	,	,	PUNCT
fcis-9422	113	23	in	in	ADP
fcis-9422	113	24	order	order	NOUN
fcis-9422	113	25	to	to	PART
fcis-9422	113	26	improve	improve	VERB
fcis-9422	113	27	the	the	DET
fcis-9422	113	28	generalization	generalization	NOUN
fcis-9422	113	29	ability	ability	NOUN
fcis-9422	113	30	of	of	ADP
fcis-9422	113	31	the	the	DET
fcis-9422	113	32	model	model	NOUN
fcis-9422	113	33	and	and	CCONJ
fcis-9422	113	34	make	make	VERB
fcis-9422	113	35	it	it	PRON
fcis-9422	113	36	more	more	ADV
fcis-9422	113	37	robust	robust	ADJ
fcis-9422	113	38	in	in	ADP
fcis-9422	113	39	new	new	ADJ
fcis-9422	113	40	datasets	dataset	NOUN
fcis-9422	113	41	and	and	CCONJ
fcis-9422	113	42	scenarios	scenario	NOUN
fcis-9422	113	43	,	,	PUNCT
fcis-9422	113	44	this	this	DET
fcis-9422	113	45	paper	paper	NOUN
fcis-9422	113	46	treats	treat	VERB
fcis-9422	113	47	it	it	PRON
fcis-9422	113	48	as	as	ADP
fcis-9422	113	49	a	a	DET
fcis-9422	113	50	binary	binary	ADJ
fcis-9422	113	51	classification	classification	NOUN
fcis-9422	113	52	problem	problem	NOUN
fcis-9422	113	53	.	.	PUNCT
fcis-9422	114	1	this	this	DET
fcis-9422	114	2	paper	paper	NOUN
fcis-9422	114	3	adds	add	VERB
fcis-9422	114	4	position	position	NOUN
fcis-9422	114	5	markers	marker	NOUN
fcis-9422	114	6	"	"	PUNCT
fcis-9422	114	7	#	#	NOUN
fcis-9422	114	8	"	"	PUNCT
fcis-9422	114	9	to	to	ADP
fcis-9422	114	10	each	each	DET
fcis-9422	114	11	entity	entity	NOUN
fcis-9422	114	12	mention	mention	VERB
fcis-9422	114	13	to	to	PART
fcis-9422	114	14	indicate	indicate	VERB
fcis-9422	114	15	the	the	DET
fcis-9422	114	16	position	position	NOUN
fcis-9422	114	17	of	of	ADP
fcis-9422	114	18	the	the	DET
fcis-9422	114	19	entity	entity	NOUN
fcis-9422	114	20	to	to	PART
fcis-9422	114	21	be	be	AUX
fcis-9422	114	22	disambiguated	disambiguate	VERB
fcis-9422	114	23	.	.	PUNCT
fcis-9422	115	1	the	the	DET
fcis-9422	115	2	short	short	ADJ
fcis-9422	115	3	text	text	NOUN
fcis-9422	115	4	with	with	ADP
fcis-9422	115	5	entity	entity	NOUN
fcis-9422	115	6	mentions	mention	VERB
fcis-9422	115	7	position	position	NOUN
fcis-9422	115	8	markers	marker	NOUN
fcis-9422	115	9	and	and	CCONJ
fcis-9422	115	10	the	the	DET
fcis-9422	115	11	candidate	candidate	NOUN
fcis-9422	115	12	entity	entity	NOUN
fcis-9422	115	13	description	description	NOUN
fcis-9422	115	14	text	text	NOUN
fcis-9422	115	15	are	be	AUX
fcis-9422	115	16	combined	combine	VERB
fcis-9422	115	17	and	and	CCONJ
fcis-9422	115	18	input	input	NOUN
fcis-9422	115	19	into	into	ADP
fcis-9422	115	20	the	the	DET
fcis-9422	115	21	model	model	NOUN
fcis-9422	115	22	as	as	ADP
fcis-9422	115	23	a	a	DET
fcis-9422	115	24	text	text	NOUN
fcis-9422	115	25	pair	pair	NOUN
fcis-9422	115	26	.	.	PUNCT
fcis-9422	116	1	the	the	DET
fcis-9422	116	2	roformer	roformer	NOUN
fcis-9422	116	3	-	-	PUNCT
fcis-9422	116	4	sim	sim	NOUN
fcis-9422	116	5	model	model	NOUN
fcis-9422	116	6	is	be	AUX
fcis-9422	116	7	used	use	VERB
fcis-9422	116	8	to	to	PART
fcis-9422	116	9	extract	extract	VERB
fcis-9422	116	10	contextual	contextual	ADJ
fcis-9422	116	11	features	feature	NOUN
fcis-9422	116	12	,	,	PUNCT
fcis-9422	116	13	and	and	CCONJ
fcis-9422	116	14	then	then	ADV
fcis-9422	116	15	the	the	DET
fcis-9422	116	16	feature	feature	NOUN
fcis-9422	116	17	vector	vector	NOUN
fcis-9422	116	18	of	of	ADP
fcis-9422	116	19	the	the	DET
fcis-9422	116	20	cls	cls	NOUN
fcis-9422	116	21	position	position	NOUN
fcis-9422	116	22	output	output	NOUN
fcis-9422	116	23	by	by	ADP
fcis-9422	116	24	the	the	DET
fcis-9422	116	25	roformer	roformer	NOUN
fcis-9422	116	26	-	-	PUNCT
fcis-9422	116	27	sim	sim	NOUN
fcis-9422	116	28	model	model	NOUN
fcis-9422	116	29	is	be	AUX
fcis-9422	116	30	input	input	NOUN
fcis-9422	116	31	into	into	ADP
fcis-9422	116	32	a	a	DET
fcis-9422	116	33	fully	fully	ADV
fcis-9422	116	34	connected	connect	VERB
fcis-9422	116	35	layer	layer	NOUN
fcis-9422	116	36	.	.	PUNCT
fcis-9422	117	1	then	then	ADV
fcis-9422	117	2	,	,	PUNCT
fcis-9422	117	3	the	the	DET
fcis-9422	117	4	candidate	candidate	NOUN
fcis-9422	117	5	entity	entity	NOUN
fcis-9422	117	6	similarity	similarity	NOUN
fcis-9422	117	7	score	score	NOUN
fcis-9422	117	8	is	be	AUX
fcis-9422	117	9	mapped	map	VERB
fcis-9422	117	10	to	to	ADP
fcis-9422	117	11	the	the	DET
fcis-9422	117	12	range	range	NOUN
fcis-9422	117	13	of	of	ADP
fcis-9422	117	14	[	[	X
fcis-9422	117	15	-1	-1	X
fcis-9422	117	16	,	,	PUNCT
fcis-9422	117	17	1	1	X
fcis-9422	117	18	]	]	PUNCT
fcis-9422	117	19	using	use	VERB
fcis-9422	117	20	the	the	DET
fcis-9422	117	21	tanh	tanh	NOUN
fcis-9422	117	22	function	function	NOUN
fcis-9422	117	23	after	after	ADP
fcis-9422	117	24	the	the	DET
fcis-9422	117	25	dropout	dropout	NOUN
fcis-9422	117	26	layer	layer	NOUN
fcis-9422	117	27	,	,	PUNCT
fcis-9422	117	28	so	so	SCONJ
fcis-9422	117	29	that	that	SCONJ
fcis-9422	117	30	the	the	DET
fcis-9422	117	31	similarity	similarity	NOUN
fcis-9422	117	32	scores	score	NOUN
fcis-9422	117	33	can	can	AUX
fcis-9422	117	34	be	be	AUX
fcis-9422	117	35	normalized	normalize	VERB
fcis-9422	117	36	and	and	CCONJ
fcis-9422	117	37	the	the	DET
fcis-9422	117	38	similarity	similarity	NOUN
fcis-9422	117	39	between	between	ADP
fcis-9422	117	40	different	different	ADJ
fcis-9422	117	41	entity	entity	NOUN
fcis-9422	117	42	mentions	mention	NOUN
fcis-9422	117	43	and	and	CCONJ
fcis-9422	117	44	candidate	candidate	NOUN
fcis-9422	117	45	entities	entity	NOUN
fcis-9422	117	46	can	can	AUX
fcis-9422	117	47	be	be	AUX
fcis-9422	117	48	compared	compare	VERB
fcis-9422	117	49	more	more	ADV
fcis-9422	117	50	fairly	fairly	ADV
fcis-9422	117	51	.	.	PUNCT
fcis-9422	118	1	the	the	DET
fcis-9422	118	2	model	model	NOUN
fcis-9422	118	3	then	then	ADV
fcis-9422	118	4	outputs	output	VERB
fcis-9422	118	5	the	the	DET
fcis-9422	118	6	scores	score	NOUN
fcis-9422	118	7	of	of	ADP
fcis-9422	118	8	each	each	DET
fcis-9422	118	9	candidate	candidate	NOUN
fcis-9422	118	10	entity	entity	NOUN
fcis-9422	118	11	and	and	CCONJ
fcis-9422	118	12	sorts	sort	NOUN
fcis-9422	118	13	them	they	PRON
fcis-9422	118	14	,	,	PUNCT
fcis-9422	118	15	and	and	CCONJ
fcis-9422	118	16	the	the	DET
fcis-9422	118	17	candidate	candidate	NOUN
fcis-9422	118	18	entity	entity	NOUN
fcis-9422	118	19	with	with	ADP
fcis-9422	118	20	the	the	DET
fcis-9422	118	21	highest	high	ADJ
fcis-9422	118	22	score	score	NOUN
fcis-9422	118	23	is	be	AUX
fcis-9422	118	24	the	the	DET
fcis-9422	118	25	object	object	NOUN
fcis-9422	118	26	referred	refer	VERB
fcis-9422	118	27	to	to	ADP
fcis-9422	118	28	by	by	ADP
fcis-9422	118	29	the	the	DET
fcis-9422	118	30	entity	entity	NOUN
fcis-9422	118	31	to	to	PART
fcis-9422	118	32	be	be	AUX
fcis-9422	118	33	disambiguated	disambiguate	VERB
fcis-9422	118	34	.	.	PUNCT
fcis-9422	119	1	49	49	NUM
fcis-9422	119	2	3	3	NUM
fcis-9422	119	3	.	.	PUNCT
fcis-9422	119	4	experimental	experimental	ADJ
fcis-9422	119	5	results	result	NOUN
fcis-9422	119	6	3.1	3.1	NUM
fcis-9422	119	7	.	.	PUNCT
fcis-9422	120	1	experimental	experimental	ADJ
fcis-9422	120	2	dataset	dataset	NOUN
fcis-9422	120	3	this	this	DET
fcis-9422	120	4	article	article	NOUN
fcis-9422	120	5	uses	use	VERB
fcis-9422	120	6	the	the	DET
fcis-9422	120	7	dataset	dataset	NOUN
fcis-9422	120	8	provided	provide	VERB
fcis-9422	120	9	by	by	ADP
fcis-9422	120	10	the	the	DET
fcis-9422	120	11	ccks2020	ccks2020	PROPN
fcis-9422	120	12	(	(	PUNCT
fcis-9422	120	13	the	the	DET
fcis-9422	120	14	2020	2020	NUM
fcis-9422	120	15	national	national	ADJ
fcis-9422	120	16	knowledge	knowledge	NOUN
fcis-9422	120	17	graph	graph	NOUN
fcis-9422	120	18	and	and	CCONJ
fcis-9422	120	19	semantic	semantic	ADJ
fcis-9422	120	20	computing	computing	NOUN
fcis-9422	120	21	conference	conference	NOUN
fcis-9422	120	22	)	)	PUNCT
fcis-9422	120	23	short	short	ADJ
fcis-9422	120	24	text	text	NOUN
fcis-9422	120	25	entity	entity	NOUN
fcis-9422	120	26	linking	link	VERB
fcis-9422	120	27	evaluation	evaluation	NOUN
fcis-9422	120	28	.	.	PUNCT
fcis-9422	121	1	the	the	DET
fcis-9422	121	2	ccks2020	ccks2020	PROPN
fcis-9422	121	3	evaluation	evaluation	NOUN
fcis-9422	121	4	dataset	dataset	NOUN
fcis-9422	121	5	includes	include	VERB
fcis-9422	121	6	a	a	DET
fcis-9422	121	7	standard	standard	ADJ
fcis-9422	121	8	dataset	dataset	NOUN
fcis-9422	121	9	and	and	CCONJ
fcis-9422	121	10	a	a	DET
fcis-9422	121	11	knowledge	knowledge	NOUN
fcis-9422	121	12	base	base	NOUN
fcis-9422	121	13	.	.	PUNCT
fcis-9422	122	1	the	the	DET
fcis-9422	122	2	knowledge	knowledge	NOUN
fcis-9422	122	3	base	base	NOUN
fcis-9422	122	4	comes	come	VERB
fcis-9422	122	5	from	from	ADP
fcis-9422	122	6	the	the	DET
fcis-9422	122	7	baidu	baidu	NOUN
fcis-9422	122	8	encyclopedia	encyclopedia	NOUN
fcis-9422	122	9	knowledge	knowledge	NOUN
fcis-9422	122	10	base	base	NOUN
fcis-9422	122	11	.	.	PUNCT
fcis-9422	123	1	each	each	DET
fcis-9422	123	2	entity	entity	NOUN
fcis-9422	123	3	in	in	ADP
fcis-9422	123	4	the	the	DET
fcis-9422	123	5	knowledge	knowledge	NOUN
fcis-9422	123	6	base	base	NOUN
fcis-9422	123	7	contains	contain	VERB
fcis-9422	123	8	a	a	DET
fcis-9422	123	9	subject_id	subject_id	PROPN
fcis-9422	123	10	(	(	PUNCT
fcis-9422	123	11	knowledge	knowledge	NOUN
fcis-9422	123	12	base	base	NOUN
fcis-9422	123	13	i	i	PROPN
fcis-9422	123	14	d	d	PROPN
fcis-9422	123	15	)	)	PUNCT
fcis-9422	123	16	,	,	PUNCT
fcis-9422	123	17	a	a	DET
fcis-9422	123	18	subject	subject	ADJ
fcis-9422	123	19	name	name	NOUN
fcis-9422	123	20	,	,	PUNCT
fcis-9422	123	21	entity	entity	NOUN
fcis-9422	123	22	aliases	alias	NOUN
fcis-9422	123	23	,	,	PUNCT
fcis-9422	123	24	corresponding	correspond	VERB
fcis-9422	123	25	concept	concept	NOUN
fcis-9422	123	26	types	type	NOUN
fcis-9422	123	27	,	,	PUNCT
fcis-9422	123	28	and	and	CCONJ
fcis-9422	123	29	a	a	DET
fcis-9422	123	30	series	series	NOUN
fcis-9422	123	31	of	of	ADP
fcis-9422	123	32	binary	binary	PROPN
fcis-9422	123	33	tuples	tuple	NOUN
fcis-9422	123	34	<	<	X
fcis-9422	123	35	predicate	predicate	NOUN
fcis-9422	123	36	,	,	PUNCT
fcis-9422	123	37	object	object	VERB
fcis-9422	123	38	>	>	X
fcis-9422	123	39	(	(	PUNCT
fcis-9422	123	40	<	<	X
fcis-9422	123	41	attribute	attribute	NOUN
fcis-9422	123	42	,	,	PUNCT
fcis-9422	123	43	attribute	attribute	NOUN
fcis-9422	123	44	value	value	NOUN
fcis-9422	123	45	>	>	PUNCT
fcis-9422	123	46	)	)	PUNCT
fcis-9422	123	47	information	information	NOUN
fcis-9422	123	48	related	relate	VERB
fcis-9422	123	49	to	to	ADP
fcis-9422	123	50	the	the	DET
fcis-9422	123	51	entity	entity	NOUN
fcis-9422	123	52	.	.	PUNCT
fcis-9422	124	1	the	the	DET
fcis-9422	124	2	standard	standard	ADJ
fcis-9422	124	3	dataset	dataset	VERB
fcis-9422	124	4	consists	consist	NOUN
fcis-9422	124	5	of	of	ADP
fcis-9422	124	6	training	training	NOUN
fcis-9422	124	7	set	set	NOUN
fcis-9422	124	8	,	,	PUNCT
fcis-9422	124	9	validation	validation	NOUN
fcis-9422	124	10	set	set	NOUN
fcis-9422	124	11	,	,	PUNCT
fcis-9422	124	12	and	and	CCONJ
fcis-9422	124	13	test	test	NOUN
fcis-9422	124	14	set	set	VERB
fcis-9422	124	15	,	,	PUNCT
fcis-9422	124	16	with	with	ADP
fcis-9422	124	17	a	a	DET
fcis-9422	124	18	total	total	NOUN
fcis-9422	124	19	of	of	ADP
fcis-9422	124	20	about	about	ADV
fcis-9422	124	21	90,000	90,000	NUM
fcis-9422	124	22	annotated	annotate	VERB
fcis-9422	124	23	data	datum	NOUN
fcis-9422	124	24	points	point	NOUN
fcis-9422	124	25	,	,	PUNCT
fcis-9422	124	26	distributed	distribute	VERB
fcis-9422	124	27	in	in	ADP
fcis-9422	124	28	a	a	DET
fcis-9422	124	29	7:1:1	7:1:1	NUM
fcis-9422	124	30	ratio	ratio	NOUN
fcis-9422	124	31	.	.	PUNCT
fcis-9422	125	1	all	all	DET
fcis-9422	125	2	three	three	NUM
fcis-9422	125	3	sets	set	NOUN
fcis-9422	125	4	of	of	ADP
fcis-9422	125	5	data	datum	NOUN
fcis-9422	125	6	were	be	AUX
fcis-9422	125	7	generated	generate	VERB
fcis-9422	125	8	through	through	ADP
fcis-9422	125	9	crowdsourcing	crowdsource	VERB
fcis-9422	125	10	by	by	ADP
fcis-9422	125	11	baidu	baidu	NOUN
fcis-9422	125	12	.	.	PUNCT
fcis-9422	126	1	the	the	DET
fcis-9422	126	2	annotated	annotate	VERB
fcis-9422	126	3	data	datum	NOUN
fcis-9422	126	4	set	set	VERB
fcis-9422	126	5	mainly	mainly	ADV
fcis-9422	126	6	comes	come	VERB
fcis-9422	126	7	from	from	ADP
fcis-9422	126	8	real	real	ADJ
fcis-9422	126	9	internet	internet	NOUN
fcis-9422	126	10	web	web	NOUN
fcis-9422	126	11	page	page	NOUN
fcis-9422	126	12	titles	title	NOUN
fcis-9422	126	13	,	,	PUNCT
fcis-9422	126	14	video	video	NOUN
fcis-9422	126	15	titles	title	NOUN
fcis-9422	126	16	,	,	PUNCT
fcis-9422	126	17	and	and	CCONJ
fcis-9422	126	18	user	user	NOUN
fcis-9422	126	19	search	search	NOUN
fcis-9422	126	20	queries	query	NOUN
fcis-9422	126	21	.	.	PUNCT
fcis-9422	127	1	each	each	DET
fcis-9422	127	2	annotated	annotate	VERB
fcis-9422	127	3	data	data	NOUN
fcis-9422	127	4	point	point	NOUN
fcis-9422	127	5	contains	contain	VERB
fcis-9422	127	6	a	a	DET
fcis-9422	127	7	piece	piece	NOUN
fcis-9422	127	8	of	of	ADP
fcis-9422	127	9	text	text	NOUN
fcis-9422	127	10	,	,	PUNCT
fcis-9422	127	11	an	an	DET
fcis-9422	127	12	entity	entity	NOUN
fcis-9422	127	13	mentions	mention	VERB
fcis-9422	127	14	in	in	ADP
fcis-9422	127	15	the	the	DET
fcis-9422	127	16	text	text	NOUN
fcis-9422	127	17	,	,	PUNCT
fcis-9422	127	18	and	and	CCONJ
fcis-9422	127	19	the	the	DET
fcis-9422	127	20	target	target	NOUN
fcis-9422	127	21	entity	entity	NOUN
fcis-9422	127	22	corresponding	correspond	VERB
fcis-9422	127	23	to	to	ADP
fcis-9422	127	24	the	the	DET
fcis-9422	127	25	entity	entity	NOUN
fcis-9422	127	26	mention	mention	NOUN
fcis-9422	127	27	in	in	ADP
fcis-9422	127	28	the	the	DET
fcis-9422	127	29	given	give	VERB
fcis-9422	127	30	knowledge	knowledge	NOUN
fcis-9422	127	31	base	base	NOUN
fcis-9422	127	32	.	.	PUNCT
fcis-9422	128	1	one	one	NUM
fcis-9422	128	2	training	training	NOUN
fcis-9422	128	3	data	datum	NOUN
fcis-9422	128	4	point	point	NOUN
fcis-9422	128	5	may	may	AUX
fcis-9422	128	6	contain	contain	VERB
fcis-9422	128	7	multiple	multiple	ADJ
fcis-9422	128	8	entity	entity	NOUN
fcis-9422	128	9	mentions	mention	NOUN
fcis-9422	128	10	.	.	PUNCT
fcis-9422	129	1	the	the	DET
fcis-9422	129	2	specific	specific	ADJ
fcis-9422	129	3	annotated	annotated	ADJ
fcis-9422	129	4	data	datum	NOUN
fcis-9422	129	5	is	be	AUX
fcis-9422	129	6	shown	show	VERB
fcis-9422	129	7	in	in	ADP
fcis-9422	129	8	table	table	NOUN
fcis-9422	129	9	1	1	NUM
fcis-9422	129	10	.	.	PUNCT
fcis-9422	129	11	table	table	NOUN
fcis-9422	129	12	1	1	NUM
fcis-9422	129	13	.	.	PUNCT
fcis-9422	130	1	examples	example	NOUN
fcis-9422	130	2	of	of	ADP
fcis-9422	130	3	specific	specific	ADJ
fcis-9422	130	4	annotation	annotation	NOUN
fcis-9422	130	5	data	data	NOUN
fcis-9422	130	6	number	number	NOUN
fcis-9422	130	7	text	text	NOUN
fcis-9422	130	8	mention	mention	VERB
fcis-9422	130	9	information	information	NOUN
fcis-9422	130	10	2484	2484	NUM
fcis-9422	130	11	医疗保险占工	医疗保险占工	VERB
fcis-9422	130	12	资的比例	资的比例	NOUN
fcis-9422	130	13	{	{	PUNCT
fcis-9422	130	14	"	"	PUNCT
fcis-9422	130	15	kb_id	kb_id	PROPN
fcis-9422	130	16	"	"	PUNCT
fcis-9422	130	17	:	:	PUNCT
fcis-9422	130	18	"	"	PUNCT
fcis-9422	130	19	17855	17855	NUM
fcis-9422	130	20	"	"	PUNCT
fcis-9422	130	21	,	,	PUNCT
fcis-9422	130	22	"	"	PUNCT
fcis-9422	130	23	mention	mention	NOUN
fcis-9422	130	24	"	"	PUNCT
fcis-9422	130	25	:	:	PUNCT
fcis-9422	130	26	"	"	PUNCT
fcis-9422	130	27	医疗保险	医疗保险	NOUN
fcis-9422	130	28	"	"	PUNCT
fcis-9422	130	29	,	,	PUNCT
fcis-9422	130	30	"	"	PUNCT
fcis-9422	130	31	offset	offset	NOUN
fcis-9422	130	32	"	"	PUNCT
fcis-9422	130	33	:	:	PUNCT
fcis-9422	130	34	"	"	PUNCT
fcis-9422	130	35	0	0	X
fcis-9422	130	36	"	"	PUNCT
fcis-9422	130	37	}	}	PUNCT
fcis-9422	130	38	{	{	PUNCT
fcis-9422	130	39	"	"	PUNCT
fcis-9422	130	40	kb_id	kb_id	PROPN
fcis-9422	130	41	"	"	PUNCT
fcis-9422	130	42	:	:	PUNCT
fcis-9422	130	43	"	"	PUNCT
fcis-9422	130	44	55578	55578	NUM
fcis-9422	130	45	"	"	PUNCT
fcis-9422	130	46	,	,	PUNCT
fcis-9422	130	47	"	"	PUNCT
fcis-9422	130	48	mention	mention	NOUN
fcis-9422	130	49	"	"	PUNCT
fcis-9422	130	50	:	:	PUNCT
fcis-9422	130	51	"	"	PUNCT
fcis-9422	130	52	工资	工资	X
fcis-9422	130	53	"	"	PUNCT
fcis-9422	130	54	,	,	PUNCT
fcis-9422	130	55	"	"	PUNCT
fcis-9422	130	56	offset	offset	NOUN
fcis-9422	130	57	"	"	PUNCT
fcis-9422	130	58	:	:	PUNCT
fcis-9422	130	59	"	"	PUNCT
fcis-9422	130	60	5	5	NUM
fcis-9422	130	61	"	"	PUNCT
fcis-9422	130	62	}	}	PUNCT
fcis-9422	130	63	{	{	PUNCT
fcis-9422	130	64	"	"	PUNCT
fcis-9422	130	65	kb_id	kb_id	PROPN
fcis-9422	130	66	"	"	PUNCT
fcis-9422	130	67	:	:	PUNCT
fcis-9422	130	68	"	"	PUNCT
fcis-9422	130	69	117247	117247	NUM
fcis-9422	130	70	"	"	PUNCT
fcis-9422	130	71	,	,	PUNCT
fcis-9422	130	72	"	"	PUNCT
fcis-9422	130	73	mention	mention	NOUN
fcis-9422	130	74	"	"	PUNCT
fcis-9422	130	75	:	:	PUNCT
fcis-9422	130	76	"	"	PUNCT
fcis-9422	130	77	比例	比例	NOUN
fcis-9422	130	78	"	"	PUNCT
fcis-9422	130	79	,	,	PUNCT
fcis-9422	130	80	"	"	PUNCT
fcis-9422	130	81	offset	offset	NOUN
fcis-9422	130	82	"	"	PUNCT
fcis-9422	130	83	:	:	PUNCT
fcis-9422	130	84	"	"	PUNCT
fcis-9422	130	85	8	8	NUM
fcis-9422	130	86	"	"	PUNCT
fcis-9422	130	87	}	}	PUNCT
fcis-9422	130	88	2519	2519	NUM
fcis-9422	130	89	《	《	X
fcis-9422	130	90	刘宽》答案	刘宽》答案	PROPN
fcis-9422	130	91	解析及翻译	解析及翻译	PROPN
fcis-9422	130	92	{	{	PUNCT
fcis-9422	130	93	"	"	PUNCT
fcis-9422	130	94	kb_id	kb_id	PROPN
fcis-9422	130	95	"	"	PUNCT
fcis-9422	130	96	:	:	PUNCT
fcis-9422	130	97	"	"	PUNCT
fcis-9422	130	98	119182	119182	NUM
fcis-9422	130	99	"	"	PUNCT
fcis-9422	130	100	,	,	PUNCT
fcis-9422	130	101	"	"	PUNCT
fcis-9422	130	102	mention	mention	NOUN
fcis-9422	130	103	"	"	PUNCT
fcis-9422	130	104	:	:	PUNCT
fcis-9422	130	105	"	"	PUNCT
fcis-9422	130	106	刘宽	刘宽	NOUN
fcis-9422	130	107	"	"	PUNCT
fcis-9422	130	108	,	,	PUNCT
fcis-9422	130	109	"	"	PUNCT
fcis-9422	130	110	offset	offset	NOUN
fcis-9422	130	111	"	"	PUNCT
fcis-9422	130	112	:	:	PUNCT
fcis-9422	130	113	"	"	PUNCT
fcis-9422	130	114	1	1	NUM
fcis-9422	130	115	"	"	PUNCT
fcis-9422	130	116	}	}	PUNCT
fcis-9422	130	117	,	,	PUNCT
fcis-9422	130	118	{	{	PUNCT
fcis-9422	130	119	"	"	PUNCT
fcis-9422	130	120	kb_id	kb_id	PROPN
fcis-9422	130	121	"	"	PUNCT
fcis-9422	130	122	:	:	PUNCT
fcis-9422	130	123	"	"	PUNCT
fcis-9422	130	124	156318	156318	NUM
fcis-9422	130	125	"	"	PUNCT
fcis-9422	130	126	,	,	PUNCT
fcis-9422	130	127	"	"	PUNCT
fcis-9422	130	128	mention	mention	NOUN
fcis-9422	130	129	"	"	PUNCT
fcis-9422	130	130	:	:	PUNCT
fcis-9422	130	131	"	"	PUNCT
fcis-9422	130	132	翻译	翻译	NOUN
fcis-9422	130	133	"	"	PUNCT
fcis-9422	130	134	,	,	PUNCT
fcis-9422	130	135	"	"	PUNCT
fcis-9422	130	136	offset	offset	NOUN
fcis-9422	130	137	"	"	PUNCT
fcis-9422	130	138	:	:	PUNCT
fcis-9422	130	139	"	"	PUNCT
fcis-9422	130	140	9	9	NUM
fcis-9422	130	141	"	"	PUNCT
fcis-9422	130	142	}	}	SYM
fcis-9422	130	143	2540	2540	NUM
fcis-9422	130	144	新疆拉面的做	新疆拉面的做	NOUN
fcis-9422	130	145	法视频	法视频	PROPN
fcis-9422	130	146	{	{	PUNCT
fcis-9422	130	147	"	"	PUNCT
fcis-9422	130	148	kb_id	kb_id	PROPN
fcis-9422	130	149	"	"	PUNCT
fcis-9422	130	150	:	:	PUNCT
fcis-9422	130	151	"	"	PUNCT
fcis-9422	130	152	48047	48047	NUM
fcis-9422	130	153	"	"	PUNCT
fcis-9422	130	154	,	,	PUNCT
fcis-9422	130	155	"	"	PUNCT
fcis-9422	130	156	mention	mention	NOUN
fcis-9422	130	157	"	"	PUNCT
fcis-9422	130	158	:	:	PUNCT
fcis-9422	130	159	"	"	PUNCT
fcis-9422	130	160	新疆	新疆	NOUN
fcis-9422	130	161	"	"	PUNCT
fcis-9422	130	162	,	,	PUNCT
fcis-9422	130	163	"	"	PUNCT
fcis-9422	130	164	offset	offset	NOUN
fcis-9422	130	165	"	"	PUNCT
fcis-9422	130	166	:	:	PUNCT
fcis-9422	130	167	"	"	PUNCT
fcis-9422	130	168	0	0	X
fcis-9422	130	169	"	"	PUNCT
fcis-9422	130	170	}	}	PUNCT
fcis-9422	130	171	{	{	PUNCT
fcis-9422	130	172	"	"	PUNCT
fcis-9422	130	173	kb_id	kb_id	PROPN
fcis-9422	130	174	"	"	PUNCT
fcis-9422	130	175	:	:	PUNCT
fcis-9422	130	176	"	"	PUNCT
fcis-9422	130	177	219021	219021	NUM
fcis-9422	130	178	"	"	PUNCT
fcis-9422	130	179	,	,	PUNCT
fcis-9422	130	180	"	"	PUNCT
fcis-9422	130	181	mention	mention	NOUN
fcis-9422	130	182	"	"	PUNCT
fcis-9422	130	183	:	:	PUNCT
fcis-9422	130	184	"	"	PUNCT
fcis-9422	130	185	拉面	拉面	PROPN
fcis-9422	130	186	"	"	PUNCT
fcis-9422	130	187	,	,	PUNCT
fcis-9422	130	188	"	"	PUNCT
fcis-9422	130	189	offset	offset	NOUN
fcis-9422	130	190	"	"	PUNCT
fcis-9422	130	191	:	:	PUNCT
fcis-9422	130	192	"	"	PUNCT
fcis-9422	130	193	2	2	NUM
fcis-9422	130	194	"	"	PUNCT
fcis-9422	130	195	}	}	PUNCT
fcis-9422	130	196	{	{	PUNCT
fcis-9422	130	197	"	"	PUNCT
fcis-9422	130	198	kb_id	kb_id	PROPN
fcis-9422	130	199	"	"	PUNCT
fcis-9422	130	200	:	:	PUNCT
fcis-9422	130	201	"	"	PUNCT
fcis-9422	130	202	136981	136981	NUM
fcis-9422	130	203	"	"	PUNCT
fcis-9422	130	204	,	,	PUNCT
fcis-9422	130	205	"	"	PUNCT
fcis-9422	130	206	mention	mention	NOUN
fcis-9422	130	207	"	"	PUNCT
fcis-9422	130	208	:	:	PUNCT
fcis-9422	130	209	"	"	PUNCT
fcis-9422	130	210	做法	做法	NOUN
fcis-9422	130	211	"	"	PUNCT
fcis-9422	130	212	,	,	PUNCT
fcis-9422	130	213	"	"	PUNCT
fcis-9422	130	214	offset	offset	NOUN
fcis-9422	130	215	"	"	PUNCT
fcis-9422	130	216	:	:	PUNCT
fcis-9422	130	217	"	"	PUNCT
fcis-9422	130	218	5	5	NUM
fcis-9422	130	219	"	"	PUNCT
fcis-9422	130	220	}	}	PUNCT
fcis-9422	130	221	{	{	PUNCT
fcis-9422	130	222	"	"	PUNCT
fcis-9422	130	223	kb_id	kb_id	PROPN
fcis-9422	130	224	"	"	PUNCT
fcis-9422	130	225	:	:	PUNCT
fcis-9422	130	226	"	"	PUNCT
fcis-9422	130	227	37477	37477	NUM
fcis-9422	130	228	"	"	PUNCT
fcis-9422	130	229	,	,	PUNCT
fcis-9422	130	230	"	"	PUNCT
fcis-9422	130	231	mention	mention	NOUN
fcis-9422	130	232	"	"	PUNCT
fcis-9422	130	233	:	:	PUNCT
fcis-9422	130	234	"	"	PUNCT
fcis-9422	130	235	视频	视频	NOUN
fcis-9422	130	236	"	"	PUNCT
fcis-9422	130	237	,	,	PUNCT
fcis-9422	130	238	"	"	PUNCT
fcis-9422	130	239	offset	offset	NOUN
fcis-9422	130	240	"	"	PUNCT
fcis-9422	130	241	:	:	PUNCT
fcis-9422	130	242	"	"	PUNCT
fcis-9422	130	243	7	7	NUM
fcis-9422	130	244	"	"	PUNCT
fcis-9422	130	245	}	}	PUNCT
fcis-9422	130	246	3.2	3.2	NUM
fcis-9422	130	247	.	.	PUNCT
fcis-9422	131	1	evaluation	evaluation	NOUN
fcis-9422	131	2	metrics	metric	NOUN
fcis-9422	131	3	given	give	VERB
fcis-9422	131	4	a	a	DET
fcis-9422	131	5	short	short	ADJ
fcis-9422	131	6	text	text	NOUN
fcis-9422	131	7	input	input	NOUN
fcis-9422	131	8	text	text	NOUN
fcis-9422	131	9	with	with	ADP
fcis-9422	131	10	n	n	PRON
fcis-9422	131	11	entity	entity	NOUN
fcis-9422	131	12	mentions	mention	NOUN
fcis-9422	131	13	:	:	PUNCT
fcis-9422	131	14	mtext=	mtext=	PRON
fcis-9422	131	15	{	{	PUNCT
fcis-9422	131	16	m1	m1	PROPN
fcis-9422	131	17	,	,	PUNCT
fcis-9422	131	18	m2	m2	PROPN
fcis-9422	131	19	,	,	PUNCT
fcis-9422	131	20	…	…	PUNCT
fcis-9422	131	21	,	,	PUNCT
fcis-9422	131	22	mn	mn	PROPN
fcis-9422	131	23	}	}	PUNCT
fcis-9422	131	24	,	,	PUNCT
fcis-9422	131	25	where	where	SCONJ
fcis-9422	131	26	each	each	DET
fcis-9422	131	27	entity	entity	NOUN
fcis-9422	131	28	mention	mention	VERB
fcis-9422	131	29	is	be	AUX
fcis-9422	131	30	linked	link	VERB
fcis-9422	131	31	to	to	ADP
fcis-9422	131	32	an	an	DET
fcis-9422	131	33	entity	entity	NOUN
fcis-9422	131	34	i	i	NOUN
fcis-9422	131	35	d	d	PROPN
fcis-9422	131	36	in	in	ADP
fcis-9422	131	37	the	the	DET
fcis-9422	131	38	knowledge	knowledge	NOUN
fcis-9422	131	39	base	base	NOUN
fcis-9422	131	40	etext=	etext=	NUM
fcis-9422	131	41	{	{	PUNCT
fcis-9422	131	42	e1	e1	PROPN
fcis-9422	131	43	,	,	PUNCT
fcis-9422	131	44	e2	e2	PROPN
fcis-9422	131	45	,	,	PUNCT
fcis-9422	131	46	…	…	PUNCT
fcis-9422	131	47	,	,	PUNCT
fcis-9422	131	48	en	en	ADP
fcis-9422	131	49	}	}	PUNCT
fcis-9422	131	50	,	,	PUNCT
fcis-9422	131	51	and	and	CCONJ
fcis-9422	131	52	the	the	DET
fcis-9422	131	53	output	output	NOUN
fcis-9422	131	54	of	of	ADP
fcis-9422	131	55	entity	entity	NOUN
fcis-9422	131	56	linking	link	VERB
fcis-9422	131	57	is	be	AUX
fcis-9422	131	58	e’text={e’1,e’2,	e’text={e’1,e’2,	ADJ
fcis-9422	131	59	…	…	PUNCT
fcis-9422	131	60	,e’n	,e’n	PUNCT
fcis-9422	131	61	}	}	PUNCT
fcis-9422	131	62	,	,	PUNCT
fcis-9422	131	63	the	the	DET
fcis-9422	131	64	precision	precision	NOUN
fcis-9422	131	65	,	,	PUNCT
fcis-9422	131	66	recall	recall	NOUN
fcis-9422	131	67	,	,	PUNCT
fcis-9422	131	68	and	and	CCONJ
fcis-9422	131	69	f1	f1	NOUN
fcis-9422	131	70	-	-	PUNCT
fcis-9422	131	71	score	score	NOUN
fcis-9422	131	72	of	of	ADP
fcis-9422	131	73	entity	entity	NOUN
fcis-9422	131	74	linking	link	VERB
fcis-9422	131	75	are	be	AUX
fcis-9422	131	76	defined	define	VERB
fcis-9422	131	77	as	as	SCONJ
fcis-9422	131	78	follows	follow	VERB
fcis-9422	131	79	:	:	PUNCT
fcis-9422	131	80			X
fcis-9422	131	81			PROPN
fcis-9422	131	82			PROPN
fcis-9422	131	83			NOUN
fcis-9422	131	84			PUNCT
fcis-9422	131	85			NUM
fcis-9422	131	86	n	n	CCONJ
fcis-9422	131	87	n	n	CCONJ
fcis-9422	131	88	n	n	CCONJ
fcis-9422	131	89	nn	nn	PROPN
fcis-9422	131	90	e	e	X
fcis-9422	131	91	ee	ee	PROPN
fcis-9422	131	92	p	p	PROPN
fcis-9422	131	93	n	n	PROPN
fcis-9422	131	94	n	n	NOUN
fcis-9422	131	95	'	'	PUNCT
fcis-9422	131	96	'	'	PUNCT
fcis-9422	131	97	(	(	PUNCT
fcis-9422	131	98	2	2	NUM
fcis-9422	131	99	)	)	PUNCT
fcis-9422	131	100			X
fcis-9422	131	101			X
fcis-9422	131	102			PROPN
fcis-9422	131	103			NOUN
fcis-9422	131	104			PUNCT
fcis-9422	131	105			NUM
fcis-9422	131	106	n	n	CCONJ
fcis-9422	131	107	n	n	CCONJ
fcis-9422	131	108	n	n	CCONJ
fcis-9422	131	109	nn	nn	PROPN
fcis-9422	131	110	e	e	X
fcis-9422	131	111	ee	ee	PROPN
fcis-9422	131	112	r	r	PROPN
fcis-9422	131	113	n	n	CCONJ
fcis-9422	131	114	n	n	NOUN
fcis-9422	131	115	'	'	PUNCT
fcis-9422	131	116	(	(	PUNCT
fcis-9422	131	117	3	3	NUM
fcis-9422	131	118	)	)	PUNCT
fcis-9422	131	119	rp	rp	NOUN
fcis-9422	131	120	rp	rp	PROPN
fcis-9422	131	121	f	f	PROPN
fcis-9422	131	122			PROPN
fcis-9422	131	123			PROPN
fcis-9422	131	124			PROPN
fcis-9422	131	125	)	)	PUNCT
fcis-9422	131	126	(	(	PUNCT
fcis-9422	131	127	2	2	NUM
fcis-9422	131	128	1	1	NUM
fcis-9422	131	129	(	(	PUNCT
fcis-9422	131	130	4	4	NUM
fcis-9422	131	131	)	)	PUNCT
fcis-9422	131	132	since	since	SCONJ
fcis-9422	131	133	the	the	DET
fcis-9422	131	134	entity	entity	NOUN
fcis-9422	131	135	mentions	mention	VERB
fcis-9422	131	136	in	in	ADP
fcis-9422	131	137	the	the	DET
fcis-9422	131	138	standard	standard	ADJ
fcis-9422	131	139	dataset	dataset	NOUN
fcis-9422	131	140	are	be	AUX
fcis-9422	131	141	given	give	VERB
fcis-9422	131	142	,	,	PUNCT
fcis-9422	131	143	we	we	PRON
fcis-9422	131	144	have	have	VERB
fcis-9422	131	145	=	=	NOUN
fcis-9422	131	146	,	,	PUNCT
fcis-9422	131	147	p	p	NOUN
fcis-9422	131	148	=	=	NOUN
fcis-9422	131	149	r	r	NOUN
fcis-9422	131	150	=	=	SYM
fcis-9422	131	151	f1	f1	NOUN
fcis-9422	131	152	,	,	PUNCT
fcis-9422	131	153	and	and	CCONJ
fcis-9422	131	154	use	use	VERB
fcis-9422	131	155	f1	f1	NOUN
fcis-9422	131	156	score	score	NOUN
fcis-9422	131	157	as	as	ADP
fcis-9422	131	158	the	the	DET
fcis-9422	131	159	main	main	ADJ
fcis-9422	131	160	evaluation	evaluation	NOUN
fcis-9422	131	161	metric	metric	NOUN
fcis-9422	131	162	for	for	ADP
fcis-9422	131	163	the	the	DET
fcis-9422	131	164	experimental	experimental	ADJ
fcis-9422	131	165	results	result	NOUN
fcis-9422	131	166	.	.	PUNCT
fcis-9422	132	1	3.3	3.3	NUM
fcis-9422	132	2	.	.	PUNCT
fcis-9422	132	3	experimental	experimental	ADJ
fcis-9422	132	4	environment	environment	NOUN
fcis-9422	132	5	and	and	CCONJ
fcis-9422	132	6	parameter	parameter	NOUN
fcis-9422	132	7	settings	setting	NOUN
fcis-9422	132	8	in	in	ADP
fcis-9422	132	9	this	this	DET
fcis-9422	132	10	experiment	experiment	NOUN
fcis-9422	132	11	,	,	PUNCT
fcis-9422	132	12	we	we	PRON
fcis-9422	132	13	rely	rely	VERB
fcis-9422	132	14	on	on	ADP
fcis-9422	132	15	the	the	DET
fcis-9422	132	16	python	python	NOUN
fcis-9422	132	17	programming	programming	NOUN
fcis-9422	132	18	language	language	NOUN
fcis-9422	132	19	,	,	PUNCT
fcis-9422	132	20	which	which	PRON
fcis-9422	132	21	contains	contain	VERB
fcis-9422	132	22	a	a	DET
fcis-9422	132	23	wealth	wealth	NOUN
fcis-9422	132	24	of	of	ADP
fcis-9422	132	25	deep	deep	ADJ
fcis-9422	132	26	learning	learning	NOUN
fcis-9422	132	27	libraries	library	NOUN
fcis-9422	132	28	and	and	CCONJ
fcis-9422	132	29	is	be	AUX
fcis-9422	132	30	easy	easy	ADJ
fcis-9422	132	31	to	to	PART
fcis-9422	132	32	debug	debug	VERB
fcis-9422	132	33	.	.	PUNCT
fcis-9422	133	1	we	we	PRON
fcis-9422	133	2	use	use	VERB
fcis-9422	133	3	the	the	DET
fcis-9422	133	4	pytorch	pytorch	NOUN
fcis-9422	133	5	deep	deep	ADJ
fcis-9422	133	6	learning	learning	NOUN
fcis-9422	133	7	framework	framework	NOUN
fcis-9422	133	8	to	to	PART
fcis-9422	133	9	build	build	VERB
fcis-9422	133	10	the	the	DET
fcis-9422	133	11	overall	overall	ADJ
fcis-9422	133	12	network	network	NOUN
fcis-9422	133	13	.	.	PUNCT
fcis-9422	134	1	in	in	ADP
fcis-9422	134	2	addition	addition	NOUN
fcis-9422	134	3	,	,	PUNCT
fcis-9422	134	4	we	we	PRON
fcis-9422	134	5	use	use	VERB
fcis-9422	134	6	a	a	DET
fcis-9422	134	7	single	single	ADJ
fcis-9422	134	8	gpu	gpu	NOUN
fcis-9422	134	9	to	to	PART
fcis-9422	134	10	train	train	VERB
fcis-9422	134	11	the	the	DET
fcis-9422	134	12	model	model	NOUN
fcis-9422	134	13	and	and	CCONJ
fcis-9422	134	14	operate	operate	VERB
fcis-9422	134	15	on	on	ADP
fcis-9422	134	16	the	the	DET
fcis-9422	134	17	vscode	vscode	NOUN
fcis-9422	134	18	compilation	compilation	NOUN
fcis-9422	134	19	platform	platform	NOUN
fcis-9422	134	20	based	base	VERB
fcis-9422	134	21	on	on	ADP
fcis-9422	134	22	the	the	DET
fcis-9422	134	23	ubuntu18.04	ubuntu18.04	ADJ
fcis-9422	134	24	operating	operating	NOUN
fcis-9422	134	25	system	system	NOUN
fcis-9422	134	26	.	.	PUNCT
fcis-9422	135	1	the	the	DET
fcis-9422	135	2	specific	specific	ADJ
fcis-9422	135	3	experimental	experimental	ADJ
fcis-9422	135	4	hardware	hardware	NOUN
fcis-9422	135	5	and	and	CCONJ
fcis-9422	135	6	software	software	NOUN
fcis-9422	135	7	environment	environment	NOUN
fcis-9422	135	8	are	be	AUX
fcis-9422	135	9	shown	show	VERB
fcis-9422	135	10	in	in	ADP
fcis-9422	135	11	table	table	NOUN
fcis-9422	135	12	2	2	NUM
fcis-9422	135	13	.	.	PUNCT
fcis-9422	135	14	table	table	NOUN
fcis-9422	135	15	2	2	NUM
fcis-9422	135	16	.	.	PUNCT
fcis-9422	135	17	experimental	experimental	ADJ
fcis-9422	135	18	hardware	hardware	NOUN
fcis-9422	135	19	and	and	CCONJ
fcis-9422	135	20	software	software	NOUN
fcis-9422	135	21	environment	environment	NOUN
fcis-9422	135	22	device	device	NOUN
fcis-9422	135	23	configuration	configuration	NOUN
fcis-9422	135	24	operating	operating	NOUN
fcis-9422	135	25	system	system	NOUN
fcis-9422	135	26	ubuntu18.04	ubuntu18.04	PROPN
fcis-9422	135	27	gpu	gpu	PROPN
fcis-9422	135	28	geforce	geforce	NOUN
fcis-9422	135	29	rtx	rtx	PROPN
fcis-9422	135	30	2080ti	2080ti	PROPN
fcis-9422	135	31	deep	deep	ADJ
fcis-9422	135	32	learning	learning	NOUN
fcis-9422	135	33	framework	framework	NOUN
fcis-9422	135	34	bert4torch	bert4torch	NOUN
fcis-9422	135	35	programming	programming	NOUN
fcis-9422	135	36	language	language	NOUN
fcis-9422	135	37	python3.7	python3.7	NOUN
fcis-9422	135	38	compilation	compilation	NOUN
fcis-9422	135	39	platform	platform	NOUN
fcis-9422	135	40	vscode	vscode	NOUN
fcis-9422	135	41	the	the	DET
fcis-9422	135	42	detailed	detailed	ADJ
fcis-9422	135	43	parameters	parameter	NOUN
fcis-9422	135	44	of	of	ADP
fcis-9422	135	45	the	the	DET
fcis-9422	135	46	proposed	propose	VERB
fcis-9422	135	47	entity	entity	NOUN
fcis-9422	135	48	linking	link	VERB
fcis-9422	135	49	model	model	NOUN
fcis-9422	135	50	in	in	ADP
fcis-9422	135	51	this	this	DET
fcis-9422	135	52	paper	paper	NOUN
fcis-9422	135	53	are	be	AUX
fcis-9422	135	54	as	as	SCONJ
fcis-9422	135	55	follows	follow	VERB
fcis-9422	135	56	:	:	PUNCT
fcis-9422	135	57	the	the	DET
fcis-9422	135	58	learning	learning	NOUN
fcis-9422	135	59	rate	rate	NOUN
fcis-9422	135	60	is	be	AUX
fcis-9422	135	61	2×10−5	2×10−5	NUM
fcis-9422	135	62	,	,	PUNCT
fcis-9422	135	63	the	the	DET
fcis-9422	135	64	maximum	maximum	ADJ
fcis-9422	135	65	sequence	sequence	NOUN
fcis-9422	135	66	length	length	NOUN
fcis-9422	135	67	is	be	AUX
fcis-9422	135	68	512	512	NUM
fcis-9422	135	69	,	,	PUNCT
fcis-9422	135	70	the	the	DET
fcis-9422	135	71	training	training	NOUN
fcis-9422	135	72	batch	batch	NOUN
fcis-9422	135	73	size	size	NOUN
fcis-9422	135	74	is	be	AUX
fcis-9422	135	75	8	8	NUM
fcis-9422	135	76	,	,	PUNCT
fcis-9422	135	77	and	and	CCONJ
fcis-9422	135	78	the	the	DET
fcis-9422	135	79	adam	adam	PROPN
fcis-9422	135	80	optimization	optimization	NOUN
fcis-9422	135	81	algorithm	algorithm	NOUN
fcis-9422	135	82	is	be	AUX
fcis-9422	135	83	used	use	VERB
fcis-9422	135	84	.	.	PUNCT
fcis-9422	136	1	3.4	3.4	NUM
fcis-9422	136	2	.	.	PUNCT
fcis-9422	136	3	result	result	VERB
fcis-9422	136	4	analysis	analysis	NOUN
fcis-9422	136	5	to	to	PART
fcis-9422	136	6	demonstrate	demonstrate	VERB
fcis-9422	136	7	the	the	DET
fcis-9422	136	8	superior	superior	ADJ
fcis-9422	136	9	performance	performance	NOUN
fcis-9422	136	10	of	of	ADP
fcis-9422	136	11	the	the	DET
fcis-9422	136	12	roformersim	roformersim	ADJ
fcis-9422	136	13	model	model	NOUN
fcis-9422	136	14	,	,	PUNCT
fcis-9422	136	15	experiments	experiment	NOUN
fcis-9422	136	16	were	be	AUX
fcis-9422	136	17	conducted	conduct	VERB
fcis-9422	136	18	using	use	VERB
fcis-9422	136	19	different	different	ADJ
fcis-9422	136	20	pretrained	pretraine	VERB
fcis-9422	136	21	models	model	NOUN
fcis-9422	136	22	on	on	ADP
fcis-9422	136	23	the	the	DET
fcis-9422	136	24	same	same	ADJ
fcis-9422	136	25	dataset	dataset	NOUN
fcis-9422	136	26	and	and	CCONJ
fcis-9422	136	27	their	their	PRON
fcis-9422	136	28	effects	effect	NOUN
fcis-9422	136	29	were	be	AUX
fcis-9422	136	30	compared	compare	VERB
fcis-9422	136	31	.	.	PUNCT
fcis-9422	137	1	specifically	specifically	ADV
fcis-9422	137	2	,	,	PUNCT
fcis-9422	137	3	the	the	DET
fcis-9422	137	4	semantic	semantic	ADJ
fcis-9422	137	5	information	information	NOUN
fcis-9422	137	6	of	of	ADP
fcis-9422	137	7	the	the	DET
fcis-9422	137	8	text	text	NOUN
fcis-9422	137	9	was	be	AUX
fcis-9422	137	10	first	first	ADV
fcis-9422	137	11	extracted	extract	VERB
fcis-9422	137	12	using	use	VERB
fcis-9422	137	13	a	a	DET
fcis-9422	137	14	pre	pre	ADJ
fcis-9422	137	15	-	-	ADJ
fcis-9422	137	16	trained	train	VERB
fcis-9422	137	17	model	model	NOUN
fcis-9422	137	18	,	,	PUNCT
fcis-9422	137	19	and	and	CCONJ
fcis-9422	137	20	then	then	ADV
fcis-9422	137	21	the	the	DET
fcis-9422	137	22	semantic	semantic	ADJ
fcis-9422	137	23	information	information	NOUN
fcis-9422	137	24	was	be	AUX
fcis-9422	137	25	fed	feed	VERB
fcis-9422	137	26	into	into	ADP
fcis-9422	137	27	a	a	DET
fcis-9422	137	28	fully	fully	ADV
fcis-9422	137	29	connected	connect	VERB
fcis-9422	137	30	layer	layer	NOUN
fcis-9422	137	31	for	for	ADP
fcis-9422	137	32	prediction	prediction	NOUN
fcis-9422	137	33	.	.	PUNCT
fcis-9422	138	1	the	the	DET
fcis-9422	138	2	batch_size	batch_size	NOUN
fcis-9422	138	3	of	of	ADP
fcis-9422	138	4	the	the	DET
fcis-9422	138	5	model	model	NOUN
fcis-9422	138	6	training	training	NOUN
fcis-9422	138	7	was	be	AUX
fcis-9422	138	8	set	set	VERB
fcis-9422	138	9	to	to	ADP
fcis-9422	138	10	8	8	NUM
fcis-9422	138	11	,	,	PUNCT
fcis-9422	138	12	the	the	DET
fcis-9422	138	13	initial	initial	ADJ
fcis-9422	138	14	learning	learning	NOUN
fcis-9422	138	15	rate	rate	NOUN
fcis-9422	138	16	was	be	AUX
fcis-9422	138	17	2e-5	2e-5	PROPN
fcis-9422	138	18	,	,	PUNCT
fcis-9422	138	19	the	the	DET
fcis-9422	138	20	maximum	maximum	ADJ
fcis-9422	138	21	sequence	sequence	NOUN
fcis-9422	138	22	length	length	NOUN
fcis-9422	138	23	was	be	AUX
fcis-9422	138	24	512	512	NUM
fcis-9422	138	25	,	,	PUNCT
fcis-9422	138	26	and	and	CCONJ
fcis-9422	138	27	a	a	DET
fcis-9422	138	28	learning	learning	NOUN
fcis-9422	138	29	rate	rate	NOUN
fcis-9422	138	30	decay	decay	NOUN
fcis-9422	138	31	strategy	strategy	NOUN
fcis-9422	138	32	based	base	VERB
fcis-9422	138	33	on	on	ADP
fcis-9422	138	34	exponential	exponential	ADJ
fcis-9422	138	35	decay	decay	NOUN
fcis-9422	138	36	was	be	AUX
fcis-9422	138	37	used	use	VERB
fcis-9422	138	38	.	.	PUNCT
fcis-9422	139	1	the	the	DET
fcis-9422	139	2	f1	f1	PROPN
fcis-9422	139	3	values	value	NOUN
fcis-9422	139	4	of	of	ADP
fcis-9422	139	5	the	the	DET
fcis-9422	139	6	models	model	NOUN
fcis-9422	139	7	obtained	obtain	VERB
fcis-9422	139	8	using	use	VERB
fcis-9422	139	9	different	different	ADJ
fcis-9422	139	10	pre	pre	ADJ
fcis-9422	139	11	-	-	ADJ
fcis-9422	139	12	trained	train	VERB
fcis-9422	139	13	models	model	NOUN
fcis-9422	139	14	are	be	AUX
fcis-9422	139	15	shown	show	VERB
fcis-9422	139	16	in	in	ADP
fcis-9422	139	17	table	table	NOUN
fcis-9422	139	18	3	3	NUM
fcis-9422	139	19	.	.	PUNCT
fcis-9422	139	20	table	table	NOUN
fcis-9422	139	21	3	3	NUM
fcis-9422	139	22	.	.	PUNCT
fcis-9422	140	1	f1	f1	NOUN
fcis-9422	140	2	scores	score	NOUN
fcis-9422	140	3	using	use	VERB
fcis-9422	140	4	different	different	ADJ
fcis-9422	140	5	pre	pre	ADJ
fcis-9422	140	6	-	-	ADJ
fcis-9422	140	7	trained	train	VERB
fcis-9422	140	8	models	model	NOUN
fcis-9422	140	9	model	model	VERB
fcis-9422	140	10	f1	f1	PROPN
fcis-9422	140	11	bert	bert	PROPN
fcis-9422	140	12	0.7644	0.7644	NUM
fcis-9422	140	13	roberta	roberta	PROPN
fcis-9422	140	14	0.8139	0.8139	NUM
fcis-9422	140	15	ernie1.0	ernie1.0	PROPN
fcis-9422	140	16	0.8501	0.8501	NUM
fcis-9422	140	17	roformer	roformer	NOUN
fcis-9422	140	18	-	-	PUNCT
fcis-9422	140	19	sim	sim	NOUN
fcis-9422	140	20	0.8851	0.8851	NUM
fcis-9422	140	21	based	base	VERB
fcis-9422	140	22	on	on	ADP
fcis-9422	140	23	the	the	DET
fcis-9422	140	24	experimental	experimental	ADJ
fcis-9422	140	25	results	result	NOUN
fcis-9422	140	26	,	,	PUNCT
fcis-9422	140	27	the	the	DET
fcis-9422	140	28	model	model	NOUN
fcis-9422	140	29	performance	performance	NOUN
fcis-9422	140	30	is	be	AUX
fcis-9422	140	31	in	in	ADP
fcis-9422	140	32	the	the	DET
fcis-9422	140	33	order	order	NOUN
fcis-9422	140	34	of	of	ADP
fcis-9422	140	35	roformer	roformer	NOUN
fcis-9422	140	36	-	-	PUNCT
fcis-9422	140	37	sim	sim	NOUN
fcis-9422	140	38	>	>	X
fcis-9422	140	39	ernie1.0	ernie1.0	X
fcis-9422	140	40	>	>	X
fcis-9422	140	41	roberta	roberta	PROPN
fcis-9422	140	42	>	>	X
fcis-9422	140	43	bert	bert	PROPN
fcis-9422	140	44	.	.	PUNCT
fcis-9422	141	1	the	the	DET
fcis-9422	141	2	f1	f1	PROPN
fcis-9422	141	3	scores	score	NOUN
fcis-9422	141	4	of	of	ADP
fcis-9422	141	5	roformer	roformer	NOUN
fcis-9422	141	6	-	-	PUNCT
fcis-9422	141	7	sim	sim	NOUN
fcis-9422	141	8	model	model	NOUN
fcis-9422	141	9	are	be	AUX
fcis-9422	141	10	0.1207	0.1207	NUM
fcis-9422	141	11	,	,	PUNCT
fcis-9422	141	12	0.0712	0.0712	NUM
fcis-9422	141	13	,	,	PUNCT
fcis-9422	141	14	and	and	CCONJ
fcis-9422	141	15	0.035	0.035	NUM
fcis-9422	141	16	higher	high	ADJ
fcis-9422	141	17	than	than	ADP
fcis-9422	141	18	those	those	PRON
fcis-9422	141	19	of	of	ADP
fcis-9422	141	20	bert	bert	PROPN
fcis-9422	141	21	,	,	PUNCT
fcis-9422	141	22	roberta	roberta	PROPN
fcis-9422	141	23	,	,	PUNCT
fcis-9422	141	24	and	and	CCONJ
fcis-9422	141	25	ernie1.0	ernie1.0	PROPN
fcis-9422	141	26	models	model	NOUN
fcis-9422	141	27	,	,	PUNCT
fcis-9422	141	28	respectively	respectively	ADV
fcis-9422	141	29	.	.	PUNCT
fcis-9422	142	1	the	the	DET
fcis-9422	142	2	essence	essence	NOUN
fcis-9422	142	3	of	of	ADP
fcis-9422	142	4	the	the	DET
fcis-9422	142	5	entity	entity	NOUN
fcis-9422	142	6	disambiguation	disambiguation	NOUN
fcis-9422	142	7	task	task	NOUN
fcis-9422	142	8	is	be	AUX
fcis-9422	142	9	to	to	PART
fcis-9422	142	10	calculate	calculate	VERB
fcis-9422	142	11	text	text	NOUN
fcis-9422	142	12	similarity	similarity	NOUN
fcis-9422	142	13	,	,	PUNCT
fcis-9422	142	14	and	and	CCONJ
fcis-9422	142	15	roformer	roformer	NOUN
fcis-9422	142	16	-	-	PUNCT
fcis-9422	142	17	sim	sim	NOUN
fcis-9422	142	18	model	model	NOUN
fcis-9422	142	19	is	be	AUX
fcis-9422	142	20	trained	train	VERB
fcis-9422	142	21	on	on	ADP
fcis-9422	142	22	corpora	corpora	PROPN
fcis-9422	142	23	that	that	PRON
fcis-9422	142	24	include	include	VERB
fcis-9422	142	25	both	both	DET
fcis-9422	142	26	question	question	NOUN
fcis-9422	142	27	-	-	PUNCT
fcis-9422	142	28	type	type	NOUN
fcis-9422	142	29	similar	similar	ADJ
fcis-9422	142	30	sentences	sentence	NOUN
fcis-9422	142	31	and	and	CCONJ
fcis-9422	142	32	general	general	ADJ
fcis-9422	142	33	-	-	PUNCT
fcis-9422	142	34	type	type	NOUN
fcis-9422	142	35	similar	similar	ADJ
fcis-9422	142	36	50	50	NUM
fcis-9422	142	37	sentences	sentence	NOUN
fcis-9422	142	38	,	,	PUNCT
fcis-9422	142	39	which	which	PRON
fcis-9422	142	40	can	can	AUX
fcis-9422	142	41	provide	provide	VERB
fcis-9422	142	42	prior	prior	ADJ
fcis-9422	142	43	knowledge	knowledge	NOUN
fcis-9422	142	44	for	for	ADP
fcis-9422	142	45	entity	entity	NOUN
fcis-9422	142	46	disambiguation	disambiguation	NOUN
fcis-9422	142	47	tasks	task	NOUN
fcis-9422	142	48	.	.	PUNCT
fcis-9422	143	1	therefore	therefore	ADV
fcis-9422	143	2	,	,	PUNCT
fcis-9422	143	3	roformer	roformer	NOUN
fcis-9422	143	4	-	-	PUNCT
fcis-9422	143	5	sim	sim	NOUN
fcis-9422	143	6	can	can	AUX
fcis-9422	143	7	improve	improve	VERB
fcis-9422	143	8	the	the	DET
fcis-9422	143	9	performance	performance	NOUN
fcis-9422	143	10	of	of	ADP
fcis-9422	143	11	entity	entity	NOUN
fcis-9422	143	12	linking	link	VERB
fcis-9422	143	13	models	model	NOUN
fcis-9422	143	14	.	.	PUNCT
fcis-9422	144	1	4	4	X
fcis-9422	144	2	.	.	X
fcis-9422	144	3	conclusion	conclusion	NOUN
fcis-9422	144	4	this	this	DET
fcis-9422	144	5	paper	paper	NOUN
fcis-9422	144	6	aims	aim	VERB
fcis-9422	144	7	to	to	PART
fcis-9422	144	8	construct	construct	VERB
fcis-9422	144	9	an	an	DET
fcis-9422	144	10	entity	entity	NOUN
fcis-9422	144	11	linking	link	VERB
fcis-9422	144	12	model	model	NOUN
fcis-9422	144	13	based	base	VERB
fcis-9422	144	14	on	on	ADP
fcis-9422	144	15	the	the	DET
fcis-9422	144	16	roformer	roformer	NOUN
fcis-9422	144	17	-	-	PUNCT
fcis-9422	144	18	sim	sim	ADJ
fcis-9422	144	19	pre	pre	ADJ
fcis-9422	144	20	-	-	ADJ
fcis-9422	144	21	training	training	ADJ
fcis-9422	144	22	model	model	NOUN
fcis-9422	144	23	for	for	ADP
fcis-9422	144	24	chinese	chinese	ADJ
fcis-9422	144	25	short	short	ADJ
fcis-9422	144	26	texts	text	NOUN
fcis-9422	144	27	and	and	CCONJ
fcis-9422	144	28	verifies	verifie	NOUN
fcis-9422	144	29	it	it	PRON
fcis-9422	144	30	on	on	ADP
fcis-9422	144	31	the	the	DET
fcis-9422	144	32	dataset	dataset	NOUN
fcis-9422	144	33	provided	provide	VERB
fcis-9422	144	34	by	by	ADP
fcis-9422	144	35	the	the	DET
fcis-9422	144	36	ccks2020	ccks2020	PROPN
fcis-9422	144	37	(	(	PUNCT
fcis-9422	144	38	national	national	ADJ
fcis-9422	144	39	knowledge	knowledge	NOUN
fcis-9422	144	40	graph	graph	NOUN
fcis-9422	144	41	and	and	CCONJ
fcis-9422	144	42	semantic	semantic	ADJ
fcis-9422	144	43	computing	computing	NOUN
fcis-9422	144	44	conference	conference	NOUN
fcis-9422	144	45	2020	2020	NUM
fcis-9422	144	46	)	)	PUNCT
fcis-9422	144	47	short	short	ADJ
fcis-9422	144	48	text	text	NOUN
fcis-9422	144	49	entity	entity	NOUN
fcis-9422	144	50	linking	link	VERB
fcis-9422	144	51	evaluation	evaluation	NOUN
fcis-9422	144	52	.	.	PUNCT
fcis-9422	145	1	the	the	DET
fcis-9422	145	2	experimental	experimental	ADJ
fcis-9422	145	3	results	result	NOUN
fcis-9422	145	4	show	show	VERB
fcis-9422	145	5	that	that	SCONJ
fcis-9422	145	6	the	the	DET
fcis-9422	145	7	roformer	roformer	NOUN
fcis-9422	145	8	-	-	PUNCT
fcis-9422	145	9	sim	sim	NOUN
fcis-9422	145	10	pretraining	pretraine	VERB
fcis-9422	145	11	model	model	NOUN
fcis-9422	145	12	can	can	AUX
fcis-9422	145	13	provide	provide	VERB
fcis-9422	145	14	prior	prior	ADJ
fcis-9422	145	15	knowledge	knowledge	NOUN
fcis-9422	145	16	in	in	ADP
fcis-9422	145	17	the	the	DET
fcis-9422	145	18	process	process	NOUN
fcis-9422	145	19	of	of	ADP
fcis-9422	145	20	entity	entity	NOUN
fcis-9422	145	21	disambiguation	disambiguation	NOUN
fcis-9422	145	22	,	,	PUNCT
fcis-9422	145	23	which	which	PRON
fcis-9422	145	24	improves	improve	VERB
fcis-9422	145	25	the	the	DET
fcis-9422	145	26	linking	link	VERB
fcis-9422	145	27	effect	effect	NOUN
fcis-9422	145	28	.	.	PUNCT
fcis-9422	146	1	in	in	ADP
fcis-9422	146	2	the	the	DET
fcis-9422	146	3	future	future	NOUN
fcis-9422	146	4	,	,	PUNCT
fcis-9422	146	5	the	the	DET
fcis-9422	146	6	effectiveness	effectiveness	NOUN
fcis-9422	146	7	of	of	ADP
fcis-9422	146	8	the	the	DET
fcis-9422	146	9	proposed	propose	VERB
fcis-9422	146	10	model	model	NOUN
fcis-9422	146	11	will	will	AUX
fcis-9422	146	12	be	be	AUX
fcis-9422	146	13	verified	verify	VERB
fcis-9422	146	14	on	on	ADP
fcis-9422	146	15	other	other	ADJ
fcis-9422	146	16	public	public	ADJ
fcis-9422	146	17	entity	entity	NOUN
fcis-9422	146	18	linking	link	VERB
fcis-9422	146	19	datasets	dataset	NOUN
fcis-9422	146	20	,	,	PUNCT
fcis-9422	146	21	and	and	CCONJ
fcis-9422	146	22	the	the	DET
fcis-9422	146	23	methods	method	NOUN
fcis-9422	146	24	for	for	ADP
fcis-9422	146	25	entity	entity	NOUN
fcis-9422	146	26	recognition	recognition	NOUN
fcis-9422	146	27	and	and	CCONJ
fcis-9422	146	28	entity	entity	NOUN
fcis-9422	146	29	disambiguation	disambiguation	NOUN
fcis-9422	146	30	will	will	AUX
fcis-9422	146	31	be	be	AUX
fcis-9422	146	32	optimized	optimize	VERB
fcis-9422	146	33	to	to	PART
fcis-9422	146	34	further	far	ADV
fcis-9422	146	35	improve	improve	VERB
fcis-9422	146	36	the	the	DET
fcis-9422	146	37	accuracy	accuracy	NOUN
fcis-9422	146	38	of	of	ADP
fcis-9422	146	39	chinese	chinese	ADJ
fcis-9422	146	40	short	short	ADJ
fcis-9422	146	41	text	text	NOUN
fcis-9422	146	42	entity	entity	NOUN
fcis-9422	146	43	linking	linking	NOUN
fcis-9422	146	44	.	.	PUNCT
fcis-9422	147	1	acknowledgments	acknowledgment	NOUN
fcis-9422	147	2	this	this	DET
fcis-9422	147	3	project	project	NOUN
fcis-9422	147	4	is	be	AUX
fcis-9422	147	5	supported	support	VERB
fcis-9422	147	6	by	by	ADP
fcis-9422	147	7	the	the	DET
fcis-9422	147	8	southwest	southwest	PROPN
fcis-9422	147	9	university	university	PROPN
fcis-9422	147	10	for	for	ADP
fcis-9422	147	11	nationalities	nationality	NOUN
fcis-9422	147	12	graduate	graduate	VERB
fcis-9422	147	13	innovative	innovative	ADJ
fcis-9422	147	14	scientific	scientific	ADJ
fcis-9422	147	15	research	research	NOUN
fcis-9422	147	16	project	project	NOUN
fcis-9422	147	17	(	(	PUNCT
fcis-9422	147	18	project	project	NOUN
fcis-9422	147	19	no	no	INTJ
fcis-9422	147	20	.	.	PUNCT
fcis-9422	148	1	yb2023145	yb2023145	PROPN
fcis-9422	148	2	)	)	PUNCT
fcis-9422	148	3	.	.	PUNCT
fcis-9422	149	1	references	reference	NOUN
fcis-9422	149	2	[	[	X
fcis-9422	149	3	1	1	NUM
fcis-9422	149	4	]	]	X
fcis-9422	149	5	zhan	zhan	PROPN
fcis-9422	149	6	fei	fei	PROPN
fcis-9422	149	7	,	,	PUNCT
fcis-9422	149	8	zhu	zhu	PROPN
fcis-9422	149	9	yanhui	yanhui	PROPN
fcis-9422	149	10	,	,	PUNCT
fcis-9422	149	11	liang	liang	PROPN
fcis-9422	149	12	wentong	wentong	PROPN
fcis-9422	149	13	,	,	PUNCT
fcis-9422	149	14	et	et	PROPN
fcis-9422	149	15	al	al	PROPN
fcis-9422	149	16	.	.	PUNCT
fcis-9422	150	1	multi	multi	ADJ
fcis-9422	150	2	-	-	ADJ
fcis-9422	150	3	task	task	ADJ
fcis-9422	150	4	learning	learning	NOUN
fcis-9422	150	5	-	-	PUNCT
fcis-9422	150	6	based	base	VERB
fcis-9422	150	7	short	short	ADJ
fcis-9422	150	8	text	text	NOUN
fcis-9422	150	9	entity	entity	NOUN
fcis-9422	150	10	linking	link	VERB
fcis-9422	150	11	method	method	NOUN
fcis-9422	150	12	.	.	PUNCT
fcis-9422	151	1	computer	computer	NOUN
fcis-9422	151	2	engineering	engineering	NOUN
fcis-9422	151	3	,	,	PUNCT
fcis-9422	151	4	2022	2022	NUM
fcis-9422	151	5	,	,	PUNCT
fcis-9422	151	6	48(3	48(3	NUM
fcis-9422	151	7	):	):	PUNCT
fcis-9422	151	8	315	315	NUM
fcis-9422	151	9	-	-	SYM
fcis-9422	151	10	320	320	NUM
fcis-9422	151	11	.	.	PUNCT
fcis-9422	152	1	[	[	X
fcis-9422	152	2	2	2	NUM
fcis-9422	152	3	]	]	X
fcis-9422	152	4	zhang	zhang	PROPN
fcis-9422	152	5	shengqi	shengqi	PROPN
fcis-9422	152	6	,	,	PUNCT
fcis-9422	152	7	wang	wang	PROPN
fcis-9422	152	8	yuanlong	yuanlong	PROPN
fcis-9422	152	9	,	,	PUNCT
fcis-9422	152	10	li	li	PROPN
fcis-9422	152	11	ru	ru	PROPN
fcis-9422	152	12	,	,	PUNCT
fcis-9422	152	13	et	et	PROPN
fcis-9422	152	14	al	al	PROPN
fcis-9422	152	15	.	.	PUNCT
fcis-9422	153	1	chinese	chinese	ADJ
fcis-9422	153	2	short	short	ADJ
fcis-9422	153	3	text	text	NOUN
fcis-9422	153	4	entity	entity	NOUN
fcis-9422	153	5	linking	link	VERB
fcis-9422	153	6	based	base	VERB
fcis-9422	153	7	on	on	ADP
fcis-9422	153	8	local	local	ADJ
fcis-9422	153	9	attention	attention	NOUN
fcis-9422	153	10	mechanism	mechanism	NOUN
fcis-9422	154	1	[	[	X
fcis-9422	154	2	j	j	X
fcis-9422	154	3	]	]	X
fcis-9422	154	4	.	.	PUNCT
fcis-9422	155	1	computer	computer	NOUN
fcis-9422	155	2	engineering	engineering	NOUN
fcis-9422	155	3	,	,	PUNCT
fcis-9422	155	4	2021	2021	NUM
fcis-9422	155	5	,	,	PUNCT
fcis-9422	155	6	47(11	47(11	NUM
fcis-9422	155	7	):	):	PUNCT
fcis-9422	155	8	77	77	NUM
fcis-9422	155	9	-	-	SYM
fcis-9422	155	10	83	83	NUM
fcis-9422	155	11	,	,	PUNCT
fcis-9422	155	12	92	92	NUM
fcis-9422	155	13	.	.	PUNCT
fcis-9422	156	1	[	[	X
fcis-9422	156	2	3	3	NUM
fcis-9422	156	3	]	]	X
fcis-9422	156	4	lample	lample	NOUN
fcis-9422	156	5	g	g	PROPN
fcis-9422	156	6	,	,	PUNCT
fcis-9422	156	7	ballesteros	ballesteros	PROPN
fcis-9422	156	8	m	m	PROPN
fcis-9422	156	9	,	,	PUNCT
fcis-9422	156	10	subramanian	subramanian	PROPN
fcis-9422	156	11	s	s	PROPN
fcis-9422	156	12	,	,	PUNCT
fcis-9422	156	13	et	et	PROPN
fcis-9422	156	14	al	al	PROPN
fcis-9422	156	15	.	.	PROPN
fcis-9422	157	1	neural	neural	PROPN
fcis-9422	157	2	architectures	architecture	NOUN
fcis-9422	157	3	for	for	ADP
fcis-9422	157	4	named	name	VERB
fcis-9422	157	5	entity	entity	NOUN
fcis-9422	157	6	recognition[c	recognition[c	NOUN
fcis-9422	157	7	]	]	PUNCT
fcis-9422	157	8	.	.	PUNCT
fcis-9422	158	1	north	north	NOUN
fcis-9422	158	2	american	american	ADJ
fcis-9422	158	3	chapter	chapter	NOUN
fcis-9422	158	4	of	of	ADP
fcis-9422	158	5	the	the	DET
fcis-9422	158	6	association	association	NOUN
fcis-9422	158	7	for	for	ADP
fcis-9422	158	8	computational	computational	ADJ
fcis-9422	158	9	linguistics	linguistic	NOUN
fcis-9422	158	10	,	,	PUNCT
fcis-9422	158	11	2016	2016	NUM
fcis-9422	158	12	:	:	PUNCT
fcis-9422	158	13	260	260	NUM
fcis-9422	158	14	-	-	SYM
fcis-9422	158	15	270	270	NUM
fcis-9422	158	16	.	.	PUNCT
fcis-9422	159	1	[	[	X
fcis-9422	159	2	4	4	X
fcis-9422	159	3	]	]	X
fcis-9422	159	4	peters	peters	PROPN
fcis-9422	159	5	m	m	PROPN
fcis-9422	159	6	e	e	PROPN
fcis-9422	159	7	,	,	PUNCT
fcis-9422	159	8	neumann	neumann	PROPN
fcis-9422	159	9	m	m	PROPN
fcis-9422	159	10	,	,	PUNCT
fcis-9422	159	11	iyyer	iyyer	PROPN
fcis-9422	159	12	m	m	PROPN
fcis-9422	159	13	,	,	PUNCT
fcis-9422	159	14	et	et	PROPN
fcis-9422	159	15	al	al	PROPN
fcis-9422	159	16	.	.	PUNCT
fcis-9422	159	17	deep	deep	ADJ
fcis-9422	159	18	contextualized	contextualized	ADJ
fcis-9422	159	19	word	word	NOUN
fcis-9422	159	20	representations	representation	NOUN
fcis-9422	159	21	.	.	PUNCT
fcis-9422	160	1	naacl	naacl	PROPN
fcis-9422	160	2	-	-	PUNCT
fcis-9422	160	3	hlt[j	hlt[j	X
fcis-9422	160	4	]	]	PUNCT
fcis-9422	160	5	.	.	PUNCT
fcis-9422	161	1	2018	2018	NUM
fcis-9422	161	2	.	.	PUNCT
fcis-9422	162	1	[	[	X
fcis-9422	162	2	j	j	X
fcis-9422	162	3	]	]	X
fcis-9422	162	4	.	.	PUNCT
fcis-9422	163	1	ar	ar	PROPN
fcis-9422	163	2	xiv	xiv	PROPN
fcis-9422	163	3	preprint	preprint	PROPN
fcis-9422	163	4	ar	ar	PROPN
fcis-9422	163	5	xiv:1802.05365,2018	xiv:1802.05365,2018	PROPN
fcis-9422	163	6	.	.	PUNCT
fcis-9422	164	1	[	[	X
fcis-9422	164	2	5	5	NUM
fcis-9422	164	3	]	]	SYM
fcis-9422	164	4	lecun	lecun	ADJ
fcis-9422	164	5	,	,	PUNCT
fcis-9422	164	6	yann	yann	PROPN
fcis-9422	164	7	,	,	PUNCT
fcis-9422	164	8	et	et	PROPN
fcis-9422	164	9	al	al	PROPN
fcis-9422	164	10	.	.	PUNCT
fcis-9422	165	1	"	"	PUNCT
fcis-9422	165	2	gradient	gradient	NOUN
fcis-9422	165	3	-	-	PUNCT
fcis-9422	165	4	based	base	VERB
fcis-9422	165	5	learning	learning	NOUN
fcis-9422	165	6	applied	apply	VERB
fcis-9422	165	7	to	to	ADP
fcis-9422	165	8	document	document	NOUN
fcis-9422	165	9	recognition	recognition	NOUN
fcis-9422	165	10	.	.	PUNCT
fcis-9422	165	11	"	"	PUNCT
fcis-9422	166	1	proceedings	proceeding	NOUN
fcis-9422	166	2	of	of	ADP
fcis-9422	166	3	the	the	DET
fcis-9422	166	4	ieee	ieee	NOUN
fcis-9422	166	5	86.11	86.11	NUM
fcis-9422	166	6	(	(	PUNCT
fcis-9422	166	7	1998	1998	NUM
fcis-9422	166	8	):	):	PUNCT
fcis-9422	166	9	2278	2278	NUM
fcis-9422	166	10	-	-	SYM
fcis-9422	166	11	2324	2324	NUM
fcis-9422	166	12	.	.	PUNCT
fcis-9422	167	1	[	[	X
fcis-9422	167	2	6	6	NUM
fcis-9422	167	3	]	]	X
fcis-9422	167	4	lipton	lipton	PROPN
fcis-9422	167	5	,	,	PUNCT
fcis-9422	167	6	zachary	zachary	PROPN
fcis-9422	167	7	c.	c.	PROPN
fcis-9422	167	8	,	,	PUNCT
fcis-9422	167	9	john	john	PROPN
fcis-9422	167	10	berkowitz	berkowitz	PROPN
fcis-9422	167	11	,	,	PUNCT
fcis-9422	167	12	and	and	CCONJ
fcis-9422	167	13	charles	charles	PROPN
fcis-9422	167	14	elkan	elkan	PROPN
fcis-9422	167	15	.	.	PUNCT
fcis-9422	168	1	"	"	PUNCT
fcis-9422	168	2	a	a	DET
fcis-9422	168	3	critical	critical	ADJ
fcis-9422	168	4	review	review	NOUN
fcis-9422	168	5	of	of	ADP
fcis-9422	168	6	recurrent	recurrent	ADJ
fcis-9422	168	7	neural	neural	ADJ
fcis-9422	168	8	networks	network	NOUN
fcis-9422	168	9	for	for	ADP
fcis-9422	168	10	sequence	sequence	NOUN
fcis-9422	168	11	learning	learning	NOUN
fcis-9422	168	12	.	.	PUNCT
fcis-9422	168	13	"	"	PUNCT
fcis-9422	169	1	arxiv	arxiv	PROPN
fcis-9422	169	2	preprint	preprint	VERB
fcis-9422	169	3	arxiv:1506.00019	arxiv:1506.00019	NOUN
fcis-9422	169	4	(	(	PUNCT
fcis-9422	169	5	2015	2015	NUM
fcis-9422	169	6	)	)	PUNCT
fcis-9422	169	7	.	.	PUNCT
fcis-9422	170	1	[	[	X
fcis-9422	170	2	7	7	X
fcis-9422	170	3	]	]	X
fcis-9422	170	4	devlin	devlin	PROPN
fcis-9422	170	5	j	j	PROPN
fcis-9422	170	6	,	,	PUNCT
fcis-9422	170	7	chang	chang	PROPN
fcis-9422	170	8	m	m	PROPN
fcis-9422	170	9	w	w	PROPN
fcis-9422	170	10	,	,	PUNCT
fcis-9422	170	11	leek	leek	ADJ
fcis-9422	170	12	,	,	PUNCT
fcis-9422	170	13	et	et	PROPN
fcis-9422	170	14	al	al	PROPN
fcis-9422	170	15	.	.	PUNCT
fcis-9422	170	16	bert	bert	PROPN
fcis-9422	170	17	:	:	PUNCT
fcis-9422	170	18	pretraining	pretraine	VERB
fcis-9422	170	19	of	of	ADP
fcis-9422	170	20	deep	deep	ADJ
fcis-9422	170	21	bidirectional	bidirectional	ADJ
fcis-9422	170	22	transformers	transformer	NOUN
fcis-9422	170	23	for	for	ADP
fcis-9422	170	24	language	language	NOUN
fcis-9422	170	25	understanding	understanding	NOUN
fcis-9422	170	26	[	[	X
fcis-9422	170	27	j	j	X
fcis-9422	170	28	]	]	X
fcis-9422	170	29	.	.	PUNCT
fcis-9422	171	1	arxiv	arxiv	PROPN
fcis-9422	171	2	preprint	preprint	PROPN
fcis-9422	171	3	,	,	PUNCT
fcis-9422	171	4	arxiv	arxiv	PROPN
fcis-9422	171	5	:	:	PUNCT
fcis-9422	171	6	1810	1810	NUM
fcis-9422	171	7	.	.	PUNCT
fcis-9422	172	1	04805,2018	04805,2018	NOUN
fcis-9422	172	2	.	.	PUNCT
fcis-9422	173	1	[	[	X
fcis-9422	173	2	8	8	NUM
fcis-9422	173	3	]	]	X
fcis-9422	173	4	ilić	ilić	PROPN
fcis-9422	173	5	,	,	PUNCT
fcis-9422	173	6	suzana	suzana	PROPN
fcis-9422	173	7	,	,	PUNCT
fcis-9422	173	8	et	et	PROPN
fcis-9422	173	9	al	al	PROPN
fcis-9422	173	10	.	.	PUNCT
fcis-9422	174	1	"	"	PUNCT
fcis-9422	174	2	deep	deep	ADJ
fcis-9422	174	3	contextualized	contextualized	ADJ
fcis-9422	174	4	word	word	NOUN
fcis-9422	174	5	representations	representation	NOUN
fcis-9422	174	6	for	for	ADP
fcis-9422	174	7	detecting	detect	VERB
fcis-9422	174	8	sarcasm	sarcasm	NOUN
fcis-9422	174	9	and	and	CCONJ
fcis-9422	174	10	irony	irony	NOUN
fcis-9422	174	11	.	.	PUNCT
fcis-9422	174	12	"	"	PUNCT
fcis-9422	175	1	arxiv	arxiv	PROPN
fcis-9422	175	2	preprint	preprint	NOUN
fcis-9422	175	3	arxiv	arxiv	PROPN
fcis-9422	175	4	:	:	PUNCT
fcis-9422	175	5	1809	1809	NUM
fcis-9422	175	6	.	.	PUNCT
fcis-9422	176	1	09795	09795	NUM
fcis-9422	176	2	(	(	PUNCT
fcis-9422	176	3	2018	2018	NUM
fcis-9422	176	4	)	)	PUNCT
fcis-9422	176	5	.	.	PUNCT
fcis-9422	177	1	[	[	X
fcis-9422	177	2	9	9	NUM
fcis-9422	177	3	]	]	X
fcis-9422	177	4	zhou	zhou	PROPN
fcis-9422	177	5	pengcheng	pengcheng	PROPN
fcis-9422	177	6	,	,	PUNCT
fcis-9422	177	7	wu	wu	PROPN
fcis-9422	177	8	chuan	chuan	PROPN
fcis-9422	177	9	,	,	PUNCT
fcis-9422	177	10	lu	lu	PROPN
fcis-9422	177	11	wei	wei	PROPN
fcis-9422	177	12	.	.	PUNCT
fcis-9422	178	1	short	short	ADJ
fcis-9422	178	2	text	text	NOUN
fcis-9422	178	3	entity	entity	NOUN
fcis-9422	178	4	linking	link	VERB
fcis-9422	178	5	based	base	VERB
fcis-9422	178	6	on	on	ADP
fcis-9422	178	7	multiple	multiple	ADJ
fcis-9422	178	8	knowledge	knowledge	NOUN
fcis-9422	178	9	bases	basis	NOUN
fcis-9422	178	10	:	:	PUNCT
fcis-9422	178	11	a	a	DET
fcis-9422	178	12	case	case	NOUN
fcis-9422	178	13	study	study	NOUN
fcis-9422	178	14	of	of	ADP
fcis-9422	178	15	wikipedia	wikipedia	PROPN
fcis-9422	178	16	and	and	CCONJ
fcis-9422	178	17	freebase[j	freebase[j	PROPN
fcis-9422	178	18	]	]	PUNCT
fcis-9422	178	19	.	.	PUNCT
fcis-9422	179	1	new	new	ADJ
fcis-9422	179	2	technology	technology	NOUN
fcis-9422	179	3	of	of	ADP
fcis-9422	179	4	library	library	NOUN
fcis-9422	179	5	and	and	CCONJ
fcis-9422	179	6	information	information	NOUN
fcis-9422	179	7	service	service	NOUN
fcis-9422	179	8	,	,	PUNCT
fcis-9422	179	9	2016	2016	NUM
fcis-9422	179	10	,	,	PUNCT
fcis-9422	179	11	32(06	32(06	NUM
fcis-9422	179	12	):	):	PUNCT
fcis-9422	179	13	1	1	NUM
fcis-9422	179	14	-	-	SYM
fcis-9422	179	15	11	11	NUM
fcis-9422	179	16	.	.	PUNCT
fcis-9422	180	1	[	[	X
fcis-9422	180	2	10	10	NUM
fcis-9422	180	3	]	]	X
fcis-9422	180	4	zeng	zeng	PROPN
fcis-9422	180	5	w	w	PROPN
fcis-9422	180	6	,	,	PUNCT
fcis-9422	180	7	tang	tang	PROPN
fcis-9422	180	8	j	j	PROPN
fcis-9422	180	9	,	,	PUNCT
fcis-9422	180	10	zhao	zhao	PROPN
fcis-9422	180	11	x.	x.	NOUN
fcis-9422	180	12	entity	entity	NOUN
fcis-9422	180	13	linking	link	VERB
fcis-9422	180	14	on	on	ADP
fcis-9422	180	15	chinese	chinese	ADJ
fcis-9422	180	16	microblogs	microblog	NOUN
fcis-9422	180	17	via	via	ADP
fcis-9422	180	18	deep	deep	ADJ
fcis-9422	180	19	neural	neural	ADJ
fcis-9422	180	20	network[j	network[j	PROPN
fcis-9422	180	21	]	]	PUNCT
fcis-9422	180	22	.	.	PUNCT
fcis-9422	181	1	ieee	ieee	NOUN
fcis-9422	181	2	access	access	NOUN
fcis-9422	181	3	,	,	PUNCT
fcis-9422	181	4	2018	2018	NUM
fcis-9422	181	5	,	,	PUNCT
fcis-9422	181	6	6	6	NUM
fcis-9422	181	7	:	:	SYM
fcis-9422	181	8	2590825920	2590825920	NUM
fcis-9422	181	9	.	.	PUNCT
fcis-9422	182	1	[	[	X
fcis-9422	182	2	11	11	NUM
fcis-9422	182	3	]	]	SYM
fcis-9422	182	4	hu	hu	PROPN
fcis-9422	182	5	s	s	PROPN
fcis-9422	182	6	,	,	PUNCT
fcis-9422	182	7	tan	tan	PROPN
fcis-9422	182	8	z	z	PROPN
fcis-9422	182	9	,	,	PUNCT
fcis-9422	182	10	zeng	zeng	PROPN
fcis-9422	182	11	w	w	PROPN
fcis-9422	182	12	,	,	PUNCT
fcis-9422	182	13	et	et	PROPN
fcis-9422	182	14	al	al	PROPN
fcis-9422	182	15	.	.	PUNCT
fcis-9422	182	16	entity	entity	NOUN
fcis-9422	182	17	linking	link	VERB
fcis-9422	182	18	via	via	ADP
fcis-9422	182	19	symmetrical	symmetrical	ADJ
fcis-9422	182	20	attention	attention	NOUN
fcis-9422	182	21	-	-	PUNCT
fcis-9422	182	22	based	base	VERB
fcis-9422	182	23	neural	neural	ADJ
fcis-9422	182	24	network	network	NOUN
fcis-9422	182	25	and	and	CCONJ
fcis-9422	182	26	entity	entity	NOUN
fcis-9422	182	27	structural	structural	ADJ
fcis-9422	182	28	features[j	features[j	NOUN
fcis-9422	182	29	]	]	PUNCT
fcis-9422	182	30	.	.	PUNCT
fcis-9422	183	1	symmetry	symmetry	PROPN
fcis-9422	183	2	,	,	PUNCT
fcis-9422	183	3	2019	2019	NUM
fcis-9422	183	4	,	,	PUNCT
fcis-9422	183	5	11(4	11(4	NUM
fcis-9422	183	6	):	):	PUNCT
fcis-9422	183	7	453	453	NUM
fcis-9422	183	8	.	.	PUNCT
fcis-9422	184	1	[	[	X
fcis-9422	184	2	12	12	NUM
fcis-9422	184	3	]	]	X
fcis-9422	184	4	devlin	devlin	PROPN
fcis-9422	184	5	j	j	PROPN
fcis-9422	184	6	,	,	PUNCT
fcis-9422	184	7	chang	chang	PROPN
fcis-9422	184	8	m	m	PROPN
fcis-9422	184	9	w	w	PROPN
fcis-9422	184	10	,	,	PUNCT
fcis-9422	184	11	lee	lee	PROPN
fcis-9422	184	12	k	k	PROPN
fcis-9422	184	13	,	,	PUNCT
fcis-9422	184	14	et	et	PROPN
fcis-9422	184	15	al	al	PROPN
fcis-9422	184	16	.	.	PUNCT
fcis-9422	184	17	bert	bert	PROPN
fcis-9422	184	18	:	:	PUNCT
fcis-9422	184	19	pre	pre	ADJ
fcis-9422	184	20	-	-	NOUN
fcis-9422	184	21	training	training	NOUN
fcis-9422	184	22	of	of	ADP
fcis-9422	184	23	deep	deep	ADJ
fcis-9422	184	24	bidirectional	bidirectional	ADJ
fcis-9422	184	25	transformers	transformer	NOUN
fcis-9422	184	26	for	for	ADP
fcis-9422	184	27	language	language	NOUN
fcis-9422	184	28	understanding[j	understanding[j	NOUN
fcis-9422	184	29	]	]	PUNCT
fcis-9422	184	30	.	.	PUNCT
fcis-9422	185	1	ar	ar	PROPN
fcis-9422	185	2	xiv	xiv	PROPN
fcis-9422	185	3	preprint	preprint	PROPN
fcis-9422	185	4	ar	ar	PROPN
fcis-9422	185	5	xiv:1810.04805	xiv:1810.04805	PROPN
fcis-9422	185	6	,	,	PUNCT
fcis-9422	185	7	2018	2018	NUM
fcis-9422	185	8	.	.	PUNCT
fcis-9422	186	1	[	[	X
fcis-9422	186	2	13	13	NUM
fcis-9422	186	3	]	]	X
fcis-9422	186	4	chen	chen	PROPN
fcis-9422	186	5	s	s	PROPN
fcis-9422	186	6	,	,	PUNCT
fcis-9422	186	7	wang	wang	PROPN
fcis-9422	186	8	j	j	PROPN
fcis-9422	186	9	,	,	PUNCT
fcis-9422	186	10	jiang	jiang	PROPN
fcis-9422	186	11	f	f	PROPN
fcis-9422	186	12	,	,	PUNCT
fcis-9422	186	13	et	et	PROPN
fcis-9422	186	14	al	al	PROPN
fcis-9422	186	15	.	.	PUNCT
fcis-9422	187	1	improving	improve	VERB
fcis-9422	187	2	entity	entity	NOUN
fcis-9422	187	3	linking	link	VERB
fcis-9422	187	4	by	by	ADP
fcis-9422	187	5	modeling	model	VERB
fcis-9422	187	6	latent	latent	NOUN
fcis-9422	187	7	entity	entity	NOUN
fcis-9422	187	8	type	type	NOUN
fcis-9422	187	9	information[c]//proceedings	information[c]//proceeding	NOUN
fcis-9422	187	10	of	of	ADP
fcis-9422	187	11	the	the	DET
fcis-9422	187	12	advancement	advancement	NOUN
fcis-9422	187	13	of	of	ADP
fcis-9422	187	14	artificial	artificial	ADJ
fcis-9422	187	15	intelligence(aaai	intelligence(aaai	PROPN
fcis-9422	187	16	)	)	PUNCT
fcis-9422	187	17	.	.	PUNCT
fcis-9422	188	1	2020	2020	NUM
fcis-9422	188	2	,	,	PUNCT
fcis-9422	188	3	34(05	34(05	NUM
fcis-9422	188	4	):	):	PUNCT
fcis-9422	188	5	7529	7529	NUM
fcis-9422	188	6	-	-	SYM
fcis-9422	188	7	7537	7537	NUM
fcis-9422	188	8	.	.	PUNCT
fcis-9422	189	1	[	[	X
fcis-9422	189	2	14	14	NUM
fcis-9422	189	3	]	]	X
fcis-9422	189	4	cheng	cheng	PROPN
fcis-9422	189	5	j	j	PROPN
fcis-9422	189	6	,	,	PUNCT
fcis-9422	189	7	pan	pan	PROPN
fcis-9422	189	8	c	c	PROPN
fcis-9422	189	9	,	,	PUNCT
fcis-9422	189	10	dang	dang	PROPN
fcis-9422	189	11	j	j	PROPN
fcis-9422	189	12	,	,	PUNCT
fcis-9422	189	13	et	et	PROPN
fcis-9422	189	14	al	al	PROPN
fcis-9422	189	15	.	.	PUNCT
fcis-9422	189	16	entity	entity	NOUN
fcis-9422	189	17	linking	link	VERB
fcis-9422	189	18	for	for	ADP
fcis-9422	189	19	chinese	chinese	ADJ
fcis-9422	189	20	short	short	ADJ
fcis-9422	189	21	texts	text	NOUN
fcis-9422	189	22	based	base	VERB
fcis-9422	189	23	on	on	ADP
fcis-9422	189	24	bert	bert	NOUN
fcis-9422	189	25	and	and	CCONJ
fcis-9422	189	26	entity	entity	NOUN
fcis-9422	189	27	name	name	NOUN
fcis-9422	189	28	embeddings[c]//china	embeddings[c]//china	PROPN
fcis-9422	189	29	conference	conference	NOUN
fcis-9422	189	30	on	on	ADP
fcis-9422	189	31	knowledge	knowledge	NOUN
fcis-9422	189	32	graph	graph	NOUN
fcis-9422	189	33	and	and	CCONJ
fcis-9422	189	34	semantic	semantic	ADJ
fcis-9422	189	35	computing	computing	NOUN
fcis-9422	189	36	.	.	PUNCT
fcis-9422	190	1	2019	2019	NUM
fcis-9422	190	2	,	,	PUNCT
fcis-9422	190	3	2	2	NUM
fcis-9422	190	4	:	:	SYM
fcis-9422	190	5	1	1	NUM
fcis-9422	190	6	-	-	SYM
fcis-9422	190	7	12	12	NUM
fcis-9422	190	8	.	.	PUNCT
fcis-9422	191	1	[	[	X
fcis-9422	191	2	15	15	NUM
fcis-9422	191	3	]	]	X
fcis-9422	191	4	zhao	zhao	PROPN
fcis-9422	191	5	y	y	PROPN
fcis-9422	191	6	,	,	PUNCT
fcis-9422	191	7	wang	wang	PROPN
fcis-9422	191	8	y	y	PROPN
fcis-9422	191	9	,	,	PUNCT
fcis-9422	191	10	yang	yang	PROPN
fcis-9422	191	11	n.	n.	PROPN
fcis-9422	191	12	chinese	chinese	PROPN
fcis-9422	191	13	short	short	ADJ
fcis-9422	191	14	text	text	NOUN
fcis-9422	191	15	entity	entity	NOUN
fcis-9422	191	16	linking	link	VERB
fcis-9422	191	17	based	base	VERB
fcis-9422	191	18	on	on	ADP
fcis-9422	191	19	semantic	semantic	ADJ
fcis-9422	191	20	similarity	similarity	NOUN
fcis-9422	191	21	and	and	CCONJ
fcis-9422	191	22	entity	entity	NOUN
fcis-9422	191	23	correlation[c]//2020	correlation[c]//2020	ADP
fcis-9422	191	24	ieee	ieee	PROPN
fcis-9422	191	25	32nd	32nd	PROPN
fcis-9422	191	26	international	international	ADJ
fcis-9422	191	27	conference	conference	NOUN
fcis-9422	191	28	on	on	ADP
fcis-9422	191	29	tools	tool	NOUN
fcis-9422	191	30	with	with	ADP
fcis-9422	191	31	artificial	artificial	ADJ
fcis-9422	191	32	intelligence	intelligence	NOUN
fcis-9422	191	33	(	(	PUNCT
fcis-9422	191	34	ictai	ictai	PROPN
fcis-9422	191	35	)	)	PUNCT
fcis-9422	191	36	.	.	PUNCT
fcis-9422	192	1	ieee	ieee	PROPN
fcis-9422	192	2	,	,	PUNCT
fcis-9422	192	3	2020	2020	NUM
fcis-9422	192	4	:	:	PUNCT
fcis-9422	192	5	426	426	NUM
fcis-9422	192	6	-	-	SYM
fcis-9422	192	7	431	431	NUM
fcis-9422	192	8	.	.	PUNCT
fcis-9422	193	1	[	[	X
fcis-9422	193	2	16	16	NUM
fcis-9422	193	3	]	]	X
fcis-9422	193	4	berant	berant	PROPN
fcis-9422	193	5	j	j	PROPN
fcis-9422	193	6	,	,	PUNCT
fcis-9422	193	7	liang	liang	PROPN
fcis-9422	193	8	p.	p.	PROPN
fcis-9422	193	9	semantic	semantic	ADJ
fcis-9422	193	10	parsing	parsing	NOUN
fcis-9422	193	11	via	via	ADP
fcis-9422	193	12	paraphrasing[c]//	paraphrasing[c]//	PROPN
fcis-9422	193	13	52nd	52nd	ADJ
fcis-9422	193	14	annual	annual	ADJ
fcis-9422	193	15	meeting	meeting	NOUN
fcis-9422	193	16	of	of	ADP
fcis-9422	193	17	the	the	DET
fcis-9422	193	18	association	association	NOUN
fcis-9422	193	19	for	for	ADP
fcis-9422	193	20	computational	computational	ADJ
fcis-9422	193	21	linguistics	linguistic	NOUN
fcis-9422	193	22	.	.	PUNCT
fcis-9422	194	1	baltimore	baltimore	PROPN
fcis-9422	194	2	:	:	PUNCT
fcis-9422	194	3	association	association	NOUN
fcis-9422	194	4	for	for	ADP
fcis-9422	194	5	computational	computational	ADJ
fcis-9422	194	6	linguistics	linguistic	NOUN
fcis-9422	194	7	,	,	PUNCT
fcis-9422	194	8	2014	2014	NUM
fcis-9422	194	9	:	:	PUNCT
fcis-9422	194	10	1415	1415	NUM
fcis-9422	194	11	-	-	SYM
fcis-9422	194	12	1425	1425	NUM
fcis-9422	194	13	.	.	PUNCT
fcis-9422	195	1	[	[	X
fcis-9422	195	2	17	17	NUM
fcis-9422	195	3	]	]	X
fcis-9422	195	4	li	li	PROPN
fcis-9422	195	5	fei	fei	PROPN
fcis-9422	195	6	.	.	PUNCT
fcis-9422	196	1	research	research	NOUN
fcis-9422	196	2	on	on	ADP
fcis-9422	196	3	entity	entity	NOUN
fcis-9422	196	4	disambiguation	disambiguation	NOUN
fcis-9422	196	5	method	method	NOUN
fcis-9422	196	6	for	for	ADP
fcis-9422	196	7	knowledge	knowledge	NOUN
fcis-9422	196	8	graph	graph	NOUN
fcis-9422	196	9	based	base	VERB
fcis-9422	196	10	on	on	ADP
fcis-9422	196	11	deep	deep	ADJ
fcis-9422	196	12	learning	learning	NOUN
fcis-9422	197	1	[	[	X
fcis-9422	197	2	d	d	X
fcis-9422	197	3	]	]	X
fcis-9422	197	4	.	.	PUNCT
fcis-9422	197	5	chang'an	chang'an	PROPN
fcis-9422	197	6	university	university	NOUN
fcis-9422	197	7	,	,	PUNCT
fcis-9422	197	8	2021	2021	NUM
fcis-9422	197	9	.	.	PUNCT
fcis-9422	198	1	doi	doi	NOUN
fcis-9422	198	2	:	:	PUNCT
fcis-9422	198	3	10.26976	10.26976	NUM
fcis-9422	198	4	/	/	SYM
fcis-9422	198	5	d.cnki.gchau.2021.001439	d.cnki.gchau.2021.001439	NOUN
fcis-9422	198	6	.	.	PUNCT