id	sid	tid	token	lemma	pos
fcis-12817	1	1	frontiers	frontier	NOUN
fcis-12817	1	2	in	in	ADP
fcis-12817	1	3	computing	computing	NOUN
fcis-12817	1	4	and	and	CCONJ
fcis-12817	1	5	intelligent	intelligent	ADJ
fcis-12817	1	6	systems	system	NOUN
fcis-12817	1	7	issn	issn	VERB
fcis-12817	1	8	:	:	PUNCT
fcis-12817	1	9	2832	2832	NUM
fcis-12817	1	10	-	-	SYM
fcis-12817	1	11	6024	6024	NUM
fcis-12817	1	12	|	|	NOUN
fcis-12817	1	13	vol	vol	NOUN
fcis-12817	1	14	.	.	PROPN
fcis-12817	2	1	5	5	NUM
fcis-12817	2	2	,	,	PUNCT
fcis-12817	2	3	no	no	INTJ
fcis-12817	2	4	.	.	NOUN
fcis-12817	2	5	2	2	NUM
fcis-12817	2	6	,	,	PUNCT
fcis-12817	3	1	2023	2023	NUM
fcis-12817	3	2	76	76	NUM
fcis-12817	3	3	chinese	chinese	ADJ
fcis-12817	3	4	e‐commerce	e‐commerce	PUNCT
fcis-12817	3	5	ner	ner	NOUN
fcis-12817	3	6	using	use	VERB
fcis-12817	3	7	roberta‐wmm	roberta‐wmm	NOUN
fcis-12817	3	8	under	under	ADP
fcis-12817	3	9	the	the	DET
fcis-12817	3	10	machine	machine	NOUN
fcis-12817	3	11	reading	reading	NOUN
fcis-12817	3	12	comprehension	comprehension	NOUN
fcis-12817	3	13	paradigm	paradigm	NOUN
fcis-12817	3	14	mengpei	mengpei	PROPN
fcis-12817	3	15	li	li	PROPN
fcis-12817	3	16	,	,	PUNCT
fcis-12817	3	17	jun	jun	PROPN
fcis-12817	3	18	pan	pan	PROPN
fcis-12817	3	19	*	*	PROPN
fcis-12817	3	20	school	school	NOUN
fcis-12817	3	21	of	of	ADP
fcis-12817	3	22	science	science	NOUN
fcis-12817	3	23	,	,	PUNCT
fcis-12817	3	24	zhejiang	zhejiang	PROPN
fcis-12817	3	25	university	university	PROPN
fcis-12817	3	26	of	of	ADP
fcis-12817	3	27	science	science	NOUN
fcis-12817	3	28	and	and	CCONJ
fcis-12817	3	29	technology	technology	NOUN
fcis-12817	3	30	,	,	PUNCT
fcis-12817	3	31	hangzhou	hangzhou	PROPN
fcis-12817	3	32	310023	310023	NUM
fcis-12817	3	33	,	,	PUNCT
fcis-12817	3	34	china	china	PROPN
fcis-12817	3	35	*	*	PUNCT
fcis-12817	3	36	corresponding	correspond	VERB
fcis-12817	3	37	author	author	NOUN
fcis-12817	3	38	:	:	PUNCT
fcis-12817	3	39	jun	jun	PROPN
fcis-12817	3	40	pan	pan	PROPN
fcis-12817	3	41	(	(	PUNCT
fcis-12817	3	42	email	email	NOUN
fcis-12817	3	43	:	:	PUNCT
fcis-12817	3	44	panjun78@qq.com	panjun78@qq.com	NOUN
fcis-12817	3	45	)	)	PUNCT
fcis-12817	3	46	abstract	abstract	NOUN
fcis-12817	3	47	:	:	PUNCT
fcis-12817	3	48	in	in	ADP
fcis-12817	3	49	practical	practical	ADJ
fcis-12817	3	50	applications	application	NOUN
fcis-12817	3	51	within	within	ADP
fcis-12817	3	52	the	the	DET
fcis-12817	3	53	e	e	NOUN
fcis-12817	3	54	-	-	NOUN
fcis-12817	3	55	commerce	commerce	NOUN
fcis-12817	3	56	domain	domain	NOUN
fcis-12817	3	57	,	,	PUNCT
fcis-12817	3	58	there	there	PRON
fcis-12817	3	59	is	be	VERB
fcis-12817	3	60	often	often	ADV
fcis-12817	3	61	a	a	DET
fcis-12817	3	62	requirement	requirement	NOUN
fcis-12817	3	63	to	to	PART
fcis-12817	3	64	identify	identify	VERB
fcis-12817	3	65	product	product	NOUN
fcis-12817	3	66	entities	entity	NOUN
fcis-12817	3	67	and	and	CCONJ
fcis-12817	3	68	their	their	PRON
fcis-12817	3	69	corresponding	correspond	VERB
fcis-12817	3	70	brand	brand	NOUN
fcis-12817	3	71	entities	entity	NOUN
fcis-12817	3	72	,	,	PUNCT
fcis-12817	3	73	based	base	VERB
fcis-12817	3	74	on	on	ADP
fcis-12817	3	75	their	their	PRON
fcis-12817	3	76	descriptions	description	NOUN
fcis-12817	3	77	.	.	PUNCT
fcis-12817	4	1	however	however	ADV
fcis-12817	4	2	,	,	PUNCT
fcis-12817	4	3	there	there	PRON
fcis-12817	4	4	has	have	AUX
fcis-12817	4	5	been	be	AUX
fcis-12817	4	6	a	a	DET
fcis-12817	4	7	relatively	relatively	ADV
fcis-12817	4	8	limited	limited	ADJ
fcis-12817	4	9	focus	focus	NOUN
fcis-12817	4	10	on	on	ADP
fcis-12817	4	11	studies	study	NOUN
fcis-12817	4	12	addressing	address	VERB
fcis-12817	4	13	the	the	DET
fcis-12817	4	14	named	name	VERB
fcis-12817	4	15	entities	entity	NOUN
fcis-12817	4	16	recognition	recognition	NOUN
fcis-12817	4	17	in	in	ADP
fcis-12817	4	18	the	the	DET
fcis-12817	4	19	e	e	PROPN
fcis-12817	4	20	-	-	NOUN
fcis-12817	4	21	commerce	commerce	NOUN
fcis-12817	4	22	domain	domain	NOUN
fcis-12817	4	23	.	.	PUNCT
fcis-12817	5	1	we	we	PRON
fcis-12817	5	2	crawled	crawl	VERB
fcis-12817	5	3	data	datum	NOUN
fcis-12817	5	4	from	from	ADP
fcis-12817	5	5	e	e	NOUN
fcis-12817	5	6	-	-	NOUN
fcis-12817	5	7	commerce	commerce	NOUN
fcis-12817	5	8	websites	website	NOUN
fcis-12817	5	9	and	and	CCONJ
fcis-12817	5	10	transformed	transform	VERB
fcis-12817	5	11	them	they	PRON
fcis-12817	5	12	into	into	ADP
fcis-12817	5	13	a	a	DET
fcis-12817	5	14	named	name	VERB
fcis-12817	5	15	entity	entity	NOUN
fcis-12817	5	16	recognition	recognition	NOUN
fcis-12817	5	17	dataset	dataset	NOUN
fcis-12817	5	18	,	,	PUNCT
fcis-12817	5	19	which	which	PRON
fcis-12817	5	20	is	be	AUX
fcis-12817	5	21	suitable	suitable	ADJ
fcis-12817	5	22	for	for	ADP
fcis-12817	5	23	machine	machine	NOUN
fcis-12817	5	24	reading	reading	NOUN
fcis-12817	5	25	comprehension	comprehension	NOUN
fcis-12817	5	26	.	.	PUNCT
fcis-12817	6	1	since	since	SCONJ
fcis-12817	6	2	the	the	DET
fcis-12817	6	3	questions	question	NOUN
fcis-12817	6	4	machine	machine	NOUN
fcis-12817	6	5	reading	reading	NOUN
fcis-12817	6	6	comprehension	comprehension	NOUN
fcis-12817	6	7	contain	contain	VERB
fcis-12817	6	8	a	a	DET
fcis-12817	6	9	priori	priori	ADJ
fcis-12817	6	10	semantic	semantic	ADJ
fcis-12817	6	11	information	information	NOUN
fcis-12817	6	12	about	about	ADP
fcis-12817	6	13	the	the	DET
fcis-12817	6	14	types	type	NOUN
fcis-12817	6	15	of	of	ADP
fcis-12817	6	16	the	the	DET
fcis-12817	6	17	entities	entity	NOUN
fcis-12817	6	18	,	,	PUNCT
fcis-12817	6	19	we	we	PRON
fcis-12817	6	20	propose	propose	VERB
fcis-12817	6	21	a	a	DET
fcis-12817	6	22	model	model	NOUN
fcis-12817	6	23	that	that	PRON
fcis-12817	6	24	uses	use	VERB
fcis-12817	6	25	the	the	DET
fcis-12817	6	26	mrc	mrc	NOUN
fcis-12817	6	27	modeling	modeling	NOUN
fcis-12817	6	28	paradigm	paradigm	NOUN
fcis-12817	6	29	to	to	PART
fcis-12817	6	30	solve	solve	VERB
fcis-12817	6	31	the	the	DET
fcis-12817	6	32	task	task	NOUN
fcis-12817	6	33	of	of	ADP
fcis-12817	6	34	recognizing	recognize	VERB
fcis-12817	6	35	brand	brand	NOUN
fcis-12817	6	36	entities	entity	NOUN
fcis-12817	6	37	as	as	ADV
fcis-12817	6	38	well	well	ADV
fcis-12817	6	39	as	as	ADP
fcis-12817	6	40	commodity	commodity	NOUN
fcis-12817	6	41	entities	entity	NOUN
fcis-12817	6	42	in	in	ADP
fcis-12817	6	43	the	the	DET
fcis-12817	6	44	e	e	PROPN
fcis-12817	6	45	-	-	NOUN
fcis-12817	6	46	commerce	commerce	NOUN
fcis-12817	6	47	domain	domain	NOUN
fcis-12817	6	48	.	.	PUNCT
fcis-12817	7	1	the	the	DET
fcis-12817	7	2	model	model	NOUN
fcis-12817	7	3	encodes	encode	VERB
fcis-12817	7	4	the	the	DET
fcis-12817	7	5	contexts	context	NOUN
fcis-12817	7	6	and	and	CCONJ
fcis-12817	7	7	the	the	DET
fcis-12817	7	8	corresponding	corresponding	ADJ
fcis-12817	7	9	questions	question	NOUN
fcis-12817	7	10	using	use	VERB
fcis-12817	7	11	the	the	DET
fcis-12817	7	12	roberta	roberta	PROPN
fcis-12817	7	13	-	-	PUNCT
fcis-12817	7	14	wwm	wwm	PROPN
fcis-12817	7	15	model	model	NOUN
fcis-12817	7	16	,	,	PUNCT
fcis-12817	7	17	and	and	CCONJ
fcis-12817	7	18	then	then	ADV
fcis-12817	7	19	further	further	ADJ
fcis-12817	7	20	extracts	extract	NOUN
fcis-12817	7	21	the	the	DET
fcis-12817	7	22	semantic	semantic	ADJ
fcis-12817	7	23	information	information	NOUN
fcis-12817	7	24	of	of	ADP
fcis-12817	7	25	the	the	DET
fcis-12817	7	26	contexts	contexts	NOUN
fcis-12817	7	27	using	use	VERB
fcis-12817	7	28	an	an	DET
fcis-12817	7	29	attention	attention	NOUN
fcis-12817	7	30	network	network	NOUN
fcis-12817	7	31	.	.	PUNCT
fcis-12817	8	1	we	we	PRON
fcis-12817	8	2	utilize	utilize	VERB
fcis-12817	8	3	softmax	softmax	NOUN
fcis-12817	8	4	as	as	ADP
fcis-12817	8	5	the	the	DET
fcis-12817	8	6	decoding	decode	VERB
fcis-12817	8	7	layer	layer	NOUN
fcis-12817	8	8	to	to	PART
fcis-12817	8	9	get	get	VERB
fcis-12817	8	10	the	the	DET
fcis-12817	8	11	head	head	NOUN
fcis-12817	8	12	index	index	NOUN
fcis-12817	8	13	and	and	CCONJ
fcis-12817	8	14	tail	tail	NOUN
fcis-12817	8	15	index	index	NOUN
fcis-12817	8	16	of	of	ADP
fcis-12817	8	17	the	the	DET
fcis-12817	8	18	entity	entity	NOUN
fcis-12817	8	19	,	,	PUNCT
fcis-12817	8	20	and	and	CCONJ
fcis-12817	8	21	finally	finally	ADV
fcis-12817	8	22	use	use	VERB
fcis-12817	8	23	the	the	DET
fcis-12817	8	24	matching	matching	NOUN
fcis-12817	8	25	module	module	NOUN
fcis-12817	8	26	to	to	PART
fcis-12817	8	27	get	get	VERB
fcis-12817	8	28	the	the	DET
fcis-12817	8	29	entity	entity	NOUN
fcis-12817	8	30	index	index	NOUN
fcis-12817	8	31	.	.	PUNCT
fcis-12817	9	1	through	through	ADP
fcis-12817	9	2	experiments	experiment	NOUN
fcis-12817	9	3	on	on	ADP
fcis-12817	9	4	two	two	NUM
fcis-12817	9	5	e	e	NOUN
fcis-12817	9	6	-	-	NOUN
fcis-12817	9	7	commerce	commerce	NOUN
fcis-12817	9	8	datasets	dataset	NOUN
fcis-12817	9	9	,	,	PUNCT
fcis-12817	9	10	the	the	DET
fcis-12817	9	11	results	result	NOUN
fcis-12817	9	12	show	show	VERB
fcis-12817	9	13	that	that	SCONJ
fcis-12817	9	14	the	the	DET
fcis-12817	9	15	new	new	ADJ
fcis-12817	9	16	method	method	NOUN
fcis-12817	9	17	can	can	AUX
fcis-12817	9	18	significantly	significantly	ADV
fcis-12817	9	19	improve	improve	VERB
fcis-12817	9	20	the	the	DET
fcis-12817	9	21	recognition	recognition	NOUN
fcis-12817	9	22	effect	effect	NOUN
fcis-12817	9	23	of	of	ADP
fcis-12817	9	24	chinese	chinese	ADJ
fcis-12817	9	25	ner	ner	NOUN
fcis-12817	9	26	in	in	ADP
fcis-12817	9	27	e	e	PROPN
fcis-12817	9	28	-	-	NOUN
fcis-12817	9	29	commerce	commerce	NOUN
fcis-12817	9	30	domain	domain	NOUN
fcis-12817	9	31	.	.	PUNCT
fcis-12817	10	1	keywords	keyword	NOUN
fcis-12817	10	2	:	:	PUNCT
fcis-12817	10	3	deep	deep	ADJ
fcis-12817	10	4	learning	learning	NOUN
fcis-12817	10	5	;	;	PUNCT
fcis-12817	10	6	machine	machine	NOUN
fcis-12817	10	7	reading	reading	NOUN
fcis-12817	10	8	comprehension	comprehension	NOUN
fcis-12817	10	9	;	;	PUNCT
fcis-12817	10	10	named	name	VERB
fcis-12817	10	11	entity	entity	NOUN
fcis-12817	10	12	recognition	recognition	NOUN
fcis-12817	10	13	.	.	PUNCT
fcis-12817	11	1	1	1	X
fcis-12817	11	2	.	.	X
fcis-12817	11	3	introduction	introduction	NOUN
fcis-12817	11	4	named	name	VERB
fcis-12817	11	5	entity	entity	NOUN
fcis-12817	11	6	recognition	recognition	NOUN
fcis-12817	11	7	aims	aim	VERB
fcis-12817	11	8	to	to	PART
fcis-12817	11	9	identify	identify	VERB
fcis-12817	11	10	certain	certain	ADJ
fcis-12817	11	11	entities	entity	NOUN
fcis-12817	11	12	with	with	ADP
fcis-12817	11	13	specific	specific	ADJ
fcis-12817	11	14	meanings	meaning	NOUN
fcis-12817	11	15	from	from	ADP
fcis-12817	11	16	the	the	DET
fcis-12817	11	17	text	text	NOUN
fcis-12817	11	18	and	and	CCONJ
fcis-12817	11	19	categorize	categorize	VERB
fcis-12817	11	20	the	the	DET
fcis-12817	11	21	entities	entity	NOUN
fcis-12817	11	22	into	into	ADP
fcis-12817	11	23	different	different	ADJ
fcis-12817	11	24	classes	class	NOUN
fcis-12817	11	25	.	.	PUNCT
fcis-12817	12	1	named	name	VERB
fcis-12817	12	2	entity	entity	NOUN
fcis-12817	12	3	recognition	recognition	NOUN
fcis-12817	12	4	is	be	AUX
fcis-12817	12	5	useful	useful	ADJ
fcis-12817	12	6	for	for	ADP
fcis-12817	12	7	question	question	NOUN
fcis-12817	12	8	-	-	PUNCT
fcis-12817	12	9	answer	answer	NOUN
fcis-12817	12	10	systems	system	NOUN
fcis-12817	12	11	[	[	X
fcis-12817	12	12	1	1	NUM
fcis-12817	12	13	]	]	PUNCT
fcis-12817	12	14	,	,	PUNCT
fcis-12817	12	15	machine	machine	NOUN
fcis-12817	12	16	translation	translation	NOUN
fcis-12817	13	1	[	[	X
fcis-12817	13	2	2	2	NUM
fcis-12817	13	3	]	]	PUNCT
fcis-12817	13	4	,	,	PUNCT
fcis-12817	13	5	event	event	NOUN
fcis-12817	13	6	extraction	extraction	NOUN
fcis-12817	13	7	[	[	X
fcis-12817	13	8	3	3	NUM
fcis-12817	13	9	]	]	PUNCT
fcis-12817	13	10	and	and	CCONJ
fcis-12817	13	11	entity	entity	NOUN
fcis-12817	13	12	linking	link	VERB
fcis-12817	13	13	[	[	X
fcis-12817	13	14	4	4	NUM
fcis-12817	13	15	]	]	PUNCT
fcis-12817	13	16	and	and	CCONJ
fcis-12817	13	17	other	other	ADJ
fcis-12817	13	18	downstream	downstream	ADJ
fcis-12817	13	19	tasks	task	NOUN
fcis-12817	13	20	.	.	PUNCT
fcis-12817	14	1	named	name	VERB
fcis-12817	14	2	entity	entity	NOUN
fcis-12817	14	3	recognition	recognition	NOUN
fcis-12817	14	4	is	be	AUX
fcis-12817	14	5	often	often	ADV
fcis-12817	14	6	regarded	regard	VERB
fcis-12817	14	7	as	as	ADP
fcis-12817	14	8	a	a	DET
fcis-12817	14	9	sequence	sequence	NOUN
fcis-12817	14	10	labeling	labeling	NOUN
fcis-12817	14	11	problem	problem	NOUN
fcis-12817	14	12	or	or	CCONJ
fcis-12817	14	13	a	a	DET
fcis-12817	14	14	machine	machine	NOUN
fcis-12817	14	15	reading	reading	NOUN
fcis-12817	14	16	comprehension	comprehension	NOUN
fcis-12817	14	17	(	(	PUNCT
fcis-12817	14	18	mrc	mrc	NOUN
fcis-12817	14	19	)	)	PUNCT
fcis-12817	14	20	task	task	NOUN
fcis-12817	14	21	.	.	PUNCT
fcis-12817	15	1	the	the	DET
fcis-12817	15	2	process	process	NOUN
fcis-12817	15	3	of	of	ADP
fcis-12817	15	4	mrc	mrc	NOUN
fcis-12817	15	5	involves	involve	VERB
fcis-12817	15	6	enabling	enable	VERB
fcis-12817	15	7	the	the	DET
fcis-12817	15	8	model	model	NOUN
fcis-12817	15	9	to	to	PART
fcis-12817	15	10	comprehend	comprehend	VERB
fcis-12817	15	11	the	the	DET
fcis-12817	15	12	context	context	NOUN
fcis-12817	15	13	of	of	ADP
fcis-12817	15	14	a	a	DET
fcis-12817	15	15	provided	provide	VERB
fcis-12817	15	16	text	text	NOUN
fcis-12817	15	17	and	and	CCONJ
fcis-12817	15	18	the	the	DET
fcis-12817	15	19	questions	question	NOUN
fcis-12817	15	20	,	,	PUNCT
fcis-12817	15	21	followed	follow	VERB
fcis-12817	15	22	by	by	ADP
fcis-12817	15	23	extracting	extract	VERB
fcis-12817	15	24	the	the	DET
fcis-12817	15	25	relevant	relevant	ADJ
fcis-12817	15	26	information	information	NOUN
fcis-12817	15	27	from	from	ADP
fcis-12817	15	28	the	the	DET
fcis-12817	15	29	text	text	NOUN
fcis-12817	15	30	that	that	PRON
fcis-12817	15	31	can	can	AUX
fcis-12817	15	32	answer	answer	VERB
fcis-12817	15	33	the	the	DET
fcis-12817	15	34	given	give	VERB
fcis-12817	15	35	questions	question	NOUN
fcis-12817	15	36	.	.	PUNCT
fcis-12817	16	1	in	in	ADP
fcis-12817	16	2	the	the	DET
fcis-12817	16	3	mrc	mrc	NOUN
fcis-12817	16	4	modeling	modeling	NOUN
fcis-12817	16	5	paradigm	paradigm	NOUN
fcis-12817	16	6	,	,	PUNCT
fcis-12817	16	7	we	we	PRON
fcis-12817	16	8	use	use	VERB
fcis-12817	16	9	the	the	DET
fcis-12817	16	10	question	question	NOUN
fcis-12817	16	11	to	to	PART
fcis-12817	16	12	represent	represent	VERB
fcis-12817	16	13	the	the	DET
fcis-12817	16	14	type	type	NOUN
fcis-12817	16	15	of	of	ADP
fcis-12817	16	16	entity	entity	NOUN
fcis-12817	16	17	,	,	PUNCT
fcis-12817	16	18	and	and	CCONJ
fcis-12817	16	19	the	the	DET
fcis-12817	16	20	model	model	NOUN
fcis-12817	16	21	understands	understand	VERB
fcis-12817	16	22	the	the	DET
fcis-12817	16	23	text	text	NOUN
fcis-12817	16	24	and	and	CCONJ
fcis-12817	16	25	then	then	ADV
fcis-12817	16	26	gives	give	VERB
fcis-12817	16	27	the	the	DET
fcis-12817	16	28	corresponding	correspond	VERB
fcis-12817	16	29	answer	answer	NOUN
fcis-12817	16	30	according	accord	VERB
fcis-12817	16	31	to	to	ADP
fcis-12817	16	32	the	the	DET
fcis-12817	16	33	question	question	NOUN
fcis-12817	16	34	.	.	PUNCT
fcis-12817	17	1	for	for	ADP
fcis-12817	17	2	example	example	NOUN
fcis-12817	17	3	,	,	PUNCT
fcis-12817	17	4	when	when	SCONJ
fcis-12817	17	5	we	we	PRON
fcis-12817	17	6	need	need	VERB
fcis-12817	17	7	to	to	PART
fcis-12817	17	8	extract	extract	VERB
fcis-12817	17	9	the	the	DET
fcis-12817	17	10	product	product	NOUN
fcis-12817	17	11	entities	entity	NOUN
fcis-12817	17	12	in	in	ADP
fcis-12817	17	13	the	the	DET
fcis-12817	17	14	sentence	sentence	NOUN
fcis-12817	17	15	,	,	PUNCT
fcis-12817	17	16	the	the	DET
fcis-12817	17	17	task	task	NOUN
fcis-12817	17	18	is	be	AUX
fcis-12817	17	19	transformed	transform	VERB
fcis-12817	17	20	into	into	ADP
fcis-12817	17	21	answering	answer	VERB
fcis-12817	17	22	the	the	DET
fcis-12817	17	23	question	question	NOUN
fcis-12817	17	24	"	"	PUNCT
fcis-12817	17	25	find	find	VERB
fcis-12817	17	26	out	out	ADP
fcis-12817	17	27	which	which	DET
fcis-12817	17	28	products	product	NOUN
fcis-12817	17	29	are	be	AUX
fcis-12817	17	30	mentioned	mention	VERB
fcis-12817	17	31	in	in	ADP
fcis-12817	17	32	the	the	DET
fcis-12817	17	33	sentence	sentence	NOUN
fcis-12817	17	34	"	"	PUNCT
fcis-12817	17	35	.	.	PUNCT
fcis-12817	18	1	when	when	SCONJ
fcis-12817	18	2	we	we	PRON
fcis-12817	18	3	need	need	VERB
fcis-12817	18	4	to	to	PART
fcis-12817	18	5	extract	extract	VERB
fcis-12817	18	6	brands	brand	NOUN
fcis-12817	18	7	,	,	PUNCT
fcis-12817	18	8	the	the	DET
fcis-12817	18	9	task	task	NOUN
fcis-12817	18	10	translates	translate	VERB
fcis-12817	18	11	into	into	ADP
fcis-12817	18	12	answering	answer	VERB
fcis-12817	18	13	the	the	DET
fcis-12817	18	14	question	question	NOUN
fcis-12817	18	15	"	"	PUNCT
fcis-12817	18	16	find	find	VERB
fcis-12817	18	17	out	out	ADP
fcis-12817	18	18	which	which	DET
fcis-12817	18	19	brands	brand	NOUN
fcis-12817	18	20	are	be	AUX
fcis-12817	18	21	mentioned	mention	VERB
fcis-12817	18	22	in	in	ADP
fcis-12817	18	23	the	the	DET
fcis-12817	18	24	sentence	sentence	NOUN
fcis-12817	18	25	"	"	PUNCT
fcis-12817	18	26	.	.	PUNCT
fcis-12817	19	1	compared	compare	VERB
fcis-12817	19	2	with	with	ADP
fcis-12817	19	3	the	the	DET
fcis-12817	19	4	mrc	mrc	NOUN
fcis-12817	19	5	-	-	PUNCT
fcis-12817	19	6	based	base	VERB
fcis-12817	19	7	approach	approach	NOUN
fcis-12817	19	8	,	,	PUNCT
fcis-12817	19	9	the	the	DET
fcis-12817	19	10	sequencelabeling	sequencelabeling	NOUN
fcis-12817	19	11	-	-	PUNCT
fcis-12817	19	12	based	base	VERB
fcis-12817	19	13	approach	approach	NOUN
fcis-12817	19	14	treats	treat	VERB
fcis-12817	19	15	the	the	DET
fcis-12817	19	16	entity	entity	NOUN
fcis-12817	19	17	categories	category	NOUN
fcis-12817	19	18	as	as	ADP
fcis-12817	19	19	class	class	NOUN
fcis-12817	19	20	indexes	index	NOUN
fcis-12817	19	21	and	and	CCONJ
fcis-12817	19	22	loses	lose	VERB
fcis-12817	19	23	the	the	DET
fcis-12817	19	24	linguistic	linguistic	ADJ
fcis-12817	19	25	information	information	NOUN
fcis-12817	19	26	of	of	ADP
fcis-12817	19	27	the	the	DET
fcis-12817	19	28	categories	category	NOUN
fcis-12817	19	29	.	.	PUNCT
fcis-12817	20	1	for	for	ADP
fcis-12817	20	2	example	example	NOUN
fcis-12817	20	3	,	,	PUNCT
fcis-12817	20	4	the	the	DET
fcis-12817	20	5	two	two	NUM
fcis-12817	20	6	entity	entity	NOUN
fcis-12817	20	7	categories	category	NOUN
fcis-12817	20	8	of	of	ADP
fcis-12817	20	9	"	"	PUNCT
fcis-12817	20	10	product	product	NOUN
fcis-12817	20	11	"	"	PUNCT
fcis-12817	20	12	and	and	CCONJ
fcis-12817	20	13	"	"	PUNCT
fcis-12817	20	14	brand	brand	NOUN
fcis-12817	20	15	"	"	PUNCT
fcis-12817	20	16	are	be	AUX
fcis-12817	20	17	only	only	ADV
fcis-12817	20	18	represented	represent	VERB
fcis-12817	20	19	by	by	ADP
fcis-12817	20	20	"	"	PUNCT
fcis-12817	20	21	0	0	NUM
fcis-12817	20	22	"	"	PUNCT
fcis-12817	20	23	and	and	CCONJ
fcis-12817	20	24	"	"	PUNCT
fcis-12817	20	25	1	1	NUM
fcis-12817	20	26	"	"	PUNCT
fcis-12817	20	27	to	to	PART
fcis-12817	20	28	represent	represent	VERB
fcis-12817	20	29	their	their	PRON
fcis-12817	20	30	corresponding	correspond	VERB
fcis-12817	20	31	classes	class	NOUN
fcis-12817	20	32	and	and	CCONJ
fcis-12817	20	33	do	do	AUX
fcis-12817	20	34	not	not	PART
fcis-12817	20	35	have	have	VERB
fcis-12817	20	36	semantic	semantic	ADJ
fcis-12817	20	37	information	information	NOUN
fcis-12817	20	38	.	.	PUNCT
fcis-12817	21	1	the	the	DET
fcis-12817	21	2	questions	question	NOUN
fcis-12817	21	3	constructed	construct	VERB
fcis-12817	21	4	in	in	ADP
fcis-12817	21	5	mrc	mrc	NOUN
fcis-12817	21	6	contain	contain	VERB
fcis-12817	21	7	a	a	DET
fcis-12817	21	8	priori	priori	ADJ
fcis-12817	21	9	information	information	NOUN
fcis-12817	21	10	about	about	ADP
fcis-12817	21	11	the	the	DET
fcis-12817	21	12	entity	entity	NOUN
fcis-12817	21	13	categories	category	NOUN
fcis-12817	21	14	,	,	PUNCT
fcis-12817	21	15	for	for	ADP
fcis-12817	21	16	example	example	NOUN
fcis-12817	21	17	,	,	PUNCT
fcis-12817	21	18	the	the	DET
fcis-12817	21	19	question	question	NOUN
fcis-12817	21	20	"	"	PUNCT
fcis-12817	21	21	find	find	VERB
fcis-12817	21	22	the	the	DET
fcis-12817	21	23	products	product	NOUN
fcis-12817	21	24	in	in	ADP
fcis-12817	21	25	the	the	DET
fcis-12817	21	26	text	text	NOUN
fcis-12817	21	27	"	"	PUNCT
fcis-12817	21	28	will	will	AUX
fcis-12817	21	29	provide	provide	VERB
fcis-12817	21	30	corresponding	corresponding	ADJ
fcis-12817	21	31	semantics	semantic	NOUN
fcis-12817	21	32	of	of	ADP
fcis-12817	21	33	"	"	PUNCT
fcis-12817	21	34	product	product	NOUN
fcis-12817	21	35	"	"	PUNCT
fcis-12817	21	36	.	.	PUNCT
fcis-12817	22	1	the	the	DET
fcis-12817	22	2	question	question	NOUN
fcis-12817	22	3	"	"	PUNCT
fcis-12817	22	4	find	find	VERB
fcis-12817	22	5	the	the	DET
fcis-12817	22	6	brands	brand	NOUN
fcis-12817	22	7	in	in	ADP
fcis-12817	22	8	the	the	DET
fcis-12817	22	9	text	text	NOUN
fcis-12817	22	10	"	"	PUNCT
fcis-12817	22	11	will	will	AUX
fcis-12817	22	12	provide	provide	VERB
fcis-12817	22	13	the	the	DET
fcis-12817	22	14	corresponding	corresponding	ADJ
fcis-12817	22	15	semantics	semantic	NOUN
fcis-12817	22	16	of	of	ADP
fcis-12817	22	17	"	"	PUNCT
fcis-12817	22	18	brand	brand	NOUN
fcis-12817	22	19	"	"	PUNCT
fcis-12817	22	20	.	.	PUNCT
fcis-12817	23	1	therefore	therefore	ADV
fcis-12817	23	2	,	,	PUNCT
fcis-12817	23	3	the	the	DET
fcis-12817	23	4	overall	overall	ADJ
fcis-12817	23	5	ner	ner	NOUN
fcis-12817	23	6	performance	performance	NOUN
fcis-12817	23	7	is	be	AUX
fcis-12817	23	8	improved	improve	VERB
fcis-12817	23	9	.	.	PUNCT
fcis-12817	24	1	since	since	SCONJ
fcis-12817	24	2	the	the	DET
fcis-12817	24	3	constructed	construct	VERB
fcis-12817	24	4	questions	question	NOUN
fcis-12817	24	5	contain	contain	VERB
fcis-12817	24	6	a	a	DET
fcis-12817	24	7	priori	priori	ADJ
fcis-12817	24	8	information	information	NOUN
fcis-12817	24	9	,	,	PUNCT
fcis-12817	24	10	the	the	DET
fcis-12817	24	11	accuracy	accuracy	NOUN
fcis-12817	24	12	of	of	ADP
fcis-12817	24	13	the	the	DET
fcis-12817	24	14	a	a	DET
fcis-12817	24	15	priori	priori	ADJ
fcis-12817	24	16	information	information	NOUN
fcis-12817	24	17	will	will	AUX
fcis-12817	24	18	have	have	VERB
fcis-12817	24	19	an	an	DET
fcis-12817	24	20	impact	impact	NOUN
fcis-12817	24	21	on	on	ADP
fcis-12817	24	22	the	the	DET
fcis-12817	24	23	answers	answer	NOUN
fcis-12817	24	24	.	.	PUNCT
fcis-12817	25	1	therefore	therefore	ADV
fcis-12817	25	2	,	,	PUNCT
fcis-12817	25	3	we	we	PRON
fcis-12817	25	4	conducted	conduct	VERB
fcis-12817	25	5	experiments	experiment	NOUN
fcis-12817	25	6	on	on	ADP
fcis-12817	25	7	the	the	DET
fcis-12817	25	8	effect	effect	NOUN
fcis-12817	25	9	of	of	ADP
fcis-12817	25	10	several	several	ADJ
fcis-12817	25	11	different	different	ADJ
fcis-12817	25	12	question	question	NOUN
fcis-12817	25	13	templates	template	NOUN
fcis-12817	25	14	.	.	PUNCT
fcis-12817	26	1	2	2	X
fcis-12817	26	2	.	.	X
fcis-12817	26	3	related	relate	VERB
fcis-12817	26	4	work	work	NOUN
fcis-12817	26	5	the	the	DET
fcis-12817	26	6	reading	reading	NOUN
fcis-12817	26	7	comprehension	comprehension	NOUN
fcis-12817	26	8	modeling	modeling	NOUN
fcis-12817	26	9	paradigm	paradigm	NOUN
fcis-12817	26	10	is	be	AUX
fcis-12817	26	11	widely	widely	ADV
fcis-12817	26	12	used	use	VERB
fcis-12817	26	13	with	with	ADP
fcis-12817	26	14	answering	answer	VERB
fcis-12817	26	15	systems	system	NOUN
fcis-12817	26	16	in	in	ADP
fcis-12817	26	17	the	the	DET
fcis-12817	26	18	field	field	NOUN
fcis-12817	26	19	of	of	ADP
fcis-12817	26	20	nlp	nlp	NOUN
fcis-12817	26	21	[	[	X
fcis-12817	26	22	5	5	NUM
fcis-12817	26	23	]	]	PUNCT
fcis-12817	26	24	,	,	PUNCT
fcis-12817	26	25	question	question	NOUN
fcis-12817	26	26	answering	answer	VERB
fcis-12817	26	27	[	[	X
fcis-12817	26	28	6	6	NUM
fcis-12817	26	29	]	]	PUNCT
fcis-12817	26	30	,	,	PUNCT
fcis-12817	26	31	information	information	NOUN
fcis-12817	26	32	retrieval	retrieval	NOUN
fcis-12817	27	1	[	[	X
fcis-12817	27	2	7	7	X
fcis-12817	27	3	]	]	PUNCT
fcis-12817	27	4	and	and	CCONJ
fcis-12817	27	5	other	other	ADJ
fcis-12817	27	6	domains	domain	NOUN
fcis-12817	27	7	.	.	PUNCT
fcis-12817	28	1	as	as	ADP
fcis-12817	28	2	an	an	DET
fcis-12817	28	3	upstream	upstream	ADJ
fcis-12817	28	4	task	task	NOUN
fcis-12817	28	5	of	of	ADP
fcis-12817	28	6	nlp	nlp	NOUN
fcis-12817	28	7	,	,	PUNCT
fcis-12817	28	8	ner	ner	NOUN
fcis-12817	28	9	tasks	task	NOUN
fcis-12817	28	10	can	can	AUX
fcis-12817	28	11	also	also	ADV
fcis-12817	28	12	be	be	AUX
fcis-12817	28	13	modeled	model	VERB
fcis-12817	28	14	as	as	ADP
fcis-12817	28	15	machine	machine	NOUN
fcis-12817	28	16	reading	reading	NOUN
fcis-12817	28	17	comprehension	comprehension	NOUN
fcis-12817	28	18	tasks	task	NOUN
fcis-12817	28	19	[	[	X
fcis-12817	28	20	8	8	NUM
fcis-12817	28	21	]	]	PUNCT
fcis-12817	28	22	.	.	PUNCT
fcis-12817	29	1	li	li	PROPN
fcis-12817	29	2	et	et	PROPN
fcis-12817	29	3	al	al	PROPN
fcis-12817	29	4	.	.	PUNCT
fcis-12817	30	1	[	[	X
fcis-12817	30	2	9	9	NUM
fcis-12817	30	3	]	]	PUNCT
fcis-12817	30	4	proposed	propose	VERB
fcis-12817	30	5	a	a	DET
fcis-12817	30	6	unified	unified	ADJ
fcis-12817	30	7	framework	framework	NOUN
fcis-12817	30	8	for	for	ADP
fcis-12817	30	9	ner	ner	NOUN
fcis-12817	30	10	based	base	VERB
fcis-12817	30	11	on	on	ADP
fcis-12817	30	12	the	the	DET
fcis-12817	30	13	reading	reading	NOUN
fcis-12817	30	14	comprehension	comprehension	NOUN
fcis-12817	30	15	paradigm	paradigm	NOUN
fcis-12817	30	16	that	that	PRON
fcis-12817	30	17	can	can	AUX
fcis-12817	30	18	handle	handle	VERB
fcis-12817	30	19	both	both	CCONJ
fcis-12817	30	20	flat	flat	ADJ
fcis-12817	30	21	and	and	CCONJ
fcis-12817	30	22	nested	nested	ADJ
fcis-12817	30	23	entities	entity	NOUN
fcis-12817	30	24	.	.	PUNCT
fcis-12817	31	1	presently	presently	ADV
fcis-12817	31	2	,	,	PUNCT
fcis-12817	31	3	ner	ner	NOUN
fcis-12817	31	4	models	model	NOUN
fcis-12817	31	5	that	that	PRON
fcis-12817	31	6	employ	employ	VERB
fcis-12817	31	7	the	the	DET
fcis-12817	31	8	mrc	mrc	NOUN
fcis-12817	31	9	paradigm	paradigm	NOUN
fcis-12817	31	10	are	be	AUX
fcis-12817	31	11	primarily	primarily	ADV
fcis-12817	31	12	utilized	utilize	VERB
fcis-12817	31	13	to	to	PART
fcis-12817	31	14	address	address	VERB
fcis-12817	31	15	ner	ner	NOUN
fcis-12817	31	16	challenges	challenge	NOUN
fcis-12817	31	17	in	in	ADP
fcis-12817	31	18	the	the	DET
fcis-12817	31	19	fields	field	NOUN
fcis-12817	31	20	of	of	ADP
fcis-12817	31	21	biology	biology	NOUN
fcis-12817	31	22	[	[	X
fcis-12817	31	23	10	10	NUM
fcis-12817	31	24	]	]	PUNCT
fcis-12817	31	25	and	and	CCONJ
fcis-12817	31	26	finance	finance	NOUN
fcis-12817	31	27	[	[	X
fcis-12817	31	28	11	11	NUM
fcis-12817	31	29	]	]	PUNCT
fcis-12817	31	30	.	.	PUNCT
fcis-12817	32	1	in	in	ADP
fcis-12817	32	2	contrast	contrast	NOUN
fcis-12817	32	3	,	,	PUNCT
fcis-12817	32	4	there	there	PRON
fcis-12817	32	5	has	have	AUX
fcis-12817	32	6	been	be	AUX
fcis-12817	32	7	relatively	relatively	ADV
fcis-12817	32	8	less	less	ADJ
fcis-12817	32	9	research	research	NOUN
fcis-12817	32	10	focused	focus	VERB
fcis-12817	32	11	on	on	ADP
fcis-12817	32	12	named	name	VERB
fcis-12817	32	13	entity	entity	NOUN
fcis-12817	32	14	recognition	recognition	NOUN
fcis-12817	32	15	within	within	ADP
fcis-12817	32	16	the	the	DET
fcis-12817	32	17	e	e	PROPN
fcis-12817	32	18	-	-	NOUN
fcis-12817	32	19	commerce	commerce	NOUN
fcis-12817	32	20	domain	domain	NOUN
fcis-12817	32	21	using	use	VERB
fcis-12817	32	22	the	the	DET
fcis-12817	32	23	mrc	mrc	NOUN
fcis-12817	32	24	paradigm	paradigm	NOUN
fcis-12817	32	25	.	.	PUNCT
fcis-12817	33	1	3	3	X
fcis-12817	33	2	.	.	PUNCT
fcis-12817	33	3	model	model	NOUN
fcis-12817	33	4	architecture	architecture	NOUN
fcis-12817	33	5	in	in	ADP
fcis-12817	33	6	the	the	DET
fcis-12817	33	7	mrc	mrc	NOUN
fcis-12817	33	8	-	-	PUNCT
fcis-12817	33	9	based	base	VERB
fcis-12817	33	10	chinese	chinese	ADJ
fcis-12817	33	11	entity	entity	NOUN
fcis-12817	33	12	recognition	recognition	NOUN
fcis-12817	33	13	task	task	NOUN
fcis-12817	33	14	,	,	PUNCT
fcis-12817	33	15	when	when	SCONJ
fcis-12817	33	16	given	give	VERB
fcis-12817	33	17	a	a	DET
fcis-12817	33	18	sequence	sequence	NOUN
fcis-12817	33	19	of	of	ADP
fcis-12817	33	20	input	input	NOUN
fcis-12817	33	21	segments	segment	NOUN
fcis-12817	34	1			PROPN
fcis-12817	34	2	1	1	NOUN
fcis-12817	34	3	2	2	NUM
fcis-12817	34	4	,	,	PUNCT
fcis-12817	34	5	,	,	PUNCT
fcis-12817	34	6	,	,	PUNCT
fcis-12817	34	7	nx	nx	X
fcis-12817	34	8	x	x	SYM
fcis-12817	34	9	x	x	X
fcis-12817	34	10	x	x	PROPN
fcis-12817	34	11			NOUN
fcis-12817	34	12	,	,	PUNCT
fcis-12817	34	13	where	where	SCONJ
fcis-12817	34	14	x	x	PRON
fcis-12817	34	15	represents	represent	VERB
fcis-12817	34	16	the	the	DET
fcis-12817	34	17	words	word	NOUN
fcis-12817	34	18	that	that	PRON
fcis-12817	34	19	constitute	constitute	VERB
fcis-12817	34	20	a	a	DET
fcis-12817	34	21	sentence	sentence	NOUN
fcis-12817	34	22	and	and	CCONJ
fcis-12817	34	23	n	n	CCONJ
fcis-12817	34	24	represents	represent	VERB
fcis-12817	34	25	the	the	DET
fcis-12817	34	26	length	length	NOUN
fcis-12817	34	27	of	of	ADP
fcis-12817	34	28	the	the	DET
fcis-12817	34	29	sentence	sentence	NOUN
fcis-12817	34	30	.	.	PUNCT
fcis-12817	35	1	the	the	DET
fcis-12817	35	2	purpose	purpose	NOUN
fcis-12817	35	3	of	of	ADP
fcis-12817	35	4	the	the	DET
fcis-12817	35	5	mrc	mrc	NOUN
fcis-12817	35	6	-	-	PUNCT
fcis-12817	35	7	based	base	VERB
fcis-12817	35	8	entity	entity	NOUN
fcis-12817	35	9	recognition	recognition	NOUN
fcis-12817	35	10	modeling	modeling	NOUN
fcis-12817	35	11	paradigm	paradigm	NOUN
fcis-12817	35	12	is	be	AUX
fcis-12817	35	13	to	to	PART
fcis-12817	35	14	first	first	ADV
fcis-12817	35	15	construct	construct	VERB
fcis-12817	35	16	the	the	DET
fcis-12817	35	17	corresponding	corresponding	ADJ
fcis-12817	35	18	problem	problem	NOUN
fcis-12817	35	19			PROPN
fcis-12817	35	20	1	1	NOUN
fcis-12817	35	21	2	2	NUM
fcis-12817	35	22	,	,	PUNCT
fcis-12817	35	23	,	,	PUNCT
fcis-12817	35	24	,	,	PUNCT
fcis-12817	35	25	mq	mq	PROPN
fcis-12817	35	26	q	q	PROPN
fcis-12817	35	27	q	q	PROPN
fcis-12817	35	28	q	q	PROPN
fcis-12817	35	29			NOUN
fcis-12817	35	30	according	accord	VERB
fcis-12817	35	31	to	to	ADP
fcis-12817	35	32	the	the	DET
fcis-12817	35	33	entity	entity	NOUN
fcis-12817	35	34	types	type	NOUN
fcis-12817	35	35	,	,	PUNCT
fcis-12817	35	36	where	where	SCONJ
fcis-12817	35	37	each	each	DET
fcis-12817	35	38	entity	entity	NOUN
fcis-12817	35	39	type	type	NOUN
fcis-12817	35	40	corresponds	correspond	VERB
fcis-12817	35	41	to	to	ADP
fcis-12817	35	42	one	one	NUM
fcis-12817	35	43	problem	problem	NOUN
fcis-12817	35	44	.	.	PUNCT
fcis-12817	36	1	the	the	DET
fcis-12817	36	2	model	model	NOUN
fcis-12817	36	3	obtains	obtain	VERB
fcis-12817	36	4	the	the	DET
fcis-12817	36	5	answer	answer	NOUN
fcis-12817	36	6	to	to	ADP
fcis-12817	36	7	the	the	DET
fcis-12817	36	8	question	question	NOUN
fcis-12817	36	9	from	from	ADP
fcis-12817	36	10	s	s	PRON
fcis-12817	36	11	based	base	VERB
fcis-12817	36	12	on	on	ADP
fcis-12817	36	13	q	q	PROPN
fcis-12817	36	14	as	as	ADP
fcis-12817	36	15	a	a	DET
fcis-12817	36	16	text	text	NOUN
fcis-12817	36	17	fragment	fragment	NOUN
fcis-12817	36	18			PROPN
fcis-12817	36	19	1	1	NOUN
fcis-12817	36	20	2	2	NUM
fcis-12817	36	21	,	,	PUNCT
fcis-12817	36	22	,	,	PUNCT
fcis-12817	36	23	,	,	PUNCT
fcis-12817	36	24	,	,	PUNCT
fcis-12817	36	25	,	,	PUNCT
fcis-12817	36	26	...	...	PUNCT
fcis-12817	36	27	,	,	PUNCT
fcis-12817	36	28	j	j	PROPN
fcis-12817	36	29	head	head	NOUN
fcis-12817	36	30	tail	tail	NOUN
fcis-12817	36	31	head	head	NOUN
fcis-12817	36	32	tail	tail	NOUN
fcis-12817	36	33	head	head	NOUN
fcis-12817	37	1	taila	taila	ADV
fcis-12817	37	2	x	x	X
fcis-12817	37	3	x	x	X
fcis-12817	37	4	x	x	PROPN
fcis-12817	37	5	in	in	ADP
fcis-12817	37	6	s	s	PRON
fcis-12817	37	7	,	,	PUNCT
fcis-12817	37	8	where	where	SCONJ
fcis-12817	37	9	j	j	PROPN
fcis-12817	37	10	represents	represent	VERB
fcis-12817	37	11	the	the	DET
fcis-12817	37	12	number	number	NOUN
fcis-12817	37	13	of	of	ADP
fcis-12817	37	14	entities	entity	NOUN
fcis-12817	37	15	obtained	obtain	VERB
fcis-12817	37	16	from	from	ADP
fcis-12817	37	17	the	the	DET
fcis-12817	37	18	question	question	NOUN
fcis-12817	37	19	q	q	X
fcis-12817	37	20	,	,	PUNCT
fcis-12817	37	21	and	and	CCONJ
fcis-12817	37	22	head	head	NOUN
fcis-12817	37	23	and	and	CCONJ
fcis-12817	37	24	tail	tail	NOUN
fcis-12817	37	25	represent	represent	VERB
fcis-12817	37	26	the	the	DET
fcis-12817	37	27	corresponding	correspond	VERB
fcis-12817	37	28	positions	position	NOUN
fcis-12817	37	29	of	of	ADP
fcis-12817	37	30	the	the	DET
fcis-12817	37	31	entities	entity	NOUN
fcis-12817	37	32	in	in	ADP
fcis-12817	37	33	the	the	DET
fcis-12817	37	34	original	original	ADJ
fcis-12817	37	35	input	input	NOUN
fcis-12817	37	36	segment	segment	NOUN
fcis-12817	37	37	s	s	PART
fcis-12817	37	38	.	.	PUNCT
fcis-12817	38	1	the	the	DET
fcis-12817	38	2	overall	overall	ADJ
fcis-12817	38	3	structure	structure	NOUN
fcis-12817	38	4	of	of	ADP
fcis-12817	38	5	our	our	PRON
fcis-12817	38	6	proposed	propose	VERB
fcis-12817	38	7	model	model	NOUN
fcis-12817	38	8	is	be	AUX
fcis-12817	38	9	shown	show	VERB
fcis-12817	38	10	in	in	ADP
fcis-12817	38	11	77	77	NUM
fcis-12817	38	12	figure	figure	NOUN
fcis-12817	38	13	1	1	NUM
fcis-12817	38	14	.	.	PUNCT
fcis-12817	39	1	first	first	ADV
fcis-12817	39	2	,	,	PUNCT
fcis-12817	39	3	the	the	DET
fcis-12817	39	4	roberta	roberta	PROPN
fcis-12817	39	5	-	-	PUNCT
fcis-12817	39	6	wwm	wwm	PROPN
fcis-12817	39	7	model	model	NOUN
fcis-12817	39	8	is	be	AUX
fcis-12817	39	9	used	use	VERB
fcis-12817	39	10	to	to	PART
fcis-12817	39	11	encode	encode	VERB
fcis-12817	39	12	the	the	DET
fcis-12817	39	13	context	context	NOUN
fcis-12817	39	14	and	and	CCONJ
fcis-12817	39	15	the	the	DET
fcis-12817	39	16	corresponding	corresponding	ADJ
fcis-12817	39	17	problem	problem	NOUN
fcis-12817	39	18	,	,	PUNCT
fcis-12817	39	19	and	and	CCONJ
fcis-12817	39	20	only	only	ADV
fcis-12817	39	21	the	the	DET
fcis-12817	39	22	output	output	NOUN
fcis-12817	39	23	u	u	NOUN
fcis-12817	39	24	corresponding	correspond	VERB
fcis-12817	39	25	to	to	ADP
fcis-12817	39	26	input	input	NOUN
fcis-12817	39	27	sequence	sequence	NOUN
fcis-12817	39	28	x	x	PRON
fcis-12817	39	29	is	be	AUX
fcis-12817	39	30	retained	retain	VERB
fcis-12817	39	31	,	,	PUNCT
fcis-12817	39	32	and	and	CCONJ
fcis-12817	39	33	then	then	ADV
fcis-12817	39	34	the	the	DET
fcis-12817	39	35	u	u	NOUN
fcis-12817	39	36	is	be	AUX
fcis-12817	39	37	inputted	inputte	VERB
fcis-12817	39	38	to	to	ADP
fcis-12817	39	39	the	the	DET
fcis-12817	39	40	attention	attention	NOUN
fcis-12817	39	41	layer	layer	NOUN
fcis-12817	39	42	,	,	PUNCT
fcis-12817	39	43	and	and	CCONJ
fcis-12817	39	44	the	the	DET
fcis-12817	39	45	corresponding	correspond	VERB
fcis-12817	39	46	context	context	NOUN
fcis-12817	39	47	feature	feature	NOUN
fcis-12817	39	48	h	h	NOUN
fcis-12817	39	49	will	will	AUX
fcis-12817	39	50	be	be	AUX
fcis-12817	39	51	obtained	obtain	VERB
fcis-12817	39	52	finally	finally	ADV
fcis-12817	39	53	.	.	PUNCT
fcis-12817	40	1	then	then	ADV
fcis-12817	40	2	the	the	DET
fcis-12817	40	3	context	context	NOUN
fcis-12817	40	4	featureh	featureh	NOUN
fcis-12817	40	5	is	be	AUX
fcis-12817	40	6	used	use	VERB
fcis-12817	40	7	to	to	PART
fcis-12817	40	8	predict	predict	VERB
fcis-12817	40	9	the	the	DET
fcis-12817	40	10	head	head	NOUN
fcis-12817	40	11	index	index	NOUN
fcis-12817	40	12	position	position	NOUN
fcis-12817	40	13	and	and	CCONJ
fcis-12817	40	14	tail	tail	NOUN
fcis-12817	40	15	index	index	NOUN
fcis-12817	40	16	position	position	NOUN
fcis-12817	40	17	of	of	ADP
fcis-12817	40	18	the	the	DET
fcis-12817	40	19	corresponding	corresponding	ADJ
fcis-12817	40	20	entity	entity	NOUN
fcis-12817	40	21	in	in	ADP
fcis-12817	40	22	x	x	X
fcis-12817	40	23	.	.	PUNCT
fcis-12817	41	1	the	the	DET
fcis-12817	41	2	matching	matching	NOUN
fcis-12817	41	3	module	module	NOUN
fcis-12817	41	4	is	be	AUX
fcis-12817	41	5	used	use	VERB
fcis-12817	41	6	to	to	PART
fcis-12817	41	7	match	match	VERB
fcis-12817	41	8	the	the	DET
fcis-12817	41	9	predicted	predict	VERB
fcis-12817	41	10	head	head	NOUN
fcis-12817	41	11	index	index	NOUN
fcis-12817	41	12	position	position	NOUN
fcis-12817	41	13	and	and	CCONJ
fcis-12817	41	14	tail	tail	NOUN
fcis-12817	41	15	index	index	NOUN
fcis-12817	41	16	position	position	NOUN
fcis-12817	41	17	to	to	PART
fcis-12817	41	18	finally	finally	ADV
fcis-12817	41	19	get	get	VERB
fcis-12817	41	20	the	the	DET
fcis-12817	41	21	entity	entity	NOUN
fcis-12817	41	22	position	position	NOUN
fcis-12817	41	23	corresponding	correspond	VERB
fcis-12817	41	24	to	to	ADP
fcis-12817	41	25	the	the	DET
fcis-12817	41	26	question	question	NOUN
fcis-12817	41	27	in	in	ADP
fcis-12817	41	28	x	x	X
fcis-12817	41	29	.	.	PUNCT
fcis-12817	42	1	figure	figure	NOUN
fcis-12817	42	2	1	1	NUM
fcis-12817	42	3	.	.	PUNCT
fcis-12817	42	4	model	model	NOUN
fcis-12817	42	5	structure	structure	NOUN
fcis-12817	42	6	3.1	3.1	NUM
fcis-12817	42	7	.	.	PUNCT
fcis-12817	43	1	input	input	NOUN
fcis-12817	43	2	representation	representation	NOUN
fcis-12817	43	3	layer	layer	NOUN
fcis-12817	43	4	for	for	ADP
fcis-12817	43	5	a	a	DET
fcis-12817	43	6	given	give	VERB
fcis-12817	43	7	sequence	sequence	NOUN
fcis-12817	43	8	of	of	ADP
fcis-12817	43	9	inputs	input	NOUN
fcis-12817	43	10			PROPN
fcis-12817	43	11	1	1	NOUN
fcis-12817	43	12	2	2	NUM
fcis-12817	43	13	,	,	PUNCT
fcis-12817	43	14	,	,	PUNCT
fcis-12817	43	15	,	,	PUNCT
fcis-12817	43	16	nx	nx	X
fcis-12817	43	17	x	x	SYM
fcis-12817	43	18	x	x	X
fcis-12817	43	19	x	x	PUNCT
fcis-12817	43	20			NOUN
fcis-12817	43	21	and	and	CCONJ
fcis-12817	43	22	a	a	DET
fcis-12817	43	23	problem	problem	NOUN
fcis-12817	43	24	constructed	construct	VERB
fcis-12817	43	25	based	base	VERB
fcis-12817	43	26	on	on	ADP
fcis-12817	43	27	the	the	DET
fcis-12817	43	28	entity	entity	NOUN
fcis-12817	43	29	types	type	NOUN
fcis-12817	44	1			PRON
fcis-12817	44	2	1	1	NOUN
fcis-12817	44	3	2	2	NUM
fcis-12817	44	4	,	,	PUNCT
fcis-12817	44	5	,	,	PUNCT
fcis-12817	44	6	,	,	PUNCT
fcis-12817	44	7	mq	mq	PROPN
fcis-12817	44	8	q	q	PROPN
fcis-12817	44	9	q	q	PROPN
fcis-12817	44	10	q	q	PROPN
fcis-12817	44	11			NOUN
fcis-12817	44	12	,	,	PUNCT
fcis-12817	44	13	we	we	PRON
fcis-12817	44	14	utilize	utilize	VERB
fcis-12817	44	15	both	both	PRON
fcis-12817	44	16	to	to	PART
fcis-12817	44	17	construct	construct	VERB
fcis-12817	44	18	the	the	DET
fcis-12817	44	19	input	input	NOUN
fcis-12817	44	20	representation	representation	NOUN
fcis-12817	44	21			PROPN
fcis-12817	44	22			ADP
fcis-12817	44	23			ADJ
fcis-12817	44	24			NOUN
fcis-12817	44	25	1	1	NOUN
fcis-12817	44	26	2	2	NUM
fcis-12817	44	27	1	1	NUM
fcis-12817	44	28	2	2	NUM
fcis-12817	44	29	,	,	PUNCT
fcis-12817	44	30	,	,	PUNCT
fcis-12817	44	31	,	,	PUNCT
fcis-12817	44	32	,	,	PUNCT
fcis-12817	44	33	,	,	PUNCT
fcis-12817	44	34	,	,	PUNCT
fcis-12817	44	35	,	,	PUNCT
fcis-12817	44	36	,	,	PUNCT
fcis-12817	44	37	...	...	PUNCT
fcis-12817	44	38	,	,	PUNCT
fcis-12817	44	39	n	n	PROPN
fcis-12817	44	40	mi	mi	PROPN
fcis-12817	44	41	cls	cls	NOUN
fcis-12817	44	42	x	x	PUNCT
fcis-12817	44	43	x	x	PUNCT
fcis-12817	44	44	x	x	PUNCT
fcis-12817	44	45	sep	sep	PROPN
fcis-12817	44	46	q	q	PROPN
fcis-12817	44	47	q	q	PROPN
fcis-12817	44	48	q	q	PROPN
fcis-12817	44	49			NOUN
fcis-12817	44	50	to	to	ADP
fcis-12817	44	51	the	the	DET
fcis-12817	44	52	roberta	roberta	PROPN
fcis-12817	44	53	-	-	PUNCT
fcis-12817	44	54	wmm	wmm	PROPN
fcis-12817	44	55	pre	pre	ADJ
fcis-12817	44	56	-	-	ADJ
fcis-12817	44	57	trained	train	VERB
fcis-12817	44	58	model	model	NOUN
fcis-12817	44	59	,	,	PUNCT
fcis-12817	44	60	where	where	SCONJ
fcis-12817	44	61	[	[	X
fcis-12817	44	62	cls	cls	NOUN
fcis-12817	44	63	]	]	PUNCT
fcis-12817	44	64	and	and	CCONJ
fcis-12817	44	65	[	[	X
fcis-12817	44	66	sep	sep	X
fcis-12817	44	67	]	]	X
fcis-12817	44	68	denote	denote	VERB
fcis-12817	44	69	special	special	ADJ
fcis-12817	44	70	tokens	token	NOUN
fcis-12817	44	71	.	.	PUNCT
fcis-12817	45	1	3.2	3.2	NUM
fcis-12817	45	2	.	.	X
fcis-12817	45	3	encoding	encode	VERB
fcis-12817	45	4	layer	layer	NOUN
fcis-12817	45	5	after	after	ADP
fcis-12817	45	6	roberta	roberta	PROPN
fcis-12817	45	7	-	-	PUNCT
fcis-12817	45	8	wmm	wmm	PROPN
fcis-12817	45	9	encoding	encoding	NOUN
fcis-12817	45	10	,	,	PUNCT
fcis-12817	45	11	we	we	PRON
fcis-12817	45	12	get	get	VERB
fcis-12817	45	13	the	the	DET
fcis-12817	45	14	corresponding	corresponding	ADJ
fcis-12817	45	15	outputv	outputv	NOUN
fcis-12817	45	16	that	that	PRON
fcis-12817	45	17	contains	contain	VERB
fcis-12817	45	18	the	the	DET
fcis-12817	45	19	understanding	understanding	NOUN
fcis-12817	45	20	of	of	ADP
fcis-12817	45	21	the	the	DET
fcis-12817	45	22	problem	problem	NOUN
fcis-12817	45	23	.	.	PUNCT
fcis-12817	46	1	we	we	PRON
fcis-12817	46	2	only	only	ADV
fcis-12817	46	3	keep	keep	VERB
fcis-12817	46	4	the	the	DET
fcis-12817	46	5	features	feature	NOUN
fcis-12817	46	6	n	n	X
fcis-12817	46	7	du	du	VERB
fcis-12817	46	8			PROPN
fcis-12817	46	9	that	that	PRON
fcis-12817	46	10	are	be	AUX
fcis-12817	46	11	related	relate	VERB
fcis-12817	46	12	to	to	ADP
fcis-12817	46	13	the	the	DET
fcis-12817	46	14	context	context	NOUN
fcis-12817	46	15	of	of	ADP
fcis-12817	46	16	the	the	DET
fcis-12817	46	17	input	input	NOUN
fcis-12817	46	18	sequence	sequence	NOUN
fcis-12817	46	19	in	in	ADP
fcis-12817	46	20	v.	v.	NOUN
fcis-12817	46	21	to	to	PART
fcis-12817	46	22	further	far	ADV
fcis-12817	46	23	utilize	utilize	VERB
fcis-12817	46	24	the	the	DET
fcis-12817	46	25	contextual	contextual	ADJ
fcis-12817	46	26	information	information	NOUN
fcis-12817	46	27	,	,	PUNCT
fcis-12817	46	28	the	the	DET
fcis-12817	46	29	u	u	NOUN
fcis-12817	46	30	is	be	AUX
fcis-12817	46	31	sent	send	VERB
fcis-12817	46	32	to	to	ADP
fcis-12817	46	33	the	the	DET
fcis-12817	46	34	attention	attention	NOUN
fcis-12817	46	35	network	network	NOUN
fcis-12817	46	36	,	,	PUNCT
fcis-12817	46	37	where	where	SCONJ
fcis-12817	46	38	the	the	DET
fcis-12817	46	39	formula	formula	NOUN
fcis-12817	46	40	for	for	ADP
fcis-12817	46	41	attention	attention	NOUN
fcis-12817	46	42	is	be	AUX
fcis-12817	46	43	shown	show	VERB
fcis-12817	46	44	below	below	ADP
fcis-12817	46	45	:	:	PUNCT
fcis-12817	46	46			NUM
fcis-12817	46	47	attention	attention	NOUN
fcis-12817	46	48	,	,	PUNCT
fcis-12817	46	49	,	,	PUNCT
fcis-12817	46	50	t	t	PROPN
fcis-12817	47	1	k	k	PROPN
fcis-12817	47	2	qk	qk	PROPN
fcis-12817	47	3	q	q	PROPN
fcis-12817	47	4	k	k	PROPN
fcis-12817	47	5	v	v	X
fcis-12817	47	6	softmax	softmax	NOUN
fcis-12817	47	7	v	v	NOUN
fcis-12817	48	1	d	d	PROPN
fcis-12817	48	2			NOUN
fcis-12817	48	3			NOUN
fcis-12817	48	4			PROPN
fcis-12817	48	5			PROPN
fcis-12817	49	1			NOUN
fcis-12817	50	1			PROPN
fcis-12817	51	1			ADJ
fcis-12817	51	2			NOUN
fcis-12817	51	3	(	(	PUNCT
fcis-12817	51	4	1	1	NUM
fcis-12817	51	5	)	)	PUNCT
fcis-12817	51	6	q	q	NOUN
fcis-12817	51	7	,	,	PUNCT
fcis-12817	51	8	k	k	NOUN
fcis-12817	51	9	,	,	PUNCT
fcis-12817	51	10	v	v	NUM
fcis-12817	51	11	refer	refer	VERB
fcis-12817	51	12	to	to	ADP
fcis-12817	51	13	the	the	DET
fcis-12817	51	14	query	query	NOUN
fcis-12817	51	15	matrix	matrix	NOUN
fcis-12817	51	16	,	,	PUNCT
fcis-12817	51	17	key	key	ADJ
fcis-12817	51	18	matrix	matrix	NOUN
fcis-12817	51	19	,	,	PUNCT
fcis-12817	51	20	and	and	CCONJ
fcis-12817	51	21	value	value	NOUN
fcis-12817	51	22	matrix	matrix	NOUN
fcis-12817	51	23	obtained	obtain	VERB
fcis-12817	51	24	fromu	fromu	NOUN
fcis-12817	51	25	after	after	ADP
fcis-12817	51	26	transformation	transformation	NOUN
fcis-12817	51	27	,	,	PUNCT
fcis-12817	51	28	respectively	respectively	ADV
fcis-12817	51	29	.	.	PUNCT
fcis-12817	52	1	kd	kd	PROPN
fcis-12817	52	2	represents	represent	VERB
fcis-12817	52	3	the	the	DET
fcis-12817	52	4	dimension	dimension	NOUN
fcis-12817	52	5	of	of	ADP
fcis-12817	52	6	the	the	DET
fcis-12817	52	7	input	input	NOUN
fcis-12817	52	8	vector	vector	NOUN
fcis-12817	52	9	.	.	PUNCT
fcis-12817	53	1	eventually	eventually	ADV
fcis-12817	53	2	,	,	PUNCT
fcis-12817	53	3	after	after	ADP
fcis-12817	53	4	encoding	encode	VERB
fcis-12817	53	5	the	the	DET
fcis-12817	53	6	context	context	NOUN
fcis-12817	53	7	features	feature	NOUN
fcis-12817	53	8	by	by	ADP
fcis-12817	53	9	using	use	VERB
fcis-12817	53	10	attention	attention	NOUN
fcis-12817	53	11	to	to	ADP
fcis-12817	53	12	them	they	PRON
fcis-12817	53	13	,	,	PUNCT
fcis-12817	53	14	we	we	PRON
fcis-12817	53	15	obtain	obtain	VERB
fcis-12817	53	16	the	the	DET
fcis-12817	53	17	relevant	relevant	ADJ
fcis-12817	53	18	encoded	encode	VERB
fcis-12817	53	19	representation	representation	NOUN
fcis-12817	53	20	of	of	ADP
fcis-12817	53	21	the	the	DET
fcis-12817	53	22	sentence	sentence	NOUN
fcis-12817	53	23			PROPN
fcis-12817	53	24	1	1	NOUN
fcis-12817	53	25	2	2	NUM
fcis-12817	53	26	,	,	PUNCT
fcis-12817	53	27	,	,	PUNCT
fcis-12817	53	28	,	,	PUNCT
fcis-12817	53	29	nh	nh	PROPN
fcis-12817	53	30	h	h	NOUN
fcis-12817	53	31	h	h	NOUN
fcis-12817	53	32	h	h	PROPN
fcis-12817	53	33			NOUN
fcis-12817	53	34	,	,	PUNCT
fcis-12817	53	35	where	where	SCONJ
fcis-12817	53	36	n	n	VERB
fcis-12817	53	37	dh	dh	NOUN
fcis-12817	53	38			X
fcis-12817	53	39	.	.	PUNCT
fcis-12817	54	1	3.3	3.3	NUM
fcis-12817	54	2	.	.	PUNCT
fcis-12817	55	1	entity	entity	NOUN
fcis-12817	55	2	head	head	NOUN
fcis-12817	55	3	prediction	prediction	NOUN
fcis-12817	55	4	and	and	CCONJ
fcis-12817	55	5	tail	tail	NOUN
fcis-12817	55	6	prediction	prediction	NOUN
fcis-12817	55	7	layers	layer	NOUN
fcis-12817	55	8	the	the	DET
fcis-12817	55	9	model	model	NOUN
fcis-12817	55	10	utilizes	utilize	VERB
fcis-12817	55	11	the	the	DET
fcis-12817	55	12	contextual	contextual	ADJ
fcis-12817	55	13	featureh	featureh	NOUN
fcis-12817	55	14	to	to	PART
fcis-12817	55	15	predict	predict	VERB
fcis-12817	55	16	the	the	DET
fcis-12817	55	17	probability	probability	NOUN
fcis-12817	55	18	of	of	ADP
fcis-12817	55	19	each	each	PRON
fcis-12817	55	20	token	token	VERB
fcis-12817	55	21	in	in	ADP
fcis-12817	55	22	the	the	DET
fcis-12817	55	23	entire	entire	ADJ
fcis-12817	55	24	input	input	NOUN
fcis-12817	55	25	sequence	sequence	NOUN
fcis-12817	55	26	as	as	ADP
fcis-12817	55	27	the	the	DET
fcis-12817	55	28	head	head	NOUN
fcis-12817	55	29	or	or	CCONJ
fcis-12817	55	30	the	the	DET
fcis-12817	55	31	tail	tail	NOUN
fcis-12817	55	32	of	of	ADP
fcis-12817	55	33	the	the	DET
fcis-12817	55	34	entity	entity	NOUN
fcis-12817	55	35	,	,	PUNCT
fcis-12817	55	36	respectively	respectively	ADV
fcis-12817	55	37	.	.	PUNCT
fcis-12817	56	1	the	the	DET
fcis-12817	56	2	specific	specific	ADJ
fcis-12817	56	3	formula	formula	NOUN
fcis-12817	56	4	is	be	AUX
fcis-12817	56	5	shown	show	VERB
fcis-12817	56	6	below	below	ADP
fcis-12817	56	7	.	.	PUNCT
fcis-12817	57	1	1	1	NUM
fcis-12817	57	2	1z	1z	NUM
fcis-12817	57	3	hw	hw	NOUN
fcis-12817	57	4	(	(	PUNCT
fcis-12817	57	5	2	2	NUM
fcis-12817	57	6	)	)	SYM
fcis-12817	57	7	2	2	NUM
fcis-12817	57	8	2z	2z	NUM
fcis-12817	57	9	hw	hw	NOUN
fcis-12817	57	10	(	(	PUNCT
fcis-12817	57	11	3	3	NUM
fcis-12817	57	12	)	)	PUNCT
fcis-12817	57	13	where	where	SCONJ
fcis-12817	57	14	2	2	NUM
fcis-12817	57	15	1	1	NUM
fcis-12817	57	16	dw	dw	NOUN
fcis-12817	57	17			PROPN
fcis-12817	57	18	and	and	CCONJ
fcis-12817	57	19	2	2	NUM
fcis-12817	57	20	2	2	NUM
fcis-12817	57	21	dw	dw	NOUN
fcis-12817	57	22			NOUN
fcis-12817	57	23	are	be	AUX
fcis-12817	57	24	trainable	trainable	ADJ
fcis-12817	57	25	parameters	parameter	NOUN
fcis-12817	57	26	,	,	PUNCT
fcis-12817	57	27	2	2	NUM
fcis-12817	57	28	1	1	NUM
fcis-12817	57	29	nz	nz	NOUN
fcis-12817	57	30			PRON
fcis-12817	57	31	and	and	CCONJ
fcis-12817	57	32	2	2	NUM
fcis-12817	57	33	2	2	NUM
fcis-12817	57	34	nz	nz	NOUN
fcis-12817	57	35			NOUN
fcis-12817	57	36	are	be	AUX
fcis-12817	57	37	entity	entity	NOUN
fcis-12817	57	38	head	head	NOUN
fcis-12817	57	39	prediction	prediction	NOUN
fcis-12817	57	40	features	feature	NOUN
fcis-12817	57	41	and	and	CCONJ
fcis-12817	57	42	entity	entity	NOUN
fcis-12817	57	43	tail	tail	NOUN
fcis-12817	57	44	prediction	prediction	NOUN
fcis-12817	57	45	features	feature	NOUN
fcis-12817	57	46	,	,	PUNCT
fcis-12817	57	47	respectively	respectively	ADV
fcis-12817	57	48	.	.	PUNCT
fcis-12817	58	1	for	for	ADP
fcis-12817	58	2	each	each	DET
fcis-12817	58	3	line	line	NOUN
fcis-12817	58	4	of	of	ADP
fcis-12817	58	5	1z	1z	NOUN
fcis-12817	58	6	and	and	CCONJ
fcis-12817	58	7	2z	2z	NUM
fcis-12817	58	8	,	,	PUNCT
fcis-12817	58	9	softmax	softmax	NOUN
fcis-12817	58	10	function	function	NOUN
fcis-12817	58	11	is	be	AUX
fcis-12817	58	12	used	use	VERB
fcis-12817	58	13	to	to	PART
fcis-12817	58	14	get	get	VERB
fcis-12817	58	15	the	the	DET
fcis-12817	58	16	probability	probability	NOUN
fcis-12817	58	17	of	of	ADP
fcis-12817	58	18	entity	entity	NOUN
fcis-12817	58	19	head	head	NOUN
fcis-12817	58	20	prediction	prediction	NOUN
fcis-12817	58	21	headp	headp	NOUN
fcis-12817	58	22	and	and	CCONJ
fcis-12817	58	23	entity	entity	NOUN
fcis-12817	58	24	tail	tail	NOUN
fcis-12817	58	25	prediction	prediction	NOUN
fcis-12817	58	26	tailp	tailp	PROPN
fcis-12817	58	27	.	.	PUNCT
fcis-12817	59	1			NOUN
fcis-12817	59	2			PROPN
fcis-12817	59	3	row	row	NOUN
fcis-12817	59	4	1head	1head	NUM
fcis-12817	59	5	eachp	eachp	NOUN
fcis-12817	59	6	softmax	softmax	PROPN
fcis-12817	59	7	z	z	PROPN
fcis-12817	59	8	(	(	PUNCT
fcis-12817	59	9	4	4	NUM
fcis-12817	59	10	)	)	PUNCT
fcis-12817	59	11			NOUN
fcis-12817	59	12			PROPN
fcis-12817	59	13	row	row	NOUN
fcis-12817	59	14	2tail	2tail	NUM
fcis-12817	59	15	eachp	eachp	NOUN
fcis-12817	59	16	softmax	softmax	NOUN
fcis-12817	59	17	z	z	NOUN
fcis-12817	59	18	(	(	PUNCT
fcis-12817	59	19	5	5	NUM
fcis-12817	59	20	)	)	PUNCT
fcis-12817	59	21	eventually	eventually	ADV
fcis-12817	59	22	we	we	PRON
fcis-12817	59	23	get	get	VERB
fcis-12817	59	24	the	the	DET
fcis-12817	59	25	probability	probability	NOUN
fcis-12817	59	26	that	that	SCONJ
fcis-12817	59	27	each	each	PRON
fcis-12817	59	28	token	token	VERB
fcis-12817	59	29	in	in	ADP
fcis-12817	59	30	the	the	DET
fcis-12817	59	31	input	input	NOUN
fcis-12817	59	32	sequence	sequence	NOUN
fcis-12817	59	33	corresponds	correspond	VERB
fcis-12817	59	34	to	to	ADP
fcis-12817	59	35	the	the	DET
fcis-12817	59	36	head	head	NOUN
fcis-12817	59	37	and	and	CCONJ
fcis-12817	59	38	tail	tail	NOUN
fcis-12817	59	39	of	of	ADP
fcis-12817	59	40	the	the	DET
fcis-12817	59	41	entity	entity	NOUN
fcis-12817	59	42	.	.	PUNCT
fcis-12817	60	1	in	in	ADP
fcis-12817	60	2	order	order	NOUN
fcis-12817	60	3	to	to	PART
fcis-12817	60	4	obtain	obtain	VERB
fcis-12817	60	5	the	the	DET
fcis-12817	60	6	head	head	NOUN
fcis-12817	60	7	position	position	NOUN
fcis-12817	60	8	prediction	prediction	NOUN
fcis-12817	60	9	of	of	ADP
fcis-12817	60	10	the	the	DET
fcis-12817	60	11	entity	entity	NOUN
fcis-12817	60	12	and	and	CCONJ
fcis-12817	60	13	the	the	DET
fcis-12817	60	14	tail	tail	NOUN
fcis-12817	60	15	prediction	prediction	NOUN
fcis-12817	60	16	result	result	NOUN
fcis-12817	60	17	of	of	ADP
fcis-12817	60	18	the	the	DET
fcis-12817	60	19	entity	entity	NOUN
fcis-12817	60	20	,	,	PUNCT
fcis-12817	60	21	we	we	PRON
fcis-12817	60	22	apply	apply	VERB
fcis-12817	60	23	the	the	DET
fcis-12817	60	24	following	following	ADJ
fcis-12817	60	25	formulas	formula	NOUN
fcis-12817	60	26	on	on	ADP
fcis-12817	60	27	the	the	DET
fcis-12817	60	28	corresponding	correspond	VERB
fcis-12817	60	29	prediction	prediction	NOUN
fcis-12817	60	30	probabilities	probability	NOUN
fcis-12817	60	31	of	of	ADP
fcis-12817	60	32	each	each	PRON
fcis-12817	60	33	token	token	VERB
fcis-12817	60	34	separately	separately	ADV
fcis-12817	60	35	:	:	PUNCT
fcis-12817	60	36			NOUN
fcis-12817	60	37			PUNCT
fcis-12817	60	38	}	}	PUNCT
fcis-12817	60	39	ˆ	ˆ	PROPN
fcis-12817	60	40	{	{	PUNCT
fcis-12817	60	41	|	|	ADV
fcis-12817	60	42	arg	arg	NOUN
fcis-12817	60	43	,	,	PUNCT
fcis-12817	60	44	1	1	NUM
fcis-12817	60	45	,	,	PUNCT
fcis-12817	60	46	...	...	PUNCT
fcis-12817	60	47	,	,	PUNCT
fcis-12817	60	48	i	i	PRON
fcis-12817	60	49	head	head	VERB
fcis-12817	60	50	headk	headk	NOUN
fcis-12817	61	1	i	i	PRON
fcis-12817	61	2	max	max	PROPN
fcis-12817	62	1	p	p	PROPN
fcis-12817	62	2	i	i	PRON
fcis-12817	62	3	n	n	NOUN
fcis-12817	63	1			NOUN
fcis-12817	63	2	(	(	PUNCT
fcis-12817	63	3	6	6	NUM
fcis-12817	63	4	)	)	PUNCT
fcis-12817	63	5			NOUN
fcis-12817	63	6			PUNCT
fcis-12817	63	7	}	}	PUNCT
fcis-12817	63	8	ˆ	ˆ	PROPN
fcis-12817	63	9	{	{	PUNCT
fcis-12817	63	10	|	|	ADV
fcis-12817	63	11	arg	arg	NOUN
fcis-12817	63	12	,	,	PUNCT
fcis-12817	63	13	1	1	NUM
fcis-12817	63	14	,	,	PUNCT
fcis-12817	63	15	...	...	PUNCT
fcis-12817	63	16	,	,	PUNCT
fcis-12817	63	17	j	j	PROPN
fcis-12817	63	18	tail	tail	PROPN
fcis-12817	63	19	tailk	tailk	NOUN
fcis-12817	64	1	j	j	PROPN
fcis-12817	64	2	max	max	PROPN
fcis-12817	64	3	p	p	PROPN
fcis-12817	64	4	j	j	PROPN
fcis-12817	65	1	n	n	NOUN
fcis-12817	65	2			PROPN
fcis-12817	65	3	(	(	PUNCT
fcis-12817	65	4	7	7	NUM
fcis-12817	65	5	)	)	PUNCT
fcis-12817	65	6	where	where	SCONJ
fcis-12817	65	7	i	i	PRON
fcis-12817	65	8	headp	headp	NOUN
fcis-12817	65	9	represents	represent	VERB
fcis-12817	65	10	the	the	DET
fcis-12817	65	11	i	i	PROPN
fcis-12817	65	12	row	row	NOUN
fcis-12817	65	13	of	of	ADP
fcis-12817	65	14	the	the	DET
fcis-12817	65	15	headp	headp	ADJ
fcis-12817	65	16	probability	probability	NOUN
fcis-12817	65	17	matrix	matrix	NOUN
fcis-12817	65	18	and	and	CCONJ
fcis-12817	65	19	j	j	PROPN
fcis-12817	65	20	tailp	tailp	PROPN
fcis-12817	65	21	represents	represent	VERB
fcis-12817	65	22	the	the	DET
fcis-12817	65	23	j	j	PROPN
fcis-12817	65	24	row	row	NOUN
fcis-12817	65	25	of	of	ADP
fcis-12817	65	26	the	the	DET
fcis-12817	65	27	tailp	tailp	ADJ
fcis-12817	65	28	probability	probability	NOUN
fcis-12817	65	29	matrix	matrix	NOUN
fcis-12817	65	30	.	.	PUNCT
fcis-12817	66	1	ˆ	ˆ	ADJ
fcis-12817	66	2	headk	headk	NOUN
fcis-12817	66	3	represents	represent	VERB
fcis-12817	66	4	the	the	DET
fcis-12817	66	5	index	index	NOUN
fcis-12817	66	6	of	of	ADP
fcis-12817	66	7	the	the	DET
fcis-12817	66	8	position	position	NOUN
fcis-12817	66	9	in	in	ADP
fcis-12817	66	10	the	the	DET
fcis-12817	66	11	sentence	sentence	NOUN
fcis-12817	66	12	that	that	PRON
fcis-12817	66	13	is	be	AUX
fcis-12817	66	14	likely	likely	ADJ
fcis-12817	66	15	to	to	PART
fcis-12817	66	16	be	be	AUX
fcis-12817	66	17	the	the	DET
fcis-12817	66	18	head	head	NOUN
fcis-12817	66	19	of	of	ADP
fcis-12817	66	20	the	the	DET
fcis-12817	66	21	entity	entity	NOUN
fcis-12817	66	22	,	,	PUNCT
fcis-12817	66	23	and	and	CCONJ
fcis-12817	66	24	ˆ	ˆ	ADJ
fcis-12817	66	25	headk	headk	NOUN
fcis-12817	66	26	represents	represent	VERB
fcis-12817	66	27	the	the	DET
fcis-12817	66	28	index	index	NOUN
fcis-12817	66	29	of	of	ADP
fcis-12817	66	30	the	the	DET
fcis-12817	66	31	position	position	NOUN
fcis-12817	66	32	in	in	ADP
fcis-12817	66	33	the	the	DET
fcis-12817	66	34	input	input	NOUN
fcis-12817	66	35	sentence	sentence	NOUN
fcis-12817	66	36	that	that	PRON
fcis-12817	66	37	is	be	AUX
fcis-12817	66	38	likely	likely	ADJ
fcis-12817	66	39	to	to	PART
fcis-12817	66	40	be	be	AUX
fcis-12817	66	41	the	the	DET
fcis-12817	66	42	tail	tail	NOUN
fcis-12817	66	43	of	of	ADP
fcis-12817	66	44	the	the	DET
fcis-12817	66	45	entity	entity	NOUN
fcis-12817	66	46	.	.	PUNCT
fcis-12817	67	1	3.4	3.4	NUM
fcis-12817	67	2	.	.	PUNCT
fcis-12817	67	3	entity	entity	NOUN
fcis-12817	67	4	head	head	NOUN
fcis-12817	67	5	and	and	CCONJ
fcis-12817	67	6	tail	tail	NOUN
fcis-12817	67	7	matching	matching	NOUN
fcis-12817	67	8	layer	layer	NOUN
fcis-12817	67	9	in	in	ADP
fcis-12817	67	10	this	this	DET
fcis-12817	67	11	stage	stage	NOUN
fcis-12817	67	12	,	,	PUNCT
fcis-12817	67	13	we	we	PRON
fcis-12817	67	14	combine	combine	VERB
fcis-12817	67	15	and	and	CCONJ
fcis-12817	67	16	match	match	VERB
fcis-12817	67	17	all	all	DET
fcis-12817	67	18	the	the	DET
fcis-12817	67	19	head	head	NOUN
fcis-12817	67	20	indexes	index	VERB
fcis-12817	67	21	headi	headi	ADJ
fcis-12817	67	22	in	in	ADP
fcis-12817	67	23	the	the	DET
fcis-12817	67	24	entity	entity	NOUN
fcis-12817	67	25	head	head	NOUN
fcis-12817	67	26	collection	collection	NOUN
fcis-12817	67	27	ˆ	ˆ	ADJ
fcis-12817	67	28	headk	headk	NOUN
fcis-12817	67	29	and	and	CCONJ
fcis-12817	67	30	all	all	DET
fcis-12817	67	31	the	the	DET
fcis-12817	67	32	tail	tail	NOUN
fcis-12817	67	33	indexes	index	NOUN
fcis-12817	67	34	tailj	tailj	VERB
fcis-12817	67	35	in	in	ADP
fcis-12817	67	36	the	the	DET
fcis-12817	67	37	entity	entity	NOUN
fcis-12817	67	38	tail	tail	NOUN
fcis-12817	67	39	index	index	NOUN
fcis-12817	67	40	collection	collection	NOUN
fcis-12817	67	41	ˆ	ˆ	ADJ
fcis-12817	67	42	tailk	tailk	NOUN
fcis-12817	67	43	.	.	PUNCT
fcis-12817	68	1	to	to	PART
fcis-12817	68	2	get	get	VERB
fcis-12817	68	3	the	the	DET
fcis-12817	68	4	probability	probability	NOUN
fcis-12817	68	5	of	of	ADP
fcis-12817	68	6	predicted	predict	VERB
fcis-12817	68	7	entity	entity	NOUN
fcis-12817	68	8	head	head	NOUN
fcis-12817	68	9	and	and	CCONJ
fcis-12817	68	10	tail	tail	NOUN
fcis-12817	68	11	matching	matching	NOUN
fcis-12817	68	12	,	,	PUNCT
fcis-12817	68	13	we	we	PRON
fcis-12817	68	14	use	use	VERB
fcis-12817	68	15	binary	binary	ADJ
fcis-12817	68	16	classification	classification	NOUN
fcis-12817	68	17	as	as	SCONJ
fcis-12817	68	18	follows	follow	VERB
fcis-12817	68	19	:	:	PUNCT
fcis-12817	69	1			PROPN
fcis-12817	69	2			PROPN
fcis-12817	69	3			PROPN
fcis-12817	69	4	,	,	PUNCT
fcis-12817	69	5	,	,	PUNCT
fcis-12817	69	6	head	head	NOUN
fcis-12817	69	7	tail	tail	NOUN
fcis-12817	69	8	head	head	NOUN
fcis-12817	69	9	taili	taili	NOUN
fcis-12817	69	10	j	j	PROPN
fcis-12817	70	1	i	i	PRON
fcis-12817	70	2	jp	jp	VERB
fcis-12817	70	3	sigmoid	sigmoid	VERB
fcis-12817	70	4	w	w	PROPN
fcis-12817	70	5	concat	concat	PROPN
fcis-12817	70	6	h	h	NOUN
fcis-12817	70	7	h	h	ADJ
fcis-12817	70	8			PROPN
fcis-12817	70	9	(	(	PUNCT
fcis-12817	70	10	8)	8)	NUM
fcis-12817	70	11	1	1	NUM
fcis-12817	70	12	2dw	2dw	ADJ
fcis-12817	70	13			NOUN
fcis-12817	70	14	are	be	AUX
fcis-12817	70	15	the	the	DET
fcis-12817	70	16	trainable	trainable	ADJ
fcis-12817	70	17	weights	weight	NOUN
fcis-12817	70	18	.	.	PUNCT
fcis-12817	71	1	4	4	X
fcis-12817	71	2	.	.	X
fcis-12817	71	3	training	training	NOUN
fcis-12817	71	4	and	and	CCONJ
fcis-12817	71	5	testing	test	VERB
fcis-12817	71	6	4.1	4.1	NUM
fcis-12817	71	7	.	.	PUNCT
fcis-12817	72	1	training	training	NOUN
fcis-12817	72	2	in	in	ADP
fcis-12817	72	3	the	the	DET
fcis-12817	72	4	training	training	NOUN
fcis-12817	72	5	phase	phase	NOUN
fcis-12817	72	6	,	,	PUNCT
fcis-12817	72	7	when	when	SCONJ
fcis-12817	72	8	given	give	VERB
fcis-12817	72	9	s	s	PRON
fcis-12817	72	10	training	training	NOUN
fcis-12817	72	11	samples	sample	NOUN
fcis-12817	72	12			NOUN
fcis-12817	72	13			SYM
fcis-12817	72	14			NOUN
fcis-12817	72	15			SYM
fcis-12817	72	16			NOUN
fcis-12817	72	17			SYM
fcis-12817	72	18			NOUN
fcis-12817	72	19			NUM
fcis-12817	72	20	'	'	PUNCT
fcis-12817	72	21	'	'	PUNCT
fcis-12817	73	1	[	[	X
fcis-12817	73	2	(	(	PUNCT
fcis-12817	73	3	;	;	PUNCT
fcis-12817	73	4	)	)	PUNCT
fcis-12817	73	5	,	,	PUNCT
fcis-12817	73	6	,	,	PUNCT
fcis-12817	73	7	(	(	PUNCT
fcis-12817	73	8	;	;	PUNCT
fcis-12817	73	9	)	)	PUNCT
fcis-12817	73	10	]	]	X
fcis-12817	73	11	i	i	PRON
fcis-12817	73	12	i	i	PRON
fcis-12817	73	13	s	s	VERB
fcis-12817	73	14	sx	sx	PROPN
fcis-12817	73	15	y	y	PROPN
fcis-12817	73	16	x	x	PROPN
fcis-12817	73	17	y	y	PROPN
fcis-12817	73	18	.	.	PUNCT
fcis-12817	74	1	we	we	PRON
fcis-12817	74	2	use	use	VERB
fcis-12817	74	3	the	the	DET
fcis-12817	74	4	cross	cross	ADJ
fcis-12817	74	5	-	-	ADJ
fcis-12817	74	6	entropy	entropy	ADJ
fcis-12817	74	7	function	function	NOUN
fcis-12817	74	8	as	as	ADP
fcis-12817	74	9	the	the	DET
fcis-12817	74	10	loss	loss	NOUN
fcis-12817	74	11	function	function	NOUN
fcis-12817	74	12	.	.	PUNCT
fcis-12817	75	1	the	the	DET
fcis-12817	75	2	cross	cross	ADJ
fcis-12817	75	3	-	-	ADJ
fcis-12817	75	4	entropy	entropy	ADJ
fcis-12817	75	5	loss	loss	NOUN
fcis-12817	75	6	function	function	NOUN
fcis-12817	75	7	is	be	AUX
fcis-12817	75	8	shown	show	VERB
fcis-12817	75	9	as	as	SCONJ
fcis-12817	75	10	follows	follow	VERB
fcis-12817	75	11	.	.	PUNCT
fcis-12817	76	1	1	1	NUM
fcis-12817	76	2	[	[	PUNCT
fcis-12817	76	3	log	log	NOUN
fcis-12817	76	4	(	(	PUNCT
fcis-12817	76	5	)	)	PUNCT
fcis-12817	76	6	(	(	PUNCT
fcis-12817	76	7	1	1	X
fcis-12817	76	8	)	)	PUNCT
fcis-12817	76	9	log(1	log(1	NOUN
fcis-12817	76	10	)	)	PUNCT
fcis-12817	77	1	]	]	PUNCT
fcis-12817	77	2	i	i	PRON
fcis-12817	78	1	i	i	PRON
fcis-12817	78	2	i	i	PRON
fcis-12817	79	1	i	i	PRON
fcis-12817	80	1	i	i	VERB
fcis-12817	80	2	l	l	VERB
fcis-12817	81	1	y	y	PROPN
fcis-12817	81	2	p	p	X
fcis-12817	81	3	y	y	PROPN
fcis-12817	81	4	p	p	X
fcis-12817	81	5	s	s	PART
fcis-12817	81	6			PROPN
fcis-12817	81	7			PROPN
fcis-12817	81	8			PROPN
fcis-12817	81	9			PUNCT
fcis-12817	81	10			NOUN
fcis-12817	81	11			PROPN
fcis-12817	81	12			NOUN
fcis-12817	81	13	(	(	PUNCT
fcis-12817	81	14	9	9	NUM
fcis-12817	81	15	)	)	PUNCT
fcis-12817	81	16	we	we	PRON
fcis-12817	81	17	utilize	utilize	VERB
fcis-12817	81	18	the	the	DET
fcis-12817	81	19	cross	cross	ADJ
fcis-12817	81	20	-	-	ADJ
fcis-12817	81	21	entropy	entropy	ADJ
fcis-12817	81	22	loss	loss	NOUN
fcis-12817	81	23	function	function	NOUN
fcis-12817	81	24	to	to	PART
fcis-12817	81	25	calculate	calculate	VERB
fcis-12817	81	26	the	the	DET
fcis-12817	81	27	difference	difference	NOUN
fcis-12817	81	28	between	between	ADP
fcis-12817	81	29	the	the	DET
fcis-12817	81	30	predicted	predict	VERB
fcis-12817	81	31	probability	probability	NOUN
fcis-12817	81	32	and	and	CCONJ
fcis-12817	81	33	the	the	DET
fcis-12817	81	34	true	true	ADJ
fcis-12817	81	35	label	label	NOUN
fcis-12817	81	36	,	,	PUNCT
fcis-12817	81	37	the	the	DET
fcis-12817	81	38	formula	formula	NOUN
fcis-12817	81	39	for	for	ADP
fcis-12817	81	40	this	this	DET
fcis-12817	81	41	process	process	NOUN
fcis-12817	81	42	is	be	AUX
fcis-12817	81	43	shown	show	VERB
fcis-12817	81	44	below	below	ADP
fcis-12817	81	45	:	:	PUNCT
fcis-12817	81	46			PROPN
fcis-12817	81	47	,head	,head	NOUN
fcis-12817	81	48	head	head	NOUN
fcis-12817	81	49	headl	headl	NOUN
fcis-12817	81	50	l	l	PROPN
fcis-12817	81	51	p	p	PROPN
fcis-12817	81	52	y	y	PROPN
fcis-12817	81	53	(	(	PUNCT
fcis-12817	81	54	10	10	NUM
fcis-12817	81	55	)	)	SYM
fcis-12817	81	56	78	78	NUM
fcis-12817	81	57			NOUN
fcis-12817	81	58	,tail	,tail	NOUN
fcis-12817	81	59	tail	tail	NOUN
fcis-12817	81	60	taill	taill	NOUN
fcis-12817	82	1	l	l	PROPN
fcis-12817	82	2	p	p	PROPN
fcis-12817	82	3	y	y	PROPN
fcis-12817	82	4	(	(	PUNCT
fcis-12817	82	5	11	11	NUM
fcis-12817	82	6	)	)	PUNCT
fcis-12817	82	7			NOUN
fcis-12817	82	8			PROPN
fcis-12817	82	9	:	:	PUNCT
fcis-12817	82	10	:	:	PUNCT
fcis-12817	82	11	:	:	PUNCT
fcis-12817	82	12	,	,	PUNCT
fcis-12817	82	13	head	head	NOUN
fcis-12817	82	14	tail	tail	NOUN
fcis-12817	82	15	head	head	NOUN
fcis-12817	82	16	tail	tail	NOUN
fcis-12817	82	17	head	head	NOUN
fcis-12817	82	18	taill	taill	NOUN
fcis-12817	83	1	l	l	PROPN
fcis-12817	83	2	p	p	PROPN
fcis-12817	83	3	y	y	PROPN
fcis-12817	83	4	(	(	PUNCT
fcis-12817	83	5	12	12	NUM
fcis-12817	83	6	)	)	PUNCT
fcis-12817	83	7	where	where	SCONJ
fcis-12817	83	8	heady	heady	ADJ
fcis-12817	83	9	,	,	PUNCT
fcis-12817	83	10	taily	taily	ADV
fcis-12817	83	11	,	,	PUNCT
fcis-12817	83	12	:	:	PUNCT
fcis-12817	83	13	head	head	NOUN
fcis-12817	83	14	taily	taily	ADV
fcis-12817	83	15	represent	represent	VERB
fcis-12817	83	16	the	the	DET
fcis-12817	83	17	real	real	ADJ
fcis-12817	83	18	labels	label	NOUN
fcis-12817	83	19	about	about	ADP
fcis-12817	83	20	entity	entity	NOUN
fcis-12817	83	21	head	head	NOUN
fcis-12817	83	22	,	,	PUNCT
fcis-12817	83	23	tail	tail	NOUN
fcis-12817	83	24	and	and	CCONJ
fcis-12817	83	25	entity	entity	NOUN
fcis-12817	83	26	matching	matching	NOUN
fcis-12817	83	27	span	span	NOUN
fcis-12817	83	28	respectively	respectively	ADV
fcis-12817	83	29	.	.	PUNCT
fcis-12817	84	1	headl	headl	NOUN
fcis-12817	84	2	,	,	PUNCT
fcis-12817	84	3	taill	taill	NOUN
fcis-12817	84	4	,	,	PUNCT
fcis-12817	84	5	:	:	PUNCT
fcis-12817	84	6	head	head	NOUN
fcis-12817	84	7	taill	taill	NOUN
fcis-12817	84	8	represent	represent	VERB
fcis-12817	84	9	the	the	DET
fcis-12817	84	10	loss	loss	NOUN
fcis-12817	84	11	function	function	NOUN
fcis-12817	84	12	for	for	ADP
fcis-12817	84	13	entity	entity	NOUN
fcis-12817	84	14	head	head	NOUN
fcis-12817	84	15	position	position	NOUN
fcis-12817	84	16	prediction	prediction	NOUN
fcis-12817	84	17	,	,	PUNCT
fcis-12817	84	18	the	the	DET
fcis-12817	84	19	loss	loss	NOUN
fcis-12817	84	20	function	function	NOUN
fcis-12817	84	21	for	for	ADP
fcis-12817	84	22	tail	tail	NOUN
fcis-12817	84	23	position	position	NOUN
fcis-12817	84	24	prediction	prediction	NOUN
fcis-12817	84	25	,	,	PUNCT
fcis-12817	84	26	and	and	CCONJ
fcis-12817	84	27	the	the	DET
fcis-12817	84	28	loss	loss	NOUN
fcis-12817	84	29	function	function	NOUN
fcis-12817	84	30	for	for	ADP
fcis-12817	84	31	entity	entity	NOUN
fcis-12817	84	32	span	span	NOUN
fcis-12817	84	33	matching	match	VERB
fcis-12817	84	34	prediction	prediction	NOUN
fcis-12817	84	35	,	,	PUNCT
fcis-12817	84	36	respectively	respectively	ADV
fcis-12817	84	37	.	.	PUNCT
fcis-12817	85	1	the	the	DET
fcis-12817	85	2	final	final	ADJ
fcis-12817	85	3	objective	objective	ADJ
fcis-12817	85	4	loss	loss	NOUN
fcis-12817	85	5	function	function	NOUN
fcis-12817	85	6	of	of	ADP
fcis-12817	85	7	the	the	DET
fcis-12817	85	8	model	model	NOUN
fcis-12817	85	9	is	be	AUX
fcis-12817	85	10	shown	show	VERB
fcis-12817	85	11	below	below	ADP
fcis-12817	85	12	:	:	PUNCT
fcis-12817	85	13	:	:	PUNCT
fcis-12817	85	14	total	total	ADJ
fcis-12817	85	15	head	head	NOUN
fcis-12817	85	16	tail	tail	NOUN
fcis-12817	85	17	head	head	NOUN
fcis-12817	85	18	taill	taill	NOUN
fcis-12817	85	19	l	l	PROPN
fcis-12817	85	20	l	l	PROPN
fcis-12817	85	21	l	l	NUM
fcis-12817	85	22			ADV
fcis-12817	85	23			PUNCT
fcis-12817	85	24	(	(	PUNCT
fcis-12817	85	25	13	13	NUM
fcis-12817	85	26	)	)	PUNCT
fcis-12817	85	27	where	where	NOUN
fcis-12817	85	28	is	be	AUX
fcis-12817	85	29	the	the	DET
fcis-12817	85	30	hyperparameter	hyperparameter	NOUN
fcis-12817	85	31	.	.	PUNCT
fcis-12817	86	1	4.2	4.2	NUM
fcis-12817	86	2	.	.	PUNCT
fcis-12817	87	1	testing	testing	NOUN
fcis-12817	87	2	in	in	ADP
fcis-12817	87	3	the	the	DET
fcis-12817	87	4	testing	testing	NOUN
fcis-12817	87	5	phase	phase	NOUN
fcis-12817	87	6	,	,	PUNCT
fcis-12817	87	7	in	in	ADP
fcis-12817	87	8	order	order	NOUN
fcis-12817	87	9	to	to	PART
fcis-12817	87	10	get	get	VERB
fcis-12817	87	11	the	the	DET
fcis-12817	87	12	final	final	ADJ
fcis-12817	87	13	answer	answer	NOUN
fcis-12817	87	14	,	,	PUNCT
fcis-12817	87	15	we	we	PRON
fcis-12817	87	16	first	first	ADV
fcis-12817	87	17	get	get	VERB
fcis-12817	87	18	the	the	DET
fcis-12817	87	19	head	head	NOUN
fcis-12817	87	20	index	index	NOUN
fcis-12817	87	21	and	and	CCONJ
fcis-12817	87	22	tail	tail	NOUN
fcis-12817	87	23	index	index	NOUN
fcis-12817	87	24	of	of	ADP
fcis-12817	87	25	the	the	DET
fcis-12817	87	26	entity	entity	NOUN
fcis-12817	87	27	in	in	ADP
fcis-12817	87	28	the	the	DET
fcis-12817	87	29	entity	entity	NOUN
fcis-12817	87	30	head	head	NOUN
fcis-12817	87	31	set	set	VERB
fcis-12817	87	32	ˆ	ˆ	ADJ
fcis-12817	87	33	headk	headk	NOUN
fcis-12817	87	34	and	and	CCONJ
fcis-12817	87	35	the	the	DET
fcis-12817	87	36	entity	entity	NOUN
fcis-12817	87	37	tail	tail	NOUN
fcis-12817	87	38	index	index	NOUN
fcis-12817	87	39	set	set	VERB
fcis-12817	87	40	ˆ	ˆ	DET
fcis-12817	87	41	tailk	tailk	NOUN
fcis-12817	87	42	.	.	PUNCT
fcis-12817	88	1	subsequently	subsequently	ADV
fcis-12817	88	2	,	,	PUNCT
fcis-12817	88	3	we	we	PRON
fcis-12817	88	4	use	use	VERB
fcis-12817	88	5	the	the	DET
fcis-12817	88	6	entity	entity	NOUN
fcis-12817	88	7	head	head	NOUN
fcis-12817	88	8	-	-	PUNCT
fcis-12817	88	9	tail	tail	NOUN
fcis-12817	88	10	pairing	pair	VERB
fcis-12817	88	11	model	model	NOUN
fcis-12817	88	12	to	to	PART
fcis-12817	88	13	match	match	VERB
fcis-12817	88	14	them	they	PRON
fcis-12817	88	15	and	and	CCONJ
fcis-12817	88	16	get	get	VERB
fcis-12817	88	17	the	the	DET
fcis-12817	88	18	final	final	ADJ
fcis-12817	88	19	entity	entity	NOUN
fcis-12817	88	20	index	index	NOUN
fcis-12817	88	21	.	.	PUNCT
fcis-12817	89	1	5	5	X
fcis-12817	89	2	.	.	X
fcis-12817	89	3	experiments	experiment	NOUN
fcis-12817	89	4	5.1	5.1	NUM
fcis-12817	89	5	.	.	PUNCT
fcis-12817	90	1	data	datum	NOUN
fcis-12817	90	2	sets	set	VERB
fcis-12817	90	3	the	the	DET
fcis-12817	90	4	two	two	NUM
fcis-12817	90	5	datasets	dataset	NOUN
fcis-12817	90	6	used	use	VERB
fcis-12817	90	7	for	for	ADP
fcis-12817	90	8	the	the	DET
fcis-12817	90	9	experiments	experiment	NOUN
fcis-12817	90	10	are	be	AUX
fcis-12817	90	11	from	from	ADP
fcis-12817	90	12	the	the	DET
fcis-12817	90	13	ecommerce	ecommerce	NOUN
fcis-12817	90	14	domain	domain	NOUN
fcis-12817	90	15	and	and	CCONJ
fcis-12817	90	16	both	both	PRON
fcis-12817	90	17	contain	contain	VERB
fcis-12817	90	18	entity	entity	NOUN
fcis-12817	90	19	types	type	NOUN
fcis-12817	90	20	of	of	ADP
fcis-12817	90	21	product	product	NOUN
fcis-12817	90	22	names	name	NOUN
fcis-12817	90	23	,	,	PUNCT
fcis-12817	90	24	brand	brand	NOUN
fcis-12817	90	25	names	name	NOUN
fcis-12817	90	26	.	.	PUNCT
fcis-12817	91	1	jd	jd	AUX
fcis-12817	91	2	-	-	PUNCT
fcis-12817	91	3	e	e	NOUN
fcis-12817	91	4	-	-	NOUN
fcis-12817	91	5	commerce	commerce	NOUN
fcis-12817	91	6	:	:	PUNCT
fcis-12817	91	7	ding	ding	NOUN
fcis-12817	91	8	et	et	PROPN
fcis-12817	91	9	al	al	PROPN
fcis-12817	91	10	.	.	PUNCT
fcis-12817	92	1	[	[	X
fcis-12817	92	2	12	12	NUM
fcis-12817	92	3	]	]	PUNCT
fcis-12817	92	4	have	have	AUX
fcis-12817	92	5	published	publish	VERB
fcis-12817	92	6	a	a	DET
fcis-12817	92	7	dataset	dataset	NOUN
fcis-12817	92	8	e	e	NOUN
fcis-12817	92	9	-	-	NOUN
fcis-12817	92	10	commerce	commerce	NOUN
fcis-12817	92	11	in	in	ADP
fcis-12817	92	12	the	the	DET
fcis-12817	92	13	field	field	NOUN
fcis-12817	92	14	of	of	ADP
fcis-12817	92	15	e	e	NOUN
fcis-12817	92	16	-	-	NOUN
fcis-12817	92	17	commerce	commerce	NOUN
fcis-12817	92	18	,	,	PUNCT
fcis-12817	92	19	which	which	PRON
fcis-12817	92	20	contains	contain	VERB
fcis-12817	92	21	two	two	NUM
fcis-12817	92	22	types	type	NOUN
fcis-12817	92	23	of	of	ADP
fcis-12817	92	24	entities	entity	NOUN
fcis-12817	92	25	:	:	PUNCT
fcis-12817	92	26	brands	brand	NOUN
fcis-12817	92	27	and	and	CCONJ
fcis-12817	92	28	products	product	NOUN
fcis-12817	92	29	.	.	PUNCT
fcis-12817	93	1	however	however	ADV
fcis-12817	93	2	,	,	PUNCT
fcis-12817	93	3	considering	consider	VERB
fcis-12817	93	4	that	that	SCONJ
fcis-12817	93	5	the	the	DET
fcis-12817	93	6	e	e	PROPN
fcis-12817	93	7	-	-	NOUN
fcis-12817	93	8	commerce	commerce	NOUN
fcis-12817	93	9	dataset	dataset	NOUN
fcis-12817	93	10	is	be	AUX
fcis-12817	93	11	small	small	ADJ
fcis-12817	93	12	,	,	PUNCT
fcis-12817	93	13	in	in	ADP
fcis-12817	93	14	order	order	NOUN
fcis-12817	93	15	to	to	PART
fcis-12817	93	16	be	be	AUX
fcis-12817	93	17	able	able	ADJ
fcis-12817	93	18	to	to	PART
fcis-12817	93	19	further	far	ADV
fcis-12817	93	20	demonstrate	demonstrate	VERB
fcis-12817	93	21	the	the	DET
fcis-12817	93	22	validity	validity	NOUN
fcis-12817	93	23	of	of	ADP
fcis-12817	93	24	the	the	DET
fcis-12817	93	25	model	model	NOUN
fcis-12817	93	26	,	,	PUNCT
fcis-12817	93	27	we	we	PRON
fcis-12817	93	28	additionally	additionally	ADV
fcis-12817	93	29	collected	collect	VERB
fcis-12817	93	30	relevant	relevant	ADJ
fcis-12817	93	31	data	datum	NOUN
fcis-12817	93	32	from	from	ADP
fcis-12817	93	33	e	e	NOUN
fcis-12817	93	34	-	-	NOUN
fcis-12817	93	35	commerce	commerce	NOUN
fcis-12817	93	36	websites	website	NOUN
fcis-12817	93	37	to	to	PART
fcis-12817	93	38	construct	construct	VERB
fcis-12817	93	39	the	the	DET
fcis-12817	93	40	jd	jd	PROPN
fcis-12817	93	41	-	-	PUNCT
fcis-12817	93	42	e	e	PROPN
fcis-12817	93	43	-	-	NOUN
fcis-12817	93	44	commerce	commerce	NOUN
fcis-12817	93	45	dataset	dataset	NOUN
fcis-12817	93	46	.	.	PUNCT
fcis-12817	94	1	we	we	PRON
fcis-12817	94	2	use	use	VERB
fcis-12817	94	3	web	web	NOUN
fcis-12817	94	4	crawler	crawler	NOUN
fcis-12817	94	5	to	to	PART
fcis-12817	94	6	get	get	VERB
fcis-12817	94	7	the	the	DET
fcis-12817	94	8	text	text	NOUN
fcis-12817	94	9	related	relate	VERB
fcis-12817	94	10	to	to	ADP
fcis-12817	94	11	the	the	DET
fcis-12817	94	12	title	title	NOUN
fcis-12817	94	13	content	content	NOUN
fcis-12817	94	14	of	of	ADP
fcis-12817	94	15	the	the	DET
fcis-12817	94	16	product	product	NOUN
fcis-12817	94	17	from	from	ADP
fcis-12817	94	18	the	the	DET
fcis-12817	94	19	shopping	shopping	NOUN
fcis-12817	94	20	site	site	NOUN
fcis-12817	94	21	.	.	PUNCT
fcis-12817	95	1	the	the	DET
fcis-12817	95	2	crawling	crawling	NOUN
fcis-12817	95	3	technique	technique	NOUN
fcis-12817	95	4	is	be	AUX
fcis-12817	95	5	based	base	VERB
fcis-12817	95	6	on	on	ADP
fcis-12817	95	7	the	the	DET
fcis-12817	95	8	python	python	NOUN
fcis-12817	95	9	and	and	CCONJ
fcis-12817	95	10	utilizes	utilize	VERB
fcis-12817	95	11	the	the	DET
fcis-12817	95	12	selenium	selenium	NOUN
fcis-12817	95	13	and	and	CCONJ
fcis-12817	95	14	requests	request	NOUN
fcis-12817	95	15	modules	module	NOUN
fcis-12817	95	16	.	.	PUNCT
fcis-12817	96	1	finally	finally	ADV
fcis-12817	96	2	,	,	PUNCT
fcis-12817	96	3	we	we	PRON
fcis-12817	96	4	collected	collect	VERB
fcis-12817	96	5	a	a	DET
fcis-12817	96	6	total	total	NOUN
fcis-12817	96	7	of	of	ADP
fcis-12817	96	8	10,456	10,456	NUM
fcis-12817	96	9	texts	text	NOUN
fcis-12817	96	10	by	by	ADP
fcis-12817	96	11	crawling	crawl	VERB
fcis-12817	96	12	.	.	PUNCT
fcis-12817	97	1	we	we	PRON
fcis-12817	97	2	use	use	VERB
fcis-12817	97	3	the	the	DET
fcis-12817	97	4	open	open	ADJ
fcis-12817	97	5	-	-	PUNCT
fcis-12817	97	6	source	source	NOUN
fcis-12817	97	7	label	label	NOUN
fcis-12817	97	8	studio	studio	NOUN
fcis-12817	97	9	annotation	annotation	NOUN
fcis-12817	97	10	tool	tool	NOUN
fcis-12817	97	11	to	to	PART
fcis-12817	97	12	manually	manually	ADV
fcis-12817	97	13	annotate	annotate	VERB
fcis-12817	97	14	entities	entity	NOUN
fcis-12817	97	15	,	,	PUNCT
fcis-12817	97	16	which	which	PRON
fcis-12817	97	17	supports	support	VERB
fcis-12817	97	18	fast	fast	ADJ
fcis-12817	97	19	matching	matching	NOUN
fcis-12817	97	20	of	of	ADP
fcis-12817	97	21	entity	entity	NOUN
fcis-12817	97	22	labels	label	NOUN
fcis-12817	97	23	through	through	ADP
fcis-12817	97	24	a	a	DET
fcis-12817	97	25	visual	visual	ADJ
fcis-12817	97	26	interface	interface	NOUN
fcis-12817	97	27	,	,	PUNCT
fcis-12817	97	28	and	and	CCONJ
fcis-12817	97	29	supports	support	VERB
fcis-12817	97	30	export	export	NOUN
fcis-12817	97	31	in	in	ADP
fcis-12817	97	32	multiple	multiple	ADJ
fcis-12817	97	33	formats	format	NOUN
fcis-12817	97	34	.	.	PUNCT
fcis-12817	98	1	we	we	PRON
fcis-12817	98	2	adopt	adopt	VERB
fcis-12817	98	3	the	the	DET
fcis-12817	98	4	bioes	bioe	NOUN
fcis-12817	98	5	annotation	annotation	NOUN
fcis-12817	98	6	scheme	scheme	NOUN
fcis-12817	98	7	to	to	PART
fcis-12817	98	8	annotate	annotate	VERB
fcis-12817	98	9	entities	entity	NOUN
fcis-12817	98	10	for	for	ADP
fcis-12817	98	11	two	two	NUM
fcis-12817	98	12	types	type	NOUN
fcis-12817	98	13	of	of	ADP
fcis-12817	98	14	entities	entity	NOUN
fcis-12817	98	15	.	.	PUNCT
fcis-12817	99	1	we	we	PRON
fcis-12817	99	2	divide	divide	VERB
fcis-12817	99	3	the	the	DET
fcis-12817	99	4	jd	jd	PROPN
fcis-12817	99	5	-	-	PUNCT
fcis-12817	99	6	e	e	PROPN
fcis-12817	99	7	-	-	NOUN
fcis-12817	99	8	commerce	commerce	NOUN
fcis-12817	99	9	dataset	dataset	NOUN
fcis-12817	99	10	into	into	ADP
fcis-12817	99	11	training	training	NOUN
fcis-12817	99	12	set	set	NOUN
fcis-12817	99	13	,	,	PUNCT
fcis-12817	99	14	validation	validation	NOUN
fcis-12817	99	15	set	set	NOUN
fcis-12817	99	16	,	,	PUNCT
fcis-12817	99	17	and	and	CCONJ
fcis-12817	99	18	test	test	NOUN
fcis-12817	99	19	set	set	VERB
fcis-12817	99	20	in	in	ADP
fcis-12817	99	21	the	the	DET
fcis-12817	99	22	ratio	ratio	NOUN
fcis-12817	99	23	of	of	ADP
fcis-12817	99	24	6:2:2	6:2:2	NUM
fcis-12817	99	25	.	.	PUNCT
fcis-12817	100	1	table	table	NOUN
fcis-12817	100	2	1	1	NUM
fcis-12817	100	3	shows	show	VERB
fcis-12817	100	4	the	the	DET
fcis-12817	100	5	detailed	detailed	ADJ
fcis-12817	100	6	statistical	statistical	ADJ
fcis-12817	100	7	information	information	NOUN
fcis-12817	100	8	of	of	ADP
fcis-12817	100	9	the	the	DET
fcis-12817	100	10	jde	jde	NOUN
fcis-12817	100	11	-	-	PUNCT
fcis-12817	100	12	commerce	commerce	NOUN
fcis-12817	100	13	dataset	dataset	NOUN
fcis-12817	100	14	.	.	PUNCT
fcis-12817	101	1	table	table	NOUN
fcis-12817	101	2	1	1	NUM
fcis-12817	101	3	.	.	PUNCT
fcis-12817	102	1	statistical	statistical	ADJ
fcis-12817	102	2	information	information	NOUN
fcis-12817	102	3	of	of	ADP
fcis-12817	102	4	jd	jd	PROPN
fcis-12817	102	5	-	-	ADJ
fcis-12817	102	6	e	e	PROPN
fcis-12817	102	7	-	-	NOUN
fcis-12817	102	8	commerce	commerce	NOUN
fcis-12817	102	9	dataset	dataset	NOUN
fcis-12817	102	10	jd	jd	PROPN
fcis-12817	102	11	-	-	ADJ
fcis-12817	102	12	e	e	ADJ
fcis-12817	102	13	-	-	ADJ
fcis-12817	102	14	commerce	commerce	NOUN
fcis-12817	102	15	dataset	dataset	NOUN
fcis-12817	102	16	training	training	NOUN
fcis-12817	102	17	set	set	VERB
fcis-12817	102	18	validation	validation	NOUN
fcis-12817	102	19	set	set	NOUN
fcis-12817	102	20	test	test	NOUN
fcis-12817	102	21	set	set	VERB
fcis-12817	102	22	sentences	sentence	NOUN
fcis-12817	102	23	6275	6275	NUM
fcis-12817	102	24	2091	2091	NUM
fcis-12817	102	25	2090	2090	NUM
fcis-12817	102	26	characters	character	NOUN
fcis-12817	102	27	339.5k	339.5k	NUM
fcis-12817	102	28	113.3k	113.3k	NUM
fcis-12817	102	29	113.1k	113.1k	NUM
fcis-12817	102	30	maximum	maximum	ADJ
fcis-12817	102	31	number	number	NOUN
fcis-12817	102	32	of	of	ADP
fcis-12817	102	33	characters	character	NOUN
fcis-12817	102	34	98	98	NUM
fcis-12817	102	35	91	91	NUM
fcis-12817	102	36	97	97	NUM
fcis-12817	102	37	minimum	minimum	ADJ
fcis-12817	102	38	number	number	NOUN
fcis-12817	102	39	of	of	ADP
fcis-12817	102	40	characters	character	NOUN
fcis-12817	102	41	8	8	NUM
fcis-12817	102	42	13	13	NUM
fcis-12817	102	43	15	15	NUM
fcis-12817	102	44	e	e	NOUN
fcis-12817	102	45	-	-	NOUN
fcis-12817	102	46	commerce	commerce	NOUN
fcis-12817	102	47	:	:	PUNCT
fcis-12817	102	48	for	for	ADP
fcis-12817	102	49	the	the	DET
fcis-12817	102	50	e	e	PROPN
fcis-12817	102	51	-	-	NOUN
fcis-12817	102	52	commerce	commerce	NOUN
fcis-12817	102	53	dataset	dataset	NOUN
fcis-12817	102	54	,	,	PUNCT
fcis-12817	102	55	we	we	PRON
fcis-12817	102	56	used	use	VERB
fcis-12817	102	57	the	the	DET
fcis-12817	102	58	same	same	ADJ
fcis-12817	102	59	method	method	NOUN
fcis-12817	102	60	as	as	ADP
fcis-12817	102	61	ding	ding	NOUN
fcis-12817	102	62	et	et	X
fcis-12817	102	63	al	al	PROPN
fcis-12817	103	1	[	[	X
fcis-12817	103	2	13	13	NUM
fcis-12817	103	3	]	]	PUNCT
fcis-12817	103	4	.	.	PUNCT
fcis-12817	104	1	table	table	NOUN
fcis-12817	104	2	2	2	NUM
fcis-12817	104	3	shows	show	VERB
fcis-12817	104	4	the	the	DET
fcis-12817	104	5	statistical	statistical	ADJ
fcis-12817	104	6	information	information	NOUN
fcis-12817	104	7	of	of	ADP
fcis-12817	104	8	the	the	DET
fcis-12817	104	9	e	e	PROPN
fcis-12817	104	10	-	-	PROPN
fcis-12817	104	11	commerce	commerce	NOUN
fcis-12817	104	12	dataset	dataset	NOUN
fcis-12817	104	13	.	.	PUNCT
fcis-12817	105	1	table	table	NOUN
fcis-12817	105	2	2	2	NUM
fcis-12817	105	3	.	.	PUNCT
fcis-12817	105	4	statistical	statistical	ADJ
fcis-12817	105	5	information	information	NOUN
fcis-12817	105	6	on	on	ADP
fcis-12817	105	7	e	e	NOUN
fcis-12817	105	8	-	-	NOUN
fcis-12817	105	9	commerce	commerce	NOUN
fcis-12817	105	10	dataset	dataset	NOUN
fcis-12817	105	11	e	e	PROPN
fcis-12817	105	12	-	-	ADJ
fcis-12817	105	13	commerce	commerce	NOUN
fcis-12817	105	14	dataset	dataset	NOUN
fcis-12817	105	15	training	training	NOUN
fcis-12817	105	16	set	set	VERB
fcis-12817	105	17	validation	validation	NOUN
fcis-12817	105	18	set	set	NOUN
fcis-12817	105	19	test	test	NOUN
fcis-12817	105	20	set	set	VERB
fcis-12817	105	21	character	character	NOUN
fcis-12817	105	22	119.1k	119.1k	NUM
fcis-12817	105	23	14.9k	14.9k	NUM
fcis-12817	105	24	14.7k	14.7k	NUM
fcis-12817	105	25	sentences	sentence	VERB
fcis-12817	105	26	3989	3989	NUM
fcis-12817	105	27	500	500	NUM
fcis-12817	105	28	498	498	NUM
fcis-12817	105	29	we	we	PRON
fcis-12817	105	30	need	need	VERB
fcis-12817	105	31	to	to	PART
fcis-12817	105	32	transform	transform	VERB
fcis-12817	105	33	the	the	DET
fcis-12817	105	34	labeled	label	VERB
fcis-12817	105	35	dataset	dataset	NOUN
fcis-12817	105	36	into	into	ADP
fcis-12817	105	37	a	a	DET
fcis-12817	105	38	dataset	dataset	NOUN
fcis-12817	105	39	suitable	suitable	ADJ
fcis-12817	105	40	for	for	ADP
fcis-12817	105	41	mrc	mrc	NOUN
fcis-12817	105	42	.	.	PUNCT
fcis-12817	106	1	rajpurkar	rajpurkar	PROPN
fcis-12817	106	2	et	et	PROPN
fcis-12817	106	3	al.[13	al.[13	NOUN
fcis-12817	106	4	]	]	PUNCT
fcis-12817	106	5	in	in	ADP
fcis-12817	106	6	2016	2016	NUM
fcis-12817	106	7	proposed	propose	VERB
fcis-12817	106	8	squad	squad	NOUN
fcis-12817	106	9	,	,	PUNCT
fcis-12817	106	10	a	a	DET
fcis-12817	106	11	stanford	stanford	PROPN
fcis-12817	106	12	question	question	NOUN
fcis-12817	106	13	and	and	CCONJ
fcis-12817	106	14	answer	answer	NOUN
fcis-12817	106	15	dataset	dataset	NOUN
fcis-12817	106	16	that	that	PRON
fcis-12817	106	17	can	can	AUX
fcis-12817	106	18	be	be	AUX
fcis-12817	106	19	used	use	VERB
fcis-12817	106	20	for	for	ADP
fcis-12817	106	21	mrc	mrc	NOUN
fcis-12817	106	22	.	.	PUNCT
fcis-12817	107	1	the	the	DET
fcis-12817	107	2	dataset	dataset	NOUN
fcis-12817	107	3	in	in	ADP
fcis-12817	107	4	this	this	DET
fcis-12817	107	5	paper	paper	NOUN
fcis-12817	107	6	was	be	AUX
fcis-12817	107	7	constructed	construct	VERB
fcis-12817	107	8	in	in	ADP
fcis-12817	107	9	a	a	DET
fcis-12817	107	10	form	form	NOUN
fcis-12817	107	11	similar	similar	ADJ
fcis-12817	107	12	to	to	ADP
fcis-12817	107	13	the	the	DET
fcis-12817	107	14	squad	squad	ADJ
fcis-12817	107	15	dataset	dataset	NOUN
fcis-12817	107	16	.	.	PUNCT
fcis-12817	108	1	we	we	PRON
fcis-12817	108	2	then	then	ADV
fcis-12817	108	3	wrote	write	VERB
fcis-12817	108	4	scripts	script	NOUN
fcis-12817	108	5	to	to	PART
fcis-12817	108	6	convert	convert	VERB
fcis-12817	108	7	the	the	DET
fcis-12817	108	8	format	format	NOUN
fcis-12817	108	9	of	of	ADP
fcis-12817	108	10	the	the	DET
fcis-12817	108	11	dataset	dataset	NOUN
fcis-12817	108	12	from	from	ADP
fcis-12817	108	13	the	the	DET
fcis-12817	108	14	bieso	bieso	NOUN
fcis-12817	108	15	labeled	label	VERB
fcis-12817	108	16	format	format	NOUN
fcis-12817	108	17	to	to	ADP
fcis-12817	108	18	the	the	DET
fcis-12817	108	19	mrc	mrc	NOUN
fcis-12817	108	20	dataset	dataset	NOUN
fcis-12817	108	21	format	format	NOUN
fcis-12817	108	22	.	.	PUNCT
fcis-12817	109	1	5.2	5.2	NUM
fcis-12817	109	2	.	.	PUNCT
fcis-12817	109	3	evaluation	evaluation	NOUN
fcis-12817	109	4	metrics	metric	NOUN
fcis-12817	109	5	there	there	PRON
fcis-12817	109	6	are	be	VERB
fcis-12817	109	7	three	three	NUM
fcis-12817	109	8	main	main	ADJ
fcis-12817	109	9	evaluation	evaluation	NOUN
fcis-12817	109	10	metrics	metric	NOUN
fcis-12817	109	11	currently	currently	ADV
fcis-12817	109	12	used	use	VERB
fcis-12817	109	13	for	for	ADP
fcis-12817	109	14	named	name	VERB
fcis-12817	109	15	entity	entity	NOUN
fcis-12817	109	16	recognition	recognition	NOUN
fcis-12817	109	17	,	,	PUNCT
fcis-12817	109	18	which	which	PRON
fcis-12817	109	19	are	be	AUX
fcis-12817	109	20	precision	precision	NOUN
fcis-12817	109	21	,	,	PUNCT
fcis-12817	109	22	recall	recall	NOUN
fcis-12817	109	23	,	,	PUNCT
fcis-12817	109	24	and	and	CCONJ
fcis-12817	109	25	f1	f1	NOUN
fcis-12817	109	26	-	-	PUNCT
fcis-12817	109	27	score	score	NOUN
fcis-12817	109	28	.	.	PUNCT
fcis-12817	110	1	precision	precision	NOUN
fcis-12817	110	2	is	be	AUX
fcis-12817	110	3	the	the	DET
fcis-12817	110	4	ratio	ratio	NOUN
fcis-12817	110	5	between	between	ADP
fcis-12817	110	6	the	the	DET
fcis-12817	110	7	correctly	correctly	ADV
fcis-12817	110	8	predicted	predict	VERB
fcis-12817	110	9	positive	positive	ADJ
fcis-12817	110	10	examples	example	NOUN
fcis-12817	110	11	and	and	CCONJ
fcis-12817	110	12	all	all	DET
fcis-12817	110	13	the	the	DET
fcis-12817	110	14	examples	example	NOUN
fcis-12817	110	15	predicted	predict	VERB
fcis-12817	110	16	as	as	ADP
fcis-12817	110	17	positive	positive	ADJ
fcis-12817	110	18	.	.	PUNCT
fcis-12817	111	1	recall	recall	PROPN
fcis-12817	111	2	is	be	AUX
fcis-12817	111	3	the	the	DET
fcis-12817	111	4	ratio	ratio	NOUN
fcis-12817	111	5	between	between	ADP
fcis-12817	111	6	correctly	correctly	ADV
fcis-12817	111	7	predicted	predict	VERB
fcis-12817	111	8	positive	positive	ADJ
fcis-12817	111	9	examples	example	NOUN
fcis-12817	111	10	and	and	CCONJ
fcis-12817	111	11	all	all	DET
fcis-12817	111	12	actual	actual	ADJ
fcis-12817	111	13	positive	positive	ADJ
fcis-12817	111	14	examples	example	NOUN
fcis-12817	111	15	.	.	PUNCT
fcis-12817	112	1	1f	1f	PROPN
fcis-12817	112	2	is	be	AUX
fcis-12817	112	3	calculation	calculation	NOUN
fcis-12817	112	4	of	of	ADP
fcis-12817	112	5	the	the	DET
fcis-12817	112	6	value	value	NOUN
fcis-12817	112	7	metric	metric	ADJ
fcis-12817	112	8	utilizing	utilize	VERB
fcis-12817	112	9	both	both	CCONJ
fcis-12817	112	10	p	p	NOUN
fcis-12817	112	11	and	and	CCONJ
fcis-12817	112	12	r	r	NOUN
fcis-12817	112	13	,	,	PUNCT
fcis-12817	112	14	which	which	PRON
fcis-12817	112	15	are	be	AUX
fcis-12817	112	16	comprehensive	comprehensive	ADJ
fcis-12817	112	17	evaluation	evaluation	NOUN
fcis-12817	112	18	metrics	metric	NOUN
fcis-12817	112	19	.	.	PUNCT
fcis-12817	113	1	1f	1f	PROPN
fcis-12817	113	2	is	be	AUX
fcis-12817	113	3	widely	widely	ADV
fcis-12817	113	4	used	use	VERB
fcis-12817	113	5	to	to	PART
fcis-12817	113	6	evaluate	evaluate	VERB
fcis-12817	113	7	the	the	DET
fcis-12817	113	8	performance	performance	NOUN
fcis-12817	113	9	of	of	ADP
fcis-12817	113	10	the	the	DET
fcis-12817	113	11	model	model	NOUN
fcis-12817	113	12	.	.	PUNCT
fcis-12817	114	1	we	we	PRON
fcis-12817	114	2	use	use	VERB
fcis-12817	114	3	1f	1f	NUM
fcis-12817	114	4	value	value	NOUN
fcis-12817	114	5	,	,	PUNCT
fcis-12817	114	6	p	p	NOUN
fcis-12817	114	7	and	and	CCONJ
fcis-12817	114	8	r	r	NOUN
fcis-12817	114	9	as	as	ADP
fcis-12817	114	10	the	the	DET
fcis-12817	114	11	evaluation	evaluation	NOUN
fcis-12817	114	12	metrics	metric	NOUN
fcis-12817	114	13	.	.	PUNCT
fcis-12817	115	1	the	the	DET
fcis-12817	115	2	calculation	calculation	NOUN
fcis-12817	115	3	of	of	ADP
fcis-12817	115	4	the	the	DET
fcis-12817	115	5	three	three	NUM
fcis-12817	115	6	metrics	metric	NOUN
fcis-12817	115	7	is	be	AUX
fcis-12817	115	8	shown	show	VERB
fcis-12817	115	9	below	below	ADP
fcis-12817	115	10	:	:	PUNCT
fcis-12817	115	11	100	100	NUM
fcis-12817	115	12	%	%	NOUN
fcis-12817	115	13	tp	tp	ADP
fcis-12817	115	14	p	p	NOUN
fcis-12817	115	15	tp	tp	ADP
fcis-12817	115	16	fp	fp	PROPN
fcis-12817	115	17			PROPN
fcis-12817	115	18			PROPN
fcis-12817	115	19			PROPN
fcis-12817	115	20	(	(	PUNCT
fcis-12817	115	21	14	14	NUM
fcis-12817	115	22	)	)	PUNCT
fcis-12817	115	23	100	100	NUM
fcis-12817	115	24	%	%	NOUN
fcis-12817	115	25	tp	tp	ADP
fcis-12817	115	26	r	r	NOUN
fcis-12817	115	27	tp	tp	NOUN
fcis-12817	115	28	fn	fn	NOUN
fcis-12817	116	1			PROPN
fcis-12817	116	2			NOUN
fcis-12817	116	3			X
fcis-12817	116	4	(	(	PUNCT
fcis-12817	116	5	15	15	NUM
fcis-12817	116	6	)	)	SYM
fcis-12817	116	7	2	2	NUM
fcis-12817	116	8	1	1	NUM
fcis-12817	116	9	100	100	NUM
fcis-12817	116	10	%	%	NOUN
fcis-12817	116	11	p	p	NOUN
fcis-12817	116	12	r	r	NOUN
fcis-12817	116	13	f	f	NOUN
fcis-12817	117	1	p	p	NOUN
fcis-12817	117	2	r	r	NOUN
fcis-12817	117	3			NOUN
fcis-12817	117	4			PROPN
fcis-12817	117	5			PROPN
fcis-12817	117	6			PROPN
fcis-12817	117	7			PROPN
fcis-12817	117	8	(	(	PUNCT
fcis-12817	117	9	16	16	NUM
fcis-12817	117	10	)	)	PUNCT
fcis-12817	117	11	where	where	SCONJ
fcis-12817	117	12	tp	tp	NOUN
fcis-12817	117	13	is	be	AUX
fcis-12817	117	14	the	the	DET
fcis-12817	117	15	sample	sample	NOUN
fcis-12817	117	16	that	that	PRON
fcis-12817	117	17	was	be	AUX
fcis-12817	117	18	correctly	correctly	ADV
fcis-12817	117	19	predicted	predict	VERB
fcis-12817	117	20	as	as	ADP
fcis-12817	117	21	positive	positive	ADJ
fcis-12817	117	22	.	.	PUNCT
fcis-12817	118	1	fp	fp	X
fcis-12817	118	2	is	be	AUX
fcis-12817	118	3	the	the	DET
fcis-12817	118	4	sample	sample	NOUN
fcis-12817	118	5	that	that	PRON
fcis-12817	118	6	was	be	AUX
fcis-12817	118	7	incorrectly	incorrectly	ADV
fcis-12817	118	8	predicted	predict	VERB
fcis-12817	118	9	as	as	ADP
fcis-12817	118	10	positive	positive	ADJ
fcis-12817	118	11	.	.	PUNCT
fcis-12817	119	1	fn	fn	NOUN
fcis-12817	119	2	is	be	AUX
fcis-12817	119	3	the	the	DET
fcis-12817	119	4	sample	sample	NOUN
fcis-12817	119	5	that	that	PRON
fcis-12817	119	6	was	be	AUX
fcis-12817	119	7	incorrectly	incorrectly	ADV
fcis-12817	119	8	predicted	predict	VERB
fcis-12817	119	9	as	as	ADP
fcis-12817	119	10	negative	negative	ADJ
fcis-12817	119	11	.	.	PUNCT
fcis-12817	120	1	5.3	5.3	NUM
fcis-12817	120	2	.	.	PUNCT
fcis-12817	120	3	results	result	NOUN
fcis-12817	120	4	and	and	CCONJ
fcis-12817	120	5	analysis	analysis	NOUN
fcis-12817	120	6	we	we	PRON
fcis-12817	120	7	demonstrate	demonstrate	VERB
fcis-12817	120	8	the	the	DET
fcis-12817	120	9	effectiveness	effectiveness	NOUN
fcis-12817	120	10	of	of	ADP
fcis-12817	120	11	our	our	PRON
fcis-12817	120	12	proposed	propose	VERB
fcis-12817	120	13	model	model	NOUN
fcis-12817	120	14	by	by	ADP
fcis-12817	120	15	comparing	compare	VERB
fcis-12817	120	16	the	the	DET
fcis-12817	120	17	experimental	experimental	ADJ
fcis-12817	120	18	results	result	NOUN
fcis-12817	120	19	with	with	ADP
fcis-12817	120	20	existing	exist	VERB
fcis-12817	120	21	models	model	NOUN
fcis-12817	120	22	on	on	ADP
fcis-12817	120	23	both	both	PRON
fcis-12817	120	24	jd	jd	PROPN
fcis-12817	120	25	-	-	ADJ
fcis-12817	120	26	e	e	ADJ
fcis-12817	120	27	-	-	NOUN
fcis-12817	120	28	commerce	commerce	NOUN
fcis-12817	120	29	dataset	dataset	NOUN
fcis-12817	120	30	and	and	CCONJ
fcis-12817	120	31	e	e	NOUN
fcis-12817	120	32	-	-	NOUN
fcis-12817	120	33	commerce	commerce	NOUN
fcis-12817	120	34	dataset	dataset	NOUN
fcis-12817	120	35	.	.	PUNCT
fcis-12817	121	1	for	for	ADP
fcis-12817	121	2	both	both	DET
fcis-12817	121	3	datasets	dataset	NOUN
fcis-12817	121	4	,	,	PUNCT
fcis-12817	121	5	we	we	PRON
fcis-12817	121	6	use	use	VERB
fcis-12817	121	7	the	the	DET
fcis-12817	121	8	following	following	ADJ
fcis-12817	121	9	model	model	NOUN
fcis-12817	121	10	as	as	ADP
fcis-12817	121	11	a	a	DET
fcis-12817	121	12	comparison	comparison	NOUN
fcis-12817	121	13	model	model	NOUN
fcis-12817	121	14	:	:	PUNCT
fcis-12817	121	15	(	(	PUNCT
fcis-12817	121	16	1	1	X
fcis-12817	121	17	)	)	PUNCT
fcis-12817	121	18	lstm+crf	lstm+crf	NOUN
fcis-12817	121	19	:	:	PUNCT
fcis-12817	121	20	the	the	DET
fcis-12817	121	21	classic	classic	ADJ
fcis-12817	121	22	sequence	sequence	NOUN
fcis-12817	121	23	labeling	labeling	NOUN
fcis-12817	121	24	based	base	VERB
fcis-12817	121	25	ner	ner	PROPN
fcis-12817	121	26	model	model	NOUN
fcis-12817	121	27	.	.	PUNCT
fcis-12817	122	1	(	(	PUNCT
fcis-12817	122	2	2	2	X
fcis-12817	122	3	)	)	PUNCT
fcis-12817	122	4	roberta	roberta	PROPN
fcis-12817	122	5	-	-	PUNCT
fcis-12817	122	6	wwm	wwm	PROPN
fcis-12817	122	7	model	model	NOUN
fcis-12817	122	8	based	base	VERB
fcis-12817	122	9	on	on	ADP
fcis-12817	122	10	sequence	sequence	NOUN
fcis-12817	122	11	labeling	labeling	NOUN
fcis-12817	122	12	:	:	PUNCT
fcis-12817	122	13	with	with	ADP
fcis-12817	122	14	the	the	DET
fcis-12817	122	15	modeling	modeling	NOUN
fcis-12817	122	16	paradigm	paradigm	NOUN
fcis-12817	122	17	of	of	ADP
fcis-12817	122	18	sequence	sequence	NOUN
fcis-12817	122	19	labeling	labeling	NOUN
fcis-12817	122	20	,	,	PUNCT
fcis-12817	122	21	the	the	DET
fcis-12817	122	22	results	result	NOUN
fcis-12817	122	23	are	be	AUX
fcis-12817	122	24	obtained	obtain	VERB
fcis-12817	122	25	by	by	ADP
fcis-12817	122	26	fine	fine	ADV
fcis-12817	122	27	-	-	PUNCT
fcis-12817	122	28	tuning	tuning	NOUN
fcis-12817	122	29	on	on	ADP
fcis-12817	122	30	the	the	DET
fcis-12817	122	31	dataset	dataset	NOUN
fcis-12817	122	32	using	use	VERB
fcis-12817	122	33	the	the	DET
fcis-12817	122	34	roberta	roberta	PROPN
fcis-12817	122	35	-	-	PUNCT
fcis-12817	122	36	wwm	wwm	PROPN
fcis-12817	122	37	model	model	NOUN
fcis-12817	122	38	.	.	PUNCT
fcis-12817	123	1	(	(	PUNCT
fcis-12817	123	2	3	3	X
fcis-12817	123	3	)	)	PUNCT
fcis-12817	123	4	roberta	roberta	PROPN
fcis-12817	123	5	-	-	PUNCT
fcis-12817	123	6	wwm	wwm	PROPN
fcis-12817	123	7	model	model	NOUN
fcis-12817	123	8	based	base	VERB
fcis-12817	123	9	on	on	ADP
fcis-12817	123	10	mrc	mrc	NOUN
fcis-12817	123	11	:	:	PUNCT
fcis-12817	123	12	mrc	mrc	NOUN
fcis-12817	123	13	is	be	AUX
fcis-12817	123	14	used	use	VERB
fcis-12817	123	15	as	as	ADP
fcis-12817	123	16	a	a	DET
fcis-12817	123	17	modeling	modeling	NOUN
fcis-12817	123	18	paradigm	paradigm	NOUN
fcis-12817	123	19	and	and	CCONJ
fcis-12817	123	20	roberta	roberta	PROPN
fcis-12817	123	21	-	-	PUNCT
fcis-12817	123	22	wwm	wwm	PROPN
fcis-12817	123	23	model	model	NOUN
fcis-12817	123	24	is	be	AUX
fcis-12817	123	25	utilized	utilize	VERB
fcis-12817	123	26	as	as	ADP
fcis-12817	123	27	an	an	DET
fcis-12817	123	28	encoder	encoder	NOUN
fcis-12817	123	29	.	.	PUNCT
fcis-12817	124	1	jd	jd	PROPN
fcis-12817	124	2	-	-	PUNCT
fcis-12817	124	3	e	e	NOUN
fcis-12817	124	4	-	-	NOUN
fcis-12817	124	5	commerce	commerce	NOUN
fcis-12817	124	6	:	:	PUNCT
fcis-12817	124	7	tables	table	NOUN
fcis-12817	124	8	3	3	NUM
fcis-12817	124	9	show	show	VERB
fcis-12817	124	10	the	the	DET
fcis-12817	124	11	performance	performance	NOUN
fcis-12817	124	12	of	of	ADP
fcis-12817	124	13	our	our	PRON
fcis-12817	124	14	model	model	NOUN
fcis-12817	124	15	on	on	ADP
fcis-12817	124	16	jd	jd	PROPN
fcis-12817	124	17	-	-	ADJ
fcis-12817	124	18	e	e	PROPN
fcis-12817	124	19	-	-	NOUN
fcis-12817	124	20	commerce	commerce	NOUN
fcis-12817	124	21	dataset	dataset	NOUN
fcis-12817	124	22	.	.	PUNCT
fcis-12817	125	1	from	from	ADP
fcis-12817	125	2	the	the	DET
fcis-12817	125	3	results	result	NOUN
fcis-12817	125	4	,	,	PUNCT
fcis-12817	125	5	it	it	PRON
fcis-12817	125	6	can	can	AUX
fcis-12817	125	7	be	be	AUX
fcis-12817	125	8	seen	see	VERB
fcis-12817	125	9	that	that	SCONJ
fcis-12817	125	10	our	our	PRON
fcis-12817	125	11	proposed	propose	VERB
fcis-12817	125	12	model	model	NOUN
fcis-12817	125	13	gets	get	VERB
fcis-12817	125	14	the	the	DET
fcis-12817	125	15	highest	high	ADJ
fcis-12817	125	16	f1	f1	NOUN
fcis-12817	125	17	value	value	NOUN
fcis-12817	125	18	compared	compare	VERB
fcis-12817	125	19	to	to	ADP
fcis-12817	125	20	other	other	ADJ
fcis-12817	125	21	models	model	NOUN
fcis-12817	125	22	.	.	PUNCT
fcis-12817	126	1	comparing	compare	VERB
fcis-12817	126	2	with	with	ADP
fcis-12817	126	3	the	the	DET
fcis-12817	126	4	classical	classical	ADJ
fcis-12817	126	5	model	model	NOUN
fcis-12817	126	6	lstm+crf	lstm+crf	NOUN
fcis-12817	126	7	,	,	PUNCT
fcis-12817	126	8	it	it	PRON
fcis-12817	126	9	improves	improve	VERB
fcis-12817	126	10	4.52	4.52	NUM
fcis-12817	126	11	%	%	NOUN
fcis-12817	126	12	in	in	ADP
fcis-12817	126	13	f1	f1	NOUN
fcis-12817	126	14	value	value	NOUN
fcis-12817	126	15	metrics	metric	NOUN
fcis-12817	126	16	.	.	PUNCT
fcis-12817	127	1	comparing	compare	VERB
fcis-12817	127	2	with	with	ADP
fcis-12817	127	3	sequence	sequence	NOUN
fcis-12817	127	4	labeling	labeling	NOUN
fcis-12817	127	5	based	base	VERB
fcis-12817	127	6	roberta	roberta	PROPN
fcis-12817	127	7	-	-	PUNCT
fcis-12817	127	8	wwm	wwm	PROPN
fcis-12817	127	9	model	model	NOUN
fcis-12817	127	10	,	,	PUNCT
fcis-12817	127	11	it	it	PRON
fcis-12817	127	12	improves	improve	VERB
fcis-12817	127	13	1.44	1.44	NUM
fcis-12817	127	14	%	%	NOUN
fcis-12817	127	15	in	in	ADP
fcis-12817	127	16	f1	f1	NOUN
fcis-12817	127	17	value	value	NOUN
fcis-12817	127	18	.	.	PUNCT
fcis-12817	128	1	the	the	DET
fcis-12817	128	2	comparison	comparison	NOUN
fcis-12817	128	3	79	79	NUM
fcis-12817	128	4	between	between	ADP
fcis-12817	128	5	sequence	sequence	NOUN
fcis-12817	128	6	labeling	labeling	NOUN
fcis-12817	128	7	based	base	VERB
fcis-12817	128	8	roberta	roberta	PROPN
fcis-12817	128	9	-	-	PUNCT
fcis-12817	128	10	wwm	wwm	PROPN
fcis-12817	128	11	model	model	NOUN
fcis-12817	128	12	and	and	CCONJ
fcis-12817	128	13	mrc	mrc	NOUN
fcis-12817	128	14	based	base	VERB
fcis-12817	128	15	roberta	roberta	PROPN
fcis-12817	128	16	-	-	PUNCT
fcis-12817	128	17	wwm	wwm	PROPN
fcis-12817	128	18	,	,	PUNCT
fcis-12817	128	19	we	we	PRON
fcis-12817	128	20	can	can	AUX
fcis-12817	128	21	prove	prove	VERB
fcis-12817	128	22	that	that	SCONJ
fcis-12817	128	23	the	the	DET
fcis-12817	128	24	modeling	modeling	NOUN
fcis-12817	128	25	approach	approach	NOUN
fcis-12817	128	26	of	of	ADP
fcis-12817	128	27	reading	reading	NOUN
fcis-12817	128	28	comprehension	comprehension	NOUN
fcis-12817	128	29	can	can	AUX
fcis-12817	128	30	effectively	effectively	ADV
fcis-12817	128	31	improve	improve	VERB
fcis-12817	128	32	the	the	DET
fcis-12817	128	33	performance	performance	NOUN
fcis-12817	128	34	of	of	ADP
fcis-12817	128	35	the	the	DET
fcis-12817	128	36	model	model	NOUN
fcis-12817	128	37	.	.	PUNCT
fcis-12817	129	1	table	table	NOUN
fcis-12817	129	2	3	3	NUM
fcis-12817	129	3	.	.	PUNCT
fcis-12817	130	1	experimental	experimental	ADJ
fcis-12817	130	2	comparison	comparison	NOUN
fcis-12817	130	3	results	result	VERB
fcis-12817	130	4	in	in	ADP
fcis-12817	130	5	the	the	DET
fcis-12817	130	6	jd	jd	PROPN
fcis-12817	130	7	-	-	PUNCT
fcis-12817	130	8	e	e	PROPN
fcis-12817	130	9	-	-	NOUN
fcis-12817	130	10	commerce	commerce	NOUN
fcis-12817	130	11	model	model	NOUN
fcis-12817	130	12	p	p	PROPN
fcis-12817	130	13	r	r	NOUN
fcis-12817	130	14	f1	f1	NOUN
fcis-12817	130	15	bi	bi	NOUN
fcis-12817	130	16	-	-	NOUN
fcis-12817	130	17	lstm+crf	lstm+crf	NOUN
fcis-12817	130	18	82.91	82.91	NUM
fcis-12817	130	19	81.40	81.40	NUM
fcis-12817	130	20	82.14	82.14	NUM
fcis-12817	130	21	roberta	roberta	PROPN
fcis-12817	130	22	-	-	PUNCT
fcis-12817	130	23	wwm	wwm	PROPN
fcis-12817	130	24	based	base	VERB
fcis-12817	130	25	on	on	ADP
fcis-12817	130	26	sequence	sequence	NOUN
fcis-12817	130	27	labeling	label	VERB
fcis-12817	130	28	84.75	84.75	NUM
fcis-12817	130	29	85.69	85.69	NUM
fcis-12817	130	30	85.22	85.22	NUM
fcis-12817	130	31	roberta	roberta	PROPN
fcis-12817	130	32	-	-	PUNCT
fcis-12817	130	33	wwm	wwm	PROPN
fcis-12817	130	34	based	base	VERB
fcis-12817	130	35	on	on	ADP
fcis-12817	130	36	mrc	mrc	NOUN
fcis-12817	130	37	86.27	86.27	NUM
fcis-12817	130	38	85.53	85.53	NUM
fcis-12817	130	39	85.90	85.90	NUM
fcis-12817	130	40	our	our	PRON
fcis-12817	130	41	model	model	NOUN
fcis-12817	130	42	87.19	87.19	NUM
fcis-12817	130	43	86.13	86.13	NUM
fcis-12817	130	44	86.66	86.66	NUM
fcis-12817	130	45	e	e	NOUN
fcis-12817	130	46	-	-	NOUN
fcis-12817	130	47	commerce	commerce	NOUN
fcis-12817	130	48	:	:	PUNCT
fcis-12817	130	49	table	table	NOUN
fcis-12817	130	50	4	4	NUM
fcis-12817	130	51	shows	show	VERB
fcis-12817	130	52	the	the	DET
fcis-12817	130	53	performance	performance	NOUN
fcis-12817	130	54	of	of	ADP
fcis-12817	130	55	our	our	PRON
fcis-12817	130	56	model	model	NOUN
fcis-12817	130	57	on	on	ADP
fcis-12817	130	58	the	the	DET
fcis-12817	130	59	e	e	PROPN
fcis-12817	130	60	-	-	NOUN
fcis-12817	130	61	commerce	commerce	NOUN
fcis-12817	130	62	dataset	dataset	NOUN
fcis-12817	130	63	.	.	PUNCT
fcis-12817	131	1	we	we	PRON
fcis-12817	131	2	use	use	VERB
fcis-12817	131	3	the	the	DET
fcis-12817	131	4	model	model	NOUN
fcis-12817	131	5	proposed	propose	VERB
fcis-12817	131	6	by	by	ADP
fcis-12817	131	7	ding	ding	NOUN
fcis-12817	131	8	et	et	PROPN
fcis-12817	131	9	al	al	PROPN
fcis-12817	132	1	[	[	X
fcis-12817	132	2	12	12	NUM
fcis-12817	132	3	]	]	PUNCT
fcis-12817	132	4	,	,	PUNCT
fcis-12817	132	5	the	the	DET
fcis-12817	132	6	producers	producer	NOUN
fcis-12817	132	7	of	of	ADP
fcis-12817	132	8	the	the	DET
fcis-12817	132	9	e	e	PROPN
fcis-12817	132	10	-	-	PROPN
fcis-12817	132	11	commerce	commerce	NOUN
fcis-12817	132	12	dataset	dataset	NOUN
fcis-12817	132	13	,	,	PUNCT
fcis-12817	132	14	as	as	ADP
fcis-12817	132	15	a	a	DET
fcis-12817	132	16	comparative	comparative	ADJ
fcis-12817	132	17	model	model	NOUN
fcis-12817	132	18	.	.	PUNCT
fcis-12817	133	1	their	their	PRON
fcis-12817	133	2	model	model	NOUN
fcis-12817	133	3	is	be	AUX
fcis-12817	133	4	based	base	VERB
fcis-12817	133	5	on	on	ADP
fcis-12817	133	6	a	a	DET
fcis-12817	133	7	graph	graph	NOUN
fcis-12817	133	8	neural	neural	ADJ
fcis-12817	133	9	network	network	NOUN
fcis-12817	133	10	that	that	PRON
fcis-12817	133	11	captures	capture	VERB
fcis-12817	133	12	the	the	DET
fcis-12817	133	13	information	information	NOUN
fcis-12817	133	14	in	in	ADP
fcis-12817	133	15	the	the	DET
fcis-12817	133	16	gazetteer	gazetteer	NOUN
fcis-12817	133	17	through	through	ADP
fcis-12817	133	18	a	a	DET
fcis-12817	133	19	multi	multi	ADJ
fcis-12817	133	20	-	-	ADJ
fcis-12817	133	21	graph	graph	NOUN
fcis-12817	133	22	structure	structure	NOUN
fcis-12817	133	23	.	.	PUNCT
fcis-12817	134	1	comparing	compare	VERB
fcis-12817	134	2	with	with	ADP
fcis-12817	134	3	the	the	DET
fcis-12817	134	4	classical	classical	ADJ
fcis-12817	134	5	bi	bi	NOUN
fcis-12817	134	6	-	-	ADJ
fcis-12817	134	7	lstm	lstm	ADJ
fcis-12817	134	8	+	+	CCONJ
fcis-12817	134	9	crf	crf	PROPN
fcis-12817	134	10	and	and	CCONJ
fcis-12817	134	11	the	the	DET
fcis-12817	134	12	model	model	NOUN
fcis-12817	134	13	proposed	propose	VERB
fcis-12817	134	14	ding	ding	PROPN
fcis-12817	134	15	et	et	PROPN
fcis-12817	134	16	al	al	PROPN
fcis-12817	134	17	,	,	PUNCT
fcis-12817	134	18	our	our	PRON
fcis-12817	134	19	model	model	NOUN
fcis-12817	134	20	improves	improve	VERB
fcis-12817	134	21	5.80	5.80	NUM
fcis-12817	134	22	%	%	NOUN
fcis-12817	134	23	and	and	CCONJ
fcis-12817	134	24	4.13	4.13	NUM
fcis-12817	134	25	%	%	NOUN
fcis-12817	134	26	in	in	ADP
fcis-12817	134	27	f1	f1	NOUN
fcis-12817	134	28	,	,	PUNCT
fcis-12817	134	29	respectively	respectively	ADV
fcis-12817	134	30	.	.	PUNCT
fcis-12817	135	1	comparing	compare	VERB
fcis-12817	135	2	with	with	ADP
fcis-12817	135	3	sequence	sequence	NOUN
fcis-12817	135	4	labeling	labeling	NOUN
fcis-12817	135	5	based	base	VERB
fcis-12817	135	6	roberta	roberta	PROPN
fcis-12817	135	7	-	-	PUNCT
fcis-12817	135	8	wwm	wwm	PROPN
fcis-12817	135	9	model	model	NOUN
fcis-12817	135	10	and	and	CCONJ
fcis-12817	135	11	mrc	mrc	NOUN
fcis-12817	135	12	-	-	PUNCT
fcis-12817	135	13	based	base	VERB
fcis-12817	135	14	roberta	roberta	PROPN
fcis-12817	135	15	-	-	PUNCT
fcis-12817	135	16	wwm	wwm	PROPN
fcis-12817	135	17	model	model	NOUN
fcis-12817	135	18	,	,	PUNCT
fcis-12817	135	19	it	it	PRON
fcis-12817	135	20	improves	improve	VERB
fcis-12817	135	21	1.17	1.17	NUM
fcis-12817	135	22	%	%	NOUN
fcis-12817	135	23	and	and	CCONJ
fcis-12817	135	24	0.51	0.51	NUM
fcis-12817	135	25	%	%	NOUN
fcis-12817	135	26	in	in	ADP
fcis-12817	135	27	f1	f1	NOUN
fcis-12817	135	28	value	value	NOUN
fcis-12817	135	29	,	,	PUNCT
fcis-12817	135	30	respectively	respectively	ADV
fcis-12817	135	31	.	.	PUNCT
fcis-12817	136	1	table	table	NOUN
fcis-12817	136	2	4	4	NUM
fcis-12817	136	3	.	.	PUNCT
fcis-12817	136	4	experimental	experimental	ADJ
fcis-12817	136	5	results	result	NOUN
fcis-12817	136	6	in	in	ADP
fcis-12817	136	7	e	e	NOUN
fcis-12817	136	8	-	-	NOUN
fcis-12817	136	9	commerce	commerce	NOUN
fcis-12817	136	10	model	model	NOUN
fcis-12817	136	11	p	p	PROPN
fcis-12817	136	12	r	r	NOUN
fcis-12817	136	13	f1	f1	NOUN
fcis-12817	136	14	bi	bi	NOUN
fcis-12817	136	15	-	-	NOUN
fcis-12817	136	16	lstm+crf	lstm+crf	NOUN
fcis-12817	136	17	76.14	76.14	NUM
fcis-12817	136	18	71.17	71.17	NUM
fcis-12817	136	19	73.57	73.57	NUM
fcis-12817	136	20	ding	ding	NOUN
fcis-12817	136	21	et	et	NOUN
fcis-12817	136	22	al.[13	al.[13	NOUN
fcis-12817	136	23	]	]	PUNCT
fcis-12817	136	24	76.21	76.21	NUM
fcis-12817	136	25	74.31	74.31	NUM
fcis-12817	136	26	75.24	75.24	NUM
fcis-12817	136	27	roberta	roberta	PROPN
fcis-12817	136	28	-	-	PUNCT
fcis-12817	136	29	wwm	wwm	PROPN
fcis-12817	136	30	based	base	VERB
fcis-12817	136	31	on	on	ADP
fcis-12817	136	32	sequence	sequence	NOUN
fcis-12817	136	33	labeling	label	VERB
fcis-12817	136	34	78.45	78.45	NUM
fcis-12817	136	35	77.97	77.97	NUM
fcis-12817	136	36	78.20	78.20	NUM
fcis-12817	136	37	roberta	roberta	PROPN
fcis-12817	136	38	-	-	PUNCT
fcis-12817	136	39	wwm	wwm	PROPN
fcis-12817	136	40	based	base	VERB
fcis-12817	136	41	on	on	ADP
fcis-12817	136	42	mrc	mrc	NOUN
fcis-12817	136	43	78.01	78.01	NUM
fcis-12817	136	44	77.72	77.72	NUM
fcis-12817	136	45	78.86	78.86	NUM
fcis-12817	136	46	our	our	PRON
fcis-12817	136	47	model	model	NOUN
fcis-12817	136	48	79.96	79.96	NUM
fcis-12817	136	49	78.79	78.79	NUM
fcis-12817	136	50	79.37	79.37	NUM
fcis-12817	136	51	5.4	5.4	NUM
fcis-12817	136	52	.	.	PUNCT
fcis-12817	137	1	different	different	ADJ
fcis-12817	137	2	question	question	NOUN
fcis-12817	137	3	templates	template	NOUN
fcis-12817	137	4	comparison	comparison	NOUN
fcis-12817	137	5	experiment	experiment	NOUN
fcis-12817	137	6	since	since	SCONJ
fcis-12817	137	7	the	the	DET
fcis-12817	137	8	question	question	NOUN
fcis-12817	137	9	contains	contain	VERB
fcis-12817	137	10	a	a	DET
fcis-12817	137	11	priori	priori	ADJ
fcis-12817	137	12	information	information	NOUN
fcis-12817	137	13	,	,	PUNCT
fcis-12817	137	14	the	the	DET
fcis-12817	137	15	accuracy	accuracy	NOUN
fcis-12817	137	16	of	of	ADP
fcis-12817	137	17	the	the	DET
fcis-12817	137	18	a	a	DET
fcis-12817	137	19	priori	priori	ADJ
fcis-12817	137	20	information	information	NOUN
fcis-12817	137	21	will	will	AUX
fcis-12817	137	22	have	have	VERB
fcis-12817	137	23	an	an	DET
fcis-12817	137	24	impact	impact	NOUN
fcis-12817	137	25	on	on	ADP
fcis-12817	137	26	the	the	DET
fcis-12817	137	27	answer	answer	NOUN
fcis-12817	137	28	.	.	PUNCT
fcis-12817	138	1	if	if	SCONJ
fcis-12817	138	2	the	the	DET
fcis-12817	138	3	a	a	DET
fcis-12817	138	4	priori	priori	ADJ
fcis-12817	138	5	information	information	NOUN
fcis-12817	138	6	is	be	AUX
fcis-12817	138	7	wrong	wrong	ADJ
fcis-12817	138	8	,	,	PUNCT
fcis-12817	138	9	it	it	PRON
fcis-12817	138	10	will	will	AUX
fcis-12817	138	11	make	make	VERB
fcis-12817	138	12	the	the	DET
fcis-12817	138	13	whole	whole	ADJ
fcis-12817	138	14	model	model	NOUN
fcis-12817	138	15	perform	perform	VERB
fcis-12817	138	16	poorly	poorly	ADV
fcis-12817	138	17	.	.	PUNCT
fcis-12817	139	1	meanwhile	meanwhile	ADV
fcis-12817	139	2	,	,	PUNCT
fcis-12817	139	3	the	the	DET
fcis-12817	139	4	richness	richness	NOUN
fcis-12817	139	5	of	of	ADP
fcis-12817	139	6	the	the	DET
fcis-12817	139	7	a	a	DET
fcis-12817	139	8	priori	priori	ADJ
fcis-12817	139	9	information	information	NOUN
fcis-12817	139	10	will	will	AUX
fcis-12817	139	11	also	also	ADV
fcis-12817	139	12	have	have	VERB
fcis-12817	139	13	an	an	DET
fcis-12817	139	14	impact	impact	NOUN
fcis-12817	139	15	on	on	ADP
fcis-12817	139	16	the	the	DET
fcis-12817	139	17	overall	overall	ADJ
fcis-12817	139	18	performance	performance	NOUN
fcis-12817	139	19	of	of	ADP
fcis-12817	139	20	the	the	DET
fcis-12817	139	21	model	model	NOUN
fcis-12817	139	22	.	.	PUNCT
fcis-12817	140	1	in	in	ADP
fcis-12817	140	2	order	order	NOUN
fcis-12817	140	3	to	to	PART
fcis-12817	140	4	explore	explore	VERB
fcis-12817	140	5	the	the	DET
fcis-12817	140	6	impact	impact	NOUN
fcis-12817	140	7	of	of	ADP
fcis-12817	140	8	different	different	ADJ
fcis-12817	140	9	question	question	NOUN
fcis-12817	140	10	templates	template	NOUN
fcis-12817	140	11	on	on	ADP
fcis-12817	140	12	the	the	DET
fcis-12817	140	13	model	model	NOUN
fcis-12817	140	14	performance	performance	NOUN
fcis-12817	140	15	,	,	PUNCT
fcis-12817	140	16	we	we	PRON
fcis-12817	140	17	conducted	conduct	VERB
fcis-12817	140	18	corresponding	corresponding	ADJ
fcis-12817	140	19	experiments	experiment	NOUN
fcis-12817	140	20	for	for	ADP
fcis-12817	140	21	several	several	ADJ
fcis-12817	140	22	different	different	ADJ
fcis-12817	140	23	question	question	NOUN
fcis-12817	140	24	template	template	NOUN
fcis-12817	140	25	about	about	ADP
fcis-12817	140	26	e	e	NOUN
fcis-12817	140	27	-	-	NOUN
fcis-12817	140	28	commerce	commerce	NOUN
fcis-12817	140	29	related	relate	VERB
fcis-12817	140	30	entities	entity	NOUN
fcis-12817	140	31	.	.	PUNCT
fcis-12817	141	1	we	we	PRON
fcis-12817	141	2	constructed	construct	VERB
fcis-12817	141	3	corresponding	correspond	VERB
fcis-12817	141	4	q&a	q&a	PROPN
fcis-12817	141	5	datasets	dataset	NOUN
fcis-12817	141	6	for	for	ADP
fcis-12817	141	7	each	each	PRON
fcis-12817	141	8	of	of	ADP
fcis-12817	141	9	the	the	DET
fcis-12817	141	10	following	follow	VERB
fcis-12817	141	11	three	three	NUM
fcis-12817	141	12	questioning	questioning	NOUN
fcis-12817	141	13	templates	template	NOUN
fcis-12817	141	14	.	.	PUNCT
fcis-12817	142	1	(	(	PUNCT
fcis-12817	142	2	1	1	X
fcis-12817	142	3	)	)	PUNCT
fcis-12817	142	4	template	template	NOUN
fcis-12817	142	5	1	1	NUM
fcis-12817	142	6	:	:	PUNCT
fcis-12817	142	7	describe	describe	VERB
fcis-12817	142	8	the	the	DET
fcis-12817	142	9	entity	entity	NOUN
fcis-12817	142	10	category	category	NOUN
fcis-12817	142	11	in	in	ADP
fcis-12817	142	12	as	as	ADV
fcis-12817	142	13	much	much	ADJ
fcis-12817	142	14	detail	detail	NOUN
fcis-12817	142	15	as	as	ADP
fcis-12817	142	16	possible	possible	ADJ
fcis-12817	142	17	by	by	ADP
fcis-12817	142	18	referring	refer	VERB
fcis-12817	142	19	to	to	ADP
fcis-12817	142	20	the	the	DET
fcis-12817	142	21	annotation	annotation	NOUN
fcis-12817	142	22	template	template	NOUN
fcis-12817	142	23	used	use	VERB
fcis-12817	142	24	to	to	PART
fcis-12817	142	25	annotate	annotate	VERB
fcis-12817	142	26	the	the	DET
fcis-12817	142	27	entity	entity	NOUN
fcis-12817	142	28	.	.	PUNCT
fcis-12817	143	1	adopt	adopt	VERB
fcis-12817	143	2	the	the	DET
fcis-12817	143	3	annotation	annotation	NOUN
fcis-12817	143	4	template	template	NOUN
fcis-12817	143	5	that	that	PRON
fcis-12817	143	6	we	we	PRON
fcis-12817	143	7	use	use	VERB
fcis-12817	143	8	when	when	SCONJ
fcis-12817	143	9	manually	manually	ADV
fcis-12817	143	10	annotating	annotate	VERB
fcis-12817	143	11	entities	entity	NOUN
fcis-12817	143	12	.	.	PUNCT
fcis-12817	144	1	(	(	PUNCT
fcis-12817	144	2	2	2	X
fcis-12817	144	3	)	)	PUNCT
fcis-12817	144	4	template	template	NOUN
fcis-12817	144	5	2	2	NUM
fcis-12817	144	6	:	:	PUNCT
fcis-12817	144	7	construct	construct	VERB
fcis-12817	144	8	questions	question	NOUN
fcis-12817	144	9	directly	directly	ADV
fcis-12817	144	10	with	with	ADP
fcis-12817	144	11	keywords	keyword	NOUN
fcis-12817	144	12	for	for	ADP
fcis-12817	144	13	entity	entity	NOUN
fcis-12817	144	14	categories	category	NOUN
fcis-12817	144	15	.	.	PUNCT
fcis-12817	145	1	(	(	PUNCT
fcis-12817	145	2	3	3	X
fcis-12817	145	3	)	)	PUNCT
fcis-12817	145	4	template	template	NOUN
fcis-12817	145	5	3	3	NUM
fcis-12817	145	6	:	:	PUNCT
fcis-12817	145	7	constructing	construct	VERB
fcis-12817	145	8	questions	question	NOUN
fcis-12817	145	9	with	with	ADP
fcis-12817	145	10	explanations	explanation	NOUN
fcis-12817	145	11	corresponding	correspond	VERB
fcis-12817	145	12	to	to	ADP
fcis-12817	145	13	entity	entity	NOUN
fcis-12817	145	14	categories	category	NOUN
fcis-12817	145	15	.	.	PUNCT
fcis-12817	146	1	we	we	PRON
fcis-12817	146	2	refer	refer	VERB
fcis-12817	146	3	to	to	ADP
fcis-12817	146	4	wikipedia	wikipedia	PROPN
fcis-12817	146	5	's	's	PART
fcis-12817	146	6	explanation	explanation	NOUN
fcis-12817	146	7	.	.	PUNCT
fcis-12817	147	1	specific	specific	ADJ
fcis-12817	147	2	details	detail	NOUN
fcis-12817	147	3	about	about	ADP
fcis-12817	147	4	the	the	DET
fcis-12817	147	5	question	question	NOUN
fcis-12817	147	6	template	template	NOUN
fcis-12817	147	7	are	be	AUX
fcis-12817	147	8	shown	show	VERB
fcis-12817	147	9	in	in	ADP
fcis-12817	147	10	table	table	NOUN
fcis-12817	147	11	5	5	NUM
fcis-12817	147	12	.	.	PUNCT
fcis-12817	147	13	table	table	NOUN
fcis-12817	147	14	5	5	NUM
fcis-12817	147	15	.	.	PUNCT
fcis-12817	148	1	different	different	ADJ
fcis-12817	148	2	questioning	questioning	NOUN
fcis-12817	148	3	templates	template	NOUN
fcis-12817	148	4	question	question	NOUN
fcis-12817	148	5	template	template	NOUN
fcis-12817	148	6	brand	brand	NOUN
fcis-12817	148	7	entity	entity	NOUN
fcis-12817	148	8	product	product	NOUN
fcis-12817	148	9	entity	entity	NOUN
fcis-12817	148	10	template	template	NOUN
fcis-12817	148	11	1	1	NUM
fcis-12817	148	12	identify	identify	VERB
fcis-12817	148	13	the	the	DET
fcis-12817	148	14	brands	brand	NOUN
fcis-12817	148	15	in	in	ADP
fcis-12817	148	16	the	the	DET
fcis-12817	148	17	sentence	sentence	NOUN
fcis-12817	148	18	,	,	PUNCT
fcis-12817	148	19	including	include	VERB
fcis-12817	148	20	companies	company	NOUN
fcis-12817	148	21	.	.	PUNCT
fcis-12817	149	1	identify	identify	VERB
fcis-12817	149	2	the	the	DET
fcis-12817	149	3	products	product	NOUN
fcis-12817	149	4	mentioned	mention	VERB
fcis-12817	149	5	in	in	ADP
fcis-12817	149	6	the	the	DET
fcis-12817	149	7	sentence	sentence	NOUN
fcis-12817	149	8	,	,	PUNCT
fcis-12817	149	9	including	include	VERB
fcis-12817	149	10	goods	good	NOUN
fcis-12817	149	11	.	.	PUNCT
fcis-12817	150	1	template	template	NOUN
fcis-12817	150	2	2	2	NUM
fcis-12817	150	3	find	find	VERB
fcis-12817	150	4	the	the	DET
fcis-12817	150	5	brands	brand	NOUN
fcis-12817	150	6	mentioned	mention	VERB
fcis-12817	150	7	in	in	ADP
fcis-12817	150	8	the	the	DET
fcis-12817	150	9	sentence	sentence	NOUN
fcis-12817	150	10	find	find	VERB
fcis-12817	150	11	the	the	DET
fcis-12817	150	12	product	product	NOUN
fcis-12817	150	13	mentioned	mention	VERB
fcis-12817	150	14	in	in	ADP
fcis-12817	150	15	the	the	DET
fcis-12817	150	16	sentence	sentence	NOUN
fcis-12817	150	17	template	template	NOUN
fcis-12817	150	18	3	3	NUM
fcis-12817	150	19	identify	identify	VERB
fcis-12817	150	20	the	the	DET
fcis-12817	150	21	symbolic	symbolic	ADJ
fcis-12817	150	22	signs	sign	NOUN
fcis-12817	150	23	in	in	ADP
fcis-12817	150	24	the	the	DET
fcis-12817	150	25	sentence	sentence	NOUN
fcis-12817	150	26	that	that	PRON
fcis-12817	150	27	indicate	indicate	VERB
fcis-12817	150	28	a	a	DET
fcis-12817	150	29	product	product	NOUN
fcis-12817	150	30	to	to	ADP
fcis-12817	150	31	the	the	DET
fcis-12817	150	32	public	public	NOUN
fcis-12817	150	33	find	find	VERB
fcis-12817	150	34	anything	anything	PRON
fcis-12817	150	35	produced	produce	VERB
fcis-12817	150	36	in	in	ADP
fcis-12817	150	37	the	the	DET
fcis-12817	150	38	sentence	sentence	NOUN
fcis-12817	150	39	we	we	PRON
fcis-12817	150	40	conducted	conduct	VERB
fcis-12817	150	41	the	the	DET
fcis-12817	150	42	corresponding	correspond	VERB
fcis-12817	150	43	experiments	experiment	NOUN
fcis-12817	150	44	on	on	ADP
fcis-12817	150	45	the	the	DET
fcis-12817	150	46	jd	jd	PROPN
fcis-12817	150	47	-	-	NOUN
fcis-12817	150	48	ecommerce	ecommerce	NOUN
fcis-12817	150	49	as	as	ADV
fcis-12817	150	50	well	well	ADV
fcis-12817	150	51	as	as	ADP
fcis-12817	150	52	e	e	NOUN
fcis-12817	150	53	-	-	NOUN
fcis-12817	150	54	commerce	commerce	NOUN
fcis-12817	150	55	datasets	dataset	NOUN
fcis-12817	150	56	by	by	ADP
fcis-12817	150	57	constructing	construct	VERB
fcis-12817	150	58	the	the	DET
fcis-12817	150	59	corresponding	correspond	VERB
fcis-12817	150	60	dataset	dataset	NOUN
fcis-12817	150	61	forms	form	NOUN
fcis-12817	150	62	with	with	ADP
fcis-12817	150	63	these	these	DET
fcis-12817	150	64	three	three	NUM
fcis-12817	150	65	question	question	NOUN
fcis-12817	150	66	construction	construction	NOUN
fcis-12817	150	67	templates	template	NOUN
fcis-12817	150	68	,	,	PUNCT
fcis-12817	150	69	respectively	respectively	ADV
fcis-12817	150	70	.	.	PUNCT
fcis-12817	150	71	table	table	NOUN
fcis-12817	150	72	6	6	NUM
fcis-12817	150	73	and	and	CCONJ
fcis-12817	150	74	7	7	NUM
fcis-12817	150	75	shows	show	VERB
fcis-12817	150	76	the	the	DET
fcis-12817	150	77	experimental	experimental	ADJ
fcis-12817	150	78	results	result	NOUN
fcis-12817	150	79	of	of	ADP
fcis-12817	150	80	the	the	DET
fcis-12817	150	81	impact	impact	NOUN
fcis-12817	150	82	of	of	ADP
fcis-12817	150	83	different	different	ADJ
fcis-12817	150	84	questioning	questioning	NOUN
fcis-12817	150	85	templates	template	NOUN
fcis-12817	150	86	on	on	ADP
fcis-12817	150	87	the	the	DET
fcis-12817	150	88	model	model	NOUN
fcis-12817	150	89	performance	performance	NOUN
fcis-12817	150	90	.	.	PUNCT
fcis-12817	151	1	table	table	NOUN
fcis-12817	151	2	6	6	NUM
fcis-12817	151	3	.	.	PUNCT
fcis-12817	152	1	different	different	ADJ
fcis-12817	152	2	questioning	questioning	NOUN
fcis-12817	152	3	templates	template	NOUN
fcis-12817	152	4	in	in	ADP
fcis-12817	152	5	jd	jd	PROPN
fcis-12817	152	6	-	-	ADJ
fcis-12817	152	7	e	e	ADJ
fcis-12817	152	8	-	-	NOUN
fcis-12817	152	9	commerce	commerce	NOUN
fcis-12817	152	10	question	question	NOUN
fcis-12817	152	11	templates	template	NOUN
fcis-12817	152	12	p	p	NOUN
fcis-12817	152	13	r	r	NOUN
fcis-12817	152	14	f1	f1	NOUN
fcis-12817	152	15	template	template	NOUN
fcis-12817	152	16	1	1	NUM
fcis-12817	152	17	87.19	87.19	NUM
fcis-12817	152	18	86.13	86.13	NUM
fcis-12817	152	19	86.66	86.66	NUM
fcis-12817	152	20	template	template	NOUN
fcis-12817	152	21	2	2	NUM
fcis-12817	152	22	86.64	86.64	NUM
fcis-12817	152	23	86.21	86.21	NUM
fcis-12817	152	24	86.42	86.42	NUM
fcis-12817	152	25	template	template	NOUN
fcis-12817	152	26	3	3	NUM
fcis-12817	152	27	85.32	85.32	NUM
fcis-12817	152	28	84.14	84.14	NUM
fcis-12817	152	29	84.72	84.72	NUM
fcis-12817	152	30	table	table	NOUN
fcis-12817	152	31	7	7	NUM
fcis-12817	152	32	.	.	PUNCT
fcis-12817	153	1	different	different	ADJ
fcis-12817	153	2	questioning	questioning	NOUN
fcis-12817	153	3	templates	template	NOUN
fcis-12817	153	4	in	in	ADP
fcis-12817	153	5	e	e	NOUN
fcis-12817	153	6	-	-	NOUN
fcis-12817	153	7	commerce	commerce	NOUN
fcis-12817	153	8	question	question	NOUN
fcis-12817	153	9	templates	template	NOUN
fcis-12817	153	10	p	p	NOUN
fcis-12817	153	11	r	r	NOUN
fcis-12817	153	12	f1	f1	NOUN
fcis-12817	153	13	template	template	NOUN
fcis-12817	153	14	1	1	NUM
fcis-12817	153	15	79.96	79.96	NUM
fcis-12817	153	16	78.79	78.79	NUM
fcis-12817	153	17	79.37	79.37	NUM
fcis-12817	153	18	template	template	NOUN
fcis-12817	153	19	2	2	NUM
fcis-12817	153	20	79.87	79.87	NUM
fcis-12817	153	21	78.64	78.64	NUM
fcis-12817	153	22	79.25	79.25	NUM
fcis-12817	153	23	template	template	NOUN
fcis-12817	153	24	3	3	NUM
fcis-12817	153	25	78.49	78.49	NUM
fcis-12817	153	26	76.88	76.88	NUM
fcis-12817	153	27	77.67	77.67	NUM
fcis-12817	153	28	from	from	ADP
fcis-12817	153	29	the	the	DET
fcis-12817	153	30	results	result	NOUN
fcis-12817	153	31	,	,	PUNCT
fcis-12817	153	32	the	the	DET
fcis-12817	153	33	model	model	NOUN
fcis-12817	153	34	can	can	AUX
fcis-12817	153	35	obtain	obtain	VERB
fcis-12817	153	36	better	well	ADJ
fcis-12817	153	37	performance	performance	NOUN
fcis-12817	153	38	in	in	ADP
fcis-12817	153	39	the	the	DET
fcis-12817	153	40	e	e	PROPN
fcis-12817	153	41	-	-	NOUN
fcis-12817	153	42	commerce	commerce	NOUN
fcis-12817	153	43	domain	domain	NOUN
fcis-12817	153	44	when	when	SCONJ
fcis-12817	153	45	we	we	PRON
fcis-12817	153	46	construct	construct	VERB
fcis-12817	153	47	the	the	DET
fcis-12817	153	48	question	question	NOUN
fcis-12817	153	49	templates	template	NOUN
fcis-12817	153	50	with	with	ADP
fcis-12817	153	51	either	either	CCONJ
fcis-12817	153	52	manual	manual	ADJ
fcis-12817	153	53	annotations	annotation	NOUN
fcis-12817	153	54	or	or	CCONJ
fcis-12817	153	55	keywords	keyword	NOUN
fcis-12817	153	56	of	of	ADP
fcis-12817	153	57	entity	entity	NOUN
fcis-12817	153	58	categories	category	NOUN
fcis-12817	153	59	.	.	PUNCT
fcis-12817	154	1	the	the	DET
fcis-12817	154	2	manual	manual	ADJ
fcis-12817	154	3	annotation	annotation	NOUN
fcis-12817	154	4	of	of	ADP
fcis-12817	154	5	the	the	DET
fcis-12817	154	6	dataset	dataset	NOUN
fcis-12817	154	7	carries	carry	VERB
fcis-12817	154	8	more	more	ADV
fcis-12817	154	9	semantically	semantically	ADV
fcis-12817	154	10	similar	similar	ADJ
fcis-12817	154	11	words	word	NOUN
fcis-12817	154	12	,	,	PUNCT
fcis-12817	154	13	which	which	PRON
fcis-12817	154	14	helps	help	VERB
fcis-12817	154	15	the	the	DET
fcis-12817	154	16	model	model	NOUN
fcis-12817	154	17	to	to	PART
fcis-12817	154	18	accurately	accurately	ADV
fcis-12817	154	19	recognize	recognize	VERB
fcis-12817	154	20	the	the	DET
fcis-12817	154	21	entity	entity	NOUN
fcis-12817	154	22	categories	category	NOUN
fcis-12817	154	23	.	.	PUNCT
fcis-12817	155	1	when	when	SCONJ
fcis-12817	155	2	we	we	PRON
fcis-12817	155	3	construct	construct	VERB
fcis-12817	155	4	the	the	DET
fcis-12817	155	5	question	question	NOUN
fcis-12817	155	6	as	as	ADP
fcis-12817	155	7	an	an	DET
fcis-12817	155	8	interpretation	interpretation	NOUN
fcis-12817	155	9	of	of	ADP
fcis-12817	155	10	the	the	DET
fcis-12817	155	11	entity	entity	NOUN
fcis-12817	155	12	category	category	NOUN
fcis-12817	155	13	,	,	PUNCT
fcis-12817	155	14	it	it	PRON
fcis-12817	155	15	leads	lead	VERB
fcis-12817	155	16	to	to	ADP
fcis-12817	155	17	a	a	DET
fcis-12817	155	18	decrease	decrease	NOUN
fcis-12817	155	19	in	in	ADP
fcis-12817	155	20	the	the	DET
fcis-12817	155	21	performance	performance	NOUN
fcis-12817	155	22	of	of	ADP
fcis-12817	155	23	entity	entity	NOUN
fcis-12817	155	24	recognition	recognition	NOUN
fcis-12817	155	25	,	,	PUNCT
fcis-12817	155	26	where	where	SCONJ
fcis-12817	155	27	the	the	DET
fcis-12817	155	28	f1	f1	ADJ
fcis-12817	155	29	value	value	NOUN
fcis-12817	155	30	of	of	ADP
fcis-12817	155	31	the	the	DET
fcis-12817	155	32	model	model	NOUN
fcis-12817	155	33	decreases	decrease	VERB
fcis-12817	155	34	to	to	ADP
fcis-12817	155	35	the	the	DET
fcis-12817	155	36	extent	extent	NOUN
fcis-12817	155	37	of	of	ADP
fcis-12817	155	38	1.94	1.94	NUM
fcis-12817	155	39	%	%	NOUN
fcis-12817	155	40	as	as	ADV
fcis-12817	155	41	well	well	ADV
fcis-12817	155	42	as	as	ADP
fcis-12817	155	43	1.7	1.7	NUM
fcis-12817	155	44	%	%	NOUN
fcis-12817	155	45	on	on	ADP
fcis-12817	155	46	the	the	DET
fcis-12817	155	47	jd	jd	PROPN
fcis-12817	155	48	-	-	PUNCT
fcis-12817	155	49	e	e	PROPN
fcis-12817	155	50	-	-	NOUN
fcis-12817	155	51	commerce	commerce	NOUN
fcis-12817	155	52	dataset	dataset	NOUN
fcis-12817	155	53	as	as	ADV
fcis-12817	155	54	well	well	ADV
fcis-12817	155	55	as	as	ADP
fcis-12817	155	56	on	on	ADP
fcis-12817	155	57	the	the	DET
fcis-12817	155	58	e	e	NOUN
fcis-12817	155	59	-	-	NOUN
fcis-12817	155	60	commerce	commerce	NOUN
fcis-12817	155	61	dataset	dataset	NOUN
fcis-12817	155	62	.	.	PUNCT
fcis-12817	156	1	this	this	DET
fcis-12817	156	2	excessive	excessive	ADJ
fcis-12817	156	3	interpretation	interpretation	NOUN
fcis-12817	156	4	prevents	prevent	VERB
fcis-12817	156	5	the	the	DET
fcis-12817	156	6	model	model	NOUN
fcis-12817	156	7	from	from	ADP
fcis-12817	156	8	using	use	VERB
fcis-12817	156	9	it	it	PRON
fcis-12817	156	10	as	as	ADV
fcis-12817	156	11	explicit	explicit	ADJ
fcis-12817	156	12	a	a	DET
fcis-12817	156	13	priori	priori	ADJ
fcis-12817	156	14	information	information	NOUN
fcis-12817	156	15	.	.	PUNCT
fcis-12817	157	1	therefore	therefore	ADV
fcis-12817	157	2	,	,	PUNCT
fcis-12817	157	3	when	when	SCONJ
fcis-12817	157	4	there	there	PRON
fcis-12817	157	5	is	be	VERB
fcis-12817	157	6	a	a	DET
fcis-12817	157	7	bias	bias	NOUN
fcis-12817	157	8	in	in	ADP
fcis-12817	157	9	the	the	DET
fcis-12817	157	10	questions	question	NOUN
fcis-12817	157	11	,	,	PUNCT
fcis-12817	157	12	it	it	PRON
fcis-12817	157	13	can	can	AUX
fcis-12817	157	14	lead	lead	VERB
fcis-12817	157	15	to	to	ADP
fcis-12817	157	16	an	an	DET
fcis-12817	157	17	overall	overall	ADJ
fcis-12817	157	18	performance	performance	NOUN
fcis-12817	157	19	degradation	degradation	NOUN
fcis-12817	157	20	.	.	PUNCT
fcis-12817	158	1	the	the	DET
fcis-12817	158	2	best	good	ADJ
fcis-12817	158	3	way	way	NOUN
fcis-12817	158	4	is	be	AUX
fcis-12817	158	5	to	to	PART
fcis-12817	158	6	construct	construct	VERB
fcis-12817	158	7	the	the	DET
fcis-12817	158	8	question	question	NOUN
fcis-12817	158	9	based	base	VERB
fcis-12817	158	10	on	on	ADP
fcis-12817	158	11	the	the	DET
fcis-12817	158	12	annotation	annotation	NOUN
fcis-12817	158	13	template	template	NOUN
fcis-12817	158	14	during	during	ADP
fcis-12817	158	15	manual	manual	ADJ
fcis-12817	158	16	annotation	annotation	NOUN
fcis-12817	158	17	,	,	PUNCT
fcis-12817	158	18	giving	give	VERB
fcis-12817	158	19	some	some	PRON
fcis-12817	158	20	of	of	ADP
fcis-12817	158	21	the	the	DET
fcis-12817	158	22	proximate	proximate	NOUN
fcis-12817	158	23	synonyms	synonym	NOUN
fcis-12817	158	24	of	of	ADP
fcis-12817	158	25	the	the	DET
fcis-12817	158	26	entity	entity	NOUN
fcis-12817	158	27	types	type	NOUN
fcis-12817	158	28	to	to	PART
fcis-12817	158	29	facilitate	facilitate	VERB
fcis-12817	158	30	semantic	semantic	ADJ
fcis-12817	158	31	matching	matching	NOUN
fcis-12817	158	32	by	by	ADP
fcis-12817	158	33	the	the	DET
fcis-12817	158	34	model	model	NOUN
fcis-12817	158	35	.	.	PUNCT
fcis-12817	159	1	5.5	5.5	NUM
fcis-12817	159	2	.	.	PUNCT
fcis-12817	159	3	effect	effect	NOUN
fcis-12817	159	4	of	of	ADP
fcis-12817	159	5	different	different	ADJ
fcis-12817	159	6	loss	loss	NOUN
fcis-12817	159	7	function	function	NOUN
fcis-12817	159	8	weights	weight	NOUN
fcis-12817	159	9	on	on	ADP
fcis-12817	159	10	the	the	DET
fcis-12817	159	11	model	model	NOUN
fcis-12817	159	12	the	the	DET
fcis-12817	159	13	target	target	NOUN
fcis-12817	159	14	loss	loss	NOUN
fcis-12817	159	15	function	function	NOUN
fcis-12817	159	16	consists	consist	VERB
fcis-12817	159	17	of	of	ADP
fcis-12817	159	18	loss	loss	NOUN
fcis-12817	159	19	functions	function	NOUN
fcis-12817	159	20	used	use	VERB
fcis-12817	159	21	to	to	PART
fcis-12817	159	22	control	control	VERB
fcis-12817	159	23	the	the	DET
fcis-12817	159	24	entity	entity	NOUN
fcis-12817	159	25	span	span	NOUN
fcis-12817	159	26	matching	matching	NOUN
fcis-12817	159	27	,	,	PUNCT
fcis-12817	159	28	entity	entity	NOUN
fcis-12817	159	29	head	head	NOUN
fcis-12817	159	30	position	position	NOUN
fcis-12817	159	31	and	and	CCONJ
fcis-12817	159	32	tail	tail	NOUN
fcis-12817	159	33	position	position	NOUN
fcis-12817	159	34	prediction	prediction	NOUN
fcis-12817	159	35	tasks	task	NOUN
fcis-12817	159	36	.	.	PUNCT
fcis-12817	160	1	the	the	DET
fcis-12817	160	2	hyperparameters	hyperparameter	NOUN
fcis-12817	160	3			X
fcis-12817	160	4	determine	determine	VERB
fcis-12817	160	5	the	the	DET
fcis-12817	160	6	contribution	contribution	NOUN
fcis-12817	160	7	of	of	ADP
fcis-12817	160	8	the	the	DET
fcis-12817	160	9	loss	loss	NOUN
fcis-12817	160	10	function	function	NOUN
fcis-12817	160	11	in	in	ADP
fcis-12817	160	12	the	the	DET
fcis-12817	160	13	entity	entity	NOUN
fcis-12817	160	14	matching	matching	NOUN
fcis-12817	160	15	task	task	NOUN
fcis-12817	160	16	to	to	ADP
fcis-12817	160	17	the	the	DET
fcis-12817	160	18	overall	overall	ADJ
fcis-12817	160	19	training	training	NOUN
fcis-12817	160	20	task	task	NOUN
fcis-12817	160	21	.	.	PUNCT
fcis-12817	161	1	we	we	PRON
fcis-12817	161	2	conducted	conduct	VERB
fcis-12817	161	3	several	several	ADJ
fcis-12817	161	4	corresponding	corresponding	ADJ
fcis-12817	161	5	experiments	experiment	NOUN
fcis-12817	161	6	on	on	ADP
fcis-12817	161	7	the	the	DET
fcis-12817	161	8	size	size	NOUN
fcis-12817	161	9	of	of	ADP
fcis-12817	161	10	the	the	DET
fcis-12817	161	11	hyperparameters	hyperparameter	NOUN
fcis-12817	161	12	with	with	ADP
fcis-12817	161	13	an	an	DET
fcis-12817	161	14	interval	interval	NOUN
fcis-12817	161	15	of	of	ADP
fcis-12817	161	16	0.1	0.1	NUM
fcis-12817	161	17	units	unit	NOUN
fcis-12817	161	18	between	between	ADP
fcis-12817	161	19	(	(	PUNCT
fcis-12817	161	20	0,1	0,1	NUM
fcis-12817	161	21	]	]	PUNCT
fcis-12817	161	22	.	.	PUNCT
fcis-12817	162	1	the	the	DET
fcis-12817	162	2	effect	effect	NOUN
fcis-12817	162	3	of	of	ADP
fcis-12817	162	4	the	the	DET
fcis-12817	162	5	loss	loss	NOUN
fcis-12817	162	6	weight	weight	NOUN
fcis-12817	162	7	on	on	ADP
fcis-12817	162	8	the	the	DET
fcis-12817	162	9	final	final	ADJ
fcis-12817	162	10	experimental	experimental	ADJ
fcis-12817	162	11	results	result	NOUN
fcis-12817	162	12	is	be	AUX
fcis-12817	162	13	shown	show	VERB
fcis-12817	162	14	in	in	ADP
fcis-12817	162	15	figures	figure	NOUN
fcis-12817	162	16	2	2	NUM
fcis-12817	162	17	.	.	NOUN
fcis-12817	162	18	80	80	NUM
fcis-12817	162	19	according	accord	VERB
fcis-12817	162	20	to	to	ADP
fcis-12817	162	21	the	the	DET
fcis-12817	162	22	results	result	NOUN
fcis-12817	162	23	shown	show	VERB
fcis-12817	162	24	in	in	ADP
fcis-12817	162	25	the	the	DET
fcis-12817	162	26	figure	figure	NOUN
fcis-12817	162	27	,	,	PUNCT
fcis-12817	162	28	for	for	ADP
fcis-12817	162	29	the	the	DET
fcis-12817	162	30	jd	jd	PROPN
fcis-12817	162	31	-	-	PUNCT
fcis-12817	162	32	ecommerce	ecommerce	NOUN
fcis-12817	162	33	dataset	dataset	NOUN
fcis-12817	162	34	,	,	PUNCT
fcis-12817	162	35	the	the	DET
fcis-12817	162	36	recognition	recognition	NOUN
fcis-12817	162	37	rate	rate	NOUN
fcis-12817	162	38	of	of	ADP
fcis-12817	162	39	entities	entity	NOUN
fcis-12817	162	40	is	be	AUX
fcis-12817	162	41	the	the	DET
fcis-12817	162	42	highest	high	ADJ
fcis-12817	162	43	when	when	SCONJ
fcis-12817	162	44	the	the	DET
fcis-12817	162	45	weight	weight	NOUN
fcis-12817	162	46	of	of	ADP
fcis-12817	162	47	the	the	DET
fcis-12817	162	48	loss	loss	NOUN
fcis-12817	162	49	function	function	NOUN
fcis-12817	162	50	is	be	AUX
fcis-12817	162	51	set	set	VERB
fcis-12817	162	52	to	to	ADP
fcis-12817	162	53	0.8	0.8	NUM
fcis-12817	162	54	.	.	PUNCT
fcis-12817	163	1	for	for	ADP
fcis-12817	163	2	the	the	DET
fcis-12817	163	3	e	e	PROPN
fcis-12817	163	4	-	-	PROPN
fcis-12817	163	5	commerce	commerce	NOUN
fcis-12817	163	6	dataset	dataset	NOUN
fcis-12817	163	7	,	,	PUNCT
fcis-12817	163	8	the	the	DET
fcis-12817	163	9	recognition	recognition	NOUN
fcis-12817	163	10	rate	rate	NOUN
fcis-12817	163	11	of	of	ADP
fcis-12817	163	12	entities	entity	NOUN
fcis-12817	163	13	is	be	AUX
fcis-12817	163	14	highest	high	ADJ
fcis-12817	163	15	when	when	SCONJ
fcis-12817	163	16	the	the	DET
fcis-12817	163	17	loss	loss	NOUN
fcis-12817	163	18	function	function	VERB
fcis-12817	163	19	weights	weight	NOUN
fcis-12817	163	20	are	be	AUX
fcis-12817	163	21	set	set	VERB
fcis-12817	163	22	to	to	ADP
fcis-12817	163	23	0.7	0.7	NUM
fcis-12817	163	24	.	.	PUNCT
fcis-12817	164	1	improper	improper	ADJ
fcis-12817	164	2	configuration	configuration	NOUN
fcis-12817	164	3	of	of	ADP
fcis-12817	164	4	the	the	DET
fcis-12817	164	5	hyperparameter	hyperparameter	NOUN
fcis-12817	164	6	often	often	ADV
fcis-12817	164	7	results	result	VERB
fcis-12817	164	8	in	in	ADP
fcis-12817	164	9	the	the	DET
fcis-12817	164	10	degradation	degradation	NOUN
fcis-12817	164	11	of	of	ADP
fcis-12817	164	12	entity	entity	NOUN
fcis-12817	164	13	recognition	recognition	NOUN
fcis-12817	164	14	performance	performance	NOUN
fcis-12817	164	15	.	.	PUNCT
fcis-12817	165	1	therefore	therefore	ADV
fcis-12817	165	2	,	,	PUNCT
fcis-12817	165	3	when	when	SCONJ
fcis-12817	165	4	performing	perform	VERB
fcis-12817	165	5	chinese	chinese	ADJ
fcis-12817	165	6	e	e	PROPN
fcis-12817	165	7	-	-	PROPN
fcis-12817	165	8	commerce	commerce	NOUN
fcis-12817	165	9	ner	ner	NOUN
fcis-12817	165	10	,	,	PUNCT
fcis-12817	165	11	we	we	PRON
fcis-12817	165	12	need	need	VERB
fcis-12817	165	13	to	to	PART
fcis-12817	165	14	choose	choose	VERB
fcis-12817	165	15	the	the	DET
fcis-12817	165	16	corresponding	corresponding	ADJ
fcis-12817	165	17	loss	loss	NOUN
fcis-12817	165	18	weights	weight	NOUN
fcis-12817	165	19	.	.	PUNCT
fcis-12817	166	1	figure	figure	NOUN
fcis-12817	166	2	2	2	NUM
fcis-12817	166	3	.	.	PUNCT
fcis-12817	166	4	effect	effect	NOUN
fcis-12817	166	5	of	of	ADP
fcis-12817	166	6	loss	loss	NOUN
fcis-12817	166	7	weights	weights	NOUN
fcis-12817	166	8	on	on	ADP
fcis-12817	166	9	experimental	experimental	ADJ
fcis-12817	166	10	results	result	NOUN
fcis-12817	166	11	6	6	NUM
fcis-12817	166	12	.	.	PUNCT
fcis-12817	167	1	conclusion	conclusion	NOUN
fcis-12817	167	2	in	in	ADP
fcis-12817	167	3	this	this	DET
fcis-12817	167	4	paper	paper	NOUN
fcis-12817	167	5	,	,	PUNCT
fcis-12817	167	6	we	we	PRON
fcis-12817	167	7	propose	propose	VERB
fcis-12817	167	8	a	a	DET
fcis-12817	167	9	model	model	NOUN
fcis-12817	167	10	using	use	VERB
fcis-12817	167	11	an	an	DET
fcis-12817	167	12	mrc	mrc	NOUN
fcis-12817	167	13	modeling	modeling	NOUN
fcis-12817	167	14	paradigm	paradigm	NOUN
fcis-12817	167	15	to	to	PART
fcis-12817	167	16	solve	solve	VERB
fcis-12817	167	17	the	the	DET
fcis-12817	167	18	problem	problem	NOUN
fcis-12817	167	19	of	of	ADP
fcis-12817	167	20	named	name	VERB
fcis-12817	167	21	entity	entity	NOUN
fcis-12817	167	22	recognition	recognition	NOUN
fcis-12817	167	23	in	in	ADP
fcis-12817	167	24	the	the	DET
fcis-12817	167	25	e	e	PROPN
fcis-12817	167	26	-	-	NOUN
fcis-12817	167	27	commerce	commerce	NOUN
fcis-12817	167	28	domain	domain	NOUN
fcis-12817	167	29	.	.	PUNCT
fcis-12817	168	1	since	since	SCONJ
fcis-12817	168	2	the	the	DET
fcis-12817	168	3	questions	question	NOUN
fcis-12817	168	4	in	in	ADP
fcis-12817	168	5	mrc	mrc	NOUN
fcis-12817	168	6	contain	contain	VERB
fcis-12817	168	7	a	a	DET
fcis-12817	168	8	priori	priori	ADJ
fcis-12817	168	9	semantic	semantic	ADJ
fcis-12817	168	10	information	information	NOUN
fcis-12817	168	11	about	about	ADP
fcis-12817	168	12	the	the	DET
fcis-12817	168	13	entity	entity	NOUN
fcis-12817	168	14	type	type	NOUN
fcis-12817	168	15	,	,	PUNCT
fcis-12817	168	16	the	the	DET
fcis-12817	168	17	model	model	NOUN
fcis-12817	168	18	performs	perform	VERB
fcis-12817	168	19	better	well	ADV
fcis-12817	168	20	compared	compare	VERB
fcis-12817	168	21	to	to	ADP
fcis-12817	168	22	models	model	NOUN
fcis-12817	168	23	based	base	VERB
fcis-12817	168	24	on	on	ADP
fcis-12817	168	25	sequence	sequence	NOUN
fcis-12817	168	26	labeling	labeling	NOUN
fcis-12817	168	27	.	.	PUNCT
fcis-12817	169	1	in	in	ADP
fcis-12817	169	2	order	order	NOUN
fcis-12817	169	3	to	to	PART
fcis-12817	169	4	prove	prove	VERB
fcis-12817	169	5	the	the	DET
fcis-12817	169	6	effectiveness	effectiveness	NOUN
fcis-12817	169	7	of	of	ADP
fcis-12817	169	8	the	the	DET
fcis-12817	169	9	model	model	NOUN
fcis-12817	169	10	,	,	PUNCT
fcis-12817	169	11	we	we	PRON
fcis-12817	169	12	collected	collect	VERB
fcis-12817	169	13	product	product	NOUN
fcis-12817	169	14	description	description	NOUN
fcis-12817	169	15	data	datum	NOUN
fcis-12817	169	16	from	from	ADP
fcis-12817	169	17	shopping	shop	VERB
fcis-12817	169	18	websites	website	NOUN
fcis-12817	169	19	through	through	ADP
fcis-12817	169	20	crawlers	crawler	NOUN
fcis-12817	169	21	and	and	CCONJ
fcis-12817	169	22	transformed	transform	VERB
fcis-12817	169	23	it	it	PRON
fcis-12817	169	24	into	into	ADP
fcis-12817	169	25	a	a	DET
fcis-12817	169	26	dataset	dataset	NOUN
fcis-12817	169	27	suitable	suitable	ADJ
fcis-12817	169	28	for	for	ADP
fcis-12817	169	29	mrc	mrc	NOUN
fcis-12817	169	30	using	use	VERB
fcis-12817	169	31	scripts	script	NOUN
fcis-12817	169	32	.	.	PUNCT
fcis-12817	170	1	we	we	PRON
fcis-12817	170	2	conducted	conduct	VERB
fcis-12817	170	3	relevant	relevant	ADJ
fcis-12817	170	4	experiments	experiment	NOUN
fcis-12817	170	5	on	on	ADP
fcis-12817	170	6	two	two	NUM
fcis-12817	170	7	e	e	NOUN
fcis-12817	170	8	-	-	NOUN
fcis-12817	170	9	commerce	commerce	NOUN
fcis-12817	170	10	entity	entity	NOUN
fcis-12817	170	11	recognition	recognition	NOUN
fcis-12817	170	12	datasets	dataset	NOUN
fcis-12817	170	13	to	to	PART
fcis-12817	170	14	demonstrate	demonstrate	VERB
fcis-12817	170	15	the	the	DET
fcis-12817	170	16	effectiveness	effectiveness	NOUN
fcis-12817	170	17	of	of	ADP
fcis-12817	170	18	the	the	DET
fcis-12817	170	19	method	method	NOUN
fcis-12817	170	20	.	.	PUNCT
fcis-12817	171	1	acknowledgments	acknowledgment	NOUN
fcis-12817	171	2	this	this	DET
fcis-12817	171	3	work	work	NOUN
fcis-12817	171	4	is	be	AUX
fcis-12817	171	5	supported	support	VERB
fcis-12817	171	6	in	in	ADP
fcis-12817	171	7	part	part	NOUN
fcis-12817	171	8	zhejiang	zhejiang	PROPN
fcis-12817	171	9	public	public	ADJ
fcis-12817	171	10	welfare	welfare	NOUN
fcis-12817	171	11	technology	technology	NOUN
fcis-12817	171	12	application	application	NOUN
fcis-12817	171	13	research	research	NOUN
fcis-12817	171	14	project	project	NOUN
fcis-12817	171	15	of	of	ADP
fcis-12817	171	16	china	china	PROPN
fcis-12817	171	17	(	(	PUNCT
fcis-12817	171	18	grant	grant	PROPN
fcis-12817	171	19	:	:	PUNCT
fcis-12817	171	20	lgn21f020003	lgn21f020003	NOUN
fcis-12817	171	21	)	)	PUNCT
fcis-12817	171	22	.	.	PUNCT
fcis-12817	172	1	references	reference	NOUN
fcis-12817	172	2	[	[	X
fcis-12817	172	3	1	1	X
fcis-12817	172	4	]	]	PUNCT
fcis-12817	172	5	diefenbach	diefenbach	PROPN
fcis-12817	172	6	d	d	PROPN
fcis-12817	172	7	,	,	PUNCT
fcis-12817	172	8	lopez	lopez	PROPN
fcis-12817	172	9	v	v	PROPN
fcis-12817	172	10	,	,	PUNCT
fcis-12817	172	11	singh	singh	PROPN
fcis-12817	172	12	k	k	PROPN
fcis-12817	172	13	,	,	PUNCT
fcis-12817	172	14	et	et	PROPN
fcis-12817	172	15	al	al	PROPN
fcis-12817	172	16	.	.	PUNCT
fcis-12817	173	1	core	core	NOUN
fcis-12817	173	2	techniques	technique	NOUN
fcis-12817	173	3	of	of	ADP
fcis-12817	173	4	question	question	NOUN
fcis-12817	173	5	answering	answering	NOUN
fcis-12817	173	6	systems	system	NOUN
fcis-12817	173	7	over	over	ADP
fcis-12817	173	8	knowledge	knowledge	NOUN
fcis-12817	173	9	bases	basis	NOUN
fcis-12817	173	10	:	:	PUNCT
fcis-12817	173	11	a	a	DET
fcis-12817	173	12	survey	survey	NOUN
fcis-12817	173	13	[	[	X
fcis-12817	173	14	j	j	X
fcis-12817	173	15	]	]	X
fcis-12817	173	16	.	.	PUNCT
fcis-12817	174	1	knowledge	knowledge	NOUN
fcis-12817	174	2	and	and	CCONJ
fcis-12817	174	3	information	information	NOUN
fcis-12817	174	4	systems	system	NOUN
fcis-12817	174	5	,	,	PUNCT
fcis-12817	174	6	2018	2018	NUM
fcis-12817	174	7	,	,	PUNCT
fcis-12817	174	8	55(3	55(3	NUM
fcis-12817	174	9	):	):	PUNCT
fcis-12817	174	10	529	529	NUM
fcis-12817	174	11	-	-	SYM
fcis-12817	174	12	569	569	NUM
fcis-12817	174	13	.	.	PUNCT
fcis-12817	175	1	[	[	X
fcis-12817	175	2	2	2	X
fcis-12817	175	3	]	]	X
fcis-12817	175	4	dandapat	dandapat	PROPN
fcis-12817	175	5	s	s	PROPN
fcis-12817	175	6	,	,	PUNCT
fcis-12817	175	7	way	way	NOUN
fcis-12817	175	8	a.	a.	NOUN
fcis-12817	175	9	improved	improved	AUX
fcis-12817	175	10	named	name	VERB
fcis-12817	175	11	entity	entity	NOUN
fcis-12817	175	12	recognition	recognition	NOUN
fcis-12817	175	13	using	use	VERB
fcis-12817	175	14	machine	machine	NOUN
fcis-12817	175	15	translation	translation	NOUN
fcis-12817	175	16	-	-	PUNCT
fcis-12817	175	17	based	base	VERB
fcis-12817	175	18	cross	cross	NOUN
fcis-12817	175	19	-	-	ADJ
fcis-12817	175	20	lingual	lingual	ADJ
fcis-12817	175	21	information[j	information[j	NOUN
fcis-12817	175	22	]	]	PUNCT
fcis-12817	175	23	.	.	PUNCT
fcis-12817	176	1	computacióny	computacióny	PROPN
fcis-12817	176	2	sistemas	sistemas	PROPN
fcis-12817	176	3	,	,	PUNCT
fcis-12817	176	4	2016	2016	NUM
fcis-12817	176	5	,	,	PUNCT
fcis-12817	176	6	20(3	20(3	NUM
fcis-12817	176	7	):	):	PUNCT
fcis-12817	176	8	495	495	NUM
fcis-12817	176	9	-	-	SYM
fcis-12817	176	10	504	504	NUM
fcis-12817	176	11	.	.	PUNCT
fcis-12817	177	1	[	[	X
fcis-12817	177	2	3	3	NUM
fcis-12817	177	3	]	]	X
fcis-12817	177	4	cheng	cheng	PROPN
fcis-12817	177	5	q	q	PROPN
fcis-12817	177	6	,	,	PUNCT
fcis-12817	177	7	liu	liu	PROPN
fcis-12817	177	8	j	j	PROPN
fcis-12817	177	9	,	,	PUNCT
fcis-12817	177	10	qu	qu	PROPN
fcis-12817	177	11	x	x	PROPN
fcis-12817	177	12	,	,	PUNCT
fcis-12817	177	13	et	et	PROPN
fcis-12817	177	14	al	al	PROPN
fcis-12817	177	15	.	.	PROPN
fcis-12817	177	16	hacred	hacre	VERB
fcis-12817	177	17	:	:	PUNCT
fcis-12817	177	18	a	a	DET
fcis-12817	177	19	large	large	ADJ
fcis-12817	177	20	-	-	PUNCT
fcis-12817	177	21	scale	scale	NOUN
fcis-12817	177	22	relation	relation	NOUN
fcis-12817	177	23	extraction	extraction	NOUN
fcis-12817	177	24	dataset	dataset	VERB
fcis-12817	177	25	toward	toward	ADP
fcis-12817	177	26	hard	hard	ADJ
fcis-12817	177	27	cases	case	NOUN
fcis-12817	177	28	in	in	ADP
fcis-12817	177	29	practical	practical	ADJ
fcis-12817	177	30	applications	application	NOUN
fcis-12817	177	31	[	[	X
fcis-12817	177	32	c	c	X
fcis-12817	177	33	]	]	PUNCT
fcis-12817	177	34	.	.	PUNCT
fcis-12817	178	1	findings	finding	NOUN
fcis-12817	178	2	of	of	ADP
fcis-12817	178	3	the	the	DET
fcis-12817	178	4	association	association	NOUN
fcis-12817	178	5	for	for	ADP
fcis-12817	178	6	computational	computational	ADJ
fcis-12817	178	7	linguistics	linguistic	NOUN
fcis-12817	178	8	:	:	PUNCT
fcis-12817	178	9	acl	acl	PROPN
fcis-12817	178	10	-	-	PUNCT
fcis-12817	178	11	ijcnlp	ijcnlp	NOUN
fcis-12817	178	12	2021	2021	NUM
fcis-12817	178	13	,	,	PUNCT
fcis-12817	178	14	2021	2021	NUM
fcis-12817	178	15	:	:	PUNCT
fcis-12817	178	16	2819	2819	NUM
fcis-12817	178	17	-	-	SYM
fcis-12817	178	18	2831	2831	NUM
fcis-12817	178	19	.	.	PUNCT
fcis-12817	179	1	[	[	X
fcis-12817	179	2	4	4	X
fcis-12817	179	3	]	]	PUNCT
fcis-12817	179	4	martins	martin	VERB
fcis-12817	179	5	p	p	NOUN
fcis-12817	179	6	h	h	NOUN
fcis-12817	179	7	,	,	PUNCT
fcis-12817	179	8	marinho	marinho	PROPN
fcis-12817	179	9	z	z	PROPN
fcis-12817	179	10	,	,	PUNCT
fcis-12817	179	11	martins	martin	VERB
fcis-12817	179	12	a	a	DET
fcis-12817	179	13	f	f	PROPN
fcis-12817	179	14	t.	t.	NOUN
fcis-12817	179	15	joint	joint	ADJ
fcis-12817	179	16	learning	learning	NOUN
fcis-12817	179	17	of	of	ADP
fcis-12817	179	18	named	name	VERB
fcis-12817	179	19	entity	entity	NOUN
fcis-12817	179	20	recognition	recognition	NOUN
fcis-12817	179	21	and	and	CCONJ
fcis-12817	179	22	entity	entity	NOUN
fcis-12817	179	23	linking[c	linking[c	NOUN
fcis-12817	179	24	]	]	PUNCT
fcis-12817	179	25	.	.	PUNCT
fcis-12817	180	1	proceedings	proceeding	NOUN
fcis-12817	180	2	of	of	ADP
fcis-12817	180	3	the	the	DET
fcis-12817	180	4	57th	57th	ADJ
fcis-12817	180	5	annual	annual	ADJ
fcis-12817	180	6	meeting	meeting	NOUN
fcis-12817	180	7	of	of	ADP
fcis-12817	180	8	the	the	DET
fcis-12817	180	9	association	association	NOUN
fcis-12817	180	10	for	for	ADP
fcis-12817	180	11	computational	computational	ADJ
fcis-12817	180	12	linguistics	linguistic	NOUN
fcis-12817	180	13	.	.	PUNCT
fcis-12817	181	1	student	student	NOUN
fcis-12817	181	2	research	research	NOUN
fcis-12817	181	3	workshop	workshop	NOUN
fcis-12817	181	4	,	,	PUNCT
fcis-12817	181	5	2019	2019	NUM
fcis-12817	181	6	:	:	PUNCT
fcis-12817	181	7	190	190	NUM
fcis-12817	181	8	-	-	SYM
fcis-12817	181	9	196	196	NUM
fcis-12817	181	10	.	.	PUNCT
fcis-12817	182	1	[	[	X
fcis-12817	182	2	5	5	NUM
fcis-12817	182	3	]	]	PUNCT
fcis-12817	182	4	levy	levy	NOUN
fcis-12817	182	5	o	o	NOUN
fcis-12817	182	6	,	,	PUNCT
fcis-12817	182	7	seo	seo	NOUN
fcis-12817	182	8	m	m	PROPN
fcis-12817	182	9	,	,	PUNCT
fcis-12817	182	10	choi	choi	PROPN
fcis-12817	182	11	e	e	PROPN
fcis-12817	182	12	,	,	PUNCT
fcis-12817	182	13	et	et	PROPN
fcis-12817	182	14	al	al	PROPN
fcis-12817	182	15	.	.	PROPN
fcis-12817	182	16	zero	zero	NUM
fcis-12817	182	17	-	-	PUNCT
fcis-12817	182	18	shot	shot	NOUN
fcis-12817	182	19	relation	relation	NOUN
fcis-12817	182	20	extraction	extraction	NOUN
fcis-12817	182	21	via	via	ADP
fcis-12817	182	22	reading	read	VERB
fcis-12817	182	23	comprehension[c	comprehension[c	PROPN
fcis-12817	182	24	]	]	PUNCT
fcis-12817	182	25	.	.	PUNCT
fcis-12817	183	1	proceedings	proceeding	NOUN
fcis-12817	183	2	of	of	ADP
fcis-12817	183	3	the	the	DET
fcis-12817	183	4	21st	21st	ADJ
fcis-12817	183	5	conference	conference	NOUN
fcis-12817	183	6	on	on	ADP
fcis-12817	183	7	computational	computational	ADJ
fcis-12817	183	8	natural	natural	ADJ
fcis-12817	183	9	language	language	NOUN
fcis-12817	183	10	learning	learning	NOUN
fcis-12817	183	11	.	.	PUNCT
fcis-12817	184	1	2017	2017	NUM
fcis-12817	184	2	:	:	PUNCT
fcis-12817	184	3	333	333	NUM
fcis-12817	184	4	-	-	SYM
fcis-12817	184	5	342	342	NUM
fcis-12817	184	6	.	.	PUNCT
fcis-12817	185	1	[	[	X
fcis-12817	185	2	6	6	NUM
fcis-12817	185	3	]	]	PUNCT
fcis-12817	185	4	mccann	mccann	PROPN
fcis-12817	185	5	b	b	PROPN
fcis-12817	185	6	,	,	PUNCT
fcis-12817	185	7	keskar	keskar	NOUN
fcis-12817	185	8	n	n	PROPN
fcis-12817	185	9	s	s	PROPN
fcis-12817	185	10	,	,	PUNCT
fcis-12817	185	11	xiong	xiong	PROPN
fcis-12817	185	12	c	c	AUX
fcis-12817	185	13	,	,	PUNCT
fcis-12817	185	14	et	et	PROPN
fcis-12817	185	15	al	al	PROPN
fcis-12817	185	16	.	.	PUNCT
fcis-12817	186	1	the	the	DET
fcis-12817	186	2	natural	natural	ADJ
fcis-12817	186	3	language	language	NOUN
fcis-12817	186	4	decathlon	decathlon	NOUN
fcis-12817	186	5	:	:	PUNCT
fcis-12817	186	6	multitask	multitask	PROPN
fcis-12817	186	7	learning	learning	NOUN
fcis-12817	186	8	as	as	ADP
fcis-12817	186	9	question	question	NOUN
fcis-12817	186	10	answering[j	answering[j	NOUN
fcis-12817	186	11	]	]	PUNCT
fcis-12817	186	12	.	.	PUNCT
fcis-12817	187	1	arxiv	arxiv	PROPN
fcis-12817	187	2	preprint	preprint	VERB
fcis-12817	187	3	arxiv:1806.08730	arxiv:1806.08730	ADV
fcis-12817	187	4	,	,	PUNCT
fcis-12817	187	5	2018	2018	NUM
fcis-12817	187	6	.	.	PUNCT
fcis-12817	188	1	[	[	X
fcis-12817	188	2	online	online	X
fcis-12817	188	3	]	]	X
fcis-12817	188	4	.	.	PUNCT
fcis-12817	189	1	available	available	ADJ
fcis-12817	189	2	:	:	PUNCT
fcis-12817	189	3	https://	https://	PROPN
fcis-12817	189	4	arxiv	arxiv	PROPN
fcis-12817	189	5	.	.	PUNCT
fcis-12817	190	1	org	org	PROPN
fcis-12817	190	2	/	/	SYM
fcis-12817	190	3	pdf/1806.08730.pdf	pdf/1806.08730.pdf	PROPN
fcis-12817	191	1	[	[	X
fcis-12817	191	2	7	7	NUM
fcis-12817	191	3	]	]	X
fcis-12817	191	4	li	li	PROPN
fcis-12817	191	5	x	x	PROPN
fcis-12817	191	6	,	,	PUNCT
fcis-12817	191	7	yin	yin	PROPN
fcis-12817	191	8	f	f	PROPN
fcis-12817	191	9	,	,	PUNCT
fcis-12817	191	10	sun	sun	PROPN
fcis-12817	191	11	z	z	PROPN
fcis-12817	191	12	,	,	PUNCT
fcis-12817	191	13	et	et	PROPN
fcis-12817	191	14	al	al	PROPN
fcis-12817	191	15	.	.	PUNCT
fcis-12817	191	16	entity	entity	NOUN
fcis-12817	191	17	-	-	PUNCT
fcis-12817	191	18	relation	relation	NOUN
fcis-12817	191	19	extraction	extraction	NOUN
fcis-12817	191	20	as	as	ADP
fcis-12817	191	21	multiturn	multiturn	NOUN
fcis-12817	191	22	question	question	NOUN
fcis-12817	191	23	answering[c	answering[c	NOUN
fcis-12817	191	24	]	]	PUNCT
fcis-12817	191	25	.	.	PUNCT
fcis-12817	192	1	proceedings	proceeding	NOUN
fcis-12817	192	2	of	of	ADP
fcis-12817	192	3	the	the	DET
fcis-12817	192	4	57th	57th	ADJ
fcis-12817	192	5	annual	annual	ADJ
fcis-12817	192	6	meeting	meeting	NOUN
fcis-12817	192	7	of	of	ADP
fcis-12817	192	8	the	the	DET
fcis-12817	192	9	association	association	NOUN
fcis-12817	192	10	for	for	ADP
fcis-12817	192	11	computational	computational	ADJ
fcis-12817	192	12	linguistics	linguistic	NOUN
fcis-12817	192	13	.	.	PUNCT
fcis-12817	193	1	2019	2019	NUM
fcis-12817	193	2	:	:	PUNCT
fcis-12817	193	3	1340	1340	NUM
fcis-12817	193	4	-	-	SYM
fcis-12817	193	5	1350	1350	NUM
fcis-12817	193	6	.	.	PUNCT
fcis-12817	194	1	[	[	X
fcis-12817	194	2	8	8	NUM
fcis-12817	194	3	]	]	X
fcis-12817	194	4	fei	fei	PROPN
fcis-12817	194	5	y	y	PROPN
fcis-12817	194	6	,	,	PUNCT
fcis-12817	194	7	xu	xu	PROPN
fcis-12817	194	8	x.	x.	PROPN
fcis-12817	194	9	gfmrc	gfmrc	PROPN
fcis-12817	194	10	:	:	PUNCT
fcis-12817	194	11	a	a	DET
fcis-12817	194	12	machine	machine	NOUN
fcis-12817	194	13	reading	reading	NOUN
fcis-12817	194	14	comprehension	comprehension	NOUN
fcis-12817	194	15	model	model	NOUN
fcis-12817	194	16	for	for	ADP
fcis-12817	194	17	named	name	VERB
fcis-12817	194	18	entity	entity	NOUN
fcis-12817	194	19	recognition[j	recognition[j	NOUN
fcis-12817	194	20	]	]	PUNCT
fcis-12817	194	21	.	.	PUNCT
fcis-12817	195	1	pattern	pattern	NOUN
fcis-12817	195	2	recognition	recognition	NOUN
fcis-12817	195	3	letters	letter	NOUN
fcis-12817	195	4	,	,	PUNCT
fcis-12817	195	5	2023	2023	NUM
fcis-12817	195	6	.	.	PUNCT
fcis-12817	196	1	[	[	X
fcis-12817	196	2	9	9	NUM
fcis-12817	196	3	]	]	SYM
fcis-12817	196	4	li	li	PROPN
fcis-12817	196	5	x	x	PROPN
fcis-12817	196	6	,	,	PUNCT
fcis-12817	196	7	feng	feng	PROPN
fcis-12817	196	8	j	j	PROPN
fcis-12817	196	9	,	,	PUNCT
fcis-12817	196	10	meng	meng	PROPN
fcis-12817	196	11	y	y	PROPN
fcis-12817	196	12	,	,	PUNCT
fcis-12817	196	13	et	et	PROPN
fcis-12817	196	14	al	al	PROPN
fcis-12817	196	15	.	.	PUNCT
fcis-12817	197	1	a	a	DET
fcis-12817	197	2	unified	unified	ADJ
fcis-12817	197	3	mrc	mrc	NOUN
fcis-12817	197	4	framework	framework	NOUN
fcis-12817	197	5	for	for	ADP
fcis-12817	197	6	named	name	VERB
fcis-12817	197	7	entity	entity	NOUN
fcis-12817	197	8	recognition[c	recognition[c	NOUN
fcis-12817	197	9	]	]	PUNCT
fcis-12817	197	10	.	.	PUNCT
fcis-12817	198	1	proceedings	proceeding	NOUN
fcis-12817	198	2	of	of	ADP
fcis-12817	198	3	the	the	DET
fcis-12817	198	4	58th	58th	ADJ
fcis-12817	198	5	annual	annual	ADJ
fcis-12817	198	6	meeting	meeting	NOUN
fcis-12817	198	7	of	of	ADP
fcis-12817	198	8	the	the	DET
fcis-12817	198	9	association	association	NOUN
fcis-12817	198	10	for	for	ADP
fcis-12817	198	11	computational	computational	ADJ
fcis-12817	198	12	linguistics	linguistic	NOUN
fcis-12817	198	13	.	.	PUNCT
fcis-12817	199	1	2020	2020	NUM
fcis-12817	199	2	:	:	PUNCT
fcis-12817	199	3	5849	5849	NUM
fcis-12817	199	4	-	-	SYM
fcis-12817	199	5	5859	5859	NUM
fcis-12817	199	6	.	.	PUNCT
fcis-12817	200	1	[	[	X
fcis-12817	200	2	10	10	NUM
fcis-12817	200	3	]	]	X
fcis-12817	200	4	chen	chen	PROPN
fcis-12817	200	5	p	p	PROPN
fcis-12817	200	6	,	,	PUNCT
fcis-12817	200	7	wang	wang	PROPN
fcis-12817	200	8	j	j	PROPN
fcis-12817	200	9	,	,	PUNCT
fcis-12817	200	10	lin	lin	PROPN
fcis-12817	200	11	h	h	PROPN
fcis-12817	200	12	,	,	PUNCT
fcis-12817	200	13	et	et	PROPN
fcis-12817	200	14	al	al	PROPN
fcis-12817	200	15	.	.	PROPN
fcis-12817	200	16	knowledge	knowledge	PROPN
fcis-12817	200	17	adaptive	adaptive	ADJ
fcis-12817	200	18	multi	multi	ADJ
fcis-12817	200	19	-	-	ADJ
fcis-12817	200	20	way	way	ADJ
fcis-12817	200	21	matching	matching	NOUN
fcis-12817	200	22	network	network	NOUN
fcis-12817	200	23	for	for	ADP
fcis-12817	200	24	biomedical	biomedical	ADJ
fcis-12817	200	25	named	name	VERB
fcis-12817	200	26	entity	entity	NOUN
fcis-12817	200	27	recognition	recognition	NOUN
fcis-12817	200	28	via	via	ADP
fcis-12817	200	29	machine	machine	NOUN
fcis-12817	200	30	reading	read	VERB
fcis-12817	200	31	comprehension[j	comprehension[j	PROPN
fcis-12817	200	32	]	]	PUNCT
fcis-12817	200	33	.	.	PUNCT
fcis-12817	201	1	ieee	ieee	NOUN
fcis-12817	201	2	transactions	transaction	NOUN
fcis-12817	201	3	on	on	ADP
fcis-12817	201	4	computational	computational	ADJ
fcis-12817	201	5	biology	biology	NOUN
fcis-12817	201	6	and	and	CCONJ
fcis-12817	201	7	bioinformatics	bioinformatics	NOUN
fcis-12817	201	8	,	,	PUNCT
fcis-12817	201	9	2023	2023	NUM
fcis-12817	201	10	.	.	PUNCT
fcis-12817	202	1	[	[	X
fcis-12817	202	2	11	11	NUM
fcis-12817	202	3	]	]	X
fcis-12817	202	4	zhang	zhang	PROPN
fcis-12817	202	5	y	y	PROPN
fcis-12817	202	6	,	,	PUNCT
fcis-12817	202	7	zhang	zhang	PROPN
fcis-12817	202	8	h.	h.	PROPN
fcis-12817	202	9	finbert	finbert	PROPN
fcis-12817	202	10	-	-	PUNCT
fcis-12817	202	11	mrc	mrc	NOUN
fcis-12817	202	12	:	:	PUNCT
fcis-12817	202	13	financial	financial	ADJ
fcis-12817	202	14	named	name	VERB
fcis-12817	202	15	entity	entity	NOUN
fcis-12817	202	16	recognition	recognition	NOUN
fcis-12817	202	17	using	use	VERB
fcis-12817	202	18	bert	bert	NOUN
fcis-12817	202	19	under	under	ADP
fcis-12817	202	20	the	the	DET
fcis-12817	202	21	machine	machine	NOUN
fcis-12817	202	22	reading	reading	NOUN
fcis-12817	202	23	comprehension	comprehension	NOUN
fcis-12817	202	24	paradigm[j	paradigm[j	NOUN
fcis-12817	202	25	]	]	PUNCT
fcis-12817	202	26	.	.	PUNCT
fcis-12817	203	1	neural	neural	ADJ
fcis-12817	203	2	processing	processing	NOUN
fcis-12817	203	3	letters	letter	NOUN
fcis-12817	203	4	,	,	PUNCT
fcis-12817	203	5	2023	2023	NUM
fcis-12817	203	6	:	:	PUNCT
fcis-12817	203	7	1	1	NUM
fcis-12817	203	8	-	-	SYM
fcis-12817	203	9	21	21	NUM
fcis-12817	203	10	.	.	PUNCT
fcis-12817	204	1	[	[	X
fcis-12817	204	2	12	12	NUM
fcis-12817	204	3	]	]	X
fcis-12817	204	4	ding	ding	NOUN
fcis-12817	204	5	r	r	PROPN
fcis-12817	204	6	,	,	PUNCT
fcis-12817	204	7	xie	xie	PROPN
fcis-12817	205	1	p	p	PROPN
fcis-12817	205	2	,	,	PUNCT
fcis-12817	205	3	zhang	zhang	PROPN
fcis-12817	205	4	x	x	PROPN
fcis-12817	205	5	,	,	PUNCT
fcis-12817	205	6	et	et	PROPN
fcis-12817	205	7	al	al	PROPN
fcis-12817	205	8	.	.	PUNCT
fcis-12817	206	1	a	a	DET
fcis-12817	206	2	neural	neural	ADJ
fcis-12817	206	3	multi	multi	ADJ
fcis-12817	206	4	-	-	ADJ
fcis-12817	206	5	digraph	digraph	ADJ
fcis-12817	206	6	model	model	NOUN
fcis-12817	206	7	for	for	ADP
fcis-12817	206	8	chinese	chinese	ADJ
fcis-12817	206	9	ner	ner	NOUN
fcis-12817	206	10	with	with	ADP
fcis-12817	206	11	gazetteers[c	gazetteers[c	PROPN
fcis-12817	206	12	]	]	PUNCT
fcis-12817	206	13	.	.	PUNCT
fcis-12817	207	1	proceedings	proceeding	NOUN
fcis-12817	207	2	of	of	ADP
fcis-12817	207	3	the	the	DET
fcis-12817	207	4	57th	57th	ADJ
fcis-12817	207	5	annual	annual	ADJ
fcis-12817	207	6	meeting	meeting	NOUN
fcis-12817	207	7	of	of	ADP
fcis-12817	207	8	the	the	DET
fcis-12817	207	9	association	association	NOUN
fcis-12817	207	10	for	for	ADP
fcis-12817	207	11	computational	computational	ADJ
fcis-12817	207	12	linguistics	linguistic	NOUN
fcis-12817	207	13	.	.	PUNCT
fcis-12817	208	1	2019:1462	2019:1462	NUM
fcis-12817	208	2	-	-	SYM
fcis-12817	208	3	1467	1467	NUM
fcis-12817	208	4	.	.	PUNCT
fcis-12817	209	1	[	[	X
fcis-12817	209	2	13	13	NUM
fcis-12817	209	3	]	]	X
fcis-12817	209	4	rajpurkar	rajpurkar	NOUN
fcis-12817	209	5	p	p	X
fcis-12817	209	6	,	,	PUNCT
fcis-12817	209	7	zhang	zhang	PROPN
fcis-12817	209	8	j	j	PROPN
fcis-12817	209	9	,	,	PUNCT
fcis-12817	209	10	lopyrev	lopyrev	PROPN
fcis-12817	209	11	k	k	PROPN
fcis-12817	209	12	,	,	PUNCT
fcis-12817	209	13	et	et	PROPN
fcis-12817	209	14	al	al	PROPN
fcis-12817	209	15	.	.	PUNCT
fcis-12817	209	16	squad	squad	PROPN
fcis-12817	209	17	:	:	PUNCT
fcis-12817	210	1	100,000	100,000	NUM
fcis-12817	210	2	+	+	NUM
fcis-12817	210	3	questions	question	NOUN
fcis-12817	210	4	for	for	ADP
fcis-12817	210	5	machine	machine	NOUN
fcis-12817	210	6	comprehension	comprehension	NOUN
fcis-12817	210	7	of	of	ADP
fcis-12817	210	8	text[c	text[c	NOUN
fcis-12817	210	9	]	]	PUNCT
fcis-12817	210	10	.	.	PUNCT
fcis-12817	211	1	proceedings	proceeding	NOUN
fcis-12817	211	2	of	of	ADP
fcis-12817	211	3	the	the	DET
fcis-12817	211	4	2016	2016	NUM
fcis-12817	211	5	conference	conference	NOUN
fcis-12817	211	6	on	on	ADP
fcis-12817	211	7	empirical	empirical	ADJ
fcis-12817	211	8	methods	method	NOUN
fcis-12817	211	9	in	in	ADP
fcis-12817	211	10	natural	natural	ADJ
fcis-12817	211	11	language	language	NOUN
fcis-12817	211	12	processing	processing	NOUN
fcis-12817	211	13	.	.	PUNCT
fcis-12817	212	1	2016	2016	NUM
fcis-12817	212	2	:	:	PUNCT
fcis-12817	212	3	2383	2383	NUM
fcis-12817	212	4	-	-	SYM
fcis-12817	212	5	2392	2392	NUM
fcis-12817	212	6	.	.	PUNCT