id	sid	tid	token	lemma	pos
ajst-19203	1	1	academic	academic	ADJ
ajst-19203	1	2	journal	journal	NOUN
ajst-19203	1	3	of	of	ADP
ajst-19203	1	4	science	science	NOUN
ajst-19203	1	5	and	and	CCONJ
ajst-19203	1	6	technology	technology	NOUN
ajst-19203	1	7	issn	issn	NOUN
ajst-19203	1	8	:	:	PUNCT
ajst-19203	1	9	2771	2771	NUM
ajst-19203	1	10	-	-	SYM
ajst-19203	1	11	3032	3032	NUM
ajst-19203	1	12	|	|	NOUN
ajst-19203	1	13	vol	vol	NOUN
ajst-19203	1	14	.	.	PROPN
ajst-19203	2	1	10	10	NUM
ajst-19203	2	2	,	,	PUNCT
ajst-19203	2	3	no	no	INTJ
ajst-19203	2	4	.	.	NOUN
ajst-19203	2	5	1	1	NUM
ajst-19203	2	6	,	,	PUNCT
ajst-19203	2	7	2024	2024	NUM
ajst-19203	2	8	284	284	NUM
ajst-19203	2	9	image	image	NOUN
ajst-19203	2	10	captioning	captioning	NOUN
ajst-19203	2	11	in	in	ADP
ajst-19203	2	12	news	news	NOUN
ajst-19203	2	13	report	report	NOUN
ajst-19203	2	14	scenario	scenario	PROPN
ajst-19203	2	15	tianrui	tianrui	PROPN
ajst-19203	2	16	liu1	liu1	PROPN
ajst-19203	2	17	,	,	PUNCT
ajst-19203	2	18	*	*	PROPN
ajst-19203	2	19	,	,	PUNCT
ajst-19203	2	20	qi	qi	PROPN
ajst-19203	2	21	cai2	cai2	PROPN
ajst-19203	2	22	,	,	PUNCT
ajst-19203	2	23	changxin	changxin	PROPN
ajst-19203	2	24	xu3	xu3	PROPN
ajst-19203	2	25	,	,	PUNCT
ajst-19203	2	26	bo	bo	PROPN
ajst-19203	2	27	hong4	hong4	PROPN
ajst-19203	2	28	,	,	PUNCT
ajst-19203	2	29	jize	jize	PROPN
ajst-19203	2	30	xiong5	xiong5	PROPN
ajst-19203	2	31	,	,	PUNCT
ajst-19203	2	32	yuxin	yuxin	PROPN
ajst-19203	2	33	qiao6	qiao6	PROPN
ajst-19203	2	34	,	,	PUNCT
ajst-19203	2	35	tsungwei	tsungwei	PROPN
ajst-19203	2	36	yang7	yang7	PROPN
ajst-19203	2	37	1	1	NUM
ajst-19203	2	38	department	department	NOUN
ajst-19203	2	39	of	of	ADP
ajst-19203	2	40	electrical	electrical	ADJ
ajst-19203	2	41	and	and	CCONJ
ajst-19203	2	42	computer	computer	NOUN
ajst-19203	2	43	engineering	engineering	NOUN
ajst-19203	2	44	,	,	PUNCT
ajst-19203	2	45	university	university	PROPN
ajst-19203	2	46	of	of	ADP
ajst-19203	2	47	california	california	PROPN
ajst-19203	2	48	san	san	PROPN
ajst-19203	2	49	diego	diego	PROPN
ajst-19203	2	50	,	,	PUNCT
ajst-19203	2	51	usa	usa	PROPN
ajst-19203	2	52	2	2	NUM
ajst-19203	2	53	computer	computer	NOUN
ajst-19203	2	54	science	science	NOUN
ajst-19203	2	55	and	and	CCONJ
ajst-19203	2	56	engineering	engineering	NOUN
ajst-19203	2	57	,	,	PUNCT
ajst-19203	2	58	university	university	NOUN
ajst-19203	2	59	of	of	ADP
ajst-19203	2	60	north	north	PROPN
ajst-19203	2	61	texas	texas	PROPN
ajst-19203	2	62	,	,	PUNCT
ajst-19203	2	63	denton	denton	PROPN
ajst-19203	2	64	,	,	PUNCT
ajst-19203	2	65	usa	usa	PROPN
ajst-19203	2	66	3	3	NUM
ajst-19203	2	67	computer	computer	NOUN
ajst-19203	2	68	information	information	NOUN
ajst-19203	2	69	technology	technology	NOUN
ajst-19203	2	70	,	,	PUNCT
ajst-19203	2	71	northern	northern	PROPN
ajst-19203	2	72	arizona	arizona	PROPN
ajst-19203	2	73	university	university	PROPN
ajst-19203	2	74	,	,	PUNCT
ajst-19203	2	75	flagstaff	flagstaff	PROPN
ajst-19203	2	76	,	,	PUNCT
ajst-19203	2	77	usa	usa	PROPN
ajst-19203	2	78	4	4	NUM
ajst-19203	2	79	computer	computer	NOUN
ajst-19203	2	80	information	information	NOUN
ajst-19203	2	81	technology	technology	NOUN
ajst-19203	2	82	,	,	PUNCT
ajst-19203	2	83	northern	northern	PROPN
ajst-19203	2	84	arizona	arizona	PROPN
ajst-19203	2	85	university	university	PROPN
ajst-19203	2	86	,	,	PUNCT
ajst-19203	2	87	flagstaff	flagstaff	PROPN
ajst-19203	2	88	,	,	PUNCT
ajst-19203	2	89	usa	usa	PROPN
ajst-19203	2	90	5	5	NUM
ajst-19203	2	91	computer	computer	NOUN
ajst-19203	2	92	information	information	NOUN
ajst-19203	2	93	technology	technology	NOUN
ajst-19203	2	94	,	,	PUNCT
ajst-19203	2	95	northern	northern	PROPN
ajst-19203	2	96	arizona	arizona	PROPN
ajst-19203	2	97	university	university	PROPN
ajst-19203	2	98	,	,	PUNCT
ajst-19203	2	99	flagstaff	flagstaff	PROPN
ajst-19203	2	100	,	,	PUNCT
ajst-19203	2	101	usa	usa	PROPN
ajst-19203	2	102	6	6	NUM
ajst-19203	2	103	computer	computer	NOUN
ajst-19203	2	104	information	information	NOUN
ajst-19203	2	105	technology	technology	NOUN
ajst-19203	2	106	,	,	PUNCT
ajst-19203	2	107	northern	northern	PROPN
ajst-19203	2	108	arizona	arizona	PROPN
ajst-19203	2	109	university	university	PROPN
ajst-19203	2	110	,	,	PUNCT
ajst-19203	2	111	flagstaff	flagstaff	PROPN
ajst-19203	2	112	,	,	PUNCT
ajst-19203	2	113	usa	usa	PROPN
ajst-19203	2	114	7	7	NUM
ajst-19203	2	115	computer	computer	NOUN
ajst-19203	2	116	science	science	NOUN
ajst-19203	2	117	,	,	PUNCT
ajst-19203	2	118	tunghai	tunghai	PROPN
ajst-19203	2	119	university	university	PROPN
ajst-19203	2	120	,	,	PUNCT
ajst-19203	2	121	taichung	taichung	PROPN
ajst-19203	2	122	,	,	PUNCT
ajst-19203	2	123	taiwan	taiwan	PROPN
ajst-19203	2	124	*	*	PUNCT
ajst-19203	2	125	corresponding	correspond	VERB
ajst-19203	2	126	author	author	NOUN
ajst-19203	2	127	:	:	PUNCT
ajst-19203	2	128	tianrui	tianrui	PROPN
ajst-19203	2	129	liu	liu	PROPN
ajst-19203	2	130	(	(	PUNCT
ajst-19203	2	131	email	email	NOUN
ajst-19203	2	132	:	:	PUNCT
ajst-19203	2	133	tianrui.liu.ml@gmail.com	tianrui.liu.ml@gmail.com	X
ajst-19203	2	134	)	)	PUNCT
ajst-19203	3	1	abstract	abstract	ADJ
ajst-19203	3	2	:	:	PUNCT
ajst-19203	3	3	image	image	PROPN
ajst-19203	3	4	captioning[1	captioning[1	PROPN
ajst-19203	3	5	]	]	PUNCT
ajst-19203	3	6	strives	strive	VERB
ajst-19203	3	7	to	to	PART
ajst-19203	3	8	generate	generate	VERB
ajst-19203	3	9	pertinent	pertinent	ADJ
ajst-19203	3	10	captions	caption	NOUN
ajst-19203	3	11	for	for	ADP
ajst-19203	3	12	specified	specify	VERB
ajst-19203	3	13	images[56	images[56	PROPN
ajst-19203	3	14	]	]	X
ajst-19203	3	15	,	,	PUNCT
ajst-19203	3	16	situating	situate	VERB
ajst-19203	3	17	itself	itself	PRON
ajst-19203	3	18	at	at	ADP
ajst-19203	3	19	the	the	DET
ajst-19203	3	20	crossroads	crossroad	NOUN
ajst-19203	3	21	of	of	ADP
ajst-19203	3	22	computer	computer	NOUN
ajst-19203	3	23	vision	vision	NOUN
ajst-19203	3	24	(	(	PUNCT
ajst-19203	3	25	cv	cv	PROPN
ajst-19203	3	26	)	)	PUNCT
ajst-19203	4	1	[	[	X
ajst-19203	4	2	2	2	NUM
ajst-19203	4	3	]	]	PUNCT
ajst-19203	4	4	and	and	CCONJ
ajst-19203	4	5	natural	natural	ADJ
ajst-19203	4	6	language	language	NOUN
ajst-19203	4	7	processing	processing	NOUN
ajst-19203	4	8	(	(	PUNCT
ajst-19203	4	9	nlp)[54	nlp)[54	NOUN
ajst-19203	4	10	]	]	PUNCT
ajst-19203	4	11	.	.	PUNCT
ajst-19203	5	1	this	this	DET
ajst-19203	5	2	endeavor	endeavor	NOUN
ajst-19203	5	3	is	be	AUX
ajst-19203	5	4	of	of	ADP
ajst-19203	5	5	paramount	paramount	ADJ
ajst-19203	5	6	importance	importance	NOUN
ajst-19203	5	7	with	with	ADP
ajst-19203	5	8	farreaching	farreache	VERB
ajst-19203	5	9	applications	application	NOUN
ajst-19203	5	10	in	in	ADP
ajst-19203	5	11	recommendation	recommendation	NOUN
ajst-19203	5	12	systems	system	NOUN
ajst-19203	5	13	,	,	PUNCT
ajst-19203	5	14	news	news	NOUN
ajst-19203	5	15	outlets	outlet	NOUN
ajst-19203	5	16	,	,	PUNCT
ajst-19203	5	17	social	social	ADJ
ajst-19203	5	18	media	medium	NOUN
ajst-19203	5	19	,	,	PUNCT
ajst-19203	5	20	and	and	CCONJ
ajst-19203	5	21	beyond	beyond	ADP
ajst-19203	5	22	.	.	PUNCT
ajst-19203	6	1	particularly	particularly	ADV
ajst-19203	6	2	within	within	ADP
ajst-19203	6	3	the	the	DET
ajst-19203	6	4	realm	realm	NOUN
ajst-19203	6	5	of	of	ADP
ajst-19203	6	6	news	news	NOUN
ajst-19203	6	7	reporting	reporting	NOUN
ajst-19203	6	8	,	,	PUNCT
ajst-19203	6	9	captions	caption	NOUN
ajst-19203	6	10	are	be	AUX
ajst-19203	6	11	expected	expect	VERB
ajst-19203	6	12	to	to	PART
ajst-19203	6	13	encompass	encompass	VERB
ajst-19203	6	14	detailed	detailed	ADJ
ajst-19203	6	15	information	information	NOUN
ajst-19203	6	16	,	,	PUNCT
ajst-19203	6	17	such	such	ADJ
ajst-19203	6	18	as	as	ADP
ajst-19203	6	19	the	the	DET
ajst-19203	6	20	identities	identity	NOUN
ajst-19203	6	21	of	of	ADP
ajst-19203	6	22	celebrities	celebrity	NOUN
ajst-19203	6	23	captured	capture	VERB
ajst-19203	6	24	in	in	ADP
ajst-19203	6	25	the	the	DET
ajst-19203	6	26	images	image	NOUN
ajst-19203	6	27	.	.	PUNCT
ajst-19203	7	1	however	however	ADV
ajst-19203	7	2	,	,	PUNCT
ajst-19203	7	3	much	much	ADJ
ajst-19203	7	4	of	of	ADP
ajst-19203	7	5	the	the	DET
ajst-19203	7	6	existing	exist	VERB
ajst-19203	7	7	body	body	NOUN
ajst-19203	7	8	of	of	ADP
ajst-19203	7	9	work	work	NOUN
ajst-19203	7	10	primarily	primarily	ADV
ajst-19203	7	11	centers	center	VERB
ajst-19203	7	12	around	around	ADP
ajst-19203	7	13	understanding	understand	VERB
ajst-19203	7	14	scenes	scene	NOUN
ajst-19203	7	15	and	and	CCONJ
ajst-19203	7	16	actions.[3	actions.[3	ADP
ajst-19203	7	17	]	]	PUNCT
ajst-19203	7	18	in	in	ADP
ajst-19203	7	19	this	this	DET
ajst-19203	7	20	paper	paper	NOUN
ajst-19203	7	21	,	,	PUNCT
ajst-19203	7	22	we	we	PRON
ajst-19203	7	23	explore	explore	VERB
ajst-19203	7	24	the	the	DET
ajst-19203	7	25	realm	realm	NOUN
ajst-19203	7	26	of	of	ADP
ajst-19203	7	27	image	image	NOUN
ajst-19203	7	28	captioning	caption	VERB
ajst-19203	7	29	specifically	specifically	ADV
ajst-19203	7	30	tailored	tailor	VERB
ajst-19203	7	31	for	for	ADP
ajst-19203	7	32	celebrity	celebrity	NOUN
ajst-19203	7	33	photographs	photograph	NOUN
ajst-19203	7	34	,	,	PUNCT
ajst-19203	7	35	illustrating	illustrate	VERB
ajst-19203	7	36	its	its	PRON
ajst-19203	7	37	broad	broad	ADJ
ajst-19203	7	38	potential	potential	NOUN
ajst-19203	7	39	for	for	ADP
ajst-19203	7	40	enhancing	enhance	VERB
ajst-19203	7	41	news	news	NOUN
ajst-19203	7	42	industry	industry	NOUN
ajst-19203	7	43	practices	practice	NOUN
ajst-19203	7	44	.	.	PUNCT
ajst-19203	8	1	this	this	DET
ajst-19203	8	2	exploration	exploration	NOUN
ajst-19203	8	3	aims	aim	VERB
ajst-19203	8	4	to	to	PART
ajst-19203	8	5	augment	augment	VERB
ajst-19203	8	6	automated	automate	VERB
ajst-19203	8	7	news	news	NOUN
ajst-19203	8	8	content	content	NOUN
ajst-19203	8	9	generation	generation	NOUN
ajst-19203	8	10	,	,	PUNCT
ajst-19203	8	11	thereby	thereby	ADV
ajst-19203	8	12	facilitating	facilitate	VERB
ajst-19203	8	13	a	a	DET
ajst-19203	8	14	more	more	ADV
ajst-19203	8	15	nuanced	nuanced	ADJ
ajst-19203	8	16	dissemination	dissemination	NOUN
ajst-19203	8	17	of	of	ADP
ajst-19203	8	18	information.[57	information.[57	PROPN
ajst-19203	8	19	]	]	X
ajst-19203	8	20	our	our	PRON
ajst-19203	8	21	endeavor	endeavor	NOUN
ajst-19203	8	22	shows	show	VERB
ajst-19203	8	23	a	a	DET
ajst-19203	8	24	broader	broad	ADJ
ajst-19203	8	25	horizon	horizon	NOUN
ajst-19203	8	26	,	,	PUNCT
ajst-19203	8	27	enriching	enrich	VERB
ajst-19203	8	28	the	the	DET
ajst-19203	8	29	narrative	narrative	NOUN
ajst-19203	8	30	in	in	ADP
ajst-19203	8	31	news	news	NOUN
ajst-19203	8	32	reporting	report	VERB
ajst-19203	8	33	through	through	ADP
ajst-19203	8	34	a	a	DET
ajst-19203	8	35	more	more	ADV
ajst-19203	8	36	intuitive	intuitive	ADJ
ajst-19203	8	37	image	image	NOUN
ajst-19203	8	38	captioning	captioning	NOUN
ajst-19203	8	39	framework	framework	NOUN
ajst-19203	8	40	.	.	PUNCT
ajst-19203	9	1	keywords	keyword	NOUN
ajst-19203	9	2	:	:	PUNCT
ajst-19203	9	3	image	image	NOUN
ajst-19203	9	4	captioning	captioning	NOUN
ajst-19203	9	5	;	;	PUNCT
ajst-19203	9	6	computer	computer	NOUN
ajst-19203	9	7	vision	vision	NOUN
ajst-19203	9	8	;	;	PUNCT
ajst-19203	9	9	natural	natural	ADJ
ajst-19203	9	10	language	language	NOUN
ajst-19203	9	11	processing	processing	NOUN
ajst-19203	9	12	;	;	PUNCT
ajst-19203	9	13	content	content	NOUN
ajst-19203	9	14	generation	generation	NOUN
ajst-19203	9	15	.	.	PUNCT
ajst-19203	10	1	1	1	X
ajst-19203	10	2	.	.	X
ajst-19203	10	3	introduction	introduction	NOUN
ajst-19203	10	4	image	image	NOUN
ajst-19203	10	5	captioning	captioning	NOUN
ajst-19203	10	6	is	be	AUX
ajst-19203	10	7	a	a	DET
ajst-19203	10	8	typical	typical	ADJ
ajst-19203	10	9	topic	topic	NOUN
ajst-19203	10	10	that	that	SCONJ
ajst-19203	10	11	bridges	bridge	NOUN
ajst-19203	10	12	computer	computer	NOUN
ajst-19203	10	13	vision[6	vision[6	NOUN
ajst-19203	10	14	]	]	PUNCT
ajst-19203	10	15	and	and	CCONJ
ajst-19203	10	16	natural	natural	ADJ
ajst-19203	10	17	language	language	NOUN
ajst-19203	10	18	processing.[10,42	processing.[10,42	NOUN
ajst-19203	10	19	]	]	PUNCT
ajst-19203	10	20	in	in	ADP
ajst-19203	10	21	image	image	NOUN
ajst-19203	10	22	captioning	captioning	NOUN
ajst-19203	10	23	task	task	NOUN
ajst-19203	10	24	,	,	PUNCT
ajst-19203	10	25	we	we	PRON
ajst-19203	10	26	aim	aim	VERB
ajst-19203	10	27	to	to	PART
ajst-19203	10	28	generate	generate	VERB
ajst-19203	10	29	relevant	relevant	ADJ
ajst-19203	10	30	captions	caption	NOUN
ajst-19203	10	31	for	for	ADP
ajst-19203	10	32	given	give	VERB
ajst-19203	10	33	images	image	NOUN
ajst-19203	10	34	.	.	PUNCT
ajst-19203	11	1	as	as	ADP
ajst-19203	11	2	a	a	DET
ajst-19203	11	3	very	very	ADV
ajst-19203	11	4	application	application	NOUN
ajst-19203	11	5	-	-	PUNCT
ajst-19203	11	6	oriented	orient	VERB
ajst-19203	11	7	task	task	NOUN
ajst-19203	11	8	,	,	PUNCT
ajst-19203	11	9	there	there	PRON
ajst-19203	11	10	has	have	AUX
ajst-19203	11	11	been	be	AUX
ajst-19203	11	12	a	a	DET
ajst-19203	11	13	wealth	wealth	NOUN
ajst-19203	11	14	of	of	ADP
ajst-19203	11	15	research	research	NOUN
ajst-19203	11	16	on	on	ADP
ajst-19203	11	17	image	image	NOUN
ajst-19203	11	18	captioning.[12	captioning.[12	PROPN
ajst-19203	11	19	]	]	X
ajst-19203	11	20	this	this	DET
ajst-19203	11	21	technique	technique	NOUN
ajst-19203	11	22	has	have	AUX
ajst-19203	11	23	been	be	AUX
ajst-19203	11	24	widely	widely	ADV
ajst-19203	11	25	used	use	VERB
ajst-19203	11	26	in	in	ADP
ajst-19203	11	27	many	many	ADJ
ajst-19203	11	28	fields	field	NOUN
ajst-19203	11	29	:	:	PUNCT
ajst-19203	11	30	traffic	traffic	NOUN
ajst-19203	11	31	detection[13,24	detection[13,24	PROPN
ajst-19203	11	32	]	]	PUNCT
ajst-19203	11	33	,	,	PUNCT
ajst-19203	11	34	material	material	NOUN
ajst-19203	11	35	analysis[22	analysis[22	NOUN
ajst-19203	11	36	]	]	PUNCT
ajst-19203	11	37	,	,	PUNCT
ajst-19203	11	38	communications	communication	NOUN
ajst-19203	11	39	deployment[35,39	deployment[35,39	NOUN
ajst-19203	11	40	]	]	PUNCT
ajst-19203	11	41	,	,	PUNCT
ajst-19203	11	42	aerial	aerial	ADJ
ajst-19203	11	43	search[37	search[37	NOUN
ajst-19203	11	44	]	]	X
ajst-19203	11	45	,	,	PUNCT
ajst-19203	11	46	language	language	NOUN
ajst-19203	11	47	model	model	NOUN
ajst-19203	11	48	designing[38	designing[38	PROPN
ajst-19203	11	49	]	]	PUNCT
ajst-19203	11	50	,	,	PUNCT
ajst-19203	11	51	embedding	embed	VERB
ajst-19203	11	52	development[40,41	development[40,41	NOUN
ajst-19203	11	53	]	]	PUNCT
ajst-19203	11	54	.	.	PUNCT
ajst-19203	12	1	meanwhile	meanwhile	ADV
ajst-19203	12	2	,	,	PUNCT
ajst-19203	12	3	there	there	PRON
ajst-19203	12	4	are	be	VERB
ajst-19203	12	5	much	much	ADV
ajst-19203	12	6	less	less	ADJ
ajst-19203	12	7	research	research	NOUN
ajst-19203	12	8	that	that	PRON
ajst-19203	12	9	considers	consider	VERB
ajst-19203	12	10	generating	generate	VERB
ajst-19203	12	11	captions	caption	NOUN
ajst-19203	12	12	for	for	ADP
ajst-19203	12	13	specific	specific	ADJ
ajst-19203	12	14	names	name	NOUN
ajst-19203	12	15	.	.	PUNCT
ajst-19203	13	1	we	we	PRON
ajst-19203	13	2	believe	believe	VERB
ajst-19203	13	3	this	this	DET
ajst-19203	13	4	problem	problem	NOUN
ajst-19203	13	5	is	be	AUX
ajst-19203	13	6	important	important	ADJ
ajst-19203	13	7	in	in	ADP
ajst-19203	13	8	the	the	DET
ajst-19203	13	9	settings	setting	NOUN
ajst-19203	13	10	of	of	ADP
ajst-19203	13	11	news	news	NOUN
ajst-19203	13	12	report	report	NOUN
ajst-19203	13	13	.	.	PUNCT
ajst-19203	14	1	in	in	ADP
ajst-19203	14	2	such	such	ADJ
ajst-19203	14	3	scenarios	scenario	NOUN
ajst-19203	14	4	,	,	PUNCT
ajst-19203	14	5	captions	caption	NOUN
ajst-19203	14	6	generated	generate	VERB
ajst-19203	14	7	by	by	ADP
ajst-19203	14	8	algorithms	algorithm	NOUN
ajst-19203	14	9	should	should	AUX
ajst-19203	14	10	take	take	VERB
ajst-19203	14	11	the	the	DET
ajst-19203	14	12	faces	face	NOUN
ajst-19203	14	13	into	into	ADP
ajst-19203	14	14	account.[4	account.[4	PRON
ajst-19203	14	15	]	]	PUNCT
ajst-19203	14	16	say	say	VERB
ajst-19203	14	17	we	we	PRON
ajst-19203	14	18	intend	intend	VERB
ajst-19203	14	19	to	to	PART
ajst-19203	14	20	generate	generate	VERB
ajst-19203	14	21	descriptions	description	NOUN
ajst-19203	14	22	like	like	ADP
ajst-19203	14	23	"	"	PUNCT
ajst-19203	14	24	obama	obama	PROPN
ajst-19203	14	25	is	be	AUX
ajst-19203	14	26	delivering	deliver	VERB
ajst-19203	14	27	a	a	DET
ajst-19203	14	28	speech".[36	speech".[36	NOUN
ajst-19203	14	29	]	]	PUNCT
ajst-19203	14	30	common	common	ADJ
ajst-19203	14	31	method	method	NOUN
ajst-19203	14	32	may	may	AUX
ajst-19203	14	33	only	only	ADV
ajst-19203	14	34	be	be	AUX
ajst-19203	14	35	able	able	ADJ
ajst-19203	14	36	to	to	PART
ajst-19203	14	37	generate	generate	VERB
ajst-19203	14	38	sentences	sentence	NOUN
ajst-19203	14	39	like	like	ADP
ajst-19203	14	40	"	"	PUNCT
ajst-19203	14	41	a	a	DET
ajst-19203	14	42	man	man	NOUN
ajst-19203	14	43	is	be	AUX
ajst-19203	14	44	delivering	deliver	VERB
ajst-19203	14	45	a	a	DET
ajst-19203	14	46	speech	speech	NOUN
ajst-19203	14	47	"	"	PUNCT
ajst-19203	14	48	and	and	CCONJ
ajst-19203	14	49	is	be	AUX
ajst-19203	14	50	unacceptable	unacceptable	ADJ
ajst-19203	14	51	apparently.[25	apparently.[25	ADP
ajst-19203	14	52	]	]	PUNCT
ajst-19203	14	53	in	in	ADP
ajst-19203	14	54	this	this	DET
ajst-19203	14	55	project	project	NOUN
ajst-19203	14	56	,	,	PUNCT
ajst-19203	14	57	we	we	PRON
ajst-19203	14	58	shed	shed	VERB
ajst-19203	14	59	light	light	NOUN
ajst-19203	14	60	on	on	ADP
ajst-19203	14	61	the	the	DET
ajst-19203	14	62	problem	problem	NOUN
ajst-19203	14	63	of	of	ADP
ajst-19203	14	64	image	image	NOUN
ajst-19203	14	65	captioning	captioning	NOUN
ajst-19203	14	66	in	in	ADP
ajst-19203	14	67	scenarios	scenario	NOUN
ajst-19203	14	68	where	where	SCONJ
ajst-19203	14	69	celebrities	celebrity	NOUN
ajst-19203	14	70	appear[50	appear[50	VERB
ajst-19203	14	71	]	]	PUNCT
ajst-19203	14	72	and	and	CCONJ
ajst-19203	14	73	propose	propose	VERB
ajst-19203	14	74	a	a	DET
ajst-19203	14	75	combined	combine	VERB
ajst-19203	14	76	method	method	NOUN
ajst-19203	14	77	to	to	PART
ajst-19203	14	78	solve	solve	VERB
ajst-19203	14	79	this	this	DET
ajst-19203	14	80	problem.[29	problem.[29	NOUN
ajst-19203	14	81	]	]	PUNCT
ajst-19203	14	82	our	our	PRON
ajst-19203	14	83	algorithms	algorithm	NOUN
ajst-19203	14	84	take	take	VERB
ajst-19203	14	85	three	three	NUM
ajst-19203	14	86	step	step	NOUN
ajst-19203	14	87	to	to	PART
ajst-19203	14	88	generate	generate	VERB
ajst-19203	14	89	the	the	DET
ajst-19203	14	90	ultimate	ultimate	ADJ
ajst-19203	14	91	sentences	sentence	NOUN
ajst-19203	14	92	:	:	PUNCT
ajst-19203	14	93	(	(	PUNCT
ajst-19203	14	94	1	1	X
ajst-19203	14	95	)	)	PUNCT
ajst-19203	14	96	for	for	ADP
ajst-19203	14	97	the	the	DET
ajst-19203	14	98	given	give	VERB
ajst-19203	14	99	image	image	NOUN
ajst-19203	14	100	(	(	PUNCT
ajst-19203	14	101	with	with	SCONJ
ajst-19203	14	102	celebrities	celebrity	NOUN
ajst-19203	14	103	appear	appear	VERB
ajst-19203	14	104	)	)	PUNCT
ajst-19203	14	105	,	,	PUNCT
ajst-19203	14	106	we	we	PRON
ajst-19203	14	107	employ	employ	VERB
ajst-19203	14	108	a	a	DET
ajst-19203	14	109	common	common	ADJ
ajst-19203	14	110	encoder	encoder	NOUN
ajst-19203	14	111	-	-	PUNCT
ajst-19203	14	112	decoder	decoder	NOUN
ajst-19203	14	113	architecture	architecture	NOUN
ajst-19203	14	114	for	for	ADP
ajst-19203	14	115	image	image	NOUN
ajst-19203	14	116	captioning	captioning	NOUN
ajst-19203	14	117	and	and	CCONJ
ajst-19203	14	118	generate	generate	VERB
ajst-19203	14	119	captions	caption	NOUN
ajst-19203	14	120	without	without	ADP
ajst-19203	14	121	names.[30	names.[30	NOUN
ajst-19203	14	122	]	]	PUNCT
ajst-19203	14	123	(	(	PUNCT
ajst-19203	14	124	2	2	X
ajst-19203	14	125	)	)	PUNCT
ajst-19203	14	126	we	we	PRON
ajst-19203	14	127	use	use	VERB
ajst-19203	14	128	mctnn	mctnn	NOUN
ajst-19203	14	129	and	and	CCONJ
ajst-19203	14	130	resnet[34	resnet[34	PROPN
ajst-19203	14	131	]	]	PUNCT
ajst-19203	14	132	to	to	PART
ajst-19203	14	133	get	get	VERB
ajst-19203	14	134	the	the	DET
ajst-19203	14	135	names	name	NOUN
ajst-19203	14	136	of	of	ADP
ajst-19203	14	137	faces	face	NOUN
ajst-19203	14	138	appearing	appear	VERB
ajst-19203	14	139	in	in	ADP
ajst-19203	14	140	the	the	DET
ajst-19203	14	141	image	image	NOUN
ajst-19203	14	142	.	.	PUNCT
ajst-19203	15	1	(	(	PUNCT
ajst-19203	15	2	3	3	X
ajst-19203	15	3	)	)	PUNCT
ajst-19203	15	4	we	we	PRON
ajst-19203	15	5	parse	parse	VERB
ajst-19203	15	6	the	the	DET
ajst-19203	15	7	output	output	NOUN
ajst-19203	15	8	sentence	sentence	NOUN
ajst-19203	15	9	in	in	ADP
ajst-19203	15	10	step(1	step(1	NOUN
ajst-19203	15	11	)	)	PUNCT
ajst-19203	15	12	and	and	CCONJ
ajst-19203	15	13	replace	replace	VERB
ajst-19203	15	14	the	the	DET
ajst-19203	15	15	parts	part	NOUN
ajst-19203	15	16	with	with	ADP
ajst-19203	15	17	celebrity	celebrity	NOUN
ajst-19203	15	18	names	name	NOUN
ajst-19203	15	19	accordingly	accordingly	ADV
ajst-19203	15	20	.	.	PUNCT
ajst-19203	16	1	extensive	extensive	ADJ
ajst-19203	16	2	experiments	experiment	NOUN
ajst-19203	16	3	show	show	VERB
ajst-19203	16	4	the	the	DET
ajst-19203	16	5	feasibility	feasibility	NOUN
ajst-19203	16	6	of	of	ADP
ajst-19203	16	7	our	our	PRON
ajst-19203	16	8	method	method	NOUN
ajst-19203	16	9	in	in	ADP
ajst-19203	16	10	many	many	ADJ
ajst-19203	16	11	simple	simple	ADJ
ajst-19203	16	12	scenarios	scenario	NOUN
ajst-19203	16	13	.	.	PUNCT
ajst-19203	17	1	the	the	DET
ajst-19203	17	2	structure	structure	NOUN
ajst-19203	17	3	of	of	ADP
ajst-19203	17	4	the	the	DET
ajst-19203	17	5	report	report	NOUN
ajst-19203	17	6	are	be	AUX
ajst-19203	17	7	as	as	SCONJ
ajst-19203	17	8	follows	follow	VERB
ajst-19203	17	9	:	:	PUNCT
ajst-19203	17	10	section	section	NOUN
ajst-19203	17	11	2	2	NUM
ajst-19203	17	12	defines	define	VERB
ajst-19203	17	13	the	the	DET
ajst-19203	17	14	subtask	subtask	NOUN
ajst-19203	17	15	of	of	ADP
ajst-19203	17	16	our	our	PRON
ajst-19203	17	17	project	project	NOUN
ajst-19203	17	18	,	,	PUNCT
ajst-19203	17	19	which	which	PRON
ajst-19203	17	20	are	be	AUX
ajst-19203	17	21	comprised	comprise	VERB
ajst-19203	17	22	of	of	ADP
ajst-19203	17	23	image	image	NOUN
ajst-19203	17	24	captioning	captioning	NOUN
ajst-19203	17	25	,	,	PUNCT
ajst-19203	17	26	face	face	NOUN
ajst-19203	17	27	recognition	recognition	NOUN
ajst-19203	17	28	,	,	PUNCT
ajst-19203	17	29	noun	noun	NOUN
ajst-19203	17	30	phrase(np	phrase(np	NOUN
ajst-19203	17	31	)	)	PUNCT
ajst-19203	17	32	chunks	chunk	VERB
ajst-19203	17	33	matching.[51	matching.[51	PROPN
ajst-19203	17	34	]	]	PUNCT
ajst-19203	17	35	section	section	NOUN
ajst-19203	17	36	3	3	NUM
ajst-19203	17	37	discusses	discuss	VERB
ajst-19203	17	38	the	the	DET
ajst-19203	17	39	details	detail	NOUN
ajst-19203	17	40	of	of	ADP
ajst-19203	17	41	our	our	PRON
ajst-19203	17	42	methods	method	NOUN
ajst-19203	17	43	and	and	CCONJ
ajst-19203	17	44	implementations	implementation	NOUN
ajst-19203	17	45	.	.	PUNCT
ajst-19203	18	1	our	our	PRON
ajst-19203	18	2	experiment	experiment	NOUN
ajst-19203	18	3	and	and	CCONJ
ajst-19203	18	4	results	result	NOUN
ajst-19203	18	5	are	be	AUX
ajst-19203	18	6	shown	show	VERB
ajst-19203	18	7	in	in	ADP
ajst-19203	18	8	section	section	NOUN
ajst-19203	18	9	4	4	NUM
ajst-19203	18	10	.	.	PUNCT
ajst-19203	19	1	in	in	ADP
ajst-19203	19	2	section	section	NOUN
ajst-19203	19	3	5	5	NUM
ajst-19203	19	4	,	,	PUNCT
ajst-19203	19	5	we	we	PRON
ajst-19203	19	6	conclude	conclude	VERB
ajst-19203	19	7	our	our	PRON
ajst-19203	19	8	approach	approach	NOUN
ajst-19203	19	9	and	and	CCONJ
ajst-19203	19	10	further	far	ADV
ajst-19203	19	11	discuss	discuss	VERB
ajst-19203	19	12	the	the	DET
ajst-19203	19	13	strength	strength	NOUN
ajst-19203	19	14	and	and	CCONJ
ajst-19203	19	15	weakness	weakness	NOUN
ajst-19203	19	16	of	of	ADP
ajst-19203	19	17	our	our	PRON
ajst-19203	19	18	methods	method	NOUN
ajst-19203	19	19	.	.	PUNCT
ajst-19203	20	1	2	2	X
ajst-19203	20	2	.	.	X
ajst-19203	20	3	problem	problem	NOUN
ajst-19203	20	4	definition	definition	NOUN
ajst-19203	20	5	2.1	2.1	NUM
ajst-19203	20	6	.	.	PUNCT
ajst-19203	21	1	image	image	NOUN
ajst-19203	21	2	captioning	caption	VERB
ajst-19203	21	3	the	the	DET
ajst-19203	21	4	goal	goal	NOUN
ajst-19203	21	5	of	of	ADP
ajst-19203	21	6	image	image	NOUN
ajst-19203	21	7	captioning	captioning	NOUN
ajst-19203	21	8	is	be	AUX
ajst-19203	21	9	to	to	PART
ajst-19203	21	10	generate	generate	VERB
ajst-19203	21	11	relevant	relevant	ADJ
ajst-19203	21	12	description	description	NOUN
ajst-19203	21	13	for	for	ADP
ajst-19203	21	14	given	give	VERB
ajst-19203	21	15	images	image	NOUN
ajst-19203	21	16	.	.	PUNCT
ajst-19203	22	1	the	the	DET
ajst-19203	22	2	problem	problem	NOUN
ajst-19203	22	3	can	can	AUX
ajst-19203	22	4	be	be	AUX
ajst-19203	22	5	thought	think	VERB
ajst-19203	22	6	two	two	NUM
ajst-19203	22	7	-	-	NOUN
ajst-19203	22	8	fold	fold	VERB
ajst-19203	22	9	as	as	SCONJ
ajst-19203	22	10	it	it	PRON
ajst-19203	22	11	connects	connect	VERB
ajst-19203	22	12	computer	computer	NOUN
ajst-19203	22	13	vision	vision	NOUN
ajst-19203	22	14	and	and	CCONJ
ajst-19203	22	15	natural	natural	ADJ
ajst-19203	22	16	language	language	NOUN
ajst-19203	22	17	processing	processing	NOUN
ajst-19203	22	18	:	:	PUNCT
ajst-19203	22	19	1	1	X
ajst-19203	22	20	)	)	PUNCT
ajst-19203	22	21	use	use	VERB
ajst-19203	22	22	an	an	DET
ajst-19203	22	23	encoder	encoder	NOUN
ajst-19203	22	24	architecture	architecture	NOUN
ajst-19203	22	25	(	(	PUNCT
ajst-19203	22	26	cnn	cnn	PROPN
ajst-19203	22	27	,	,	PUNCT
ajst-19203	22	28	transformer)[32	transformer)[32	PROPN
ajst-19203	22	29	]	]	PUNCT
ajst-19203	22	30	to	to	PART
ajst-19203	22	31	process	process	VERB
ajst-19203	22	32	the	the	DET
ajst-19203	22	33	image	image	NOUN
ajst-19203	22	34	;	;	PUNCT
ajst-19203	22	35	2	2	X
ajst-19203	22	36	)	)	PUNCT
ajst-19203	22	37	use	use	VERB
ajst-19203	22	38	a	a	DET
ajst-19203	22	39	decoder	decoder	NOUN
ajst-19203	22	40	to	to	PART
ajst-19203	22	41	decode	decode	VERB
ajst-19203	22	42	the	the	DET
ajst-19203	22	43	encoded	encode	VERB
ajst-19203	22	44	image	image	NOUN
ajst-19203	22	45	representation	representation	NOUN
ajst-19203	22	46	into	into	ADP
ajst-19203	22	47	sentences.[16	sentences.[16	PROPN
ajst-19203	22	48	]	]	PUNCT
ajst-19203	22	49	the	the	DET
ajst-19203	22	50	model	model	NOUN
ajst-19203	22	51	is	be	AUX
ajst-19203	22	52	trained	train	VERB
ajst-19203	22	53	in	in	ADP
ajst-19203	22	54	a	a	DET
ajst-19203	22	55	supervised	supervised	ADJ
ajst-19203	22	56	learning	learning	NOUN
ajst-19203	22	57	pattern[11	pattern[11	NOUN
ajst-19203	22	58	]	]	PUNCT
ajst-19203	22	59	with	with	ADP
ajst-19203	22	60	some	some	DET
ajst-19203	22	61	existing	exist	VERB
ajst-19203	22	62	image	image	NOUN
ajst-19203	22	63	to	to	ADP
ajst-19203	22	64	caption	caption	NOUN
ajst-19203	22	65	(	(	PUNCT
ajst-19203	22	66	ground	ground	NOUN
ajst-19203	22	67	truth	truth	NOUN
ajst-19203	22	68	)	)	PUNCT
ajst-19203	22	69	pairs	pair	NOUN
ajst-19203	22	70	.	.	PUNCT
ajst-19203	23	1	in	in	ADP
ajst-19203	23	2	the	the	DET
ajst-19203	23	3	training	training	NOUN
ajst-19203	23	4	stage	stage	NOUN
ajst-19203	23	5	we	we	PRON
ajst-19203	23	6	maximize	maximize	VERB
ajst-19203	23	7	the	the	DET
ajst-19203	23	8	similarity	similarity	NOUN
ajst-19203	23	9	of	of	ADP
ajst-19203	23	10	generated	generate	VERB
ajst-19203	23	11	caption	caption	NOUN
ajst-19203	23	12	with	with	ADP
ajst-19203	23	13	ground	ground	NOUN
ajst-19203	23	14	truth,[14	truth,[14	PROPN
ajst-19203	23	15	]	]	PUNCT
ajst-19203	23	16	and	and	CCONJ
ajst-19203	23	17	in	in	ADP
ajst-19203	23	18	the	the	DET
ajst-19203	23	19	testing	testing	NOUN
ajst-19203	23	20	stage	stage	NOUN
ajst-19203	23	21	we	we	PRON
ajst-19203	23	22	decode	decode	VERB
ajst-19203	23	23	the	the	DET
ajst-19203	23	24	encoded	encode	VERB
ajst-19203	23	25	image	image	NOUN
ajst-19203	23	26	representation	representation	NOUN
ajst-19203	23	27	directly	directly	ADV
ajst-19203	23	28	to	to	PART
ajst-19203	23	29	obtain	obtain	VERB
ajst-19203	23	30	the	the	DET
ajst-19203	23	31	outcome	outcome	NOUN
ajst-19203	23	32	.	.	PUNCT
ajst-19203	24	1	in	in	ADP
ajst-19203	24	2	recent	recent	ADJ
ajst-19203	24	3	years	year	NOUN
ajst-19203	24	4	,	,	PUNCT
ajst-19203	24	5	various	various	ADJ
ajst-19203	24	6	efforts	effort	NOUN
ajst-19203	24	7	have	have	AUX
ajst-19203	24	8	been	be	AUX
ajst-19203	24	9	made	make	VERB
ajst-19203	24	10	in	in	ADP
ajst-19203	24	11	the	the	DET
ajst-19203	24	12	image	image	NOUN
ajst-19203	24	13	captioning	captioning	NOUN
ajst-19203	24	14	tasks	task	NOUN
ajst-19203	24	15	and	and	CCONJ
ajst-19203	24	16	the	the	DET
ajst-19203	24	17	performance	performance	NOUN
ajst-19203	24	18	of	of	ADP
ajst-19203	24	19	caption	caption	NOUN
ajst-19203	24	20	generation	generation	NOUN
ajst-19203	24	21	has	have	AUX
ajst-19203	24	22	improved	improve	VERB
ajst-19203	24	23	significantly	significantly	ADV
ajst-19203	24	24	thanks	thank	NOUN
ajst-19203	24	25	to	to	ADP
ajst-19203	24	26	more	more	ADV
ajst-19203	24	27	efficient	efficient	ADJ
ajst-19203	24	28	encoderdecoder	encoderdecoder	NOUN
ajst-19203	24	29	architectures.[20	architectures.[20	PROPN
ajst-19203	24	30	]	]	PUNCT
ajst-19203	25	1	2.2	2.2	NUM
ajst-19203	25	2	.	.	PUNCT
ajst-19203	25	3	face	face	NOUN
ajst-19203	25	4	recognition	recognition	NOUN
ajst-19203	25	5	in	in	ADP
ajst-19203	25	6	images	image	NOUN
ajst-19203	25	7	facial	facial	ADJ
ajst-19203	25	8	recognition	recognition	NOUN
ajst-19203	25	9	is	be	AUX
ajst-19203	25	10	the	the	DET
ajst-19203	25	11	task	task	NOUN
ajst-19203	25	12	of	of	ADP
ajst-19203	25	13	making	make	VERB
ajst-19203	25	14	a	a	DET
ajst-19203	25	15	positive	positive	ADJ
ajst-19203	25	16	identification	identification	NOUN
ajst-19203	25	17	of	of	ADP
ajst-19203	25	18	a	a	DET
ajst-19203	25	19	face	face	NOUN
ajst-19203	25	20	in	in	ADP
ajst-19203	25	21	a	a	DET
ajst-19203	25	22	photo	photo	NOUN
ajst-19203	25	23	against	against	ADP
ajst-19203	25	24	a	a	DET
ajst-19203	25	25	pre	pre	ADJ
ajst-19203	25	26	-	-	ADJ
ajst-19203	25	27	existing	existing	ADJ
ajst-19203	25	28	database	database	NOUN
ajst-19203	25	29	of	of	ADP
ajst-19203	25	30	faces	face	NOUN
ajst-19203	25	31	.	.	PUNCT
ajst-19203	26	1	it	it	PRON
ajst-19203	26	2	begins	begin	VERB
ajst-19203	26	3	with	with	ADP
ajst-19203	26	4	detection	detection	NOUN
ajst-19203	26	5	distinguishing	distinguish	VERB
ajst-19203	26	6	human	human	ADJ
ajst-19203	26	7	faces	face	NOUN
ajst-19203	26	8	from	from	ADP
ajst-19203	26	9	other	other	ADJ
ajst-19203	26	10	objects	object	NOUN
ajst-19203	26	11	in	in	ADP
ajst-19203	26	12	the	the	DET
ajst-19203	26	13	image	image	NOUN
ajst-19203	26	14	and	and	CCONJ
ajst-19203	26	15	then	then	ADV
ajst-19203	26	16	works	work	VERB
ajst-19203	26	17	on	on	ADP
ajst-19203	26	18	identification	identification	NOUN
ajst-19203	26	19	of	of	ADP
ajst-19203	26	20	those	those	DET
ajst-19203	26	21	detected	detect	VERB
ajst-19203	26	22	faces	face	NOUN
ajst-19203	26	23	,	,	PUNCT
ajst-19203	26	24	which	which	PRON
ajst-19203	26	25	can	can	AUX
ajst-19203	26	26	be	be	AUX
ajst-19203	26	27	seen	see	VERB
ajst-19203	26	28	as	as	ADP
ajst-19203	26	29	a	a	DET
ajst-19203	26	30	classification	classification	NOUN
ajst-19203	26	31	problem	problem	NOUN
ajst-19203	26	32	.	.	PUNCT
ajst-19203	27	1	[	[	X
ajst-19203	27	2	15	15	NUM
ajst-19203	27	3	]	]	X
ajst-19203	27	4	in	in	ADP
ajst-19203	27	5	general	general	ADJ
ajst-19203	27	6	,	,	PUNCT
ajst-19203	27	7	face	face	NOUN
ajst-19203	27	8	recognition	recognition	NOUN
ajst-19203	27	9	is	be	AUX
ajst-19203	27	10	a	a	DET
ajst-19203	27	11	well	well	ADV
ajst-19203	27	12	-	-	PUNCT
ajst-19203	27	13	defined	define	VERB
ajst-19203	27	14	and	and	CCONJ
ajst-19203	27	15	relatively	relatively	ADV
ajst-19203	27	16	mature	mature	ADJ
ajst-19203	27	17	field	field	NOUN
ajst-19203	27	18	.	.	PUNCT
ajst-19203	28	1	there	there	PRON
ajst-19203	28	2	has	have	AUX
ajst-19203	28	3	been	be	AUX
ajst-19203	28	4	many	many	ADJ
ajst-19203	28	5	theoretical	theoretical	ADJ
ajst-19203	28	6	analysis[21	analysis[21	NOUN
ajst-19203	28	7	]	]	PUNCT
ajst-19203	28	8	and	and	CCONJ
ajst-19203	28	9	successful	successful	ADJ
ajst-19203	28	10	engineering	engineering	NOUN
ajst-19203	28	11	practice	practice	NOUN
ajst-19203	28	12	that	that	PRON
ajst-19203	28	13	obtain	obtain	VERB
ajst-19203	28	14	very	very	ADV
ajst-19203	28	15	high	high	ADJ
ajst-19203	28	16	accuracy	accuracy	NOUN
ajst-19203	28	17	.	.	PUNCT
ajst-19203	29	1	in	in	ADP
ajst-19203	29	2	this	this	DET
ajst-19203	29	3	task	task	NOUN
ajst-19203	29	4	,	,	PUNCT
ajst-19203	29	5	we	we	PRON
ajst-19203	29	6	employ	employ	VERB
ajst-19203	29	7	mtcnn	mtcnn	ADJ
ajst-19203	29	8	to	to	PART
ajst-19203	29	9	obtain	obtain	VERB
ajst-19203	29	10	the	the	DET
ajst-19203	29	11	285	285	NUM
ajst-19203	29	12	bounding	bounding	NOUN
ajst-19203	29	13	box	box	NOUN
ajst-19203	29	14	for	for	ADP
ajst-19203	29	15	faces	face	NOUN
ajst-19203	29	16	and	and	CCONJ
ajst-19203	29	17	use	use	VERB
ajst-19203	29	18	a	a	DET
ajst-19203	29	19	resnet	resnet	NOUN
ajst-19203	29	20	pretrained	pretraine	VERB
ajst-19203	29	21	with	with	ADP
ajst-19203	29	22	vggface2	vggface2	NOUN
ajst-19203	29	23	to	to	PART
ajst-19203	29	24	complete	complete	VERB
ajst-19203	29	25	the	the	DET
ajst-19203	29	26	classification	classification	NOUN
ajst-19203	29	27	tasks.[43	tasks.[43	NOUN
ajst-19203	29	28	]	]	PUNCT
ajst-19203	29	29	2.3	2.3	NUM
ajst-19203	29	30	.	.	PUNCT
ajst-19203	30	1	celebrity	celebrity	NOUN
ajst-19203	30	2	-	-	PUNCT
ajst-19203	30	3	aware	aware	ADJ
ajst-19203	30	4	image	image	NOUN
ajst-19203	30	5	captioning	caption	VERB
ajst-19203	30	6	celebrity	celebrity	NOUN
ajst-19203	30	7	-	-	PUNCT
ajst-19203	30	8	aware	aware	ADJ
ajst-19203	30	9	image	image	NOUN
ajst-19203	30	10	captioning	captioning	NOUN
ajst-19203	30	11	problem	problem	NOUN
ajst-19203	30	12	combines	combine	VERB
ajst-19203	30	13	both	both	PRON
ajst-19203	30	14	image	image	NOUN
ajst-19203	30	15	captioning	captioning	NOUN
ajst-19203	30	16	and	and	CCONJ
ajst-19203	30	17	face	face	NOUN
ajst-19203	30	18	recognition	recognition	NOUN
ajst-19203	30	19	.	.	PUNCT
ajst-19203	31	1	3	3	X
ajst-19203	31	2	.	.	X
ajst-19203	31	3	approach	approach	VERB
ajst-19203	31	4	our	our	PRON
ajst-19203	31	5	pipeline	pipeline	NOUN
ajst-19203	31	6	has	have	VERB
ajst-19203	31	7	three	three	NUM
ajst-19203	31	8	parts	part	NOUN
ajst-19203	31	9	:	:	PUNCT
ajst-19203	31	10	1	1	X
ajst-19203	31	11	.	.	X
ajst-19203	31	12	image	image	NOUN
ajst-19203	31	13	captioning	captioning	NOUN
ajst-19203	31	14	:	:	PUNCT
ajst-19203	31	15	image	image	NOUN
ajst-19203	31	16	captioning	captioning	NOUN
ajst-19203	31	17	modules	module	NOUN
ajst-19203	31	18	do	do	AUX
ajst-19203	31	19	supervised	supervise	VERB
ajst-19203	31	20	training	training	NOUN
ajst-19203	31	21	on	on	ADP
ajst-19203	31	22	datasets[49	datasets[49	PROPN
ajst-19203	31	23	]	]	X
ajst-19203	31	24	(	(	PUNCT
ajst-19203	31	25	we	we	PRON
ajst-19203	31	26	use	use	VERB
ajst-19203	31	27	flickr	flickr	PROPN
ajst-19203	31	28	8k/3k	8k/3k	NUM
ajst-19203	31	29	)	)	PUNCT
ajst-19203	31	30	and	and	CCONJ
ajst-19203	31	31	generate	generate	VERB
ajst-19203	31	32	common	common	ADJ
ajst-19203	31	33	(	(	PUNCT
ajst-19203	31	34	no	no	DET
ajst-19203	31	35	specific	specific	ADJ
ajst-19203	31	36	names	name	NOUN
ajst-19203	31	37	)	)	PUNCT
ajst-19203	31	38	captions	caption	NOUN
ajst-19203	31	39	in	in	ADP
ajst-19203	31	40	the	the	DET
ajst-19203	31	41	testing	testing	NOUN
ajst-19203	31	42	stage.[7	stage.[7	AUX
ajst-19203	31	43	]	]	PUNCT
ajst-19203	31	44	2	2	NUM
ajst-19203	31	45	.	.	NOUN
ajst-19203	31	46	face	face	NOUN
ajst-19203	31	47	recognition	recognition	NOUN
ajst-19203	31	48	:	:	PUNCT
ajst-19203	31	49	face	face	NOUN
ajst-19203	31	50	recognition	recognition	NOUN
ajst-19203	31	51	employs	employ	VERB
ajst-19203	31	52	mtcnn	mtcnn	ADJ
ajst-19203	31	53	network	network	NOUN
ajst-19203	31	54	to	to	PART
ajst-19203	31	55	obtain	obtain	VERB
ajst-19203	31	56	bounding	bounding	NOUN
ajst-19203	31	57	boxes	box	NOUN
ajst-19203	31	58	for	for	ADP
ajst-19203	31	59	all	all	DET
ajst-19203	31	60	appeared	appear	VERB
ajst-19203	31	61	faces	face	NOUN
ajst-19203	31	62	,	,	PUNCT
ajst-19203	31	63	and	and	CCONJ
ajst-19203	31	64	do	do	VERB
ajst-19203	31	65	classification	classification	NOUN
ajst-19203	31	66	using	use	VERB
ajst-19203	31	67	inception\_v1	inception\_v1	NUM
ajst-19203	31	68	pretrained	pretraine	VERB
ajst-19203	31	69	on	on	ADP
ajst-19203	31	70	vggface2.[31	vggface2.[31	NOUN
ajst-19203	31	71	]	]	X
ajst-19203	31	72	3	3	X
ajst-19203	31	73	.	.	X
ajst-19203	31	74	noun	noun	NOUN
ajst-19203	31	75	phrase	phrase	NOUN
ajst-19203	31	76	matching	matching	NOUN
ajst-19203	31	77	:	:	PUNCT
ajst-19203	31	78	noun	noun	NOUN
ajst-19203	31	79	phrase	phrase	NOUN
ajst-19203	31	80	matching	matching	NOUN
ajst-19203	31	81	module	module	NOUN
ajst-19203	31	82	uses	use	VERB
ajst-19203	31	83	nlp	nlp	NOUN
ajst-19203	31	84	packages	package	NOUN
ajst-19203	31	85	and	and	CCONJ
ajst-19203	31	86	some	some	DET
ajst-19203	31	87	rules	rule	NOUN
ajst-19203	31	88	to	to	PART
ajst-19203	31	89	obtain	obtain	VERB
ajst-19203	31	90	noun	noun	NOUN
ajst-19203	31	91	phrase	phrase	NOUN
ajst-19203	31	92	chunks	chunk	NOUN
ajst-19203	31	93	(	(	PUNCT
ajst-19203	31	94	also	also	ADV
ajst-19203	31	95	needs	need	VERB
ajst-19203	31	96	to	to	PART
ajst-19203	31	97	be	be	AUX
ajst-19203	31	98	people	people	NOUN
ajst-19203	31	99	-	-	PUNCT
ajst-19203	31	100	related	relate	VERB
ajst-19203	31	101	)	)	PUNCT
ajst-19203	31	102	like	like	ADP
ajst-19203	31	103	"	"	PUNCT
ajst-19203	31	104	a	a	DET
ajst-19203	31	105	man	man	NOUN
ajst-19203	31	106	"	"	PUNCT
ajst-19203	31	107	or	or	CCONJ
ajst-19203	31	108	"	"	PUNCT
ajst-19203	31	109	a	a	DET
ajst-19203	31	110	young	young	ADJ
ajst-19203	31	111	asian	asian	ADJ
ajst-19203	31	112	boy	boy	NOUN
ajst-19203	31	113	"	"	PUNCT
ajst-19203	31	114	and	and	CCONJ
ajst-19203	31	115	replaces	replace	VERB
ajst-19203	31	116	them	they	PRON
ajst-19203	31	117	with	with	ADP
ajst-19203	31	118	the	the	DET
ajst-19203	31	119	names	name	NOUN
ajst-19203	31	120	derived	derive	VERB
ajst-19203	31	121	in	in	ADP
ajst-19203	31	122	phrase	phrase	NOUN
ajst-19203	31	123	ii	ii	NOUN
ajst-19203	31	124	and	and	CCONJ
ajst-19203	31	125	generates	generate	VERB
ajst-19203	31	126	the	the	DET
ajst-19203	31	127	final	final	ADJ
ajst-19203	31	128	outputs.[26	outputs.[26	NOUN
ajst-19203	31	129	]	]	X
ajst-19203	31	130	the	the	DET
ajst-19203	31	131	overall	overall	ADJ
ajst-19203	31	132	architecture	architecture	NOUN
ajst-19203	31	133	of	of	ADP
ajst-19203	31	134	our	our	PRON
ajst-19203	31	135	pipeline	pipeline	NOUN
ajst-19203	31	136	is	be	AUX
ajst-19203	31	137	shown	show	VERB
ajst-19203	31	138	in	in	ADP
ajst-19203	31	139	figure	figure	NOUN
ajst-19203	31	140	1	1	NUM
ajst-19203	31	141	.	.	PUNCT
ajst-19203	31	142	figure	figure	NOUN
ajst-19203	31	143	1	1	NUM
ajst-19203	31	144	.	.	PUNCT
ajst-19203	32	1	overall	overall	ADJ
ajst-19203	32	2	image	image	NOUN
ajst-19203	32	3	captioning	caption	VERB
ajst-19203	32	4	architecture	architecture	NOUN
ajst-19203	32	5	3.1	3.1	NUM
ajst-19203	32	6	.	.	PUNCT
ajst-19203	33	1	image	image	NOUN
ajst-19203	33	2	captioning	captioning	NOUN
ajst-19203	33	3	in	in	ADP
ajst-19203	33	4	the	the	DET
ajst-19203	33	5	image	image	NOUN
ajst-19203	33	6	captioning	captioning	NOUN
ajst-19203	33	7	step	step	NOUN
ajst-19203	33	8	,	,	PUNCT
ajst-19203	33	9	we	we	PRON
ajst-19203	33	10	employ	employ	VERB
ajst-19203	33	11	the	the	DET
ajst-19203	33	12	widely	widely	ADV
ajst-19203	33	13	-	-	PUNCT
ajst-19203	33	14	used	use	VERB
ajst-19203	33	15	encoder	encoder	NOUN
ajst-19203	33	16	-	-	PUNCT
ajst-19203	33	17	decoder	decoder	NOUN
ajst-19203	33	18	image	image	NOUN
ajst-19203	33	19	captioning	captioning	NOUN
ajst-19203	33	20	architecture	architecture	NOUN
ajst-19203	33	21	.	.	PUNCT
ajst-19203	34	1	for	for	ADP
ajst-19203	34	2	the	the	DET
ajst-19203	34	3	image	image	NOUN
ajst-19203	34	4	side	side	NOUN
ajst-19203	34	5	,	,	PUNCT
ajst-19203	34	6	we	we	PRON
ajst-19203	34	7	use	use	VERB
ajst-19203	34	8	a	a	DET
ajst-19203	34	9	pretrained	pretraine	VERB
ajst-19203	34	10	resnet-50[53	resnet-50[53	NOUN
ajst-19203	34	11	]	]	PUNCT
ajst-19203	34	12	without	without	ADP
ajst-19203	34	13	the	the	DET
ajst-19203	34	14	last	last	ADJ
ajst-19203	34	15	pooling	pooling	NOUN
ajst-19203	34	16	and	and	CCONJ
ajst-19203	34	17	linear	linear	ADJ
ajst-19203	34	18	layer	layer	NOUN
ajst-19203	34	19	as	as	ADP
ajst-19203	34	20	encoder[52	encoder[52	NOUN
ajst-19203	34	21	]	]	PUNCT
ajst-19203	34	22	and	and	CCONJ
ajst-19203	34	23	obtain	obtain	VERB
ajst-19203	34	24	a	a	DET
ajst-19203	34	25	tensor	tensor	NOUN
ajst-19203	34	26	with	with	ADP
ajst-19203	34	27	shape	shape	NOUN
ajst-19203	34	28	[	[	X
ajst-19203	34	29	batch_size	batch_size	NOUN
ajst-19203	34	30	,	,	PUNCT
ajst-19203	34	31	49	49	NUM
ajst-19203	34	32	,	,	PUNCT
ajst-19203	34	33	2048	2048	NUM
ajst-19203	34	34	]	]	PUNCT
ajst-19203	34	35	.	.	PUNCT
ajst-19203	35	1	[	[	X
ajst-19203	35	2	8]for	8]for	NUM
ajst-19203	35	3	decoder	decoder	NOUN
ajst-19203	35	4	,	,	PUNCT
ajst-19203	35	5	we	we	PRON
ajst-19203	35	6	use	use	VERB
ajst-19203	35	7	a	a	DET
ajst-19203	35	8	lstm	lstm	NOUN
ajst-19203	35	9	.	.	PUNCT
ajst-19203	36	1	to	to	PART
ajst-19203	36	2	improve	improve	VERB
ajst-19203	36	3	the	the	DET
ajst-19203	36	4	performance	performance	NOUN
ajst-19203	36	5	of	of	ADP
ajst-19203	36	6	encoder	encoder	NOUN
ajst-19203	36	7	-	-	PUNCT
ajst-19203	36	8	decoder	decoder	NOUN
ajst-19203	36	9	,	,	PUNCT
ajst-19203	36	10	we	we	PRON
ajst-19203	36	11	add	add	VERB
ajst-19203	36	12	an	an	DET
ajst-19203	36	13	bahdanau	bahdanau	ADJ
ajst-19203	36	14	attention	attention	NOUN
ajst-19203	36	15	layers[9	layers[9	PROPN
ajst-19203	36	16	]	]	PUNCT
ajst-19203	36	17	to	to	PART
ajst-19203	36	18	calculate	calculate	VERB
ajst-19203	36	19	the	the	DET
ajst-19203	36	20	attention	attention	NOUN
ajst-19203	36	21	between	between	ADP
ajst-19203	36	22	encoder	encoder	NOUN
ajst-19203	36	23	outputs	output	NOUN
ajst-19203	36	24	and	and	CCONJ
ajst-19203	36	25	initial	initial	ADJ
ajst-19203	36	26	states	state	NOUN
ajst-19203	36	27	of	of	ADP
ajst-19203	36	28	the	the	DET
ajst-19203	36	29	decoder.[5	decoder.[5	X
ajst-19203	36	30	]	]	PUNCT
ajst-19203	36	31	the	the	DET
ajst-19203	36	32	model	model	NOUN
ajst-19203	36	33	is	be	AUX
ajst-19203	36	34	trained	train	VERB
ajst-19203	36	35	in	in	ADP
ajst-19203	36	36	a	a	DET
ajst-19203	36	37	supervised	supervised	ADJ
ajst-19203	36	38	learning	learning	NOUN
ajst-19203	36	39	setting	set	VERB
ajst-19203	36	40	:	:	PUNCT
ajst-19203	36	41	we	we	PRON
ajst-19203	36	42	first	first	ADV
ajst-19203	36	43	trained	train	VERB
ajst-19203	36	44	the	the	DET
ajst-19203	36	45	parameters	parameter	NOUN
ajst-19203	36	46	of	of	ADP
ajst-19203	36	47	lstm	lstm	NOUN
ajst-19203	36	48	to	to	PART
ajst-19203	36	49	minimize	minimize	VERB
ajst-19203	36	50	the	the	DET
ajst-19203	36	51	loss	loss	NOUN
ajst-19203	36	52	between	between	ADP
ajst-19203	36	53	generated	generate	VERB
ajst-19203	36	54	sentences	sentence	NOUN
ajst-19203	36	55	and	and	CCONJ
ajst-19203	36	56	the	the	DET
ajst-19203	36	57	ground	ground	NOUN
ajst-19203	36	58	true	true	ADJ
ajst-19203	36	59	captions	caption	NOUN
ajst-19203	36	60	.	.	PUNCT
ajst-19203	37	1	in	in	ADP
ajst-19203	37	2	the	the	DET
ajst-19203	37	3	testing	testing	NOUN
ajst-19203	37	4	stage	stage	NOUN
ajst-19203	37	5	,	,	PUNCT
ajst-19203	37	6	we	we	PRON
ajst-19203	37	7	generate	generate	VERB
ajst-19203	37	8	captions	caption	NOUN
ajst-19203	37	9	using	use	VERB
ajst-19203	37	10	the	the	DET
ajst-19203	37	11	optimal	optimal	ADJ
ajst-19203	37	12	parameters	parameter	NOUN
ajst-19203	37	13	and	and	CCONJ
ajst-19203	37	14	evaluate	evaluate	VERB
ajst-19203	37	15	the	the	DET
ajst-19203	37	16	results	result	NOUN
ajst-19203	37	17	by	by	ADP
ajst-19203	37	18	human	human	NOUN
ajst-19203	37	19	.	.	PUNCT
ajst-19203	38	1	the	the	DET
ajst-19203	38	2	architecture	architecture	NOUN
ajst-19203	38	3	of	of	ADP
ajst-19203	38	4	the	the	DET
ajst-19203	38	5	image	image	NOUN
ajst-19203	38	6	captioning	captioning	NOUN
ajst-19203	38	7	module	module	NOUN
ajst-19203	38	8	is	be	AUX
ajst-19203	38	9	shown	show	VERB
ajst-19203	38	10	in	in	ADP
ajst-19203	38	11	figure	figure	NOUN
ajst-19203	38	12	2.[27	2.[27	NUM
ajst-19203	38	13	]	]	PUNCT
ajst-19203	38	14	figure	figure	NOUN
ajst-19203	38	15	2	2	NUM
ajst-19203	38	16	.	.	PUNCT
ajst-19203	38	17	image	image	NOUN
ajst-19203	38	18	captioning	captioning	NOUN
ajst-19203	38	19	module	module	NOUN
ajst-19203	38	20	286	286	NUM
ajst-19203	38	21	3.2	3.2	NUM
ajst-19203	38	22	.	.	PUNCT
ajst-19203	39	1	face	face	NOUN
ajst-19203	39	2	recognition	recognition	NOUN
ajst-19203	39	3	since	since	SCONJ
ajst-19203	39	4	our	our	PRON
ajst-19203	39	5	input	input	NOUN
ajst-19203	39	6	images	image	NOUN
ajst-19203	39	7	are	be	AUX
ajst-19203	39	8	scenes	scene	NOUN
ajst-19203	39	9	of	of	ADP
ajst-19203	39	10	behaviors	behavior	NOUN
ajst-19203	39	11	instead	instead	ADV
ajst-19203	39	12	of	of	ADP
ajst-19203	39	13	some	some	DET
ajst-19203	39	14	simple	simple	ADJ
ajst-19203	39	15	human	human	ADJ
ajst-19203	39	16	faces	face	NOUN
ajst-19203	39	17	,	,	PUNCT
ajst-19203	39	18	we	we	PRON
ajst-19203	39	19	need	need	VERB
ajst-19203	39	20	to	to	PART
ajst-19203	39	21	first	first	ADV
ajst-19203	39	22	extract	extract	VERB
ajst-19203	39	23	face	face	NOUN
ajst-19203	39	24	regions	region	NOUN
ajst-19203	39	25	before	before	ADP
ajst-19203	39	26	performing	perform	VERB
ajst-19203	39	27	face	face	NOUN
ajst-19203	39	28	classification	classification	NOUN
ajst-19203	39	29	tasks	task	NOUN
ajst-19203	39	30	.	.	PUNCT
ajst-19203	40	1	we	we	PRON
ajst-19203	40	2	follow	follow	VERB
ajst-19203	40	3	the	the	DET
ajst-19203	40	4	work	work	NOUN
ajst-19203	40	5	of	of	ADP
ajst-19203	40	6	mtcnn	mtcnn	NOUN
ajst-19203	40	7	[	[	X
ajst-19203	40	8	10	10	NUM
ajst-19203	40	9	]	]	PUNCT
ajst-19203	40	10	to	to	PART
ajst-19203	40	11	complete	complete	VERB
ajst-19203	40	12	this	this	DET
ajst-19203	40	13	task	task	NOUN
ajst-19203	40	14	:	:	PUNCT
ajst-19203	40	15	we	we	PRON
ajst-19203	40	16	first	first	ADV
ajst-19203	40	17	employ	employ	VERB
ajst-19203	40	18	the	the	DET
ajst-19203	40	19	mtcnn	mtcnn	ADJ
ajst-19203	40	20	architecture	architecture	NOUN
ajst-19203	40	21	to	to	PART
ajst-19203	40	22	draw	draw	VERB
ajst-19203	40	23	faces	face	NOUN
ajst-19203	40	24	’	'	PUNCT
ajst-19203	40	25	bounding	bounding	NOUN
ajst-19203	40	26	boxes	box	NOUN
ajst-19203	40	27	and	and	CCONJ
ajst-19203	40	28	then	then	ADV
ajst-19203	40	29	use	use	VERB
ajst-19203	40	30	an	an	DET
ajst-19203	40	31	inception_resnet_v1	inception_resnet_v1	NOUN
ajst-19203	40	32	to	to	PART
ajst-19203	40	33	obtain	obtain	VERB
ajst-19203	40	34	the	the	DET
ajst-19203	40	35	name	name	NOUN
ajst-19203	40	36	classes	class	NOUN
ajst-19203	40	37	.	.	PUNCT
ajst-19203	41	1	the	the	DET
ajst-19203	41	2	architecture	architecture	NOUN
ajst-19203	41	3	of	of	ADP
ajst-19203	41	4	our	our	PRON
ajst-19203	41	5	face	face	NOUN
ajst-19203	41	6	recognition	recognition	NOUN
ajst-19203	41	7	networks	network	NOUN
ajst-19203	41	8	are	be	AUX
ajst-19203	41	9	shown	show	VERB
ajst-19203	41	10	in	in	ADP
ajst-19203	41	11	picture	picture	NOUN
ajst-19203	41	12	3.[28	3.[28	NUM
ajst-19203	41	13	]	]	PUNCT
ajst-19203	41	14	figure	figure	NOUN
ajst-19203	41	15	3	3	NUM
ajst-19203	41	16	.	.	PUNCT
ajst-19203	41	17	mtcnn	mtcnn	ADJ
ajst-19203	41	18	architecture	architecture	NOUN
ajst-19203	41	19	simply	simply	ADV
ajst-19203	41	20	speaking	speak	VERB
ajst-19203	41	21	,	,	PUNCT
ajst-19203	41	22	mtcnn	mtcnn	PROPN
ajst-19203	41	23	are	be	AUX
ajst-19203	41	24	three	three	NUM
ajst-19203	41	25	cascaded	cascade	VERB
ajst-19203	41	26	convolutional	convolutional	ADJ
ajst-19203	41	27	networks	network	NOUN
ajst-19203	41	28	:	:	PUNCT
ajst-19203	41	29	a	a	DET
ajst-19203	41	30	proposal	proposal	NOUN
ajst-19203	41	31	network	network	NOUN
ajst-19203	41	32	(	(	PUNCT
ajst-19203	41	33	p	p	NOUN
ajst-19203	41	34	-	-	PUNCT
ajst-19203	41	35	net	net	NOUN
ajst-19203	41	36	)	)	PUNCT
ajst-19203	41	37	,	,	PUNCT
ajst-19203	41	38	a	a	DET
ajst-19203	41	39	refinement	refinement	NOUN
ajst-19203	41	40	network	network	NOUN
ajst-19203	41	41	(	(	PUNCT
ajst-19203	41	42	r	r	NOUN
ajst-19203	41	43	-	-	PUNCT
ajst-19203	41	44	net	net	NOUN
ajst-19203	41	45	)	)	PUNCT
ajst-19203	41	46	and	and	CCONJ
ajst-19203	41	47	an	an	DET
ajst-19203	41	48	output	output	NOUN
ajst-19203	41	49	network	network	NOUN
ajst-19203	41	50	(	(	PUNCT
ajst-19203	41	51	o	o	NOUN
ajst-19203	41	52	-	-	NOUN
ajst-19203	41	53	net	net	ADJ
ajst-19203	41	54	)	)	PUNCT
ajst-19203	41	55	.	.	PUNCT
ajst-19203	42	1	firstly	firstly	ADV
ajst-19203	42	2	,	,	PUNCT
ajst-19203	42	3	candidate	candidate	NOUN
ajst-19203	42	4	windows	window	NOUN
ajst-19203	42	5	are	be	AUX
ajst-19203	42	6	produced	produce	VERB
ajst-19203	42	7	through	through	ADP
ajst-19203	42	8	a	a	DET
ajst-19203	42	9	fast	fast	ADJ
ajst-19203	42	10	p	p	NOUN
ajst-19203	42	11	-	-	PUNCT
ajst-19203	42	12	net	net	NOUN
ajst-19203	42	13	.	.	PUNCT
ajst-19203	43	1	after	after	ADP
ajst-19203	43	2	that	that	PRON
ajst-19203	43	3	,	,	PUNCT
ajst-19203	43	4	we	we	PRON
ajst-19203	43	5	refine	refine	VERB
ajst-19203	43	6	these	these	DET
ajst-19203	43	7	candidates	candidate	NOUN
ajst-19203	43	8	in	in	ADP
ajst-19203	43	9	the	the	DET
ajst-19203	43	10	next	next	ADJ
ajst-19203	43	11	stage	stage	NOUN
ajst-19203	43	12	through	through	ADP
ajst-19203	43	13	a	a	DET
ajst-19203	43	14	r	r	NOUN
ajst-19203	43	15	-	-	PUNCT
ajst-19203	43	16	net	net	NOUN
ajst-19203	43	17	.	.	PUNCT
ajst-19203	44	1	in	in	ADP
ajst-19203	44	2	the	the	DET
ajst-19203	44	3	final	final	ADJ
ajst-19203	44	4	stage	stage	NOUN
ajst-19203	44	5	,	,	PUNCT
ajst-19203	44	6	the	the	PRON
ajst-19203	44	7	o	o	NOUN
ajst-19203	44	8	-	-	ADJ
ajst-19203	44	9	net	net	ADJ
ajst-19203	44	10	produces	produce	VERB
ajst-19203	44	11	final	final	ADJ
ajst-19203	44	12	bounding	bounding	NOUN
ajst-19203	44	13	box	box	NOUN
ajst-19203	44	14	and	and	CCONJ
ajst-19203	44	15	facial	facial	ADJ
ajst-19203	44	16	landmarks	landmark	NOUN
ajst-19203	44	17	position	position	NOUN
ajst-19203	44	18	.	.	PUNCT
ajst-19203	45	1	3.3	3.3	NUM
ajst-19203	45	2	.	.	PUNCT
ajst-19203	46	1	np	np	INTJ
ajst-19203	46	2	chunk	chunk	PROPN
ajst-19203	46	3	matching	matching	NOUN
ajst-19203	46	4	figure	figure	NOUN
ajst-19203	46	5	4	4	NUM
ajst-19203	46	6	.	.	PUNCT
ajst-19203	47	1	np	np	PROPN
ajst-19203	47	2	chunk	chunk	PROPN
ajst-19203	47	3	matching	match	VERB
ajst-19203	47	4	the	the	DET
ajst-19203	47	5	last	last	ADJ
ajst-19203	47	6	step	step	NOUN
ajst-19203	47	7	is	be	AUX
ajst-19203	47	8	the	the	DET
ajst-19203	47	9	alignment	alignment	NOUN
ajst-19203	47	10	between	between	ADP
ajst-19203	47	11	celebrity	celebrity	NOUN
ajst-19203	47	12	names	name	NOUN
ajst-19203	47	13	and	and	CCONJ
ajst-19203	47	14	the	the	DET
ajst-19203	47	15	np	np	PROPN
ajst-19203	47	16	chunk	chunk	NOUN
ajst-19203	47	17	in	in	ADP
ajst-19203	47	18	the	the	DET
ajst-19203	47	19	generated	generate	VERB
ajst-19203	47	20	captions	caption	NOUN
ajst-19203	47	21	.	.	PUNCT
ajst-19203	48	1	we	we	PRON
ajst-19203	48	2	parse	parse	VERB
ajst-19203	48	3	the	the	DET
ajst-19203	48	4	captions	caption	NOUN
ajst-19203	48	5	with	with	ADP
ajst-19203	48	6	nltk	nltk	PROPN
ajst-19203	48	7	and	and	CCONJ
ajst-19203	48	8	spacy	spacy	NOUN
ajst-19203	48	9	packages	package	NOUN
ajst-19203	48	10	and	and	CCONJ
ajst-19203	48	11	obtain	obtain	VERB
ajst-19203	48	12	the	the	DET
ajst-19203	48	13	people	people	NOUN
ajst-19203	48	14	-	-	PUNCT
ajst-19203	48	15	related	relate	VERB
ajst-19203	48	16	np	np	ADP
ajst-19203	48	17	chunks	chunk	NOUN
ajst-19203	48	18	.	.	PUNCT
ajst-19203	49	1	in	in	ADP
ajst-19203	49	2	general	general	ADJ
ajst-19203	49	3	(	(	PUNCT
ajst-19203	49	4	sequence	sequence	NOUN
ajst-19203	49	5	exchangeable	exchangeable	ADJ
ajst-19203	49	6	)	)	PUNCT
ajst-19203	49	7	cases	case	NOUN
ajst-19203	49	8	,	,	PUNCT
ajst-19203	49	9	sequence	sequence	NOUN
ajst-19203	49	10	of	of	ADP
ajst-19203	49	11	the	the	DET
ajst-19203	49	12	chunks	chunk	NOUN
ajst-19203	49	13	are	be	AUX
ajst-19203	49	14	unimportant[55	unimportant[55	NOUN
ajst-19203	49	15	]	]	PUNCT
ajst-19203	50	1	so	so	CCONJ
ajst-19203	50	2	we	we	PRON
ajst-19203	50	3	simply	simply	ADV
ajst-19203	50	4	sequentially	sequentially	ADV
ajst-19203	50	5	align	align	VERB
ajst-19203	50	6	names	name	NOUN
ajst-19203	50	7	with	with	ADP
ajst-19203	50	8	chunks	chunk	NOUN
ajst-19203	50	9	.	.	PUNCT
ajst-19203	51	1	using	use	VERB
ajst-19203	51	2	regular	regular	ADJ
ajst-19203	51	3	expression	expression	NOUN
ajst-19203	51	4	,	,	PUNCT
ajst-19203	51	5	we	we	PRON
ajst-19203	51	6	replace	replace	VERB
ajst-19203	51	7	np	np	INTJ
ajst-19203	51	8	chunks	chunk	NOUN
ajst-19203	51	9	like	like	ADP
ajst-19203	51	10	"	"	PUNCT
ajst-19203	51	11	a	a	DET
ajst-19203	51	12	man	man	NOUN
ajst-19203	51	13	"	"	PUNCT
ajst-19203	51	14	,	,	PUNCT
ajst-19203	51	15	"	"	PUNCT
ajst-19203	51	16	the	the	DET
ajst-19203	51	17	woman	woman	NOUN
ajst-19203	51	18	"	"	PUNCT
ajst-19203	51	19	,	,	PUNCT
ajst-19203	51	20	"	"	PUNCT
ajst-19203	51	21	two	two	NUM
ajst-19203	51	22	young	young	ADJ
ajst-19203	51	23	boys	boy	NOUN
ajst-19203	51	24	"	"	PUNCT
ajst-19203	51	25	with	with	ADP
ajst-19203	51	26	celebrity	celebrity	NOUN
ajst-19203	51	27	names	name	NOUN
ajst-19203	51	28	we	we	PRON
ajst-19203	51	29	derived	derive	VERB
ajst-19203	51	30	in	in	ADP
ajst-19203	51	31	the	the	DET
ajst-19203	51	32	face	face	NOUN
ajst-19203	51	33	recognition	recognition	NOUN
ajst-19203	51	34	step	step	NOUN
ajst-19203	51	35	.	.	PUNCT
ajst-19203	52	1	some	some	DET
ajst-19203	52	2	corner	corner	NOUN
ajst-19203	52	3	cases	case	NOUN
ajst-19203	52	4	need	need	VERB
ajst-19203	52	5	to	to	PART
ajst-19203	52	6	be	be	AUX
ajst-19203	52	7	carefully	carefully	ADV
ajst-19203	52	8	coped	cope	VERB
ajst-19203	52	9	with	with	ADP
ajst-19203	52	10	in	in	ADP
ajst-19203	52	11	this	this	DET
ajst-19203	52	12	step	step	NOUN
ajst-19203	52	13	,	,	PUNCT
ajst-19203	52	14	and	and	CCONJ
ajst-19203	52	15	in	in	ADP
ajst-19203	52	16	some	some	DET
ajst-19203	52	17	cases	case	NOUN
ajst-19203	52	18	it	it	PRON
ajst-19203	52	19	may	may	AUX
ajst-19203	52	20	be	be	AUX
ajst-19203	52	21	hard	hard	ADJ
ajst-19203	52	22	to	to	PART
ajst-19203	52	23	match	match	VERB
ajst-19203	52	24	(	(	PUNCT
ajst-19203	52	25	or	or	CCONJ
ajst-19203	52	26	need	need	VERB
ajst-19203	52	27	more	more	ADV
ajst-19203	52	28	complicated	complicated	ADJ
ajst-19203	52	29	rules).[58	rules).[58	NOUN
ajst-19203	52	30	]	]	PUNCT
ajst-19203	52	31	for	for	ADP
ajst-19203	52	32	detailed	detailed	ADJ
ajst-19203	52	33	implementation	implementation	NOUN
ajst-19203	52	34	,	,	PUNCT
ajst-19203	52	35	please	please	INTJ
ajst-19203	52	36	see	see	VERB
ajst-19203	52	37	refer	refer	NOUN
ajst-19203	52	38	to	to	ADP
ajst-19203	52	39	our	our	PRON
ajst-19203	52	40	code	code	NOUN
ajst-19203	52	41	.	.	PUNCT
ajst-19203	53	1	in	in	ADP
ajst-19203	53	2	nonexchangeable	nonexchangeable	ADJ
ajst-19203	53	3	cases	case	NOUN
ajst-19203	53	4	,	,	PUNCT
ajst-19203	53	5	problems	problem	NOUN
ajst-19203	53	6	are	be	AUX
ajst-19203	53	7	much	much	ADV
ajst-19203	53	8	more	more	ADV
ajst-19203	53	9	challenging	challenging	ADJ
ajst-19203	53	10	.	.	PUNCT
ajst-19203	54	1	we	we	PRON
ajst-19203	54	2	intend	intend	VERB
ajst-19203	54	3	to	to	PART
ajst-19203	54	4	align	align	VERB
ajst-19203	54	5	the	the	DET
ajst-19203	54	6	names	name	NOUN
ajst-19203	54	7	and	and	CCONJ
ajst-19203	54	8	faces	face	NOUN
ajst-19203	54	9	by	by	ADP
ajst-19203	54	10	considering	consider	VERB
ajst-19203	54	11	the	the	DET
ajst-19203	54	12	position	position	NOUN
ajst-19203	54	13	relation	relation	NOUN
ajst-19203	54	14	between	between	ADP
ajst-19203	54	15	image2text	image2text	NOUN
ajst-19203	54	16	attention	attention	NOUN
ajst-19203	54	17	matrix	matrix	NOUN
ajst-19203	54	18	and	and	CCONJ
ajst-19203	54	19	recognized	recognize	VERB
ajst-19203	54	20	celebrity	celebrity	NOUN
ajst-19203	54	21	faces	face	NOUN
ajst-19203	54	22	.	.	PUNCT
ajst-19203	55	1	ideally	ideally	ADV
ajst-19203	55	2	,	,	PUNCT
ajst-19203	55	3	the	the	DET
ajst-19203	55	4	pixel	pixel	NOUN
ajst-19203	55	5	related	relate	VERB
ajst-19203	55	6	to	to	ADP
ajst-19203	55	7	faces	face	NOUN
ajst-19203	55	8	will	will	AUX
ajst-19203	55	9	have	have	VERB
ajst-19203	55	10	large	large	ADJ
ajst-19203	55	11	attention	attention	NOUN
ajst-19203	55	12	weight	weight	NOUN
ajst-19203	55	13	,	,	PUNCT
ajst-19203	55	14	whose	whose	DET
ajst-19203	55	15	position	position	NOUN
ajst-19203	55	16	will	will	AUX
ajst-19203	55	17	be	be	AUX
ajst-19203	55	18	helpful	helpful	ADJ
ajst-19203	55	19	for	for	ADP
ajst-19203	55	20	the	the	DET
ajst-19203	55	21	alignment	alignment	NOUN
ajst-19203	55	22	.	.	PUNCT
ajst-19203	56	1	however	however	ADV
ajst-19203	56	2	,	,	PUNCT
ajst-19203	56	3	we	we	PRON
ajst-19203	56	4	find	find	VERB
ajst-19203	56	5	that	that	SCONJ
ajst-19203	56	6	the	the	DET
ajst-19203	56	7	sequence	sequence	NOUN
ajst-19203	56	8	of	of	ADP
ajst-19203	56	9	np	np	INTJ
ajst-19203	56	10	chunks	chunk	NOUN
ajst-19203	56	11	in	in	ADP
ajst-19203	56	12	the	the	DET
ajst-19203	56	13	datasets[59	datasets[59	PROPN
ajst-19203	56	14	]	]	X
ajst-19203	56	15	are	be	AUX
ajst-19203	56	16	usually	usually	ADV
ajst-19203	56	17	not	not	PART
ajst-19203	56	18	in	in	ADP
ajst-19203	56	19	accordance	accordance	NOUN
ajst-19203	56	20	with	with	ADP
ajst-19203	56	21	the	the	DET
ajst-19203	56	22	position	position	NOUN
ajst-19203	56	23	of	of	ADP
ajst-19203	56	24	faces	face	NOUN
ajst-19203	56	25	.	.	PUNCT
ajst-19203	57	1	what	what	PRON
ajst-19203	57	2	's	be	AUX
ajst-19203	57	3	more	more	ADJ
ajst-19203	57	4	,	,	PUNCT
ajst-19203	57	5	the	the	DET
ajst-19203	57	6	attention	attention	NOUN
ajst-19203	57	7	we	we	PRON
ajst-19203	57	8	use	use	VERB
ajst-19203	57	9	in	in	ADP
ajst-19203	57	10	out	out	ADJ
ajst-19203	57	11	experiment	experiment	NOUN
ajst-19203	57	12	are	be	AUX
ajst-19203	57	13	not	not	PART
ajst-19203	57	14	precise	precise	ADJ
ajst-19203	57	15	enough.[48	enough.[48	NOUN
ajst-19203	57	16	]	]	PUNCT
ajst-19203	57	17	due	due	ADP
ajst-19203	57	18	to	to	ADP
ajst-19203	57	19	those	those	DET
ajst-19203	57	20	limitations	limitation	NOUN
ajst-19203	57	21	,	,	PUNCT
ajst-19203	57	22	our	our	PRON
ajst-19203	57	23	method	method	NOUN
ajst-19203	57	24	is	be	AUX
ajst-19203	57	25	limited	limit	VERB
ajst-19203	57	26	to	to	ADP
ajst-19203	57	27	solving	solve	VERB
ajst-19203	57	28	exchangeable	exchangeable	ADJ
ajst-19203	57	29	cases	case	NOUN
ajst-19203	57	30	,	,	PUNCT
ajst-19203	57	31	unless	unless	SCONJ
ajst-19203	57	32	we	we	PRON
ajst-19203	57	33	alter	alter	VERB
ajst-19203	57	34	the	the	DET
ajst-19203	57	35	architecture	architecture	NOUN
ajst-19203	57	36	.	.	PUNCT
ajst-19203	58	1	4	4	X
ajst-19203	58	2	.	.	X
ajst-19203	58	3	experiment	experiment	NOUN
ajst-19203	58	4	we	we	PRON
ajst-19203	58	5	use	use	VERB
ajst-19203	58	6	the	the	DET
ajst-19203	58	7	following	follow	VERB
ajst-19203	58	8	datasets	dataset	NOUN
ajst-19203	58	9	in	in	ADP
ajst-19203	58	10	our	our	PRON
ajst-19203	58	11	project	project	NOUN
ajst-19203	58	12	(	(	PUNCT
ajst-19203	58	13	due	due	ADP
ajst-19203	58	14	to	to	ADP
ajst-19203	58	15	stricter	strict	ADJ
ajst-19203	58	16	policy	policy	NOUN
ajst-19203	58	17	on	on	ADP
ajst-19203	58	18	privacy	privacy	NOUN
ajst-19203	58	19	launched	launch	VERB
ajst-19203	58	20	in	in	ADP
ajst-19203	58	21	recent	recent	ADJ
ajst-19203	58	22	years	year	NOUN
ajst-19203	58	23	,	,	PUNCT
ajst-19203	58	24	we	we	PRON
ajst-19203	58	25	do	do	AUX
ajst-19203	58	26	not	not	PART
ajst-19203	58	27	have	have	VERB
ajst-19203	58	28	access	access	NOUN
ajst-19203	58	29	to	to	ADP
ajst-19203	58	30	larger	large	ADJ
ajst-19203	58	31	datasets	dataset	NOUN
ajst-19203	58	32	.	.	PUNCT
ajst-19203	59	1	flickr	flickr	PROPN
ajst-19203	59	2	8k/30k	8k/30k	PROPN
ajst-19203	59	3	:	:	PUNCT
ajst-19203	59	4	contains	contain	VERB
ajst-19203	59	5	about	about	ADV
ajst-19203	59	6	8,000	8,000	NUM
ajst-19203	59	7	images	image	NOUN
ajst-19203	59	8	collected	collect	VERB
ajst-19203	59	9	from	from	ADP
ajst-19203	59	10	flickr	flickr	PROPN
ajst-19203	59	11	,	,	PUNCT
ajst-19203	59	12	together	together	ADV
ajst-19203	59	13	with	with	ADP
ajst-19203	59	14	5	5	NUM
ajst-19203	59	15	reference	reference	NOUN
ajst-19203	59	16	sentences	sentence	NOUN
ajst-19203	59	17	for	for	ADP
ajst-19203	59	18	each	each	DET
ajst-19203	59	19	image	image	NOUN
ajst-19203	59	20	provided	provide	VERB
ajst-19203	59	21	by	by	ADP
ajst-19203	59	22	human	human	ADJ
ajst-19203	59	23	annotators	annotator	NOUN
ajst-19203	59	24	.	.	PUNCT
ajst-19203	60	1	coco	coco	PROPN
ajst-19203	60	2	captions	caption	NOUN
ajst-19203	60	3	:	:	PUNCT
ajst-19203	60	4	contains	contain	VERB
ajst-19203	60	5	over	over	ADP
ajst-19203	60	6	one	one	NUM
ajst-19203	60	7	and	and	CCONJ
ajst-19203	60	8	a	a	DET
ajst-19203	60	9	half	half	NOUN
ajst-19203	60	10	million	million	NUM
ajst-19203	60	11	captions	caption	NOUN
ajst-19203	60	12	describing	describe	VERB
ajst-19203	60	13	over	over	ADP
ajst-19203	60	14	330,000	330,000	NUM
ajst-19203	60	15	images	image	NOUN
ajst-19203	60	16	.	.	PUNCT
ajst-19203	61	1	for	for	ADP
ajst-19203	61	2	the	the	DET
ajst-19203	61	3	training	training	NOUN
ajst-19203	61	4	and	and	CCONJ
ajst-19203	61	5	validation	validation	NOUN
ajst-19203	61	6	images	image	NOUN
ajst-19203	61	7	,	,	PUNCT
ajst-19203	61	8	five	five	NUM
ajst-19203	61	9	independent	independent	ADJ
ajst-19203	61	10	human	human	ADJ
ajst-19203	61	11	generated	generate	VERB
ajst-19203	61	12	captions	caption	NOUN
ajst-19203	61	13	are	be	AUX
ajst-19203	61	14	be	be	AUX
ajst-19203	61	15	provided	provide	VERB
ajst-19203	61	16	for	for	ADP
ajst-19203	61	17	each	each	DET
ajst-19203	61	18	image	image	NOUN
ajst-19203	61	19	.	.	PUNCT
ajst-19203	62	1	we	we	PRON
ajst-19203	62	2	carry	carry	VERB
ajst-19203	62	3	out	out	ADP
ajst-19203	62	4	supervised	supervised	ADJ
ajst-19203	62	5	learning	learning	NOUN
ajst-19203	62	6	using	use	VERB
ajst-19203	62	7	flickr	flickr	PROPN
ajst-19203	62	8	8k/30k	8k/30k	PROPN
ajst-19203	62	9	or	or	CCONJ
ajst-19203	62	10	vggface2	vggface2	NOUN
ajst-19203	62	11	for	for	ADP
ajst-19203	62	12	captioning	captioning	NOUN
ajst-19203	62	13	.	.	PUNCT
ajst-19203	63	1	overally	overally	ADV
ajst-19203	63	2	speaking	speak	VERB
ajst-19203	63	3	,	,	PUNCT
ajst-19203	63	4	our	our	PRON
ajst-19203	63	5	pipeline	pipeline	NOUN
ajst-19203	63	6	can	can	AUX
ajst-19203	63	7	obtain	obtain	VERB
ajst-19203	63	8	a	a	DET
ajst-19203	63	9	very	very	ADV
ajst-19203	63	10	good	good	ADJ
ajst-19203	63	11	results	result	NOUN
ajst-19203	63	12	with	with	ADP
ajst-19203	63	13	an	an	DET
ajst-19203	63	14	accuracy	accuracy	NOUN
ajst-19203	63	15	performance	performance	NOUN
ajst-19203	63	16	over	over	ADP
ajst-19203	63	17	287	287	NUM
ajst-19203	63	18	90	90	NUM
ajst-19203	63	19	%	%	NOUN
ajst-19203	63	20	,	,	PUNCT
ajst-19203	63	21	which	which	PRON
ajst-19203	63	22	demonstrates	demonstrate	VERB
ajst-19203	63	23	the	the	DET
ajst-19203	63	24	power	power	NOUN
ajst-19203	63	25	of	of	ADP
ajst-19203	63	26	combining	combine	VERB
ajst-19203	63	27	the	the	DET
ajst-19203	63	28	deep	deep	ADJ
ajst-19203	63	29	neural	neural	ADJ
ajst-19203	63	30	networks	network	NOUN
ajst-19203	63	31	and	and	CCONJ
ajst-19203	63	32	our	our	PRON
ajst-19203	63	33	face	face	NOUN
ajst-19203	63	34	recognition	recognition	NOUN
ajst-19203	63	35	modules.[47	modules.[47	NOUN
ajst-19203	63	36	]	]	PUNCT
ajst-19203	63	37	5	5	NUM
ajst-19203	63	38	.	.	X
ajst-19203	63	39	conclusion	conclusion	NOUN
ajst-19203	63	40	and	and	CCONJ
ajst-19203	63	41	discussion	discussion	NOUN
ajst-19203	63	42	5.1	5.1	NUM
ajst-19203	63	43	.	.	PUNCT
ajst-19203	64	1	discussion	discussion	NOUN
ajst-19203	64	2	while	while	SCONJ
ajst-19203	64	3	our	our	PRON
ajst-19203	64	4	pipeline	pipeline	NOUN
ajst-19203	64	5	can	can	AUX
ajst-19203	64	6	solve	solve	VERB
ajst-19203	64	7	the	the	DET
ajst-19203	64	8	captioning	captioning	NOUN
ajst-19203	64	9	problems	problem	NOUN
ajst-19203	64	10	in	in	ADP
ajst-19203	64	11	many	many	ADJ
ajst-19203	64	12	cases	case	NOUN
ajst-19203	64	13	,	,	PUNCT
ajst-19203	64	14	there	there	PRON
ajst-19203	64	15	are	be	VERB
ajst-19203	64	16	some	some	DET
ajst-19203	64	17	cases	case	NOUN
ajst-19203	64	18	we	we	PRON
ajst-19203	64	19	fails	fail	VERB
ajst-19203	64	20	.	.	PUNCT
ajst-19203	65	1	the	the	DET
ajst-19203	65	2	limitations	limitation	NOUN
ajst-19203	65	3	of	of	ADP
ajst-19203	65	4	our	our	PRON
ajst-19203	65	5	method	method	NOUN
ajst-19203	65	6	are	be	AUX
ajst-19203	65	7	as	as	SCONJ
ajst-19203	65	8	follows	follow	VERB
ajst-19203	65	9	:	:	PUNCT
ajst-19203	65	10	mediocre	mediocre	ADJ
ajst-19203	65	11	generation	generation	NOUN
ajst-19203	65	12	performance	performance	NOUN
ajst-19203	65	13	:	:	PUNCT
ajst-19203	65	14	the	the	DET
ajst-19203	65	15	generation	generation	NOUN
ajst-19203	65	16	performance	performance	NOUN
ajst-19203	65	17	is	be	AUX
ajst-19203	65	18	not	not	PART
ajst-19203	65	19	exceedingly	exceedingly	ADV
ajst-19203	65	20	good	good	ADJ
ajst-19203	65	21	because	because	SCONJ
ajst-19203	65	22	we	we	PRON
ajst-19203	65	23	use	use	VERB
ajst-19203	65	24	a	a	DET
ajst-19203	65	25	rather	rather	ADV
ajst-19203	65	26	small	small	ADJ
ajst-19203	65	27	encoder	encoder	NOUN
ajst-19203	65	28	decoder	decoder	NOUN
ajst-19203	65	29	architecture	architecture	NOUN
ajst-19203	65	30	in	in	ADP
ajst-19203	65	31	our	our	PRON
ajst-19203	65	32	model	model	NOUN
ajst-19203	65	33	.	.	PUNCT
ajst-19203	66	1	in	in	ADP
ajst-19203	66	2	addition	addition	NOUN
ajst-19203	66	3	,	,	PUNCT
ajst-19203	66	4	we	we	PRON
ajst-19203	66	5	employs	employ	VERB
ajst-19203	66	6	30,000	30,000	NUM
ajst-19203	66	7	images	image	NOUN
ajst-19203	66	8	together	together	ADV
ajst-19203	66	9	with	with	ADP
ajst-19203	66	10	150,000	150,000	NUM
ajst-19203	66	11	captions	caption	NOUN
ajst-19203	66	12	in	in	ADP
ajst-19203	66	13	the	the	DET
ajst-19203	66	14	training	training	NOUN
ajst-19203	66	15	process.[44	process.[44	NOUN
ajst-19203	66	16	]	]	PUNCT
ajst-19203	66	17	comparing	compare	VERB
ajst-19203	66	18	to	to	ADP
ajst-19203	66	19	the	the	DET
ajst-19203	66	20	massive	massive	ADJ
ajst-19203	66	21	datasets	dataset	NOUN
ajst-19203	66	22	used	use	VERB
ajst-19203	66	23	in	in	ADP
ajst-19203	66	24	state	state	NOUN
ajst-19203	66	25	-	-	PUNCT
ajst-19203	66	26	of	of	ADP
ajst-19203	66	27	-	-	PUNCT
ajst-19203	66	28	the	the	DET
ajst-19203	66	29	-	-	PUNCT
ajst-19203	66	30	art	art	NOUN
ajst-19203	66	31	language	language	NOUN
ajst-19203	66	32	and	and	CCONJ
ajst-19203	66	33	computer	computer	NOUN
ajst-19203	66	34	vision	vision	NOUN
ajst-19203	66	35	models	model	NOUN
ajst-19203	66	36	,	,	PUNCT
ajst-19203	66	37	this	this	PRON
ajst-19203	66	38	is	be	AUX
ajst-19203	66	39	still	still	ADV
ajst-19203	66	40	a	a	DET
ajst-19203	66	41	rather	rather	ADV
ajst-19203	66	42	standard	standard	ADJ
ajst-19203	66	43	number	number	NOUN
ajst-19203	66	44	.	.	PUNCT
ajst-19203	67	1	as	as	SCONJ
ajst-19203	67	2	our	our	PRON
ajst-19203	67	3	image	image	NOUN
ajst-19203	67	4	and	and	CCONJ
ajst-19203	67	5	caption	caption	NOUN
ajst-19203	67	6	data	datum	NOUN
ajst-19203	67	7	directly	directly	ADV
ajst-19203	67	8	decides	decide	VERB
ajst-19203	67	9	the	the	DET
ajst-19203	67	10	volume	volume	NOUN
ajst-19203	67	11	of	of	ADP
ajst-19203	67	12	the	the	DET
ajst-19203	67	13	vocabulary	vocabulary	PROPN
ajst-19203	67	14	dictionary	dictionary	PROPN
ajst-19203	67	15	,	,	PUNCT
ajst-19203	67	16	different	different	ADJ
ajst-19203	67	17	sentences	sentence	NOUN
ajst-19203	67	18	patterns	pattern	NOUN
ajst-19203	67	19	and	and	CCONJ
ajst-19203	67	20	object	object	VERB
ajst-19203	67	21	types	type	NOUN
ajst-19203	67	22	,	,	PUNCT
ajst-19203	67	23	the	the	DET
ajst-19203	67	24	poor	poor	ADJ
ajst-19203	67	25	performance	performance	NOUN
ajst-19203	67	26	obtained	obtain	VERB
ajst-19203	67	27	by	by	ADP
ajst-19203	67	28	limited	limited	ADJ
ajst-19203	67	29	datasets	dataset	NOUN
ajst-19203	67	30	is	be	AUX
ajst-19203	67	31	within	within	ADP
ajst-19203	67	32	our	our	PRON
ajst-19203	67	33	expectation	expectation	NOUN
ajst-19203	67	34	.	.	PUNCT
ajst-19203	68	1	in	in	ADP
ajst-19203	68	2	the	the	DET
ajst-19203	68	3	future	future	NOUN
ajst-19203	68	4	,	,	PUNCT
ajst-19203	68	5	we	we	PRON
ajst-19203	68	6	may	may	AUX
ajst-19203	68	7	attempt	attempt	VERB
ajst-19203	68	8	to	to	PART
ajst-19203	68	9	increase	increase	VERB
ajst-19203	68	10	the	the	DET
ajst-19203	68	11	dataset	dataset	NOUN
ajst-19203	68	12	size[19	size[19	NOUN
ajst-19203	68	13	]	]	PUNCT
ajst-19203	68	14	and	and	CCONJ
ajst-19203	68	15	use	use	VERB
ajst-19203	68	16	more	more	ADV
ajst-19203	68	17	powerful	powerful	ADJ
ajst-19203	68	18	encoder	encoder	NOUN
ajst-19203	68	19	-	-	PUNCT
ajst-19203	68	20	decoder	decoder	NOUN
ajst-19203	68	21	architectures	architecture	NOUN
ajst-19203	68	22	(	(	PUNCT
ajst-19203	68	23	like	like	ADP
ajst-19203	68	24	using	use	VERB
ajst-19203	68	25	pretrained	pretraine	VERB
ajst-19203	68	26	bert	bert	NOUN
ajst-19203	68	27	for	for	ADP
ajst-19203	68	28	the	the	DET
ajst-19203	68	29	captioning	captioning	NOUN
ajst-19203	68	30	part)[17	part)[17	VERB
ajst-19203	68	31	]	]	PUNCT
ajst-19203	68	32	inaccurate	inaccurate	ADJ
ajst-19203	68	33	np	np	PROPN
ajst-19203	68	34	chunk	chunk	NOUN
ajst-19203	68	35	matching	matching	NOUN
ajst-19203	68	36	:	:	PUNCT
ajst-19203	68	37	as	as	SCONJ
ajst-19203	68	38	we	we	PRON
ajst-19203	68	39	mentioned	mention	VERB
ajst-19203	68	40	in	in	ADP
ajst-19203	68	41	section	section	NOUN
ajst-19203	68	42	4	4	NUM
ajst-19203	68	43	,	,	PUNCT
ajst-19203	68	44	at	at	ADP
ajst-19203	68	45	present	present	ADJ
ajst-19203	68	46	our	our	PRON
ajst-19203	68	47	method	method	NOUN
ajst-19203	68	48	are	be	AUX
ajst-19203	68	49	not	not	PART
ajst-19203	68	50	capable	capable	ADJ
ajst-19203	68	51	to	to	PART
ajst-19203	68	52	deal	deal	VERB
ajst-19203	68	53	with	with	ADP
ajst-19203	68	54	immutable	immutable	ADJ
ajst-19203	68	55	type	type	NOUN
ajst-19203	68	56	celebrity	celebrity	NOUN
ajst-19203	68	57	image	image	NOUN
ajst-19203	68	58	captioning	captioning	NOUN
ajst-19203	68	59	problems	problem	NOUN
ajst-19203	68	60	due	due	ADJ
ajst-19203	68	61	to	to	ADP
ajst-19203	68	62	the	the	DET
ajst-19203	68	63	resolution	resolution	NOUN
ajst-19203	68	64	of	of	ADP
ajst-19203	68	65	attention	attention	NOUN
ajst-19203	68	66	we	we	PRON
ajst-19203	68	67	uses	use	VERB
ajst-19203	68	68	and	and	CCONJ
ajst-19203	68	69	the	the	DET
ajst-19203	68	70	inaccurate	inaccurate	ADJ
ajst-19203	68	71	word	word	NOUN
ajst-19203	68	72	sequence	sequence	NOUN
ajst-19203	68	73	of	of	ADP
ajst-19203	68	74	flickr	flickr	PROPN
ajst-19203	68	75	datasets.[45	datasets.[45	PROPN
ajst-19203	68	76	]	]	PUNCT
ajst-19203	68	77	and	and	CCONJ
ajst-19203	68	78	in	in	ADP
ajst-19203	68	79	some	some	DET
ajst-19203	68	80	cases	case	NOUN
ajst-19203	68	81	,	,	PUNCT
ajst-19203	68	82	we	we	PRON
ajst-19203	68	83	fail	fail	VERB
ajst-19203	68	84	to	to	PART
ajst-19203	68	85	perform	perform	VERB
ajst-19203	68	86	the	the	DET
ajst-19203	68	87	matching	matching	NOUN
ajst-19203	68	88	.	.	PUNCT
ajst-19203	69	1	there	there	PRON
ajst-19203	69	2	are	be	VERB
ajst-19203	69	3	potential	potential	ADJ
ajst-19203	69	4	solutions	solution	NOUN
ajst-19203	69	5	.	.	PUNCT
ajst-19203	70	1	1	1	X
ajst-19203	70	2	.	.	X
ajst-19203	70	3	use	use	VERB
ajst-19203	70	4	more	more	ADV
ajst-19203	70	5	sophisticated	sophisticated	ADJ
ajst-19203	70	6	multi	multi	ADJ
ajst-19203	70	7	-	-	ADJ
ajst-19203	70	8	modality	modality	ADJ
ajst-19203	70	9	approach	approach	NOUN
ajst-19203	70	10	.	.	PUNCT
ajst-19203	71	1	particularly	particularly	ADV
ajst-19203	71	2	there	there	PRON
ajst-19203	71	3	are	be	VERB
ajst-19203	71	4	much	much	ADJ
ajst-19203	71	5	research	research	NOUN
ajst-19203	71	6	like	like	ADP
ajst-19203	71	7	clip	clip	NOUN
ajst-19203	71	8	that	that	PRON
ajst-19203	71	9	proposes	propose	VERB
ajst-19203	71	10	to	to	PART
ajst-19203	71	11	connect	connect	VERB
ajst-19203	71	12	texts	text	NOUN
ajst-19203	71	13	with	with	ADP
ajst-19203	71	14	images	image	NOUN
ajst-19203	71	15	.	.	PUNCT
ajst-19203	72	1	2	2	X
ajst-19203	72	2	.	.	X
ajst-19203	72	3	improve	improve	VERB
ajst-19203	72	4	the	the	DET
ajst-19203	72	5	quality	quality	NOUN
ajst-19203	72	6	of	of	ADP
ajst-19203	72	7	the	the	DET
ajst-19203	72	8	datasets	dataset	NOUN
ajst-19203	72	9	.	.	PUNCT
ajst-19203	73	1	some	some	DET
ajst-19203	73	2	datasets	dataset	NOUN
ajst-19203	73	3	contains	contain	VERB
ajst-19203	73	4	anchor	anchor	NOUN
ajst-19203	73	5	boxes	box	NOUN
ajst-19203	73	6	for	for	ADP
ajst-19203	73	7	each	each	DET
ajst-19203	73	8	object	object	NOUN
ajst-19203	73	9	and	and	CCONJ
ajst-19203	73	10	person	person	NOUN
ajst-19203	73	11	in	in	ADP
ajst-19203	73	12	the	the	DET
ajst-19203	73	13	image	image	NOUN
ajst-19203	73	14	,	,	PUNCT
ajst-19203	73	15	which	which	PRON
ajst-19203	73	16	can	can	AUX
ajst-19203	73	17	be	be	AUX
ajst-19203	73	18	utilized	utilize	VERB
ajst-19203	73	19	for	for	ADP
ajst-19203	73	20	a	a	DET
ajst-19203	73	21	more	more	ADV
ajst-19203	73	22	precise	precise	ADJ
ajst-19203	73	23	matching	matching	NOUN
ajst-19203	73	24	.	.	PUNCT
ajst-19203	74	1	also	also	ADV
ajst-19203	74	2	,	,	PUNCT
ajst-19203	74	3	we	we	PRON
ajst-19203	74	4	need	need	VERB
ajst-19203	74	5	to	to	PART
ajst-19203	74	6	carefully	carefully	ADV
ajst-19203	74	7	look	look	VERB
ajst-19203	74	8	into	into	ADP
ajst-19203	74	9	the	the	DET
ajst-19203	74	10	sequence	sequence	NOUN
ajst-19203	74	11	of	of	ADP
ajst-19203	74	12	vocabularies	vocabulary	NOUN
ajst-19203	74	13	in	in	ADP
ajst-19203	74	14	the	the	DET
ajst-19203	74	15	sentence	sentence	NOUN
ajst-19203	74	16	.	.	PUNCT
ajst-19203	75	1	3	3	X
ajst-19203	75	2	.	.	X
ajst-19203	75	3	consider	consider	VERB
ajst-19203	75	4	the	the	DET
ajst-19203	75	5	entire	entire	ADJ
ajst-19203	75	6	task	task	NOUN
ajst-19203	75	7	jointly	jointly	ADV
ajst-19203	75	8	.	.	PUNCT
ajst-19203	76	1	if	if	SCONJ
ajst-19203	76	2	we	we	PRON
ajst-19203	76	3	can	can	AUX
ajst-19203	76	4	obtain	obtain	VERB
ajst-19203	76	5	image	image	NOUN
ajst-19203	76	6	captioning	captioning	NOUN
ajst-19203	76	7	datasets	dataset	NOUN
ajst-19203	76	8	containing	contain	VERB
ajst-19203	76	9	the	the	DET
ajst-19203	76	10	celebrities	celebrity	NOUN
ajst-19203	76	11	'	'	PART
ajst-19203	76	12	name	name	NOUN
ajst-19203	76	13	,	,	PUNCT
ajst-19203	76	14	we	we	PRON
ajst-19203	76	15	can	can	AUX
ajst-19203	76	16	customize	customize	VERB
ajst-19203	76	17	certain	certain	ADJ
ajst-19203	76	18	loss	loss	NOUN
ajst-19203	76	19	and	and	CCONJ
ajst-19203	76	20	improve	improve	VERB
ajst-19203	76	21	the	the	DET
ajst-19203	76	22	model	model	NOUN
ajst-19203	76	23	's	's	PART
ajst-19203	76	24	ability	ability	NOUN
ajst-19203	76	25	to	to	PART
ajst-19203	76	26	recognize	recognize	VERB
ajst-19203	76	27	faces	face	NOUN
ajst-19203	76	28	.	.	PUNCT
ajst-19203	77	1	joint	joint	ADJ
ajst-19203	77	2	model	model	NOUN
ajst-19203	77	3	may	may	AUX
ajst-19203	77	4	have	have	VERB
ajst-19203	77	5	better	well	ADJ
ajst-19203	77	6	grammatical	grammatical	ADJ
ajst-19203	77	7	accuracy	accuracy	NOUN
ajst-19203	77	8	then	then	ADV
ajst-19203	77	9	our	our	PRON
ajst-19203	77	10	separate	separate	ADJ
ajst-19203	77	11	-	-	PUNCT
ajst-19203	77	12	steps	step	NOUN
ajst-19203	77	13	pipeline.[18,46	pipeline.[18,46	NOUN
ajst-19203	77	14	]	]	PUNCT
ajst-19203	78	1	5.2	5.2	NUM
ajst-19203	78	2	.	.	PUNCT
ajst-19203	78	3	conclusion	conclusion	NOUN
ajst-19203	78	4	in	in	ADP
ajst-19203	78	5	this	this	DET
ajst-19203	78	6	paper	paper	NOUN
ajst-19203	78	7	,	,	PUNCT
ajst-19203	78	8	we	we	PRON
ajst-19203	78	9	use	use	VERB
ajst-19203	78	10	a	a	DET
ajst-19203	78	11	combined	combine	VERB
ajst-19203	78	12	pipeline	pipeline	NOUN
ajst-19203	78	13	to	to	PART
ajst-19203	78	14	conduct	conduct	VERB
ajst-19203	78	15	image	image	NOUN
ajst-19203	78	16	captioning	captioning	NOUN
ajst-19203	78	17	for	for	ADP
ajst-19203	78	18	celebrities	celebrity	NOUN
ajst-19203	78	19	.	.	PUNCT
ajst-19203	79	1	our	our	PRON
ajst-19203	79	2	architecture	architecture	NOUN
ajst-19203	79	3	uses	use	VERB
ajst-19203	79	4	a	a	DET
ajst-19203	79	5	cnn	cnn	NOUN
ajst-19203	79	6	and	and	CCONJ
ajst-19203	79	7	rnn	rnn	PROPN
ajst-19203	79	8	based	base	VERB
ajst-19203	79	9	encoder	encoder	NOUN
ajst-19203	79	10	-	-	PUNCT
ajst-19203	79	11	decoder	decoder	NOUN
ajst-19203	79	12	to	to	PART
ajst-19203	79	13	process	process	VERB
ajst-19203	79	14	the	the	DET
ajst-19203	79	15	input	input	NOUN
ajst-19203	79	16	images	image	NOUN
ajst-19203	79	17	and	and	CCONJ
ajst-19203	79	18	generate	generate	VERB
ajst-19203	79	19	captions	caption	NOUN
ajst-19203	79	20	.	.	PUNCT
ajst-19203	80	1	at	at	ADP
ajst-19203	80	2	the	the	DET
ajst-19203	80	3	same	same	ADJ
ajst-19203	80	4	time	time	NOUN
ajst-19203	80	5	,	,	PUNCT
ajst-19203	80	6	mtcnn	mtcnn	ADJ
ajst-19203	80	7	networks	network	NOUN
ajst-19203	80	8	draws	draw	VERB
ajst-19203	80	9	the	the	DET
ajst-19203	80	10	bounding	bounding	NOUN
ajst-19203	80	11	boxes	box	NOUN
ajst-19203	80	12	for	for	ADP
ajst-19203	80	13	faces	face	NOUN
ajst-19203	80	14	and	and	CCONJ
ajst-19203	80	15	inception	inception	NOUN
ajst-19203	80	16	network	network	NOUN
ajst-19203	80	17	performs	perform	VERB
ajst-19203	80	18	classification	classification	NOUN
ajst-19203	80	19	task	task	NOUN
ajst-19203	80	20	.	.	PUNCT
ajst-19203	81	1	lastly	lastly	ADV
ajst-19203	81	2	,	,	PUNCT
ajst-19203	81	3	we	we	PRON
ajst-19203	81	4	employ	employ	VERB
ajst-19203	81	5	nlp	nlp	NOUN
ajst-19203	81	6	packages	package	NOUN
ajst-19203	81	7	like	like	ADP
ajst-19203	81	8	nltk	nltk	NOUN
ajst-19203	81	9	and	and	CCONJ
ajst-19203	81	10	some	some	DET
ajst-19203	81	11	engineered	engineer	VERB
ajst-19203	81	12	rules	rule	NOUN
ajst-19203	81	13	to	to	PART
ajst-19203	81	14	implement	implement	VERB
ajst-19203	81	15	np	np	DET
ajst-19203	81	16	chunk	chunk	PROPN
ajst-19203	81	17	matching	matching	NOUN
ajst-19203	81	18	.	.	PUNCT
ajst-19203	82	1	[	[	X
ajst-19203	82	2	23	23	NUM
ajst-19203	82	3	]	]	PUNCT
ajst-19203	82	4	the	the	DET
ajst-19203	82	5	incorporation	incorporation	NOUN
ajst-19203	82	6	of	of	ADP
ajst-19203	82	7	such	such	DET
ajst-19203	82	8	a	a	DET
ajst-19203	82	9	pipeline	pipeline	NOUN
ajst-19203	82	10	can	can	AUX
ajst-19203	82	11	significantly	significantly	ADV
ajst-19203	82	12	abbreviate	abbreviate	VERB
ajst-19203	82	13	the	the	DET
ajst-19203	82	14	time	time	NOUN
ajst-19203	82	15	-	-	PUNCT
ajst-19203	82	16	to	to	ADP
ajst-19203	82	17	-	-	PUNCT
ajst-19203	82	18	market	market	NOUN
ajst-19203	82	19	,	,	PUNCT
ajst-19203	82	20	while	while	SCONJ
ajst-19203	82	21	ensuring	ensure	VERB
ajst-19203	82	22	a	a	DET
ajst-19203	82	23	high	high	ADJ
ajst-19203	82	24	standard	standard	NOUN
ajst-19203	82	25	of	of	ADP
ajst-19203	82	26	accuracy	accuracy	NOUN
ajst-19203	82	27	and	and	CCONJ
ajst-19203	82	28	relevance	relevance	NOUN
ajst-19203	82	29	in	in	ADP
ajst-19203	82	30	generated	generate	VERB
ajst-19203	82	31	content.[33	content.[33	NOUN
ajst-19203	82	32	]	]	PUNCT
ajst-19203	82	33	our	our	PRON
ajst-19203	82	34	endeavors	endeavor	NOUN
ajst-19203	82	35	lay	lie	VERB
ajst-19203	82	36	a	a	DET
ajst-19203	82	37	solid	solid	ADJ
ajst-19203	82	38	groundwork	groundwork	NOUN
ajst-19203	82	39	,	,	PUNCT
ajst-19203	82	40	beckoning	beckon	VERB
ajst-19203	82	41	a	a	DET
ajst-19203	82	42	new	new	ADJ
ajst-19203	82	43	era	era	NOUN
ajst-19203	82	44	of	of	ADP
ajst-19203	82	45	intelligent	intelligent	ADJ
ajst-19203	82	46	,	,	PUNCT
ajst-19203	82	47	automated	automate	VERB
ajst-19203	82	48	news	news	NOUN
ajst-19203	82	49	generation	generation	NOUN
ajst-19203	82	50	systems	system	NOUN
ajst-19203	82	51	attuned	attune	VERB
ajst-19203	82	52	to	to	ADP
ajst-19203	82	53	the	the	DET
ajst-19203	82	54	dynamic	dynamic	ADJ
ajst-19203	82	55	demands	demand	NOUN
ajst-19203	82	56	of	of	ADP
ajst-19203	82	57	the	the	DET
ajst-19203	82	58	modern	modern	ADJ
ajst-19203	82	59	digital	digital	ADJ
ajst-19203	82	60	media	medium	NOUN
ajst-19203	82	61	.	.	PUNCT
ajst-19203	83	1	references	reference	NOUN
ajst-19203	83	2	[	[	X
ajst-19203	83	3	1	1	NUM
ajst-19203	83	4	]	]	PUNCT
ajst-19203	83	5	vinyals	vinyal	NOUN
ajst-19203	83	6	,	,	PUNCT
ajst-19203	83	7	oriol	oriol	PROPN
ajst-19203	83	8	,	,	PUNCT
ajst-19203	83	9	et	et	PROPN
ajst-19203	83	10	al	al	PROPN
ajst-19203	83	11	.	.	PUNCT
ajst-19203	84	1	"	"	PUNCT
ajst-19203	84	2	show	show	VERB
ajst-19203	84	3	and	and	CCONJ
ajst-19203	84	4	tell	tell	VERB
ajst-19203	84	5	:	:	PUNCT
ajst-19203	84	6	a	a	DET
ajst-19203	84	7	neural	neural	ADJ
ajst-19203	84	8	image	image	NOUN
ajst-19203	84	9	caption	caption	NOUN
ajst-19203	84	10	generator	generator	NOUN
ajst-19203	84	11	.	.	PUNCT
ajst-19203	84	12	"	"	PUNCT
ajst-19203	85	1	proceedings	proceeding	NOUN
ajst-19203	85	2	of	of	ADP
ajst-19203	85	3	the	the	DET
ajst-19203	85	4	ieee	ieee	NOUN
ajst-19203	85	5	conference	conference	NOUN
ajst-19203	85	6	on	on	ADP
ajst-19203	85	7	computer	computer	NOUN
ajst-19203	85	8	vision	vision	NOUN
ajst-19203	85	9	and	and	CCONJ
ajst-19203	85	10	pattern	pattern	NOUN
ajst-19203	85	11	recognition	recognition	NOUN
ajst-19203	85	12	.	.	PUNCT
ajst-19203	86	1	2015	2015	NUM
ajst-19203	86	2	.	.	PUNCT
ajst-19203	87	1	[	[	X
ajst-19203	87	2	2	2	NUM
ajst-19203	87	3	]	]	X
ajst-19203	87	4	ke	ke	PROPN
ajst-19203	87	5	,	,	PUNCT
ajst-19203	87	6	lei	lei	PROPN
ajst-19203	87	7	,	,	PUNCT
ajst-19203	87	8	et	et	PROPN
ajst-19203	87	9	al	al	PROPN
ajst-19203	87	10	.	.	PUNCT
ajst-19203	87	11	"	"	PUNCT
ajst-19203	87	12	reflective	reflective	ADJ
ajst-19203	87	13	decoding	decode	VERB
ajst-19203	87	14	network	network	NOUN
ajst-19203	87	15	for	for	ADP
ajst-19203	87	16	image	image	NOUN
ajst-19203	87	17	captioning	captioning	NOUN
ajst-19203	87	18	.	.	PUNCT
ajst-19203	87	19	"	"	PUNCT
ajst-19203	88	1	proceedings	proceeding	NOUN
ajst-19203	88	2	of	of	ADP
ajst-19203	88	3	the	the	DET
ajst-19203	88	4	ieee	ieee	NOUN
ajst-19203	88	5	/	/	SYM
ajst-19203	88	6	cvf	cvf	NOUN
ajst-19203	88	7	international	international	ADJ
ajst-19203	88	8	conference	conference	NOUN
ajst-19203	88	9	on	on	ADP
ajst-19203	88	10	computer	computer	NOUN
ajst-19203	88	11	vision	vision	NOUN
ajst-19203	88	12	.	.	PUNCT
ajst-19203	89	1	2019	2019	NUM
ajst-19203	89	2	.	.	PUNCT
ajst-19203	90	1	[	[	X
ajst-19203	90	2	3	3	NUM
ajst-19203	90	3	]	]	X
ajst-19203	90	4	cao	cao	PROPN
ajst-19203	90	5	,	,	PUNCT
ajst-19203	90	6	qiong	qiong	PROPN
ajst-19203	90	7	,	,	PUNCT
ajst-19203	90	8	et	et	PROPN
ajst-19203	90	9	al	al	PROPN
ajst-19203	90	10	.	.	PUNCT
ajst-19203	91	1	"	"	PUNCT
ajst-19203	91	2	vggface2	vggface2	NOUN
ajst-19203	91	3	:	:	PUNCT
ajst-19203	91	4	a	a	DET
ajst-19203	91	5	dataset	dataset	NOUN
ajst-19203	91	6	for	for	ADP
ajst-19203	91	7	recognising	recognise	VERB
ajst-19203	91	8	faces	face	NOUN
ajst-19203	91	9	across	across	ADP
ajst-19203	91	10	pose	pose	NOUN
ajst-19203	91	11	and	and	CCONJ
ajst-19203	91	12	age	age	NOUN
ajst-19203	91	13	.	.	PUNCT
ajst-19203	91	14	"	"	PUNCT
ajst-19203	92	1	2018	2018	NUM
ajst-19203	92	2	13th	13th	NOUN
ajst-19203	92	3	ieee	ieee	NOUN
ajst-19203	92	4	international	international	ADJ
ajst-19203	92	5	conference	conference	NOUN
ajst-19203	92	6	on	on	ADP
ajst-19203	92	7	automatic	automatic	ADJ
ajst-19203	92	8	face	face	NOUN
ajst-19203	92	9	&	&	CCONJ
ajst-19203	92	10	gesture	gesture	PROPN
ajst-19203	92	11	recognition	recognition	PROPN
ajst-19203	92	12	(	(	PUNCT
ajst-19203	92	13	fg	fg	PROPN
ajst-19203	92	14	2018	2018	NUM
ajst-19203	92	15	)	)	PUNCT
ajst-19203	92	16	.	.	PUNCT
ajst-19203	93	1	ieee	ieee	NOUN
ajst-19203	93	2	,	,	PUNCT
ajst-19203	93	3	2018	2018	NUM
ajst-19203	93	4	.	.	PUNCT
ajst-19203	94	1	[	[	X
ajst-19203	94	2	4	4	NUM
ajst-19203	94	3	]	]	X
ajst-19203	94	4	zhang	zhang	PROPN
ajst-19203	94	5	,	,	PUNCT
ajst-19203	94	6	kaipeng	kaipeng	PROPN
ajst-19203	94	7	,	,	PUNCT
ajst-19203	94	8	et	et	PROPN
ajst-19203	94	9	al	al	PROPN
ajst-19203	94	10	.	.	PUNCT
ajst-19203	95	1	"	"	PUNCT
ajst-19203	95	2	joint	joint	ADJ
ajst-19203	95	3	face	face	NOUN
ajst-19203	95	4	detection	detection	NOUN
ajst-19203	95	5	and	and	CCONJ
ajst-19203	95	6	alignment	alignment	NOUN
ajst-19203	95	7	using	use	VERB
ajst-19203	95	8	multitask	multitask	ADJ
ajst-19203	95	9	cascaded	cascade	VERB
ajst-19203	95	10	convolutional	convolutional	ADJ
ajst-19203	95	11	networks	network	NOUN
ajst-19203	95	12	.	.	PUNCT
ajst-19203	95	13	"	"	PUNCT
ajst-19203	96	1	ieee	ieee	NOUN
ajst-19203	96	2	signal	signal	NOUN
ajst-19203	96	3	processing	processing	NOUN
ajst-19203	96	4	letters	letter	NOUN
ajst-19203	96	5	23.10	23.10	NUM
ajst-19203	96	6	(	(	PUNCT
ajst-19203	96	7	2016	2016	NUM
ajst-19203	96	8	):	):	PUNCT
ajst-19203	96	9	1499	1499	NUM
ajst-19203	96	10	-	-	SYM
ajst-19203	96	11	1503	1503	NUM
ajst-19203	96	12	.	.	PUNCT
ajst-19203	97	1	[	[	X
ajst-19203	97	2	5	5	NUM
ajst-19203	97	3	]	]	X
ajst-19203	97	4	liu	liu	PROPN
ajst-19203	97	5	,	,	PUNCT
ajst-19203	97	6	tianrui	tianrui	PROPN
ajst-19203	97	7	,	,	PUNCT
ajst-19203	97	8	et	et	PROPN
ajst-19203	97	9	al	al	PROPN
ajst-19203	97	10	.	.	PUNCT
ajst-19203	98	1	"	"	PUNCT
ajst-19203	98	2	news	news	NOUN
ajst-19203	98	3	recommendation	recommendation	NOUN
ajst-19203	98	4	with	with	ADP
ajst-19203	98	5	attention	attention	NOUN
ajst-19203	98	6	mechanism	mechanism	NOUN
ajst-19203	98	7	.	.	PUNCT
ajst-19203	98	8	"	"	PUNCT
ajst-19203	99	1	arxiv	arxiv	PROPN
ajst-19203	99	2	preprint	preprint	VERB
ajst-19203	99	3	arxiv:2402.07422	arxiv:2402.07422	NOUN
ajst-19203	99	4	(	(	PUNCT
ajst-19203	99	5	2024	2024	NUM
ajst-19203	99	6	)	)	PUNCT
ajst-19203	99	7	.	.	PUNCT
ajst-19203	100	1	[	[	X
ajst-19203	100	2	6	6	NUM
ajst-19203	100	3	]	]	X
ajst-19203	100	4	guo	guo	PROPN
ajst-19203	100	5	,	,	PUNCT
ajst-19203	100	6	cheng	cheng	PROPN
ajst-19203	100	7	,	,	PUNCT
ajst-19203	100	8	et	et	PROPN
ajst-19203	100	9	al	al	PROPN
ajst-19203	100	10	.	.	PUNCT
ajst-19203	101	1	"	"	PUNCT
ajst-19203	101	2	the	the	DET
ajst-19203	101	3	role	role	NOUN
ajst-19203	101	4	of	of	ADP
ajst-19203	101	5	machine	machine	NOUN
ajst-19203	101	6	learning	learn	VERB
ajst-19203	101	7	in	in	ADP
ajst-19203	101	8	enhancing	enhance	VERB
ajst-19203	101	9	computer	computer	NOUN
ajst-19203	101	10	vision	vision	NOUN
ajst-19203	101	11	processing	processing	NOUN
ajst-19203	101	12	.	.	PUNCT
ajst-19203	101	13	"	"	PUNCT
ajst-19203	102	1	актуальные	актуальные	NOUN
ajst-19203	102	2	вопросы	вопросы	NOUN
ajst-19203	102	3	современных	современных	PROPN
ajst-19203	102	4	научных	научных	PROPN
ajst-19203	102	5	исследований	исследований	PROPN
ajst-19203	102	6	.	.	PUNCT
ajst-19203	103	1	2023	2023	NUM
ajst-19203	103	2	.	.	PUNCT
ajst-19203	104	1	[	[	X
ajst-19203	104	2	7	7	NUM
ajst-19203	104	3	]	]	X
ajst-19203	104	4	anderson	anderson	PROPN
ajst-19203	104	5	,	,	PUNCT
ajst-19203	104	6	peter	peter	PROPN
ajst-19203	104	7	,	,	PUNCT
ajst-19203	104	8	et	et	PROPN
ajst-19203	104	9	al	al	PROPN
ajst-19203	104	10	.	.	PUNCT
ajst-19203	105	1	"	"	PUNCT
ajst-19203	105	2	bottom	bottom	ADJ
ajst-19203	105	3	-	-	PUNCT
ajst-19203	105	4	up	up	NOUN
ajst-19203	105	5	and	and	CCONJ
ajst-19203	105	6	top	top	ADJ
ajst-19203	105	7	-	-	PUNCT
ajst-19203	105	8	down	down	ADP
ajst-19203	105	9	attention	attention	NOUN
ajst-19203	105	10	for	for	ADP
ajst-19203	105	11	image	image	NOUN
ajst-19203	105	12	captioning	captioning	NOUN
ajst-19203	105	13	and	and	CCONJ
ajst-19203	105	14	visual	visual	ADJ
ajst-19203	105	15	question	question	NOUN
ajst-19203	105	16	answering	answer	VERB
ajst-19203	105	17	.	.	PUNCT
ajst-19203	105	18	"	"	PUNCT
ajst-19203	106	1	proceedings	proceeding	NOUN
ajst-19203	106	2	of	of	ADP
ajst-19203	106	3	the	the	DET
ajst-19203	106	4	ieee	ieee	NOUN
ajst-19203	106	5	conference	conference	NOUN
ajst-19203	106	6	on	on	ADP
ajst-19203	106	7	computer	computer	NOUN
ajst-19203	106	8	vision	vision	NOUN
ajst-19203	106	9	and	and	CCONJ
ajst-19203	106	10	pattern	pattern	NOUN
ajst-19203	106	11	recognition	recognition	NOUN
ajst-19203	106	12	.	.	PUNCT
ajst-19203	107	1	2018	2018	NUM
ajst-19203	107	2	.	.	PUNCT
ajst-19203	108	1	[	[	X
ajst-19203	108	2	8	8	NUM
ajst-19203	108	3	]	]	X
ajst-19203	108	4	zhao	zhao	X
ajst-19203	108	5	,	,	PUNCT
ajst-19203	108	6	yufan	yufan	PROPN
ajst-19203	108	7	,	,	PUNCT
ajst-19203	108	8	et	et	PROPN
ajst-19203	108	9	al	al	PROPN
ajst-19203	108	10	.	.	PUNCT
ajst-19203	109	1	"	"	PUNCT
ajst-19203	109	2	an	an	DET
ajst-19203	109	3	examination	examination	NOUN
ajst-19203	109	4	of	of	ADP
ajst-19203	109	5	transformer	transformer	NOUN
ajst-19203	109	6	:	:	PUNCT
ajst-19203	109	7	progress	progress	NOUN
ajst-19203	109	8	and	and	CCONJ
ajst-19203	109	9	application	application	NOUN
ajst-19203	109	10	in	in	ADP
ajst-19203	109	11	the	the	DET
ajst-19203	109	12	field	field	NOUN
ajst-19203	109	13	of	of	ADP
ajst-19203	109	14	computer	computer	NOUN
ajst-19203	109	15	vision	vision	NOUN
ajst-19203	109	16	.	.	PUNCT
ajst-19203	109	17	"	"	PUNCT
ajst-19203	110	1	современная	современная	PROPN
ajst-19203	110	2	наука	наука	PROPN
ajst-19203	110	3	:	:	PUNCT
ajst-19203	110	4	актуальные	актуальные	NOUN
ajst-19203	110	5	вопросы	вопросы	NOUN
ajst-19203	110	6	,	,	PUNCT
ajst-19203	110	7	достижения	достижения	ADJ
ajst-19203	110	8	и	и	PROPN
ajst-19203	110	9	инновации	инновации	NOUN
ajst-19203	110	10	.	.	PUNCT
ajst-19203	111	1	2023	2023	NUM
ajst-19203	111	2	.	.	PUNCT
ajst-19203	112	1	[	[	X
ajst-19203	112	2	9	9	NUM
ajst-19203	112	3	]	]	X
ajst-19203	112	4	liu	liu	PROPN
ajst-19203	112	5	,	,	PUNCT
ajst-19203	112	6	tianrui	tianrui	PROPN
ajst-19203	112	7	,	,	PUNCT
ajst-19203	112	8	et	et	PROPN
ajst-19203	112	9	al	al	PROPN
ajst-19203	112	10	.	.	PUNCT
ajst-19203	113	1	"	"	PUNCT
ajst-19203	113	2	news	news	NOUN
ajst-19203	113	3	recommendation	recommendation	NOUN
ajst-19203	113	4	with	with	ADP
ajst-19203	113	5	attention	attention	NOUN
ajst-19203	113	6	mechanism	mechanism	NOUN
ajst-19203	113	7	.	.	PUNCT
ajst-19203	113	8	"	"	PUNCT
ajst-19203	114	1	journal	journal	NOUN
ajst-19203	114	2	of	of	ADP
ajst-19203	114	3	industrial	industrial	ADJ
ajst-19203	114	4	engineering	engineering	NOUN
ajst-19203	114	5	and	and	CCONJ
ajst-19203	114	6	applied	apply	VERB
ajst-19203	114	7	science	science	NOUN
ajst-19203	114	8	2.1	2.1	NUM
ajst-19203	114	9	(	(	PUNCT
ajst-19203	114	10	2024	2024	NUM
ajst-19203	114	11	):	):	PUNCT
ajst-19203	114	12	21	21	NUM
ajst-19203	114	13	-	-	SYM
ajst-19203	114	14	26	26	NUM
ajst-19203	114	15	.	.	PUNCT
ajst-19203	115	1	[	[	X
ajst-19203	115	2	10	10	NUM
ajst-19203	115	3	]	]	X
ajst-19203	115	4	radford	radford	PROPN
ajst-19203	115	5	,	,	PUNCT
ajst-19203	115	6	alec	alec	PROPN
ajst-19203	115	7	,	,	PUNCT
ajst-19203	115	8	et	et	PROPN
ajst-19203	115	9	al	al	PROPN
ajst-19203	115	10	.	.	PUNCT
ajst-19203	116	1	"	"	PUNCT
ajst-19203	116	2	learning	learn	VERB
ajst-19203	116	3	transferable	transferable	ADJ
ajst-19203	116	4	visual	visual	ADJ
ajst-19203	116	5	models	model	NOUN
ajst-19203	116	6	from	from	ADP
ajst-19203	116	7	natural	natural	ADJ
ajst-19203	116	8	language	language	NOUN
ajst-19203	116	9	supervision	supervision	NOUN
ajst-19203	116	10	.	.	PUNCT
ajst-19203	116	11	"	"	PUNCT
ajst-19203	117	1	international	international	ADJ
ajst-19203	117	2	conference	conference	NOUN
ajst-19203	117	3	on	on	ADP
ajst-19203	117	4	machine	machine	NOUN
ajst-19203	117	5	learning	learning	NOUN
ajst-19203	117	6	.	.	PUNCT
ajst-19203	118	1	pmlr	pmlr	NOUN
ajst-19203	118	2	,	,	PUNCT
ajst-19203	118	3	2021	2021	NUM
ajst-19203	118	4	[	[	X
ajst-19203	118	5	11	11	NUM
ajst-19203	118	6	]	]	SYM
ajst-19203	118	7	li	li	PROPN
ajst-19203	118	8	,	,	PUNCT
ajst-19203	118	9	yanjie	yanjie	PROPN
ajst-19203	118	10	,	,	PUNCT
ajst-19203	118	11	et	et	PROPN
ajst-19203	118	12	al	al	PROPN
ajst-19203	118	13	.	.	PUNCT
ajst-19203	119	1	"	"	PUNCT
ajst-19203	119	2	transfer	transfer	NOUN
ajst-19203	119	3	-	-	PUNCT
ajst-19203	119	4	learning	learning	NOUN
ajst-19203	119	5	-	-	PUNCT
ajst-19203	119	6	based	base	VERB
ajst-19203	119	7	network	network	NOUN
ajst-19203	119	8	traffic	traffic	NOUN
ajst-19203	119	9	automatic	automatic	ADJ
ajst-19203	119	10	generation	generation	NOUN
ajst-19203	119	11	framework	framework	NOUN
ajst-19203	119	12	.	.	PUNCT
ajst-19203	119	13	"	"	PUNCT
ajst-19203	120	1	2021	2021	NUM
ajst-19203	120	2	6th	6th	ADJ
ajst-19203	120	3	international	international	ADJ
ajst-19203	120	4	conference	conference	NOUN
ajst-19203	120	5	on	on	ADP
ajst-19203	120	6	intelligent	intelligent	ADJ
ajst-19203	120	7	computing	computing	NOUN
ajst-19203	120	8	and	and	CCONJ
ajst-19203	120	9	signal	signal	NOUN
ajst-19203	120	10	processing	processing	NOUN
ajst-19203	120	11	(	(	PUNCT
ajst-19203	120	12	icsp	icsp	ADJ
ajst-19203	120	13	)	)	PUNCT
ajst-19203	120	14	.	.	PUNCT
ajst-19203	121	1	ieee	ieee	NOUN
ajst-19203	121	2	,	,	PUNCT
ajst-19203	121	3	2021	2021	NUM
ajst-19203	122	1	[	[	X
ajst-19203	122	2	12	12	NUM
ajst-19203	122	3	]	]	X
ajst-19203	122	4	liu	liu	PROPN
ajst-19203	122	5	,	,	PUNCT
ajst-19203	122	6	wei	wei	PROPN
ajst-19203	122	7	,	,	PUNCT
ajst-19203	122	8	et	et	PROPN
ajst-19203	122	9	al	al	PROPN
ajst-19203	122	10	.	.	PUNCT
ajst-19203	122	11	"	"	PUNCT
ajst-19203	122	12	cptr	cptr	NOUN
ajst-19203	122	13	:	:	PUNCT
ajst-19203	122	14	full	full	ADJ
ajst-19203	122	15	transformer	transformer	NOUN
ajst-19203	122	16	network	network	NOUN
ajst-19203	122	17	for	for	ADP
ajst-19203	122	18	image	image	NOUN
ajst-19203	122	19	captioning	captioning	NOUN
ajst-19203	122	20	.	.	PUNCT
ajst-19203	122	21	"	"	PUNCT
ajst-19203	123	1	arxiv	arxiv	PROPN
ajst-19203	123	2	preprint	preprint	VERB
ajst-19203	123	3	arxiv:2101.10804	arxiv:2101.10804	NOUN
ajst-19203	123	4	(	(	PUNCT
ajst-19203	123	5	2021	2021	NUM
ajst-19203	123	6	)	)	PUNCT
ajst-19203	123	7	.	.	PUNCT
ajst-19203	124	1	[	[	X
ajst-19203	124	2	13	13	NUM
ajst-19203	124	3	]	]	X
ajst-19203	124	4	liu	liu	PROPN
ajst-19203	124	5	,	,	PUNCT
ajst-19203	124	6	tianrui	tianrui	PROPN
ajst-19203	124	7	,	,	PUNCT
ajst-19203	124	8	et	et	PROPN
ajst-19203	124	9	al	al	PROPN
ajst-19203	124	10	.	.	PUNCT
ajst-19203	125	1	"	"	PUNCT
ajst-19203	125	2	particle	particle	NOUN
ajst-19203	125	3	filter	filter	NOUN
ajst-19203	125	4	slam	slam	NOUN
ajst-19203	125	5	for	for	ADP
ajst-19203	125	6	vehicle	vehicle	NOUN
ajst-19203	125	7	localization	localization	NOUN
ajst-19203	125	8	.	.	PUNCT
ajst-19203	125	9	"	"	PUNCT
ajst-19203	126	1	journal	journal	NOUN
ajst-19203	126	2	of	of	ADP
ajst-19203	126	3	industrial	industrial	ADJ
ajst-19203	126	4	engineering	engineering	NOUN
ajst-19203	126	5	and	and	CCONJ
ajst-19203	126	6	applied	apply	VERB
ajst-19203	126	7	science	science	NOUN
ajst-19203	126	8	2.1	2.1	NUM
ajst-19203	126	9	(	(	PUNCT
ajst-19203	126	10	2024	2024	NUM
ajst-19203	126	11	):	):	PUNCT
ajst-19203	126	12	27	27	NUM
ajst-19203	126	13	-	-	SYM
ajst-19203	126	14	31	31	NUM
ajst-19203	126	15	.	.	PUNCT
ajst-19203	127	1	[	[	X
ajst-19203	127	2	14	14	NUM
ajst-19203	127	3	]	]	X
ajst-19203	127	4	zhao	zhao	X
ajst-19203	127	5	,	,	PUNCT
ajst-19203	127	6	zhiming	zhime	VERB
ajst-19203	127	7	,	,	PUNCT
ajst-19203	127	8	et	et	PROPN
ajst-19203	127	9	al	al	PROPN
ajst-19203	127	10	.	.	PUNCT
ajst-19203	128	1	"	"	PUNCT
ajst-19203	128	2	enhancing	enhance	VERB
ajst-19203	128	3	e	e	NOUN
ajst-19203	128	4	-	-	NOUN
ajst-19203	128	5	commerce	commerce	NOUN
ajst-19203	128	6	recommendations	recommendation	NOUN
ajst-19203	128	7	:	:	PUNCT
ajst-19203	128	8	unveiling	unveil	VERB
ajst-19203	128	9	insights	insight	NOUN
ajst-19203	128	10	from	from	ADP
ajst-19203	128	11	customer	customer	NOUN
ajst-19203	128	12	reviews	review	NOUN
ajst-19203	128	13	with	with	ADP
ajst-19203	128	14	bertfusiondnn	bertfusiondnn	NOUN
ajst-19203	128	15	.	.	PUNCT
ajst-19203	128	16	"	"	PUNCT
ajst-19203	128	17	journal	journal	NOUN
ajst-19203	128	18	of	of	ADP
ajst-19203	128	19	theory	theory	NOUN
ajst-19203	128	20	and	and	CCONJ
ajst-19203	128	21	practice	practice	NOUN
ajst-19203	128	22	of	of	ADP
ajst-19203	128	23	engineering	engineer	VERB
ajst-19203	128	24	science	science	NOUN
ajst-19203	128	25	4.02	4.02	NUM
ajst-19203	128	26	(	(	PUNCT
ajst-19203	128	27	2024	2024	NUM
ajst-19203	128	28	):	):	PUNCT
ajst-19203	128	29	38	38	NUM
ajst-19203	128	30	-	-	SYM
ajst-19203	128	31	44	44	NUM
ajst-19203	128	32	.	.	PUNCT
ajst-19203	129	1	[	[	X
ajst-19203	129	2	15	15	NUM
ajst-19203	129	3	]	]	SYM
ajst-19203	129	4	su	su	PROPN
ajst-19203	129	5	,	,	PUNCT
ajst-19203	129	6	jing	jing	PROPN
ajst-19203	129	7	,	,	PUNCT
ajst-19203	129	8	et	et	PROPN
ajst-19203	129	9	al	al	PROPN
ajst-19203	129	10	.	.	PUNCT
ajst-19203	130	1	"	"	PUNCT
ajst-19203	130	2	large	large	ADJ
ajst-19203	130	3	language	language	NOUN
ajst-19203	130	4	models	model	NOUN
ajst-19203	130	5	for	for	ADP
ajst-19203	130	6	forecasting	forecasting	NOUN
ajst-19203	130	7	and	and	CCONJ
ajst-19203	130	8	anomaly	anomaly	NOUN
ajst-19203	130	9	detection	detection	NOUN
ajst-19203	130	10	:	:	PUNCT
ajst-19203	130	11	a	a	DET
ajst-19203	130	12	systematic	systematic	ADJ
ajst-19203	130	13	literature	literature	NOUN
ajst-19203	130	14	review	review	NOUN
ajst-19203	130	15	.	.	PUNCT
ajst-19203	130	16	"	"	PUNCT
ajst-19203	131	1	arxiv	arxiv	PROPN
ajst-19203	131	2	preprint	preprint	VERB
ajst-19203	131	3	arxiv:2402.10350	arxiv:2402.10350	NUM
ajst-19203	131	4	(	(	PUNCT
ajst-19203	131	5	2024	2024	NUM
ajst-19203	131	6	)	)	PUNCT
ajst-19203	131	7	.	.	PUNCT
ajst-19203	132	1	[	[	X
ajst-19203	132	2	16	16	NUM
ajst-19203	132	3	]	]	X
ajst-19203	132	4	xiong	xiong	PROPN
ajst-19203	132	5	,	,	PUNCT
ajst-19203	132	6	jize	jize	PROPN
ajst-19203	132	7	,	,	PUNCT
ajst-19203	132	8	et	et	PROPN
ajst-19203	132	9	al	al	PROPN
ajst-19203	132	10	.	.	PUNCT
ajst-19203	133	1	"	"	PUNCT
ajst-19203	133	2	decoding	decode	VERB
ajst-19203	133	3	sentiments	sentiment	NOUN
ajst-19203	133	4	:	:	PUNCT
ajst-19203	133	5	enhancing	enhance	VERB
ajst-19203	133	6	covid-19	covid-19	PROPN
ajst-19203	133	7	tweet	tweet	NOUN
ajst-19203	133	8	analysis	analysis	NOUN
ajst-19203	133	9	through	through	ADP
ajst-19203	133	10	bert	bert	PROPN
ajst-19203	133	11	-	-	PUNCT
ajst-19203	133	12	rcnn	rcnn	PROPN
ajst-19203	133	13	fusion	fusion	NOUN
ajst-19203	133	14	.	.	PUNCT
ajst-19203	133	15	"	"	PUNCT
ajst-19203	133	16	journal	journal	NOUN
ajst-19203	133	17	of	of	ADP
ajst-19203	133	18	theory	theory	NOUN
ajst-19203	133	19	and	and	CCONJ
ajst-19203	133	20	practice	practice	NOUN
ajst-19203	133	21	of	of	ADP
ajst-19203	133	22	engineering	engineering	NOUN
ajst-19203	133	23	science	science	NOUN
ajst-19203	133	24	4.01	4.01	NUM
ajst-19203	133	25	(	(	PUNCT
ajst-19203	133	26	2024	2024	NUM
ajst-19203	133	27	):	):	PUNCT
ajst-19203	133	28	86	86	NUM
ajst-19203	133	29	-	-	SYM
ajst-19203	133	30	93	93	NUM
ajst-19203	133	31	.	.	PUNCT
ajst-19203	134	1	[	[	X
ajst-19203	134	2	17	17	NUM
ajst-19203	134	3	]	]	X
ajst-19203	134	4	liu	liu	PROPN
ajst-19203	134	5	,	,	PUNCT
ajst-19203	134	6	shun	shun	PROPN
ajst-19203	134	7	,	,	PUNCT
ajst-19203	134	8	et	et	PROPN
ajst-19203	134	9	al	al	PROPN
ajst-19203	134	10	.	.	PUNCT
ajst-19203	135	1	"	"	PUNCT
ajst-19203	135	2	financial	financial	ADJ
ajst-19203	135	3	time	time	NOUN
ajst-19203	135	4	-	-	PUNCT
ajst-19203	135	5	series	series	NOUN
ajst-19203	135	6	forecasting	forecasting	NOUN
ajst-19203	135	7	:	:	PUNCT
ajst-19203	135	8	towards	towards	ADP
ajst-19203	135	9	synergizing	synergize	VERB
ajst-19203	135	10	performance	performance	NOUN
ajst-19203	135	11	and	and	CCONJ
ajst-19203	135	12	interpretability	interpretability	NOUN
ajst-19203	135	13	within	within	ADP
ajst-19203	135	14	a	a	DET
ajst-19203	135	15	hybrid	hybrid	ADJ
ajst-19203	135	16	machine	machine	NOUN
ajst-19203	135	17	learning	learning	NOUN
ajst-19203	135	18	approach	approach	NOUN
ajst-19203	135	19	.	.	PUNCT
ajst-19203	135	20	"	"	PUNCT
ajst-19203	136	1	arxiv	arxiv	PROPN
ajst-19203	136	2	preprint	preprint	NOUN
ajst-19203	136	3	arxiv:2401.00534	arxiv:2401.00534	NOUN
ajst-19203	136	4	(	(	PUNCT
ajst-19203	136	5	2023	2023	NUM
ajst-19203	136	6	)	)	PUNCT
ajst-19203	136	7	.	.	PUNCT
ajst-19203	137	1	[	[	X
ajst-19203	137	2	18	18	NUM
ajst-19203	137	3	]	]	PUNCT
ajst-19203	137	4	popokh	popokh	VERB
ajst-19203	137	5	,	,	PUNCT
ajst-19203	137	6	leo	leo	PROPN
ajst-19203	137	7	,	,	PUNCT
ajst-19203	137	8	et	et	PROPN
ajst-19203	137	9	al	al	PROPN
ajst-19203	137	10	.	.	PUNCT
ajst-19203	138	1	"	"	PUNCT
ajst-19203	138	2	illumicore	illumicore	NOUN
ajst-19203	138	3	:	:	PUNCT
ajst-19203	138	4	optimization	optimization	NOUN
ajst-19203	138	5	modeling	modeling	NOUN
ajst-19203	138	6	and	and	CCONJ
ajst-19203	138	7	implementation	implementation	NOUN
ajst-19203	138	8	for	for	ADP
ajst-19203	138	9	efficient	efficient	ADJ
ajst-19203	138	10	vnf	vnf	NOUN
ajst-19203	138	11	placement	placement	NOUN
ajst-19203	138	12	.	.	PUNCT
ajst-19203	138	13	"	"	PUNCT
ajst-19203	139	1	2021	2021	NUM
ajst-19203	139	2	international	international	ADJ
ajst-19203	139	3	conference	conference	NOUN
ajst-19203	139	4	on	on	ADP
ajst-19203	139	5	software	software	NOUN
ajst-19203	139	6	,	,	PUNCT
ajst-19203	139	7	telecommunications	telecommunications	NOUN
ajst-19203	139	8	and	and	CCONJ
ajst-19203	139	9	computer	computer	NOUN
ajst-19203	139	10	networks	network	NOUN
ajst-19203	139	11	(	(	PUNCT
ajst-19203	139	12	softcom	softcom	PROPN
ajst-19203	139	13	)	)	PUNCT
ajst-19203	139	14	.	.	PUNCT
ajst-19203	140	1	ieee	ieee	NOUN
ajst-19203	140	2	,	,	PUNCT
ajst-19203	140	3	2021	2021	NUM
ajst-19203	140	4	.	.	PUNCT
ajst-19203	141	1	[	[	X
ajst-19203	141	2	19	19	NUM
ajst-19203	141	3	]	]	SYM
ajst-19203	141	4	su	su	PROPN
ajst-19203	141	5	,	,	PUNCT
ajst-19203	141	6	jing	jing	PROPN
ajst-19203	141	7	,	,	PUNCT
ajst-19203	141	8	suku	suku	PROPN
ajst-19203	141	9	nair	nair	PROPN
ajst-19203	141	10	,	,	PUNCT
ajst-19203	141	11	and	and	CCONJ
ajst-19203	141	12	leo	leo	PROPN
ajst-19203	141	13	popokh	popokh	VERB
ajst-19203	141	14	.	.	PUNCT
ajst-19203	142	1	"	"	PUNCT
ajst-19203	142	2	edgegym	edgegym	ADV
ajst-19203	142	3	:	:	PUNCT
ajst-19203	142	4	a	a	DET
ajst-19203	142	5	reinforcement	reinforcement	NOUN
ajst-19203	142	6	learning	learn	VERB
ajst-19203	142	7	environment	environment	NOUN
ajst-19203	142	8	for	for	ADP
ajst-19203	142	9	constraint	constraint	NOUN
ajst-19203	142	10	-	-	PUNCT
ajst-19203	142	11	aware	aware	ADJ
ajst-19203	142	12	nfv	nfv	NOUN
ajst-19203	142	13	resource	resource	NOUN
ajst-19203	142	14	allocation	allocation	NOUN
ajst-19203	142	15	.	.	PUNCT
ajst-19203	142	16	"	"	PUNCT
ajst-19203	143	1	2023	2023	NUM
ajst-19203	143	2	ieee	ieee	NOUN
ajst-19203	143	3	2nd	2nd	PROPN
ajst-19203	143	4	international	international	ADJ
ajst-19203	143	5	conference	conference	NOUN
ajst-19203	143	6	on	on	ADP
ajst-19203	143	7	ai	ai	NOUN
ajst-19203	143	8	in	in	ADP
ajst-19203	143	9	cybersecurity	cybersecurity	NOUN
ajst-19203	143	10	(	(	PUNCT
ajst-19203	143	11	icaic	icaic	ADJ
ajst-19203	143	12	)	)	PUNCT
ajst-19203	143	13	.	.	PUNCT
ajst-19203	144	1	ieee	ieee	NOUN
ajst-19203	144	2	,	,	PUNCT
ajst-19203	144	3	2023	2023	NUM
ajst-19203	144	4	.	.	PUNCT
ajst-19203	145	1	288	288	NUM
ajst-19203	146	1	[	[	SYM
ajst-19203	146	2	20	20	NUM
ajst-19203	146	3	]	]	SYM
ajst-19203	146	4	su	su	PROPN
ajst-19203	146	5	,	,	PUNCT
ajst-19203	146	6	jing	jing	PROPN
ajst-19203	146	7	,	,	PUNCT
ajst-19203	146	8	suku	suku	PROPN
ajst-19203	146	9	nair	nair	PROPN
ajst-19203	146	10	,	,	PUNCT
ajst-19203	146	11	and	and	CCONJ
ajst-19203	146	12	leo	leo	PROPN
ajst-19203	146	13	popokh	popokh	VERB
ajst-19203	146	14	.	.	PUNCT
ajst-19203	147	1	"	"	PUNCT
ajst-19203	147	2	optimal	optimal	ADJ
ajst-19203	147	3	resource	resource	NOUN
ajst-19203	147	4	allocation	allocation	NOUN
ajst-19203	147	5	in	in	ADP
ajst-19203	147	6	sdn	sdn	PROPN
ajst-19203	147	7	/	/	SYM
ajst-19203	147	8	nfv	nfv	ADV
ajst-19203	147	9	-	-	PUNCT
ajst-19203	147	10	enabled	enable	VERB
ajst-19203	147	11	networks	network	NOUN
ajst-19203	147	12	via	via	ADP
ajst-19203	147	13	deep	deep	ADJ
ajst-19203	147	14	reinforcement	reinforcement	NOUN
ajst-19203	147	15	learning	learning	NOUN
ajst-19203	147	16	.	.	PUNCT
ajst-19203	147	17	"	"	PUNCT
ajst-19203	148	1	2022	2022	NUM
ajst-19203	148	2	ieee	ieee	NOUN
ajst-19203	148	3	ninth	ninth	ADJ
ajst-19203	148	4	international	international	ADJ
ajst-19203	148	5	conference	conference	NOUN
ajst-19203	148	6	on	on	ADP
ajst-19203	148	7	communications	communication	NOUN
ajst-19203	148	8	and	and	CCONJ
ajst-19203	148	9	networking	networking	NOUN
ajst-19203	148	10	(	(	PUNCT
ajst-19203	148	11	comnet	comnet	NOUN
ajst-19203	148	12	)	)	PUNCT
ajst-19203	148	13	.	.	PUNCT
ajst-19203	149	1	ieee	ieee	NOUN
ajst-19203	149	2	,	,	PUNCT
ajst-19203	149	3	2022	2022	NUM
ajst-19203	149	4	.	.	PUNCT
ajst-19203	150	1	[	[	X
ajst-19203	150	2	21	21	NUM
ajst-19203	150	3	]	]	X
ajst-19203	150	4	fu	fu	PROPN
ajst-19203	150	5	,	,	PUNCT
ajst-19203	150	6	zhe	zhe	PROPN
ajst-19203	150	7	,	,	PUNCT
ajst-19203	150	8	xi	xi	PROPN
ajst-19203	150	9	niu	niu	PROPN
ajst-19203	150	10	,	,	PUNCT
ajst-19203	150	11	and	and	CCONJ
ajst-19203	150	12	li	li	PROPN
ajst-19203	150	13	yu	yu	PROPN
ajst-19203	150	14	.	.	PUNCT
ajst-19203	150	15	"	"	PUNCT
ajst-19203	150	16	wisdom	wisdom	NOUN
ajst-19203	150	17	of	of	ADP
ajst-19203	150	18	crowds	crowd	NOUN
ajst-19203	150	19	and	and	CCONJ
ajst-19203	150	20	finegrained	finegraine	VERB
ajst-19203	150	21	learning	learn	VERB
ajst-19203	150	22	for	for	ADP
ajst-19203	150	23	serendipity	serendipity	NOUN
ajst-19203	150	24	recommendations	recommendation	NOUN
ajst-19203	150	25	.	.	PUNCT
ajst-19203	150	26	"	"	PUNCT
ajst-19203	151	1	proceedings	proceeding	NOUN
ajst-19203	151	2	of	of	ADP
ajst-19203	151	3	the	the	DET
ajst-19203	151	4	46th	46th	ADJ
ajst-19203	151	5	international	international	PROPN
ajst-19203	151	6	acm	acm	PROPN
ajst-19203	151	7	sigir	sigir	PROPN
ajst-19203	151	8	conference	conference	PROPN
ajst-19203	151	9	on	on	ADP
ajst-19203	151	10	research	research	NOUN
ajst-19203	151	11	and	and	CCONJ
ajst-19203	151	12	development	development	NOUN
ajst-19203	151	13	in	in	ADP
ajst-19203	151	14	information	information	NOUN
ajst-19203	151	15	retrieval	retrieval	NOUN
ajst-19203	151	16	.	.	PUNCT
ajst-19203	152	1	2023	2023	NUM
ajst-19203	152	2	.	.	PUNCT
ajst-19203	153	1	[	[	X
ajst-19203	153	2	22	22	NUM
ajst-19203	153	3	]	]	X
ajst-19203	153	4	wei	wei	PROPN
ajst-19203	153	5	,	,	PUNCT
ajst-19203	153	6	xiaojun	xiaojun	PROPN
ajst-19203	153	7	,	,	PUNCT
ajst-19203	153	8	et	et	PROPN
ajst-19203	153	9	al	al	PROPN
ajst-19203	153	10	.	.	PUNCT
ajst-19203	154	1	"	"	PUNCT
ajst-19203	154	2	narrowing	narrow	VERB
ajst-19203	154	3	signal	signal	NOUN
ajst-19203	154	4	distribution	distribution	NOUN
ajst-19203	154	5	by	by	ADP
ajst-19203	154	6	adamantane	adamantane	ADJ
ajst-19203	154	7	derivatization	derivatization	NOUN
ajst-19203	154	8	for	for	ADP
ajst-19203	154	9	amino	amino	NOUN
ajst-19203	154	10	acid	acid	NOUN
ajst-19203	154	11	identification	identification	NOUN
ajst-19203	154	12	using	use	VERB
ajst-19203	154	13	an	an	DET
ajst-19203	154	14	α	α	NOUN
ajst-19203	154	15	-	-	PUNCT
ajst-19203	154	16	hemolysin	hemolysin	NOUN
ajst-19203	154	17	nanopore	nanopore	NOUN
ajst-19203	154	18	.	.	PUNCT
ajst-19203	154	19	"	"	PUNCT
ajst-19203	155	1	nano	nano	NOUN
ajst-19203	155	2	letters	letter	NOUN
ajst-19203	155	3	(	(	PUNCT
ajst-19203	155	4	2024	2024	NUM
ajst-19203	155	5	)	)	PUNCT
ajst-19203	155	6	.	.	PUNCT
ajst-19203	156	1	[	[	X
ajst-19203	156	2	23	23	NUM
ajst-19203	156	3	]	]	X
ajst-19203	156	4	song	song	NOUN
ajst-19203	156	5	,	,	PUNCT
ajst-19203	156	6	ge	ge	PROPN
ajst-19203	156	7	,	,	PUNCT
ajst-19203	156	8	et	et	PROPN
ajst-19203	156	9	al	al	PROPN
ajst-19203	156	10	.	.	PUNCT
ajst-19203	157	1	"	"	PUNCT
ajst-19203	157	2	energy	energy	NOUN
ajst-19203	157	3	consumption	consumption	NOUN
ajst-19203	157	4	auditing	auditing	NOUN
ajst-19203	157	5	based	base	VERB
ajst-19203	157	6	on	on	ADP
ajst-19203	157	7	a	a	DET
ajst-19203	157	8	generative	generative	ADJ
ajst-19203	157	9	adversarial	adversarial	ADJ
ajst-19203	157	10	network	network	NOUN
ajst-19203	157	11	for	for	ADP
ajst-19203	157	12	anomaly	anomaly	NOUN
ajst-19203	157	13	detection	detection	NOUN
ajst-19203	157	14	of	of	ADP
ajst-19203	157	15	robotic	robotic	ADJ
ajst-19203	157	16	manipulators	manipulator	NOUN
ajst-19203	157	17	.	.	PUNCT
ajst-19203	157	18	"	"	PUNCT
ajst-19203	158	1	future	future	ADJ
ajst-19203	158	2	generation	generation	NOUN
ajst-19203	158	3	computer	computer	NOUN
ajst-19203	158	4	systems	system	NOUN
ajst-19203	158	5	149	149	NUM
ajst-19203	158	6	(	(	PUNCT
ajst-19203	158	7	2023	2023	NUM
ajst-19203	158	8	):	):	PUNCT
ajst-19203	158	9	376	376	NUM
ajst-19203	158	10	-	-	SYM
ajst-19203	158	11	389	389	NUM
ajst-19203	158	12	.	.	PUNCT
ajst-19203	159	1	[	[	X
ajst-19203	159	2	24	24	NUM
ajst-19203	159	3	]	]	SYM
ajst-19203	159	4	ou	ou	PROPN
ajst-19203	159	5	,	,	PUNCT
ajst-19203	159	6	junlin	junlin	PROPN
ajst-19203	159	7	,	,	PUNCT
ajst-19203	159	8	et	et	PROPN
ajst-19203	159	9	al	al	PROPN
ajst-19203	159	10	.	.	PUNCT
ajst-19203	160	1	"	"	PUNCT
ajst-19203	160	2	hybrid	hybrid	ADJ
ajst-19203	160	3	path	path	NOUN
ajst-19203	160	4	planning	planning	NOUN
ajst-19203	160	5	based	base	VERB
ajst-19203	160	6	on	on	ADP
ajst-19203	160	7	adaptive	adaptive	ADJ
ajst-19203	160	8	visibility	visibility	NOUN
ajst-19203	160	9	graph	graph	NOUN
ajst-19203	160	10	initialization	initialization	NOUN
ajst-19203	160	11	and	and	CCONJ
ajst-19203	160	12	edge	edge	NOUN
ajst-19203	160	13	computing	computing	NOUN
ajst-19203	160	14	for	for	ADP
ajst-19203	160	15	mobile	mobile	NOUN
ajst-19203	160	16	robots	robot	NOUN
ajst-19203	160	17	.	.	PUNCT
ajst-19203	160	18	"	"	PUNCT
ajst-19203	161	1	engineering	engineering	NOUN
ajst-19203	161	2	applications	application	NOUN
ajst-19203	161	3	of	of	ADP
ajst-19203	161	4	artificial	artificial	ADJ
ajst-19203	161	5	intelligence	intelligence	NOUN
ajst-19203	161	6	126	126	NUM
ajst-19203	161	7	(	(	PUNCT
ajst-19203	161	8	2023	2023	NUM
ajst-19203	161	9	):	):	PUNCT
ajst-19203	161	10	107110	107110	NUM
ajst-19203	161	11	.	.	PUNCT
ajst-19203	162	1	[	[	X
ajst-19203	162	2	25	25	NUM
ajst-19203	162	3	]	]	X
ajst-19203	162	4	zhou	zhou	PROPN
ajst-19203	162	5	,	,	PUNCT
ajst-19203	162	6	yucheng	yucheng	PROPN
ajst-19203	162	7	,	,	PUNCT
ajst-19203	162	8	et	et	PROPN
ajst-19203	162	9	al	al	PROPN
ajst-19203	162	10	.	.	PUNCT
ajst-19203	163	1	"	"	PUNCT
ajst-19203	163	2	visual	visual	ADJ
ajst-19203	163	3	in	in	ADP
ajst-19203	163	4	-	-	PUNCT
ajst-19203	163	5	context	context	NOUN
ajst-19203	163	6	learning	learning	NOUN
ajst-19203	163	7	for	for	ADP
ajst-19203	163	8	large	large	ADJ
ajst-19203	163	9	vision	vision	NOUN
ajst-19203	163	10	-	-	PUNCT
ajst-19203	163	11	language	language	NOUN
ajst-19203	163	12	models	model	NOUN
ajst-19203	163	13	.	.	PUNCT
ajst-19203	163	14	"	"	PUNCT
ajst-19203	164	1	arxiv	arxiv	PROPN
ajst-19203	164	2	preprint	preprint	NOUN
ajst-19203	164	3	arxiv	arxiv	PROPN
ajst-19203	164	4	:	:	PUNCT
ajst-19203	164	5	2402.11574	2402.11574	NUM
ajst-19203	164	6	(	(	PUNCT
ajst-19203	164	7	2024	2024	NUM
ajst-19203	164	8	)	)	PUNCT
ajst-19203	164	9	.	.	PUNCT
ajst-19203	165	1	[	[	X
ajst-19203	165	2	26	26	NUM
ajst-19203	165	3	]	]	X
ajst-19203	165	4	zhou	zhou	PROPN
ajst-19203	165	5	,	,	PUNCT
ajst-19203	165	6	yucheng	yucheng	PROPN
ajst-19203	165	7	,	,	PUNCT
ajst-19203	165	8	et	et	PROPN
ajst-19203	165	9	al	al	PROPN
ajst-19203	165	10	.	.	PUNCT
ajst-19203	165	11	"	"	PUNCT
ajst-19203	165	12	thread	thread	NOUN
ajst-19203	165	13	of	of	ADP
ajst-19203	165	14	thought	thought	NOUN
ajst-19203	165	15	unraveling	unravel	VERB
ajst-19203	165	16	chaotic	chaotic	ADJ
ajst-19203	165	17	contexts	contexts	NOUN
ajst-19203	165	18	.	.	PUNCT
ajst-19203	165	19	"	"	PUNCT
ajst-19203	166	1	arxiv	arxiv	PROPN
ajst-19203	166	2	preprint	preprint	VERB
ajst-19203	166	3	arxiv:2311.08734	arxiv:2311.08734	NUM
ajst-19203	166	4	(	(	PUNCT
ajst-19203	166	5	2023	2023	NUM
ajst-19203	166	6	)	)	PUNCT
ajst-19203	166	7	.	.	PUNCT
ajst-19203	167	1	[	[	X
ajst-19203	167	2	27	27	NUM
ajst-19203	167	3	]	]	X
ajst-19203	167	4	zhou	zhou	PROPN
ajst-19203	167	5	,	,	PUNCT
ajst-19203	167	6	yucheng	yucheng	PROPN
ajst-19203	167	7	,	,	PUNCT
ajst-19203	167	8	et	et	PROPN
ajst-19203	167	9	al	al	PROPN
ajst-19203	167	10	.	.	PUNCT
ajst-19203	168	1	"	"	PUNCT
ajst-19203	168	2	claret	claret	NOUN
ajst-19203	168	3	:	:	PUNCT
ajst-19203	168	4	pre	pre	ADJ
ajst-19203	168	5	-	-	VERB
ajst-19203	168	6	training	train	VERB
ajst-19203	168	7	a	a	DET
ajst-19203	168	8	correlation	correlation	NOUN
ajst-19203	168	9	-	-	PUNCT
ajst-19203	168	10	aware	aware	ADJ
ajst-19203	168	11	context	context	NOUN
ajst-19203	168	12	-	-	PUNCT
ajst-19203	168	13	to	to	ADP
ajst-19203	168	14	-	-	PUNCT
ajst-19203	168	15	event	event	NOUN
ajst-19203	168	16	transformer	transformer	NOUN
ajst-19203	168	17	for	for	ADP
ajst-19203	168	18	event	event	NOUN
ajst-19203	168	19	-	-	PUNCT
ajst-19203	168	20	centric	centric	ADJ
ajst-19203	168	21	generation	generation	NOUN
ajst-19203	168	22	and	and	CCONJ
ajst-19203	168	23	classification	classification	NOUN
ajst-19203	168	24	.	.	PUNCT
ajst-19203	168	25	"	"	PUNCT
ajst-19203	169	1	arxiv	arxiv	PROPN
ajst-19203	169	2	preprint	preprint	NOUN
ajst-19203	169	3	arxiv:2203.02225	arxiv:2203.02225	PROPN
ajst-19203	169	4	(	(	PUNCT
ajst-19203	169	5	2022	2022	NUM
ajst-19203	169	6	)	)	PUNCT
ajst-19203	169	7	.	.	PUNCT
ajst-19203	170	1	[	[	X
ajst-19203	170	2	28	28	NUM
ajst-19203	170	3	]	]	X
ajst-19203	170	4	tian	tian	PROPN
ajst-19203	170	5	,	,	PUNCT
ajst-19203	170	6	jiwei	jiwei	PROPN
ajst-19203	170	7	,	,	PUNCT
ajst-19203	170	8	et	et	PROPN
ajst-19203	170	9	al	al	PROPN
ajst-19203	170	10	.	.	PUNCT
ajst-19203	171	1	"	"	PUNCT
ajst-19203	171	2	lesson	lesson	NOUN
ajst-19203	171	3	:	:	PUNCT
ajst-19203	171	4	multi	multi	ADJ
ajst-19203	171	5	-	-	ADJ
ajst-19203	171	6	label	label	ADJ
ajst-19203	171	7	adversarial	adversarial	ADJ
ajst-19203	171	8	false	false	ADJ
ajst-19203	171	9	data	datum	NOUN
ajst-19203	171	10	injection	injection	NOUN
ajst-19203	171	11	attack	attack	NOUN
ajst-19203	171	12	for	for	ADP
ajst-19203	171	13	deep	deep	ADJ
ajst-19203	171	14	learning	learn	VERB
ajst-19203	171	15	locational	locational	ADJ
ajst-19203	171	16	detection	detection	NOUN
ajst-19203	171	17	.	.	PUNCT
ajst-19203	171	18	"	"	PUNCT
ajst-19203	172	1	ieee	ieee	NOUN
ajst-19203	172	2	transactions	transaction	NOUN
ajst-19203	172	3	on	on	ADP
ajst-19203	172	4	dependable	dependable	ADJ
ajst-19203	172	5	and	and	CCONJ
ajst-19203	172	6	secure	secure	ADJ
ajst-19203	172	7	computing	computing	NOUN
ajst-19203	172	8	(	(	PUNCT
ajst-19203	172	9	2024	2024	NUM
ajst-19203	172	10	)	)	PUNCT
ajst-19203	172	11	.	.	PUNCT
ajst-19203	173	1	[	[	X
ajst-19203	173	2	29	29	NUM
ajst-19203	173	3	]	]	X
ajst-19203	173	4	tian	tian	ADJ
ajst-19203	173	5	,	,	PUNCT
ajst-19203	173	6	jiwei	jiwei	PROPN
ajst-19203	173	7	,	,	PUNCT
ajst-19203	173	8	et	et	PROPN
ajst-19203	173	9	al	al	PROPN
ajst-19203	173	10	.	.	PUNCT
ajst-19203	174	1	"	"	PUNCT
ajst-19203	174	2	adversarial	adversarial	ADJ
ajst-19203	174	3	attacks	attack	NOUN
ajst-19203	174	4	and	and	CCONJ
ajst-19203	174	5	defenses	defense	NOUN
ajst-19203	174	6	for	for	ADP
ajst-19203	174	7	deeplearning	deeplearning	NOUN
ajst-19203	174	8	-	-	PUNCT
ajst-19203	174	9	based	base	VERB
ajst-19203	174	10	unmanned	unmanned	ADJ
ajst-19203	174	11	aerial	aerial	ADJ
ajst-19203	174	12	vehicles	vehicle	NOUN
ajst-19203	174	13	.	.	PUNCT
ajst-19203	174	14	"	"	PUNCT
ajst-19203	175	1	ieee	ieee	NOUN
ajst-19203	175	2	internet	internet	NOUN
ajst-19203	175	3	of	of	ADP
ajst-19203	175	4	things	thing	NOUN
ajst-19203	175	5	journal	journal	NOUN
ajst-19203	175	6	9.22	9.22	NUM
ajst-19203	175	7	(	(	PUNCT
ajst-19203	175	8	2021	2021	NUM
ajst-19203	175	9	):	):	PUNCT
ajst-19203	175	10	22399	22399	NUM
ajst-19203	175	11	-	-	SYM
ajst-19203	175	12	22409	22409	NUM
ajst-19203	175	13	.	.	PUNCT
ajst-19203	176	1	[	[	X
ajst-19203	176	2	30	30	NUM
ajst-19203	176	3	]	]	X
ajst-19203	176	4	wu	wu	PROPN
ajst-19203	176	5	,	,	PUNCT
ajst-19203	176	6	jing	jing	PROPN
ajst-19203	176	7	,	,	PUNCT
ajst-19203	176	8	et	et	PROPN
ajst-19203	176	9	al	al	PROPN
ajst-19203	176	10	.	.	PUNCT
ajst-19203	176	11	"	"	PUNCT
ajst-19203	176	12	switchtab	switchtab	NOUN
ajst-19203	176	13	:	:	PUNCT
ajst-19203	176	14	switched	switch	VERB
ajst-19203	176	15	autoencoders	autoencoder	NOUN
ajst-19203	176	16	are	be	AUX
ajst-19203	176	17	effective	effective	ADJ
ajst-19203	176	18	tabular	tabular	NOUN
ajst-19203	176	19	learners	learner	NOUN
ajst-19203	176	20	.	.	PUNCT
ajst-19203	176	21	"	"	PUNCT
ajst-19203	177	1	arxiv	arxiv	PROPN
ajst-19203	177	2	preprint	preprint	VERB
ajst-19203	177	3	arxiv:2401.02013	arxiv:2401.02013	NUM
ajst-19203	177	4	(	(	PUNCT
ajst-19203	177	5	2024	2024	NUM
ajst-19203	177	6	)	)	PUNCT
ajst-19203	177	7	.	.	PUNCT
ajst-19203	178	1	[	[	X
ajst-19203	178	2	31	31	NUM
ajst-19203	178	3	]	]	X
ajst-19203	178	4	chen	chen	PROPN
ajst-19203	178	5	,	,	PUNCT
ajst-19203	178	6	suiyao	suiyao	PROPN
ajst-19203	178	7	,	,	PUNCT
ajst-19203	178	8	et	et	PROPN
ajst-19203	178	9	al	al	PROPN
ajst-19203	178	10	.	.	PUNCT
ajst-19203	178	11	"	"	PUNCT
ajst-19203	178	12	recontab	recontab	NOUN
ajst-19203	178	13	:	:	PUNCT
ajst-19203	178	14	regularized	regularize	VERB
ajst-19203	178	15	contrastive	contrastive	ADJ
ajst-19203	178	16	representation	representation	NOUN
ajst-19203	178	17	learning	learn	VERB
ajst-19203	178	18	for	for	ADP
ajst-19203	178	19	tabular	tabular	PROPN
ajst-19203	178	20	data	datum	NOUN
ajst-19203	178	21	.	.	PUNCT
ajst-19203	178	22	"	"	PUNCT
ajst-19203	179	1	arxiv	arxiv	PROPN
ajst-19203	179	2	preprint	preprint	NOUN
ajst-19203	179	3	arxiv:2310.18541	arxiv:2310.18541	NOUN
ajst-19203	179	4	(	(	PUNCT
ajst-19203	179	5	2023	2023	NUM
ajst-19203	179	6	)	)	PUNCT
ajst-19203	179	7	.	.	PUNCT
ajst-19203	180	1	[	[	X
ajst-19203	180	2	32	32	NUM
ajst-19203	180	3	]	]	X
ajst-19203	180	4	chen	chen	PROPN
ajst-19203	180	5	,	,	PUNCT
ajst-19203	180	6	suiyao	suiyao	PROPN
ajst-19203	180	7	,	,	PUNCT
ajst-19203	180	8	et	et	PROPN
ajst-19203	180	9	al	al	PROPN
ajst-19203	180	10	.	.	PUNCT
ajst-19203	180	11	"	"	PUNCT
ajst-19203	180	12	claims	claim	VERB
ajst-19203	180	13	data	data	NOUN
ajst-19203	180	14	-	-	PUNCT
ajst-19203	180	15	driven	drive	VERB
ajst-19203	180	16	modeling	modeling	NOUN
ajst-19203	180	17	of	of	ADP
ajst-19203	180	18	hospital	hospital	NOUN
ajst-19203	180	19	time	time	NOUN
ajst-19203	180	20	-	-	PUNCT
ajst-19203	180	21	to	to	ADP
ajst-19203	180	22	-	-	PUNCT
ajst-19203	180	23	readmission	readmission	NOUN
ajst-19203	180	24	risk	risk	NOUN
ajst-19203	180	25	with	with	ADP
ajst-19203	180	26	latent	latent	NOUN
ajst-19203	180	27	heterogeneity	heterogeneity	NOUN
ajst-19203	180	28	.	.	PUNCT
ajst-19203	180	29	"	"	PUNCT
ajst-19203	181	1	health	health	NOUN
ajst-19203	181	2	care	care	NOUN
ajst-19203	181	3	management	management	PROPN
ajst-19203	181	4	science	science	NOUN
ajst-19203	181	5	22	22	NUM
ajst-19203	181	6	(	(	PUNCT
ajst-19203	181	7	2019	2019	NUM
ajst-19203	181	8	):	):	PUNCT
ajst-19203	181	9	156	156	NUM
ajst-19203	181	10	-	-	SYM
ajst-19203	181	11	179	179	NUM
ajst-19203	181	12	.	.	PUNCT
ajst-19203	182	1	[	[	X
ajst-19203	182	2	33	33	NUM
ajst-19203	182	3	]	]	X
ajst-19203	182	4	chen	chen	PROPN
ajst-19203	182	5	,	,	PUNCT
ajst-19203	182	6	suiyao	suiyao	PROPN
ajst-19203	182	7	,	,	PUNCT
ajst-19203	182	8	et	et	PROPN
ajst-19203	182	9	al	al	PROPN
ajst-19203	182	10	.	.	PUNCT
ajst-19203	183	1	"	"	PUNCT
ajst-19203	183	2	a	a	DET
ajst-19203	183	3	data	data	NOUN
ajst-19203	183	4	heterogeneity	heterogeneity	NOUN
ajst-19203	183	5	modeling	modeling	NOUN
ajst-19203	183	6	and	and	CCONJ
ajst-19203	183	7	quantification	quantification	NOUN
ajst-19203	183	8	approach	approach	NOUN
ajst-19203	183	9	for	for	ADP
ajst-19203	183	10	field	field	NOUN
ajst-19203	183	11	pre	pre	NOUN
ajst-19203	183	12	-	-	NOUN
ajst-19203	183	13	assessment	assessment	NOUN
ajst-19203	183	14	of	of	ADP
ajst-19203	183	15	chlorideinduced	chlorideinduced	ADJ
ajst-19203	183	16	corrosion	corrosion	NOUN
ajst-19203	183	17	in	in	ADP
ajst-19203	183	18	aging	age	VERB
ajst-19203	183	19	infrastructures	infrastructure	NOUN
ajst-19203	183	20	.	.	PUNCT
ajst-19203	183	21	"	"	PUNCT
ajst-19203	184	1	reliability	reliability	NOUN
ajst-19203	184	2	engineering	engineering	NOUN
ajst-19203	184	3	&	&	CCONJ
ajst-19203	184	4	system	system	NOUN
ajst-19203	184	5	safety	safety	NOUN
ajst-19203	184	6	171	171	NUM
ajst-19203	184	7	(	(	PUNCT
ajst-19203	184	8	2018	2018	NUM
ajst-19203	184	9	):	):	PUNCT
ajst-19203	184	10	123	123	NUM
ajst-19203	184	11	-	-	SYM
ajst-19203	184	12	135	135	NUM
ajst-19203	184	13	.	.	PUNCT
ajst-19203	185	1	[	[	X
ajst-19203	185	2	34	34	NUM
ajst-19203	185	3	]	]	X
ajst-19203	185	4	hsieh	hsieh	PROPN
ajst-19203	185	5	,	,	PUNCT
ajst-19203	185	6	yung	yung	NOUN
ajst-19203	185	7	-	-	PUNCT
ajst-19203	185	8	ting	ting	PROPN
ajst-19203	185	9	,	,	PUNCT
ajst-19203	185	10	khizar	khizar	NOUN
ajst-19203	185	11	anjum	anjum	PROPN
ajst-19203	185	12	,	,	PUNCT
ajst-19203	185	13	and	and	CCONJ
ajst-19203	185	14	dario	dario	NOUN
ajst-19203	185	15	pompili	pompili	NOUN
ajst-19203	185	16	.	.	PUNCT
ajst-19203	186	1	"	"	PUNCT
ajst-19203	186	2	ultralow	ultralow	ADJ
ajst-19203	186	3	power	power	NOUN
ajst-19203	186	4	analog	analog	NOUN
ajst-19203	186	5	recurrent	recurrent	ADJ
ajst-19203	186	6	neural	neural	ADJ
ajst-19203	186	7	network	network	NOUN
ajst-19203	186	8	design	design	NOUN
ajst-19203	186	9	approximation	approximation	NOUN
ajst-19203	186	10	for	for	ADP
ajst-19203	186	11	wireless	wireless	ADJ
ajst-19203	186	12	health	health	NOUN
ajst-19203	186	13	monitoring	monitoring	NOUN
ajst-19203	186	14	.	.	PUNCT
ajst-19203	186	15	"	"	PUNCT
ajst-19203	187	1	2022	2022	NUM
ajst-19203	187	2	ieee	ieee	NOUN
ajst-19203	187	3	19th	19th	ADJ
ajst-19203	187	4	international	international	ADJ
ajst-19203	187	5	conference	conference	NOUN
ajst-19203	187	6	on	on	ADP
ajst-19203	187	7	mobile	mobile	ADJ
ajst-19203	187	8	ad	ad	X
ajst-19203	187	9	hoc	hoc	X
ajst-19203	187	10	and	and	CCONJ
ajst-19203	187	11	smart	smart	ADJ
ajst-19203	187	12	systems	system	NOUN
ajst-19203	187	13	(	(	PUNCT
ajst-19203	187	14	mass	mass	PROPN
ajst-19203	187	15	)	)	PUNCT
ajst-19203	187	16	.	.	PUNCT
ajst-19203	188	1	ieee	ieee	NOUN
ajst-19203	188	2	,	,	PUNCT
ajst-19203	188	3	2022	2022	NUM
ajst-19203	188	4	.	.	PUNCT
ajst-19203	189	1	[	[	X
ajst-19203	189	2	35	35	NUM
ajst-19203	189	3	]	]	X
ajst-19203	189	4	hsieh	hsieh	PROPN
ajst-19203	189	5	,	,	PUNCT
ajst-19203	189	6	yung	yung	NOUN
ajst-19203	189	7	-	-	PUNCT
ajst-19203	189	8	ting	ting	PROPN
ajst-19203	189	9	,	,	PUNCT
ajst-19203	189	10	zhuoran	zhuoran	NOUN
ajst-19203	189	11	qi	qi	PROPN
ajst-19203	189	12	,	,	PUNCT
ajst-19203	189	13	and	and	CCONJ
ajst-19203	189	14	dario	dario	NOUN
ajst-19203	189	15	pompili	pompili	NOUN
ajst-19203	189	16	.	.	PUNCT
ajst-19203	190	1	"	"	PUNCT
ajst-19203	190	2	ml	ml	VERB
ajst-19203	190	3	-	-	PUNCT
ajst-19203	190	4	based	base	VERB
ajst-19203	190	5	joint	joint	ADJ
ajst-19203	190	6	doppler	doppler	NOUN
ajst-19203	190	7	estimation	estimation	NOUN
ajst-19203	190	8	and	and	CCONJ
ajst-19203	190	9	compensation	compensation	NOUN
ajst-19203	190	10	in	in	ADP
ajst-19203	190	11	underwater	underwater	ADJ
ajst-19203	190	12	acoustic	acoustic	ADJ
ajst-19203	190	13	communications	communication	NOUN
ajst-19203	190	14	.	.	PUNCT
ajst-19203	190	15	"	"	PUNCT
ajst-19203	191	1	proceedings	proceeding	NOUN
ajst-19203	191	2	of	of	ADP
ajst-19203	191	3	the	the	DET
ajst-19203	191	4	16th	16th	ADJ
ajst-19203	191	5	international	international	ADJ
ajst-19203	191	6	conference	conference	NOUN
ajst-19203	191	7	on	on	ADP
ajst-19203	191	8	underwater	underwater	ADJ
ajst-19203	191	9	networks	network	NOUN
ajst-19203	191	10	&	&	CCONJ
ajst-19203	191	11	systems	system	NOUN
ajst-19203	191	12	.	.	PUNCT
ajst-19203	192	1	2022	2022	NUM
ajst-19203	192	2	.	.	PUNCT
ajst-19203	193	1	[	[	X
ajst-19203	193	2	36	36	NUM
ajst-19203	193	3	]	]	X
ajst-19203	193	4	sun	sun	NOUN
ajst-19203	193	5	,	,	PUNCT
ajst-19203	193	6	chuanneng	chuanneng	PROPN
ajst-19203	193	7	,	,	PUNCT
ajst-19203	193	8	et	et	PROPN
ajst-19203	193	9	al	al	PROPN
ajst-19203	193	10	.	.	PUNCT
ajst-19203	194	1	"	"	PUNCT
ajst-19203	194	2	fed2kd	fed2kd	NOUN
ajst-19203	194	3	:	:	PUNCT
ajst-19203	194	4	heterogeneous	heterogeneous	ADJ
ajst-19203	194	5	federated	federated	ADJ
ajst-19203	194	6	learning	learning	NOUN
ajst-19203	194	7	for	for	ADP
ajst-19203	194	8	pandemic	pandemic	ADJ
ajst-19203	194	9	risk	risk	NOUN
ajst-19203	194	10	assessment	assessment	NOUN
ajst-19203	194	11	via	via	ADP
ajst-19203	194	12	two	two	NUM
ajst-19203	194	13	-	-	PUNCT
ajst-19203	194	14	way	way	NOUN
ajst-19203	194	15	knowledge	knowledge	NOUN
ajst-19203	194	16	distillation	distillation	NOUN
ajst-19203	194	17	.	.	PUNCT
ajst-19203	194	18	"	"	PUNCT
ajst-19203	195	1	2022	2022	NUM
ajst-19203	195	2	17th	17th	NOUN
ajst-19203	195	3	wireless	wireless	ADJ
ajst-19203	195	4	on	on	ADP
ajst-19203	195	5	-	-	PUNCT
ajst-19203	195	6	demand	demand	NOUN
ajst-19203	195	7	network	network	NOUN
ajst-19203	195	8	systems	system	NOUN
ajst-19203	195	9	and	and	CCONJ
ajst-19203	195	10	services	service	NOUN
ajst-19203	195	11	conference	conference	NOUN
ajst-19203	195	12	(	(	PUNCT
ajst-19203	195	13	wons	wons	PROPN
ajst-19203	195	14	)	)	PUNCT
ajst-19203	195	15	.	.	PUNCT
ajst-19203	196	1	ieee	ieee	PROPN
ajst-19203	196	2	,	,	PUNCT
ajst-19203	196	3	2022	2022	NUM
ajst-19203	196	4	.	.	PUNCT
ajst-19203	197	1	[	[	X
ajst-19203	197	2	37	37	NUM
ajst-19203	197	3	]	]	X
ajst-19203	197	4	sun	sun	NOUN
ajst-19203	197	5	,	,	PUNCT
ajst-19203	197	6	chuanneng	chuanneng	PROPN
ajst-19203	197	7	,	,	PUNCT
ajst-19203	197	8	songjun	songjun	PROPN
ajst-19203	197	9	huang	huang	PROPN
ajst-19203	197	10	,	,	PUNCT
ajst-19203	197	11	and	and	CCONJ
ajst-19203	197	12	dario	dario	NOUN
ajst-19203	197	13	pompili	pompili	NOUN
ajst-19203	197	14	.	.	PUNCT
ajst-19203	198	1	"	"	PUNCT
ajst-19203	198	2	hmaac	hmaac	NOUN
ajst-19203	198	3	:	:	PUNCT
ajst-19203	198	4	hierarchical	hierarchical	ADJ
ajst-19203	198	5	multi	multi	ADJ
ajst-19203	198	6	-	-	ADJ
ajst-19203	198	7	agent	agent	ADJ
ajst-19203	198	8	actor	actor	NOUN
ajst-19203	198	9	-	-	PUNCT
ajst-19203	198	10	critic	critic	NOUN
ajst-19203	198	11	for	for	ADP
ajst-19203	198	12	aerial	aerial	ADJ
ajst-19203	198	13	search	search	NOUN
ajst-19203	198	14	with	with	ADP
ajst-19203	198	15	explicit	explicit	ADJ
ajst-19203	198	16	coordination	coordination	NOUN
ajst-19203	198	17	modeling	modeling	NOUN
ajst-19203	198	18	.	.	PUNCT
ajst-19203	198	19	"	"	PUNCT
ajst-19203	199	1	2023	2023	NUM
ajst-19203	199	2	ieee	ieee	NOUN
ajst-19203	199	3	international	international	ADJ
ajst-19203	199	4	conference	conference	NOUN
ajst-19203	199	5	on	on	ADP
ajst-19203	199	6	robotics	robotic	NOUN
ajst-19203	199	7	and	and	CCONJ
ajst-19203	199	8	automation	automation	NOUN
ajst-19203	199	9	(	(	PUNCT
ajst-19203	199	10	icra	icra	PROPN
ajst-19203	199	11	)	)	PUNCT
ajst-19203	199	12	.	.	PUNCT
ajst-19203	200	1	ieee	ieee	PROPN
ajst-19203	200	2	,	,	PUNCT
ajst-19203	200	3	2023	2023	NUM
ajst-19203	200	4	.	.	PUNCT
ajst-19203	201	1	[	[	X
ajst-19203	201	2	38	38	NUM
ajst-19203	201	3	]	]	X
ajst-19203	201	4	sun	sun	NOUN
ajst-19203	201	5	,	,	PUNCT
ajst-19203	201	6	chuanneng	chuanneng	PROPN
ajst-19203	201	7	,	,	PUNCT
ajst-19203	201	8	et	et	PROPN
ajst-19203	201	9	al	al	PROPN
ajst-19203	201	10	.	.	PUNCT
ajst-19203	202	1	"	"	PUNCT
ajst-19203	202	2	contextual	contextual	ADJ
ajst-19203	202	3	biasing	biasing	NOUN
ajst-19203	202	4	of	of	ADP
ajst-19203	202	5	named	name	VERB
ajst-19203	202	6	-	-	PUNCT
ajst-19203	202	7	entities	entity	NOUN
ajst-19203	202	8	with	with	ADP
ajst-19203	202	9	large	large	ADJ
ajst-19203	202	10	language	language	NOUN
ajst-19203	202	11	models	model	NOUN
ajst-19203	202	12	.	.	PUNCT
ajst-19203	202	13	"	"	PUNCT
ajst-19203	203	1	arxiv	arxiv	PROPN
ajst-19203	203	2	preprint	preprint	NOUN
ajst-19203	203	3	arxiv	arxiv	PROPN
ajst-19203	203	4	:	:	PUNCT
ajst-19203	203	5	2309.00723	2309.00723	NUM
ajst-19203	203	6	(	(	PUNCT
ajst-19203	203	7	2023	2023	NUM
ajst-19203	203	8	)	)	PUNCT
ajst-19203	203	9	.	.	PUNCT
ajst-19203	204	1	[	[	X
ajst-19203	204	2	39	39	NUM
ajst-19203	204	3	]	]	X
ajst-19203	204	4	younis	younis	PROPN
ajst-19203	204	5	,	,	PUNCT
ajst-19203	204	6	ayman	ayman	PROPN
ajst-19203	204	7	,	,	PUNCT
ajst-19203	204	8	chuanneng	chuanneng	PROPN
ajst-19203	204	9	sun	sun	PROPN
ajst-19203	204	10	,	,	PUNCT
ajst-19203	204	11	and	and	CCONJ
ajst-19203	204	12	dario	dario	NOUN
ajst-19203	204	13	pompili	pompili	NOUN
ajst-19203	204	14	.	.	PUNCT
ajst-19203	205	1	"	"	PUNCT
ajst-19203	205	2	communication	communication	NOUN
ajst-19203	205	3	-	-	PUNCT
ajst-19203	205	4	efficient	efficient	ADJ
ajst-19203	205	5	federated	federated	ADJ
ajst-19203	205	6	learning	learning	NOUN
ajst-19203	205	7	design	design	NOUN
ajst-19203	205	8	with	with	ADP
ajst-19203	205	9	fronthaul	fronthaul	NOUN
ajst-19203	205	10	awareness	awareness	NOUN
ajst-19203	205	11	in	in	ADP
ajst-19203	205	12	ng	ng	PROPN
ajst-19203	205	13	-	-	PUNCT
ajst-19203	205	14	rans	ran	NOUN
ajst-19203	205	15	.	.	PUNCT
ajst-19203	205	16	"	"	PUNCT
ajst-19203	206	1	2022	2022	NUM
ajst-19203	206	2	ieee	ieee	NOUN
ajst-19203	206	3	19th	19th	ADJ
ajst-19203	206	4	international	international	ADJ
ajst-19203	206	5	conference	conference	NOUN
ajst-19203	206	6	on	on	ADP
ajst-19203	206	7	mobile	mobile	ADJ
ajst-19203	206	8	ad	ad	X
ajst-19203	206	9	hoc	hoc	X
ajst-19203	206	10	and	and	CCONJ
ajst-19203	206	11	smart	smart	ADJ
ajst-19203	206	12	systems	system	NOUN
ajst-19203	206	13	(	(	PUNCT
ajst-19203	206	14	mass	mass	PROPN
ajst-19203	206	15	)	)	PUNCT
ajst-19203	206	16	.	.	PUNCT
ajst-19203	207	1	ieee	ieee	NOUN
ajst-19203	207	2	,	,	PUNCT
ajst-19203	207	3	2022	2022	NUM
ajst-19203	207	4	.	.	PUNCT
ajst-19203	208	1	[	[	X
ajst-19203	208	2	40	40	NUM
ajst-19203	208	3	]	]	PUNCT
ajst-19203	208	4	flynn	flynn	PROPN
ajst-19203	208	5	,	,	PUNCT
ajst-19203	208	6	patrick	patrick	PROPN
ajst-19203	208	7	,	,	PUNCT
ajst-19203	208	8	xinyao	xinyao	NOUN
ajst-19203	208	9	yi	yi	PROPN
ajst-19203	208	10	,	,	PUNCT
ajst-19203	208	11	and	and	CCONJ
ajst-19203	208	12	yonghong	yonghong	PROPN
ajst-19203	208	13	yan	yan	PROPN
ajst-19203	208	14	.	.	PUNCT
ajst-19203	209	1	"	"	PUNCT
ajst-19203	209	2	exploring	explore	VERB
ajst-19203	209	3	source	source	NOUN
ajst-19203	209	4	-	-	PUNCT
ajst-19203	209	5	to	to	ADP
ajst-19203	209	6	-	-	PUNCT
ajst-19203	209	7	source	source	NOUN
ajst-19203	209	8	compiler	compiler	NOUN
ajst-19203	209	9	transformation	transformation	NOUN
ajst-19203	209	10	of	of	ADP
ajst-19203	209	11	openmp	openmp	PROPN
ajst-19203	209	12	simd	simd	PROPN
ajst-19203	209	13	constructs	construct	NOUN
ajst-19203	209	14	for	for	ADP
ajst-19203	209	15	intel	intel	PROPN
ajst-19203	209	16	avx	avx	PROPN
ajst-19203	209	17	and	and	CCONJ
ajst-19203	209	18	arm	arm	NOUN
ajst-19203	209	19	sve	sve	NOUN
ajst-19203	209	20	vector	vector	NOUN
ajst-19203	209	21	architectures	architecture	NOUN
ajst-19203	209	22	.	.	PUNCT
ajst-19203	209	23	"	"	PUNCT
ajst-19203	209	24	proceedings	proceeding	NOUN
ajst-19203	209	25	of	of	ADP
ajst-19203	209	26	the	the	DET
ajst-19203	209	27	thirteenth	thirteenth	ADJ
ajst-19203	209	28	international	international	ADJ
ajst-19203	209	29	workshop	workshop	NOUN
ajst-19203	209	30	on	on	ADP
ajst-19203	209	31	programming	programming	NOUN
ajst-19203	209	32	models	model	NOUN
ajst-19203	209	33	and	and	CCONJ
ajst-19203	209	34	applications	application	NOUN
ajst-19203	209	35	for	for	ADP
ajst-19203	209	36	multicores	multicore	NOUN
ajst-19203	209	37	and	and	CCONJ
ajst-19203	209	38	manycores	manycore	NOUN
ajst-19203	209	39	.	.	PUNCT
ajst-19203	210	1	2022	2022	NUM
ajst-19203	210	2	.	.	PUNCT
ajst-19203	211	1	[	[	X
ajst-19203	211	2	41	41	NUM
ajst-19203	211	3	]	]	X
ajst-19203	211	4	yi	yi	NOUN
ajst-19203	211	5	,	,	PUNCT
ajst-19203	211	6	xinyao	xinyao	PROPN
ajst-19203	211	7	,	,	PUNCT
ajst-19203	211	8	et	et	PROPN
ajst-19203	211	9	al	al	PROPN
ajst-19203	211	10	.	.	PUNCT
ajst-19203	212	1	"	"	PUNCT
ajst-19203	212	2	cudamicrobench	cudamicrobench	NOUN
ajst-19203	212	3	:	:	PUNCT
ajst-19203	212	4	microbenchmarks	microbenchmark	NOUN
ajst-19203	212	5	to	to	PART
ajst-19203	212	6	assist	assist	VERB
ajst-19203	212	7	cuda	cuda	NOUN
ajst-19203	212	8	performance	performance	NOUN
ajst-19203	212	9	programming	programming	NOUN
ajst-19203	212	10	.	.	PUNCT
ajst-19203	212	11	"	"	PUNCT
ajst-19203	213	1	2021	2021	NUM
ajst-19203	213	2	ieee	ieee	PROPN
ajst-19203	213	3	international	international	PROPN
ajst-19203	213	4	parallel	parallel	ADJ
ajst-19203	213	5	and	and	CCONJ
ajst-19203	213	6	distributed	distributed	ADJ
ajst-19203	213	7	processing	processing	NOUN
ajst-19203	213	8	symposium	symposium	NOUN
ajst-19203	213	9	workshops	workshop	NOUN
ajst-19203	213	10	(	(	PUNCT
ajst-19203	213	11	ipdpsw	ipdpsw	NOUN
ajst-19203	213	12	)	)	PUNCT
ajst-19203	213	13	.	.	PUNCT
ajst-19203	214	1	ieee	ieee	NOUN
ajst-19203	214	2	,	,	PUNCT
ajst-19203	214	3	2021	2021	NUM
ajst-19203	214	4	.	.	PUNCT
ajst-19203	215	1	[	[	X
ajst-19203	215	2	42	42	NUM
ajst-19203	215	3	]	]	SYM
ajst-19203	215	4	li	li	PROPN
ajst-19203	215	5	,	,	PUNCT
ajst-19203	215	6	xiaying	xiaying	PROPN
ajst-19203	215	7	,	,	PUNCT
ajst-19203	215	8	belle	belle	PROPN
ajst-19203	215	9	li	li	PROPN
ajst-19203	215	10	,	,	PUNCT
ajst-19203	215	11	and	and	CCONJ
ajst-19203	215	12	su	su	PROPN
ajst-19203	215	13	-	-	PUNCT
ajst-19203	215	14	je	je	PROPN
ajst-19203	215	15	cho	cho	PROPN
ajst-19203	215	16	.	.	PUNCT
ajst-19203	216	1	"	"	PUNCT
ajst-19203	216	2	empowering	empower	VERB
ajst-19203	216	3	chinese	chinese	ADJ
ajst-19203	216	4	language	language	NOUN
ajst-19203	216	5	learners	learner	NOUN
ajst-19203	216	6	from	from	ADP
ajst-19203	216	7	low	low	ADJ
ajst-19203	216	8	-	-	PUNCT
ajst-19203	216	9	income	income	NOUN
ajst-19203	216	10	families	family	NOUN
ajst-19203	216	11	to	to	PART
ajst-19203	216	12	improve	improve	VERB
ajst-19203	216	13	their	their	PRON
ajst-19203	216	14	chinese	chinese	ADJ
ajst-19203	216	15	writing	writing	NOUN
ajst-19203	216	16	with	with	ADP
ajst-19203	216	17	chatgpt	chatgpt	NOUN
ajst-19203	216	18	’s	’s	PART
ajst-19203	216	19	assistance	assistance	NOUN
ajst-19203	216	20	afterschool	afterschool	NOUN
ajst-19203	216	21	.	.	PUNCT
ajst-19203	216	22	"	"	PUNCT
ajst-19203	216	23	languages	language	VERB
ajst-19203	216	24	8.4	8.4	NUM
ajst-19203	216	25	(	(	PUNCT
ajst-19203	216	26	2023	2023	NUM
ajst-19203	216	27	):	):	PUNCT
ajst-19203	216	28	238	238	NUM
ajst-19203	216	29	.	.	PUNCT
ajst-19203	217	1	[	[	X
ajst-19203	217	2	43	43	NUM
ajst-19203	217	3	]	]	X
ajst-19203	217	4	xie	xie	PROPN
ajst-19203	217	5	,	,	PUNCT
ajst-19203	217	6	ying	ying	PROPN
ajst-19203	217	7	,	,	PUNCT
ajst-19203	217	8	et	et	PROPN
ajst-19203	217	9	al	al	PROPN
ajst-19203	217	10	.	.	PUNCT
ajst-19203	218	1	"	"	PUNCT
ajst-19203	218	2	advancing	advance	VERB
ajst-19203	218	3	legal	legal	ADJ
ajst-19203	218	4	citation	citation	NOUN
ajst-19203	218	5	text	text	NOUN
ajst-19203	218	6	classification	classification	NOUN
ajst-19203	218	7	a	a	DET
ajst-19203	218	8	conv1d	conv1d	ADV
ajst-19203	218	9	-	-	PUNCT
ajst-19203	218	10	based	base	VERB
ajst-19203	218	11	approach	approach	NOUN
ajst-19203	218	12	for	for	ADP
ajst-19203	218	13	multi	multi	ADJ
ajst-19203	218	14	-	-	ADJ
ajst-19203	218	15	class	class	ADJ
ajst-19203	218	16	classification	classification	NOUN
ajst-19203	218	17	.	.	PUNCT
ajst-19203	218	18	"	"	PUNCT
ajst-19203	218	19	journal	journal	NOUN
ajst-19203	218	20	of	of	ADP
ajst-19203	218	21	theory	theory	NOUN
ajst-19203	218	22	and	and	CCONJ
ajst-19203	218	23	practice	practice	NOUN
ajst-19203	218	24	of	of	ADP
ajst-19203	218	25	engineering	engineer	VERB
ajst-19203	218	26	science	science	NOUN
ajst-19203	218	27	4.02	4.02	NUM
ajst-19203	218	28	(	(	PUNCT
ajst-19203	218	29	2024	2024	NUM
ajst-19203	218	30	):	):	PUNCT
ajst-19203	218	31	15	15	NUM
ajst-19203	218	32	-	-	SYM
ajst-19203	218	33	22	22	NUM
ajst-19203	218	34	.	.	PUNCT
ajst-19203	219	1	[	[	X
ajst-19203	219	2	44	44	NUM
ajst-19203	219	3	]	]	SYM
ajst-19203	219	4	luo	luo	PROPN
ajst-19203	219	5	,	,	PUNCT
ajst-19203	219	6	yang	yang	PROPN
ajst-19203	219	7	,	,	PUNCT
ajst-19203	219	8	et	et	PROPN
ajst-19203	219	9	al	al	PROPN
ajst-19203	219	10	.	.	PUNCT
ajst-19203	219	11	"	"	PUNCT
ajst-19203	219	12	enhancing	enhance	VERB
ajst-19203	219	13	e	e	NOUN
ajst-19203	219	14	-	-	NOUN
ajst-19203	219	15	commerce	commerce	NOUN
ajst-19203	219	16	chatbots	chatbot	NOUN
ajst-19203	219	17	with	with	ADP
ajst-19203	219	18	falcon-7b	falcon-7b	NOUN
ajst-19203	219	19	and	and	CCONJ
ajst-19203	219	20	16	16	NUM
ajst-19203	219	21	-	-	PUNCT
ajst-19203	219	22	bit	bit	NOUN
ajst-19203	219	23	full	full	ADJ
ajst-19203	219	24	quantization	quantization	NOUN
ajst-19203	219	25	.	.	PUNCT
ajst-19203	219	26	"	"	PUNCT
ajst-19203	220	1	journal	journal	NOUN
ajst-19203	220	2	of	of	ADP
ajst-19203	220	3	theory	theory	NOUN
ajst-19203	220	4	and	and	CCONJ
ajst-19203	220	5	practice	practice	NOUN
ajst-19203	220	6	of	of	ADP
ajst-19203	220	7	engineering	engineer	VERB
ajst-19203	220	8	science	science	NOUN
ajst-19203	220	9	4.02	4.02	NUM
ajst-19203	220	10	(	(	PUNCT
ajst-19203	220	11	2024	2024	NUM
ajst-19203	220	12	):	):	PUNCT
ajst-19203	220	13	52	52	NUM
ajst-19203	220	14	-	-	SYM
ajst-19203	220	15	57	57	NUM
ajst-19203	220	16	.	.	PUNCT
ajst-19203	221	1	[	[	X
ajst-19203	221	2	45	45	NUM
ajst-19203	221	3	]	]	PUNCT
ajst-19203	221	4	pan	pan	PROPN
ajst-19203	221	5	,	,	PUNCT
ajst-19203	221	6	zhenyu	zhenyu	PROPN
ajst-19203	221	7	,	,	PUNCT
ajst-19203	221	8	et	et	PROPN
ajst-19203	221	9	al	al	PROPN
ajst-19203	221	10	.	.	PUNCT
ajst-19203	221	11	"	"	PUNCT
ajst-19203	221	12	ising	ise	VERB
ajst-19203	221	13	-	-	PUNCT
ajst-19203	221	14	traffic	traffic	NOUN
ajst-19203	221	15	:	:	PUNCT
ajst-19203	221	16	using	use	VERB
ajst-19203	221	17	ising	ise	VERB
ajst-19203	221	18	machine	machine	NOUN
ajst-19203	221	19	learning	learn	VERB
ajst-19203	221	20	to	to	PART
ajst-19203	221	21	predict	predict	VERB
ajst-19203	221	22	traffic	traffic	NOUN
ajst-19203	221	23	congestion	congestion	NOUN
ajst-19203	221	24	under	under	ADP
ajst-19203	221	25	uncertainty	uncertainty	NOUN
ajst-19203	221	26	.	.	PUNCT
ajst-19203	221	27	"	"	PUNCT
ajst-19203	221	28	proceedings	proceeding	NOUN
ajst-19203	221	29	of	of	ADP
ajst-19203	221	30	the	the	DET
ajst-19203	221	31	aaai	aaai	PROPN
ajst-19203	221	32	conference	conference	NOUN
ajst-19203	221	33	on	on	ADP
ajst-19203	221	34	artificial	artificial	ADJ
ajst-19203	221	35	intelligence	intelligence	NOUN
ajst-19203	221	36	.	.	PUNCT
ajst-19203	222	1	vol	vol	NOUN
ajst-19203	222	2	.	.	PUNCT
ajst-19203	223	1	37	37	NUM
ajst-19203	223	2	.	.	PUNCT
ajst-19203	224	1	no	no	INTJ
ajst-19203	224	2	.	.	NOUN
ajst-19203	225	1	8	8	NUM
ajst-19203	225	2	.	.	NOUN
ajst-19203	226	1	2023	2023	NUM
ajst-19203	226	2	.	.	PUNCT
ajst-19203	227	1	[	[	X
ajst-19203	227	2	46	46	NUM
ajst-19203	227	3	]	]	X
ajst-19203	227	4	pan	pan	PROPN
ajst-19203	227	5	,	,	PUNCT
ajst-19203	227	6	zhenyu	zhenyu	PROPN
ajst-19203	227	7	,	,	PUNCT
ajst-19203	227	8	et	et	PROPN
ajst-19203	227	9	al	al	PROPN
ajst-19203	227	10	.	.	PUNCT
ajst-19203	227	11	"	"	PUNCT
ajst-19203	227	12	cormf	cormf	ADJ
ajst-19203	227	13	:	:	PUNCT
ajst-19203	227	14	criticality	criticality	NOUN
ajst-19203	227	15	-	-	PUNCT
ajst-19203	227	16	ordered	order	VERB
ajst-19203	227	17	recurrent	recurrent	ADJ
ajst-19203	227	18	mean	mean	NOUN
ajst-19203	227	19	field	field	NOUN
ajst-19203	227	20	ising	ise	VERB
ajst-19203	227	21	solver	solver	ADV
ajst-19203	227	22	.	.	PUNCT
ajst-19203	227	23	"	"	PUNCT
ajst-19203	228	1	arxiv	arxiv	PROPN
ajst-19203	228	2	preprint	preprint	VERB
ajst-19203	228	3	arxiv:2403.03391	arxiv:2403.03391	ADV
ajst-19203	228	4	(	(	PUNCT
ajst-19203	228	5	2024	2024	NUM
ajst-19203	228	6	)	)	PUNCT
ajst-19203	228	7	.	.	PUNCT
ajst-19203	229	1	[	[	X
ajst-19203	229	2	47	47	NUM
ajst-19203	229	3	]	]	SYM
ajst-19203	229	4	hu	hu	PROPN
ajst-19203	229	5	,	,	PUNCT
ajst-19203	229	6	zhirui	zhirui	PROPN
ajst-19203	229	7	,	,	PUNCT
ajst-19203	229	8	et	et	PROPN
ajst-19203	229	9	al	al	PROPN
ajst-19203	229	10	.	.	PUNCT
ajst-19203	230	1	"	"	PUNCT
ajst-19203	230	2	on	on	ADP
ajst-19203	230	3	the	the	DET
ajst-19203	230	4	design	design	NOUN
ajst-19203	230	5	of	of	ADP
ajst-19203	230	6	quantum	quantum	NOUN
ajst-19203	230	7	graph	graph	NOUN
ajst-19203	230	8	convolutional	convolutional	ADJ
ajst-19203	230	9	neural	neural	ADJ
ajst-19203	230	10	network	network	NOUN
ajst-19203	230	11	in	in	ADP
ajst-19203	230	12	the	the	DET
ajst-19203	230	13	nisq	nisq	PROPN
ajst-19203	230	14	-	-	PUNCT
ajst-19203	230	15	era	era	NOUN
ajst-19203	230	16	and	and	CCONJ
ajst-19203	230	17	beyond	beyond	ADP
ajst-19203	230	18	.	.	PUNCT
ajst-19203	230	19	"	"	PUNCT
ajst-19203	230	20	2022	2022	NUM
ajst-19203	230	21	ieee	ieee	NOUN
ajst-19203	230	22	40th	40th	ADJ
ajst-19203	230	23	international	international	ADJ
ajst-19203	230	24	conference	conference	NOUN
ajst-19203	230	25	on	on	ADP
ajst-19203	230	26	computer	computer	NOUN
ajst-19203	230	27	design	design	NOUN
ajst-19203	230	28	(	(	PUNCT
ajst-19203	230	29	iccd	iccd	PROPN
ajst-19203	230	30	)	)	PUNCT
ajst-19203	230	31	.	.	PUNCT
ajst-19203	230	32	ieee	ieee	NOUN
ajst-19203	230	33	,	,	PUNCT
ajst-19203	230	34	2022	2022	NUM
ajst-19203	230	35	.	.	PUNCT
ajst-19203	231	1	[	[	X
ajst-19203	231	2	48	48	NUM
ajst-19203	231	3	]	]	X
ajst-19203	231	4	chen	chen	PROPN
ajst-19203	231	5	,	,	PUNCT
ajst-19203	231	6	yinda	yinda	PROPN
ajst-19203	231	7	,	,	PUNCT
ajst-19203	231	8	et	et	PROPN
ajst-19203	231	9	al	al	PROPN
ajst-19203	231	10	.	.	PUNCT
ajst-19203	231	11	"	"	PUNCT
ajst-19203	231	12	self	self	NOUN
ajst-19203	231	13	-	-	PUNCT
ajst-19203	231	14	supervised	supervise	VERB
ajst-19203	231	15	neuron	neuron	NOUN
ajst-19203	231	16	segmentation	segmentation	NOUN
ajst-19203	231	17	with	with	ADP
ajst-19203	231	18	multi	multi	ADJ
ajst-19203	231	19	-	-	ADJ
ajst-19203	231	20	agent	agent	ADJ
ajst-19203	231	21	reinforcement	reinforcement	NOUN
ajst-19203	231	22	learning	learning	NOUN
ajst-19203	231	23	.	.	PUNCT
ajst-19203	231	24	"	"	PUNCT
ajst-19203	232	1	arxiv	arxiv	NOUN
ajst-19203	232	2	preprint	preprint	NOUN
ajst-19203	232	3	arxiv:2310.04148	arxiv:2310.04148	NOUN
ajst-19203	232	4	(	(	PUNCT
ajst-19203	232	5	2023	2023	NUM
ajst-19203	232	6	)	)	PUNCT
ajst-19203	232	7	.	.	PUNCT
ajst-19203	233	1	[	[	X
ajst-19203	233	2	49	49	NUM
ajst-19203	233	3	]	]	X
ajst-19203	233	4	chen	chen	PROPN
ajst-19203	233	5	,	,	PUNCT
ajst-19203	233	6	yinda	yinda	PROPN
ajst-19203	233	7	,	,	PUNCT
ajst-19203	233	8	et	et	PROPN
ajst-19203	233	9	al	al	PROPN
ajst-19203	233	10	.	.	PUNCT
ajst-19203	234	1	"	"	PUNCT
ajst-19203	234	2	learning	learn	VERB
ajst-19203	234	3	multiscale	multiscale	ADJ
ajst-19203	234	4	consistency	consistency	NOUN
ajst-19203	234	5	for	for	ADP
ajst-19203	234	6	selfsupervised	selfsupervise	VERB
ajst-19203	234	7	electron	electron	NOUN
ajst-19203	234	8	microscopy	microscopy	NOUN
ajst-19203	234	9	instance	instance	NOUN
ajst-19203	234	10	segmentation	segmentation	NOUN
ajst-19203	234	11	.	.	PUNCT
ajst-19203	234	12	"	"	PUNCT
ajst-19203	235	1	arxiv	arxiv	PROPN
ajst-19203	235	2	preprint	preprint	NOUN
ajst-19203	235	3	arxiv:2308.09917	arxiv:2308.09917	NOUN
ajst-19203	235	4	(	(	PUNCT
ajst-19203	235	5	2023	2023	NUM
ajst-19203	235	6	)	)	PUNCT
ajst-19203	235	7	.	.	PUNCT
ajst-19203	236	1	[	[	X
ajst-19203	236	2	50	50	NUM
ajst-19203	236	3	]	]	X
ajst-19203	236	4	chen	chen	PROPN
ajst-19203	236	5	,	,	PUNCT
ajst-19203	236	6	yinda	yinda	PROPN
ajst-19203	236	7	,	,	PUNCT
ajst-19203	236	8	et	et	PROPN
ajst-19203	236	9	al	al	PROPN
ajst-19203	236	10	.	.	PUNCT
ajst-19203	237	1	"	"	PUNCT
ajst-19203	237	2	generative	generative	ADJ
ajst-19203	237	3	text	text	NOUN
ajst-19203	237	4	-	-	PUNCT
ajst-19203	237	5	guided	guide	VERB
ajst-19203	237	6	3d	3d	NUM
ajst-19203	237	7	vision	vision	NOUN
ajst-19203	237	8	-	-	PUNCT
ajst-19203	237	9	language	language	NOUN
ajst-19203	237	10	pretraining	pretraine	VERB
ajst-19203	237	11	for	for	ADP
ajst-19203	237	12	unified	unified	ADJ
ajst-19203	237	13	medical	medical	ADJ
ajst-19203	237	14	image	image	NOUN
ajst-19203	237	15	segmentation	segmentation	NOUN
ajst-19203	237	16	.	.	PUNCT
ajst-19203	237	17	"	"	PUNCT
ajst-19203	238	1	arxiv	arxiv	PROPN
ajst-19203	238	2	preprint	preprint	VERB
ajst-19203	238	3	arxiv:2306.04811	arxiv:2306.04811	NUM
ajst-19203	238	4	(	(	PUNCT
ajst-19203	238	5	2023	2023	NUM
ajst-19203	238	6	)	)	PUNCT
ajst-19203	238	7	.	.	PUNCT
ajst-19203	239	1	[	[	X
ajst-19203	239	2	51	51	NUM
ajst-19203	239	3	]	]	X
ajst-19203	239	4	tong	tong	PROPN
ajst-19203	239	5	,	,	PUNCT
ajst-19203	239	6	xin	xin	PROPN
ajst-19203	239	7	,	,	PUNCT
ajst-19203	239	8	et	et	PROPN
ajst-19203	239	9	al	al	PROPN
ajst-19203	239	10	.	.	PUNCT
ajst-19203	240	1	"	"	PUNCT
ajst-19203	240	2	a	a	DET
ajst-19203	240	3	deep‐learning	deep‐learne	VERB
ajst-19203	240	4	approach	approach	NOUN
ajst-19203	240	5	for	for	ADP
ajst-19203	240	6	low‐spatial‐	low‐spatial‐	ADJ
ajst-19203	240	7	coherence	coherence	NOUN
ajst-19203	240	8	imaging	imaging	NOUN
ajst-19203	240	9	in	in	ADP
ajst-19203	240	10	computer‐generated	computer‐generate	VERB
ajst-19203	240	11	holography	holography	NOUN
ajst-19203	240	12	.	.	PUNCT
ajst-19203	240	13	"	"	PUNCT
ajst-19203	241	1	advanced	advanced	ADJ
ajst-19203	241	2	photonics	photonic	NOUN
ajst-19203	241	3	research	research	VERB
ajst-19203	241	4	4.1	4.1	NUM
ajst-19203	241	5	(	(	PUNCT
ajst-19203	241	6	2023	2023	NUM
ajst-19203	241	7	):	):	PUNCT
ajst-19203	241	8	2200264	2200264	NUM
ajst-19203	241	9	.	.	PUNCT
ajst-19203	242	1	[	[	X
ajst-19203	242	2	52	52	NUM
ajst-19203	242	3	]	]	PUNCT
ajst-19203	242	4	xu	xu	PROPN
ajst-19203	242	5	,	,	PUNCT
ajst-19203	242	6	renjun	renjun	PROPN
ajst-19203	242	7	,	,	PUNCT
ajst-19203	242	8	et	et	PROPN
ajst-19203	242	9	al	al	PROPN
ajst-19203	242	10	.	.	PUNCT
ajst-19203	243	1	"	"	PUNCT
ajst-19203	243	2	$	$	SYM
ajst-19203	243	3	e	e	X
ajst-19203	243	4	(	(	PUNCT
ajst-19203	243	5	2	2	NUM
ajst-19203	243	6	)	)	PUNCT
ajst-19203	243	7	$	$	SYM
ajst-19203	243	8	-equivariant	-equivariant	ADJ
ajst-19203	243	9	vision	vision	NOUN
ajst-19203	243	10	transformer	transformer	NOUN
ajst-19203	243	11	.	.	PUNCT
ajst-19203	243	12	"	"	PUNCT
ajst-19203	243	13	uncertainty	uncertainty	NOUN
ajst-19203	243	14	in	in	ADP
ajst-19203	243	15	artificial	artificial	ADJ
ajst-19203	243	16	intelligence	intelligence	NOUN
ajst-19203	243	17	.	.	PUNCT
ajst-19203	244	1	pmlr	pmlr	NOUN
ajst-19203	244	2	,	,	PUNCT
ajst-19203	244	3	2023	2023	NUM
ajst-19203	244	4	.	.	PUNCT
ajst-19203	245	1	[	[	X
ajst-19203	245	2	53	53	NUM
ajst-19203	245	3	]	]	X
ajst-19203	245	4	gao	gao	PROPN
ajst-19203	245	5	,	,	PUNCT
ajst-19203	245	6	shangde	shangde	NOUN
ajst-19203	245	7	,	,	PUNCT
ajst-19203	245	8	et	et	PROPN
ajst-19203	245	9	al	al	PROPN
ajst-19203	245	10	.	.	PUNCT
ajst-19203	246	1	"	"	PUNCT
ajst-19203	246	2	contrastive	contrastive	ADJ
ajst-19203	246	3	knowledge	knowledge	NOUN
ajst-19203	246	4	amalgamation	amalgamation	NOUN
ajst-19203	246	5	for	for	ADP
ajst-19203	246	6	unsupervised	unsupervised	ADJ
ajst-19203	246	7	image	image	NOUN
ajst-19203	246	8	classification	classification	NOUN
ajst-19203	246	9	.	.	PUNCT
ajst-19203	246	10	"	"	PUNCT
ajst-19203	247	1	international	international	ADJ
ajst-19203	247	2	conference	conference	NOUN
ajst-19203	247	3	on	on	ADP
ajst-19203	247	4	artificial	artificial	ADJ
ajst-19203	247	5	neural	neural	ADJ
ajst-19203	247	6	networks	network	NOUN
ajst-19203	247	7	.	.	PUNCT
ajst-19203	248	1	cham	cham	PROPN
ajst-19203	248	2	:	:	PUNCT
ajst-19203	248	3	springer	springer	NOUN
ajst-19203	248	4	nature	nature	PROPN
ajst-19203	248	5	switzerland	switzerland	PROPN
ajst-19203	248	6	,	,	PUNCT
ajst-19203	248	7	2023	2023	NUM
ajst-19203	248	8	.	.	PUNCT
ajst-19203	249	1	[	[	X
ajst-19203	249	2	54	54	NUM
ajst-19203	249	3	]	]	X
ajst-19203	249	4	shangguan	shangguan	NOUN
ajst-19203	249	5	,	,	PUNCT
ajst-19203	249	6	zhongkai	zhongkai	PROPN
ajst-19203	249	7	,	,	PUNCT
ajst-19203	249	8	zihe	zihe	NOUN
ajst-19203	249	9	zheng	zheng	PROPN
ajst-19203	249	10	,	,	PUNCT
ajst-19203	249	11	and	and	CCONJ
ajst-19203	249	12	lei	lei	PROPN
ajst-19203	249	13	lin	lin	PROPN
ajst-19203	249	14	.	.	PUNCT
ajst-19203	250	1	"	"	PUNCT
ajst-19203	250	2	trend	trend	NOUN
ajst-19203	250	3	and	and	CCONJ
ajst-19203	250	4	thoughts	thought	NOUN
ajst-19203	250	5	:	:	PUNCT
ajst-19203	250	6	understanding	understand	VERB
ajst-19203	250	7	climate	climate	NOUN
ajst-19203	250	8	change	change	NOUN
ajst-19203	250	9	concern	concern	NOUN
ajst-19203	250	10	using	use	VERB
ajst-19203	250	11	289	289	NUM
ajst-19203	250	12	machine	machine	NOUN
ajst-19203	250	13	learning	learning	NOUN
ajst-19203	250	14	and	and	CCONJ
ajst-19203	250	15	social	social	ADJ
ajst-19203	250	16	media	medium	NOUN
ajst-19203	250	17	data	datum	NOUN
ajst-19203	250	18	.	.	PUNCT
ajst-19203	250	19	"	"	PUNCT
ajst-19203	251	1	arxiv	arxiv	PROPN
ajst-19203	251	2	preprint	preprint	VERB
ajst-19203	251	3	arxiv:2111.14929	arxiv:2111.14929	NOUN
ajst-19203	251	4	(	(	PUNCT
ajst-19203	251	5	2021	2021	NUM
ajst-19203	251	6	)	)	PUNCT
ajst-19203	251	7	.	.	PUNCT
ajst-19203	252	1	[	[	X
ajst-19203	252	2	55	55	NUM
ajst-19203	252	3	]	]	X
ajst-19203	252	4	shangguan	shangguan	NOUN
ajst-19203	252	5	,	,	PUNCT
ajst-19203	252	6	zhongkai	zhongkai	PROPN
ajst-19203	252	7	,	,	PUNCT
ajst-19203	252	8	et	et	PROPN
ajst-19203	252	9	al	al	PROPN
ajst-19203	252	10	.	.	PUNCT
ajst-19203	252	11	"	"	PUNCT
ajst-19203	252	12	neural	neural	ADJ
ajst-19203	252	13	process	process	NOUN
ajst-19203	252	14	for	for	ADP
ajst-19203	252	15	black	black	ADJ
ajst-19203	252	16	-	-	PUNCT
ajst-19203	252	17	box	box	NOUN
ajst-19203	252	18	model	model	NOUN
ajst-19203	252	19	optimization	optimization	NOUN
ajst-19203	252	20	under	under	ADP
ajst-19203	252	21	bayesian	bayesian	NOUN
ajst-19203	252	22	framework	framework	NOUN
ajst-19203	252	23	.	.	PUNCT
ajst-19203	252	24	"	"	PUNCT
ajst-19203	253	1	arxiv	arxiv	PROPN
ajst-19203	253	2	preprint	preprint	VERB
ajst-19203	253	3	arxiv:2104.02487	arxiv:2104.02487	NOUN
ajst-19203	253	4	(	(	PUNCT
ajst-19203	253	5	2021	2021	NUM
ajst-19203	253	6	)	)	PUNCT
ajst-19203	253	7	.	.	PUNCT
ajst-19203	254	1	[	[	X
ajst-19203	254	2	56	56	NUM
ajst-19203	254	3	]	]	X
ajst-19203	254	4	zang	zang	PROPN
ajst-19203	254	5	,	,	PUNCT
ajst-19203	254	6	hengyi	hengyi	ADJ
ajst-19203	254	7	.	.	PUNCT
ajst-19203	255	1	"	"	PUNCT
ajst-19203	255	2	precision	precision	NOUN
ajst-19203	255	3	calibration	calibration	NOUN
ajst-19203	255	4	of	of	ADP
ajst-19203	255	5	industrial	industrial	ADJ
ajst-19203	255	6	3d	3d	PROPN
ajst-19203	255	7	scanners	scanner	NOUN
ajst-19203	255	8	:	:	PUNCT
ajst-19203	255	9	an	an	DET
ajst-19203	255	10	ai	ai	ADV
ajst-19203	255	11	-	-	PUNCT
ajst-19203	255	12	enhanced	enhance	VERB
ajst-19203	255	13	approach	approach	NOUN
ajst-19203	255	14	for	for	ADP
ajst-19203	255	15	improved	improved	ADJ
ajst-19203	255	16	measurement	measurement	NOUN
ajst-19203	255	17	accuracy	accuracy	NOUN
ajst-19203	255	18	.	.	PUNCT
ajst-19203	255	19	"	"	PUNCT
ajst-19203	256	1	global	global	ADJ
ajst-19203	256	2	academic	academic	ADJ
ajst-19203	256	3	frontiers	frontier	NOUN
ajst-19203	256	4	2.1	2.1	NUM
ajst-19203	256	5	(	(	PUNCT
ajst-19203	256	6	2024	2024	NUM
ajst-19203	256	7	):	):	PUNCT
ajst-19203	256	8	27	27	NUM
ajst-19203	256	9	-	-	SYM
ajst-19203	256	10	37	37	NUM
ajst-19203	256	11	.	.	PUNCT
ajst-19203	257	1	[	[	X
ajst-19203	257	2	57	57	NUM
ajst-19203	257	3	]	]	X
ajst-19203	257	4	wang	wang	PROPN
ajst-19203	257	5	,	,	PUNCT
ajst-19203	257	6	yishuang	yishuang	PROPN
ajst-19203	257	7	,	,	PUNCT
ajst-19203	257	8	et	et	PROPN
ajst-19203	257	9	al	al	PROPN
ajst-19203	257	10	.	.	PUNCT
ajst-19203	258	1	"	"	PUNCT
ajst-19203	258	2	predicting	predict	VERB
ajst-19203	258	3	nonlinear	nonlinear	ADJ
ajst-19203	258	4	structural	structural	ADJ
ajst-19203	258	5	dynamic	dynamic	ADJ
ajst-19203	258	6	response	response	NOUN
ajst-19203	258	7	of	of	ADP
ajst-19203	258	8	ode	ode	ADJ
ajst-19203	258	9	systems	system	NOUN
ajst-19203	258	10	using	use	VERB
ajst-19203	258	11	constrained	constrain	VERB
ajst-19203	258	12	gaussian	gaussian	ADJ
ajst-19203	258	13	process	process	NOUN
ajst-19203	258	14	regression	regression	NOUN
ajst-19203	258	15	.	.	PUNCT
ajst-19203	258	16	"	"	PUNCT
ajst-19203	259	1	society	society	NOUN
ajst-19203	259	2	for	for	ADP
ajst-19203	259	3	experimental	experimental	ADJ
ajst-19203	259	4	mechanics	mechanic	NOUN
ajst-19203	259	5	annual	annual	ADJ
ajst-19203	259	6	conference	conference	NOUN
ajst-19203	259	7	and	and	CCONJ
ajst-19203	259	8	exposition	exposition	NOUN
ajst-19203	259	9	.	.	PUNCT
ajst-19203	260	1	cham	cham	PROPN
ajst-19203	260	2	:	:	PUNCT
ajst-19203	260	3	springer	springer	NOUN
ajst-19203	260	4	nature	nature	PROPN
ajst-19203	260	5	switzerland	switzerland	PROPN
ajst-19203	260	6	,	,	PUNCT
ajst-19203	260	7	2023	2023	NUM
ajst-19203	260	8	.	.	PUNCT
ajst-19203	261	1	[	[	X
ajst-19203	261	2	58	58	NUM
ajst-19203	261	3	]	]	X
ajst-19203	261	4	platz	platz	NOUN
ajst-19203	261	5	,	,	PUNCT
ajst-19203	261	6	roland	roland	PROPN
ajst-19203	261	7	,	,	PUNCT
ajst-19203	261	8	xinyue	xinyue	PROPN
ajst-19203	261	9	xu	xu	PROPN
ajst-19203	261	10	,	,	PUNCT
ajst-19203	261	11	and	and	CCONJ
ajst-19203	261	12	sez	sez	NOUN
ajst-19203	261	13	atamturktur	atamturktur	PROPN
ajst-19203	261	14	.	.	PUNCT
ajst-19203	262	1	"	"	PUNCT
ajst-19203	262	2	introducing	introduce	VERB
ajst-19203	262	3	a	a	DET
ajst-19203	262	4	round	round	ADJ
ajst-19203	262	5	-	-	PUNCT
ajst-19203	262	6	robin	robin	NOUN
ajst-19203	262	7	challenge	challenge	NOUN
ajst-19203	262	8	to	to	PART
ajst-19203	262	9	quantify	quantify	VERB
ajst-19203	262	10	model	model	NOUN
ajst-19203	262	11	form	form	NOUN
ajst-19203	262	12	uncertainty	uncertainty	NOUN
ajst-19203	262	13	in	in	ADP
ajst-19203	262	14	passive	passive	ADJ
ajst-19203	262	15	and	and	CCONJ
ajst-19203	262	16	active	active	ADJ
ajst-19203	262	17	vibration	vibration	NOUN
ajst-19203	262	18	isolation	isolation	NOUN
ajst-19203	262	19	.	.	PUNCT
ajst-19203	262	20	"	"	PUNCT
ajst-19203	263	1	society	society	NOUN
ajst-19203	263	2	for	for	ADP
ajst-19203	263	3	experimental	experimental	ADJ
ajst-19203	263	4	mechanics	mechanic	NOUN
ajst-19203	263	5	annual	annual	ADJ
ajst-19203	263	6	conference	conference	NOUN
ajst-19203	263	7	and	and	CCONJ
ajst-19203	263	8	exposition	exposition	NOUN
ajst-19203	263	9	.	.	PUNCT
ajst-19203	264	1	cham	cham	PROPN
ajst-19203	264	2	:	:	PUNCT
ajst-19203	264	3	springer	springer	NOUN
ajst-19203	264	4	nature	nature	PROPN
ajst-19203	264	5	switzerland	switzerland	PROPN
ajst-19203	264	6	,	,	PUNCT
ajst-19203	264	7	2023	2023	NUM
ajst-19203	264	8	.	.	PUNCT
ajst-19203	265	1	[	[	X
ajst-19203	265	2	59	59	NUM
ajst-19203	265	3	]	]	SYM
ajst-19203	265	4	li	li	PROPN
ajst-19203	265	5	,	,	PUNCT
ajst-19203	265	6	zhenglin	zhenglin	PROPN
ajst-19203	265	7	,	,	PUNCT
ajst-19203	265	8	et	et	PROPN
ajst-19203	265	9	al	al	PROPN
ajst-19203	265	10	.	.	PUNCT
ajst-19203	265	11	"	"	PUNCT
ajst-19203	265	12	comprehensive	comprehensive	ADJ
ajst-19203	265	13	evaluation	evaluation	NOUN
ajst-19203	265	14	of	of	ADP
ajst-19203	265	15	mal	mal	ADJ
ajst-19203	265	16	-	-	PUNCT
ajst-19203	265	17	api2019	api2019	NOUN
ajst-19203	265	18	dataset	dataset	VERB
ajst-19203	265	19	by	by	ADP
ajst-19203	265	20	machine	machine	NOUN
ajst-19203	265	21	learning	learning	NOUN
ajst-19203	265	22	in	in	ADP
ajst-19203	265	23	malware	malware	NOUN
ajst-19203	265	24	detection	detection	NOUN
ajst-19203	265	25	.	.	PUNCT
ajst-19203	265	26	"	"	PUNCT
ajst-19203	266	1	international	international	ADJ
ajst-19203	266	2	journal	journal	NOUN
ajst-19203	266	3	of	of	ADP
ajst-19203	266	4	computer	computer	NOUN
ajst-19203	266	5	science	science	NOUN
ajst-19203	266	6	and	and	CCONJ
ajst-19203	266	7	information	information	NOUN
ajst-19203	266	8	technology	technology	NOUN
ajst-19203	266	9	2.1	2.1	NUM
ajst-19203	266	10	(	(	PUNCT
ajst-19203	266	11	2024	2024	NUM
ajst-19203	266	12	):	):	PUNCT
ajst-19203	266	13	1	1	NUM
ajst-19203	266	14	-	-	SYM
ajst-19203	266	15	9	9	NUM
ajst-19203	266	16	.	.	PUNCT