id	sid	tid	token	lemma	pos
fcis-7985	1	1	frontiers	frontier	NOUN
fcis-7985	1	2	in	in	ADP
fcis-7985	1	3	computing	computing	NOUN
fcis-7985	1	4	and	and	CCONJ
fcis-7985	1	5	intelligent	intelligent	ADJ
fcis-7985	1	6	systems	system	NOUN
fcis-7985	1	7	issn	issn	VERB
fcis-7985	1	8	:	:	PUNCT
fcis-7985	1	9	2832	2832	NUM
fcis-7985	1	10	-	-	SYM
fcis-7985	1	11	6024	6024	NUM
fcis-7985	1	12	|	|	NOUN
fcis-7985	1	13	vol	vol	NOUN
fcis-7985	1	14	.	.	PROPN
fcis-7985	2	1	3	3	NUM
fcis-7985	2	2	,	,	PUNCT
fcis-7985	2	3	no	no	INTJ
fcis-7985	2	4	.	.	NOUN
fcis-7985	2	5	3	3	NUM
fcis-7985	2	6	,	,	PUNCT
fcis-7985	2	7	2023	2023	NUM
fcis-7985	2	8	15	15	NUM
fcis-7985	2	9	non‐negative	non‐negative	ADJ
fcis-7985	2	10	sparse	sparse	ADJ
fcis-7985	2	11	discriminative	discriminative	NOUN
fcis-7985	2	12	low	low	ADJ
fcis-7985	2	13	rank	rank	NOUN
fcis-7985	2	14	preserving	preserve	VERB
fcis-7985	2	15	projection	projection	NOUN
fcis-7985	2	16	for	for	ADP
fcis-7985	2	17	robust	robust	ADJ
fcis-7985	2	18	face	face	NOUN
fcis-7985	2	19	feature	feature	NOUN
fcis-7985	2	20	extraction	extraction	NOUN
fcis-7985	2	21	rui	rui	PROPN
fcis-7985	2	22	li	li	PROPN
fcis-7985	3	1	*	*	PROPN
fcis-7985	3	2	,	,	PUNCT
fcis-7985	3	3	haishuang	haishuang	PROPN
fcis-7985	3	4	yin	yin	PROPN
fcis-7985	3	5	school	school	NOUN
fcis-7985	3	6	of	of	ADP
fcis-7985	3	7	electrical	electrical	ADJ
fcis-7985	3	8	engineering	engineering	NOUN
fcis-7985	3	9	,	,	PUNCT
fcis-7985	3	10	northeast	northeast	ADJ
fcis-7985	3	11	petroleum	petroleum	NOUN
fcis-7985	3	12	university	university	NOUN
fcis-7985	3	13	,	,	PUNCT
fcis-7985	3	14	daqing	daqing	NOUN
fcis-7985	3	15	,	,	PUNCT
fcis-7985	3	16	heilongjiang	heilongjiang	PROPN
fcis-7985	3	17	,	,	PUNCT
fcis-7985	3	18	china	china	PROPN
fcis-7985	3	19	*	*	PUNCT
fcis-7985	3	20	corresponding	correspond	VERB
fcis-7985	3	21	author	author	NOUN
fcis-7985	3	22	:	:	PUNCT
fcis-7985	3	23	rui	rui	PROPN
fcis-7985	3	24	li	li	PROPN
fcis-7985	3	25	abstract	abstract	PROPN
fcis-7985	3	26	:	:	PUNCT
fcis-7985	3	27	face	face	NOUN
fcis-7985	3	28	recognition	recognition	NOUN
fcis-7985	3	29	is	be	AUX
fcis-7985	3	30	an	an	DET
fcis-7985	3	31	important	important	ADJ
fcis-7985	3	32	aspect	aspect	NOUN
fcis-7985	3	33	of	of	ADP
fcis-7985	3	34	intelligent	intelligent	ADJ
fcis-7985	3	35	security	security	NOUN
fcis-7985	3	36	for	for	ADP
fcis-7985	3	37	new	new	ADJ
fcis-7985	3	38	energy	energy	NOUN
fcis-7985	3	39	vehicles	vehicle	NOUN
fcis-7985	3	40	.	.	PUNCT
fcis-7985	4	1	existing	exist	VERB
fcis-7985	4	2	methods	method	NOUN
fcis-7985	4	3	extract	extract	VERB
fcis-7985	4	4	features	feature	NOUN
fcis-7985	4	5	without	without	ADP
fcis-7985	4	6	taking	take	VERB
fcis-7985	4	7	low	low	ADJ
fcis-7985	4	8	-	-	PUNCT
fcis-7985	4	9	rank	rank	NOUN
fcis-7985	4	10	and	and	CCONJ
fcis-7985	4	11	discriminative	discriminative	NOUN
fcis-7985	4	12	similarity	similarity	NOUN
fcis-7985	4	13	relations	relation	NOUN
fcis-7985	4	14	of	of	ADP
fcis-7985	4	15	data	datum	NOUN
fcis-7985	4	16	into	into	ADP
fcis-7985	4	17	account	account	NOUN
fcis-7985	4	18	and	and	CCONJ
fcis-7985	4	19	resulting	result	VERB
fcis-7985	4	20	in	in	ADP
fcis-7985	4	21	low	low	ADJ
fcis-7985	4	22	quality	quality	NOUN
fcis-7985	4	23	feature	feature	NOUN
fcis-7985	4	24	distribution	distribution	NOUN
fcis-7985	4	25	.	.	PUNCT
fcis-7985	5	1	in	in	ADP
fcis-7985	5	2	addition	addition	NOUN
fcis-7985	5	3	,	,	PUNCT
fcis-7985	5	4	the	the	DET
fcis-7985	5	5	coefficients	coefficient	NOUN
fcis-7985	5	6	of	of	ADP
fcis-7985	5	7	the	the	DET
fcis-7985	5	8	learned	learn	VERB
fcis-7985	5	9	representation	representation	NOUN
fcis-7985	5	10	matrix	matrix	NOUN
fcis-7985	5	11	can	can	AUX
fcis-7985	5	12	be	be	AUX
fcis-7985	5	13	negative	negative	ADJ
fcis-7985	5	14	and	and	CCONJ
fcis-7985	5	15	lack	lack	VERB
fcis-7985	5	16	interpretability	interpretability	NOUN
fcis-7985	5	17	.	.	PUNCT
fcis-7985	6	1	to	to	PART
fcis-7985	6	2	address	address	VERB
fcis-7985	6	3	the	the	DET
fcis-7985	6	4	above	above	ADJ
fcis-7985	6	5	issues	issue	NOUN
fcis-7985	6	6	,	,	PUNCT
fcis-7985	6	7	a	a	DET
fcis-7985	6	8	method	method	NOUN
fcis-7985	6	9	named	name	VERB
fcis-7985	6	10	non	non	ADJ
fcis-7985	6	11	-	-	ADJ
fcis-7985	6	12	negative	negative	ADJ
fcis-7985	6	13	sparse	sparse	ADJ
fcis-7985	6	14	discriminative	discriminative	NOUN
fcis-7985	6	15	low	low	ADJ
fcis-7985	6	16	rank	rank	NOUN
fcis-7985	6	17	preserving	preserve	VERB
fcis-7985	6	18	(	(	PUNCT
fcis-7985	6	19	nnsdlrpp	nnsdlrpp	NOUN
fcis-7985	6	20	)	)	PUNCT
fcis-7985	6	21	which	which	PRON
fcis-7985	6	22	introduces	introduce	VERB
fcis-7985	6	23	sparsity	sparsity	NOUN
fcis-7985	6	24	,	,	PUNCT
fcis-7985	6	25	non	non	ADJ
fcis-7985	6	26	-	-	ADJ
fcis-7985	6	27	negativity	negativity	ADJ
fcis-7985	6	28	and	and	CCONJ
fcis-7985	6	29	block	block	NOUN
fcis-7985	6	30	-	-	PUNCT
fcis-7985	6	31	diagonal	diagonal	ADJ
fcis-7985	6	32	regularization	regularization	NOUN
fcis-7985	6	33	is	be	AUX
fcis-7985	6	34	proposed	propose	VERB
fcis-7985	6	35	.	.	PUNCT
fcis-7985	7	1	as	as	ADP
fcis-7985	7	2	a	a	DET
fcis-7985	7	3	result	result	NOUN
fcis-7985	7	4	,	,	PUNCT
fcis-7985	7	5	nnsdlrpp	nnsdlrpp	NOUN
fcis-7985	7	6	improves	improve	VERB
fcis-7985	7	7	the	the	DET
fcis-7985	7	8	interpretability	interpretability	NOUN
fcis-7985	7	9	of	of	ADP
fcis-7985	7	10	representation	representation	NOUN
fcis-7985	7	11	features	feature	NOUN
fcis-7985	7	12	while	while	SCONJ
fcis-7985	7	13	capturing	capture	VERB
fcis-7985	7	14	discriminative	discriminative	NOUN
fcis-7985	7	15	information	information	NOUN
fcis-7985	7	16	of	of	ADP
fcis-7985	7	17	data	datum	NOUN
fcis-7985	7	18	.	.	PUNCT
fcis-7985	8	1	extensive	extensive	ADJ
fcis-7985	8	2	experiments	experiment	NOUN
fcis-7985	8	3	on	on	ADP
fcis-7985	8	4	two	two	NUM
fcis-7985	8	5	face	face	NOUN
fcis-7985	8	6	datasets	dataset	NOUN
fcis-7985	8	7	show	show	VERB
fcis-7985	8	8	that	that	SCONJ
fcis-7985	8	9	the	the	DET
fcis-7985	8	10	proposed	propose	VERB
fcis-7985	8	11	method	method	NOUN
fcis-7985	8	12	outperforms	outperform	VERB
fcis-7985	8	13	other	other	ADJ
fcis-7985	8	14	state	state	NOUN
fcis-7985	8	15	-	-	PUNCT
fcis-7985	8	16	of	of	ADP
fcis-7985	8	17	-	-	PUNCT
fcis-7985	8	18	the	the	DET
fcis-7985	8	19	-	-	PUNCT
fcis-7985	8	20	art	art	NOUN
fcis-7985	8	21	methods	method	NOUN
fcis-7985	8	22	.	.	PUNCT
fcis-7985	9	1	keywords	keyword	NOUN
fcis-7985	9	2	:	:	PUNCT
fcis-7985	9	3	face	face	NOUN
fcis-7985	9	4	recognition	recognition	NOUN
fcis-7985	9	5	;	;	PUNCT
fcis-7985	9	6	feature	feature	NOUN
fcis-7985	9	7	extraction	extraction	NOUN
fcis-7985	9	8	;	;	PUNCT
fcis-7985	9	9	low	low	ADJ
fcis-7985	9	10	rank	rank	NOUN
fcis-7985	9	11	representation	representation	NOUN
fcis-7985	9	12	.	.	PUNCT
fcis-7985	10	1	1	1	X
fcis-7985	10	2	.	.	X
fcis-7985	10	3	introduction	introduction	NOUN
fcis-7985	10	4	with	with	ADP
fcis-7985	10	5	the	the	DET
fcis-7985	10	6	rapid	rapid	ADJ
fcis-7985	10	7	development	development	NOUN
fcis-7985	10	8	of	of	ADP
fcis-7985	10	9	new	new	ADJ
fcis-7985	10	10	energy	energy	NOUN
fcis-7985	10	11	technology	technology	NOUN
fcis-7985	10	12	and	and	CCONJ
fcis-7985	10	13	artificial	artificial	ADJ
fcis-7985	10	14	intelligence	intelligence	NOUN
fcis-7985	10	15	,	,	PUNCT
fcis-7985	10	16	face	face	NOUN
fcis-7985	10	17	recognition	recognition	NOUN
fcis-7985	10	18	is	be	AUX
fcis-7985	10	19	of	of	ADP
fcis-7985	10	20	great	great	ADJ
fcis-7985	10	21	significance	significance	NOUN
fcis-7985	10	22	to	to	ADP
fcis-7985	10	23	automotive	automotive	ADJ
fcis-7985	10	24	intelligent	intelligent	ADJ
fcis-7985	10	25	security	security	NOUN
fcis-7985	10	26	.	.	PUNCT
fcis-7985	11	1	feature	feature	NOUN
fcis-7985	11	2	extraction	extraction	NOUN
fcis-7985	11	3	plays	play	VERB
fcis-7985	11	4	an	an	DET
fcis-7985	11	5	important	important	ADJ
fcis-7985	11	6	role	role	NOUN
fcis-7985	11	7	in	in	ADP
fcis-7985	11	8	face	face	NOUN
fcis-7985	11	9	recognition	recognition	NOUN
fcis-7985	11	10	.	.	PUNCT
fcis-7985	12	1	the	the	DET
fcis-7985	12	2	low	low	ADJ
fcis-7985	12	3	-	-	PUNCT
fcis-7985	12	4	rank	rank	NOUN
fcis-7985	12	5	-	-	PUNCT
fcis-7985	12	6	based	base	VERB
fcis-7985	12	7	feature	feature	NOUN
fcis-7985	12	8	extraction	extraction	NOUN
fcis-7985	12	9	methods	method	NOUN
fcis-7985	12	10	have	have	AUX
fcis-7985	12	11	received	receive	VERB
fcis-7985	12	12	widespread	widespread	ADJ
fcis-7985	12	13	attention	attention	NOUN
fcis-7985	12	14	due	due	ADP
fcis-7985	12	15	to	to	ADP
fcis-7985	12	16	their	their	PRON
fcis-7985	12	17	ability	ability	NOUN
fcis-7985	12	18	to	to	PART
fcis-7985	12	19	construct	construct	VERB
fcis-7985	12	20	robust	robust	ADJ
fcis-7985	12	21	graphs	graph	NOUN
fcis-7985	12	22	.	.	PUNCT
fcis-7985	13	1	in	in	ADP
fcis-7985	13	2	,	,	PUNCT
fcis-7985	13	3	both	both	CCONJ
fcis-7985	13	4	the	the	DET
fcis-7985	13	5	global	global	ADJ
fcis-7985	13	6	and	and	CCONJ
fcis-7985	13	7	local	local	ADJ
fcis-7985	13	8	structures	structure	NOUN
fcis-7985	13	9	are	be	AUX
fcis-7985	13	10	used	use	VERB
fcis-7985	13	11	for	for	ADP
fcis-7985	13	12	feature	feature	NOUN
fcis-7985	13	13	extraction	extraction	NOUN
fcis-7985	13	14	.	.	PUNCT
fcis-7985	14	1	we	we	PRON
fcis-7985	14	2	et	et	PROPN
fcis-7985	14	3	al	al	PROPN
fcis-7985	14	4	.	.	PROPN
fcis-7985	14	5	proposed	propose	VERB
fcis-7985	14	6	a	a	DET
fcis-7985	14	7	low	low	ADJ
fcis-7985	14	8	-	-	PUNCT
fcis-7985	14	9	rank	rank	NOUN
fcis-7985	14	10	preserving	preserve	VERB
fcis-7985	14	11	projection	projection	NOUN
fcis-7985	14	12	via	via	ADP
fcis-7985	14	13	graph	graph	NOUN
fcis-7985	14	14	regularized	regularize	VERB
fcis-7985	14	15	reconstruction	reconstruction	NOUN
fcis-7985	14	16	(	(	PUNCT
fcis-7985	14	17	lrpp_grr	lrpp_grr	PROPN
fcis-7985	14	18	)	)	PUNCT
fcis-7985	14	19	without	without	ADP
fcis-7985	14	20	increasing	increase	VERB
fcis-7985	14	21	complexity	complexity	NOUN
fcis-7985	14	22	.	.	PUNCT
fcis-7985	15	1	however	however	ADV
fcis-7985	15	2	,	,	PUNCT
fcis-7985	15	3	the	the	DET
fcis-7985	15	4	low	low	ADJ
fcis-7985	15	5	-	-	PUNCT
fcis-7985	15	6	rank	rank	NOUN
fcis-7985	15	7	-	-	PUNCT
fcis-7985	15	8	based	base	VERB
fcis-7985	15	9	methods	method	NOUN
fcis-7985	15	10	only	only	ADV
fcis-7985	15	11	perform	perform	VERB
fcis-7985	15	12	well	well	ADV
fcis-7985	15	13	when	when	SCONJ
fcis-7985	15	14	the	the	DET
fcis-7985	15	15	data	data	NOUN
fcis-7985	15	16	is	be	AUX
fcis-7985	15	17	strictly	strictly	ADV
fcis-7985	15	18	sampled	sample	VERB
fcis-7985	15	19	in	in	ADP
fcis-7985	15	20	its	its	PRON
fcis-7985	15	21	original	original	ADJ
fcis-7985	15	22	subspace	subspace	NOUN
fcis-7985	15	23	,	,	PUNCT
fcis-7985	15	24	which	which	PRON
fcis-7985	15	25	means	mean	VERB
fcis-7985	15	26	that	that	SCONJ
fcis-7985	15	27	the	the	DET
fcis-7985	15	28	discriminant	discriminant	NOUN
fcis-7985	15	29	features	feature	NOUN
fcis-7985	15	30	are	be	AUX
fcis-7985	15	31	orderly	orderly	ADV
fcis-7985	15	32	distributed	distribute	VERB
fcis-7985	15	33	.	.	PUNCT
fcis-7985	16	1	in	in	ADP
fcis-7985	16	2	other	other	ADJ
fcis-7985	16	3	words	word	NOUN
fcis-7985	16	4	,	,	PUNCT
fcis-7985	16	5	a	a	DET
fcis-7985	16	6	sample	sample	NOUN
fcis-7985	16	7	should	should	AUX
fcis-7985	16	8	be	be	AUX
fcis-7985	16	9	adaptively	adaptively	ADV
fcis-7985	16	10	represented	represent	VERB
fcis-7985	16	11	by	by	ADP
fcis-7985	16	12	its	its	PRON
fcis-7985	16	13	intraclass	intraclass	NOUN
fcis-7985	16	14	samples	sample	NOUN
fcis-7985	16	15	.	.	PUNCT
fcis-7985	17	1	so	so	ADV
fcis-7985	17	2	the	the	DET
fcis-7985	17	3	ideal	ideal	ADJ
fcis-7985	17	4	representation	representation	NOUN
fcis-7985	17	5	of	of	ADP
fcis-7985	17	6	the	the	DET
fcis-7985	17	7	data	data	NOUN
fcis-7985	17	8	is	be	AUX
fcis-7985	17	9	the	the	DET
fcis-7985	17	10	best	good	ADJ
fcis-7985	17	11	to	to	PART
fcis-7985	17	12	have	have	VERB
fcis-7985	17	13	blockdiagonal	blockdiagonal	ADJ
fcis-7985	17	14	structure	structure	NOUN
fcis-7985	17	15	.	.	PUNCT
fcis-7985	18	1	otherwise	otherwise	ADV
fcis-7985	18	2	,	,	PUNCT
fcis-7985	18	3	samples	sample	NOUN
fcis-7985	18	4	of	of	ADP
fcis-7985	18	5	different	different	ADJ
fcis-7985	18	6	classes	class	NOUN
fcis-7985	18	7	may	may	AUX
fcis-7985	18	8	be	be	AUX
fcis-7985	18	9	close	close	ADJ
fcis-7985	18	10	to	to	ADP
fcis-7985	18	11	each	each	DET
fcis-7985	18	12	other	other	ADJ
fcis-7985	18	13	,	,	PUNCT
fcis-7985	18	14	resulting	result	VERB
fcis-7985	18	15	in	in	ADP
fcis-7985	18	16	a	a	DET
fcis-7985	18	17	decrease	decrease	NOUN
fcis-7985	18	18	in	in	ADP
fcis-7985	18	19	classification	classification	NOUN
fcis-7985	18	20	accuracy	accuracy	NOUN
fcis-7985	18	21	.	.	PUNCT
fcis-7985	19	1	in	in	ADP
fcis-7985	19	2	addition	addition	NOUN
fcis-7985	19	3	,	,	PUNCT
fcis-7985	19	4	the	the	DET
fcis-7985	19	5	learned	learn	VERB
fcis-7985	19	6	representation	representation	NOUN
fcis-7985	19	7	coefficient	coefficient	NOUN
fcis-7985	19	8	matrix	matrix	NOUN
fcis-7985	19	9	may	may	AUX
fcis-7985	19	10	be	be	AUX
fcis-7985	19	11	negative	negative	ADJ
fcis-7985	19	12	and	and	CCONJ
fcis-7985	19	13	lack	lack	VERB
fcis-7985	19	14	interpretability	interpretability	NOUN
fcis-7985	19	15	,	,	PUNCT
fcis-7985	19	16	while	while	SCONJ
fcis-7985	19	17	non	non	ADJ
fcis-7985	19	18	-	-	ADJ
fcis-7985	19	19	negativity	negativity	ADJ
fcis-7985	19	20	is	be	AUX
fcis-7985	19	21	more	more	ADJ
fcis-7985	19	22	in	in	ADP
fcis-7985	19	23	accordance	accordance	NOUN
fcis-7985	19	24	with	with	ADP
fcis-7985	19	25	practical	practical	ADJ
fcis-7985	19	26	data	datum	NOUN
fcis-7985	19	27	representation	representation	NOUN
fcis-7985	19	28	.	.	PUNCT
fcis-7985	20	1	finally	finally	ADV
fcis-7985	20	2	,	,	PUNCT
fcis-7985	20	3	these	these	DET
fcis-7985	20	4	methods	method	NOUN
fcis-7985	20	5	often	often	ADV
fcis-7985	20	6	result	result	VERB
fcis-7985	20	7	in	in	ADP
fcis-7985	20	8	a	a	DET
fcis-7985	20	9	dense	dense	ADJ
fcis-7985	20	10	graph	graph	NOUN
fcis-7985	20	11	,	,	PUNCT
fcis-7985	20	12	which	which	PRON
fcis-7985	20	13	leads	lead	VERB
fcis-7985	20	14	to	to	ADP
fcis-7985	20	15	unnecessary	unnecessary	ADJ
fcis-7985	20	16	connections	connection	NOUN
fcis-7985	20	17	.	.	PUNCT
fcis-7985	21	1	inspired	inspire	VERB
fcis-7985	21	2	by	by	ADP
fcis-7985	21	3	the	the	DET
fcis-7985	21	4	above	above	ADJ
fcis-7985	21	5	insights	insight	NOUN
fcis-7985	21	6	,	,	PUNCT
fcis-7985	21	7	we	we	PRON
fcis-7985	21	8	construct	construct	VERB
fcis-7985	21	9	a	a	DET
fcis-7985	21	10	non	non	ADJ
fcis-7985	21	11	-	-	ADJ
fcis-7985	21	12	negative	negative	ADJ
fcis-7985	21	13	sparse	sparse	ADJ
fcis-7985	21	14	discriminative	discriminative	NOUN
fcis-7985	21	15	low	low	ADJ
fcis-7985	21	16	-	-	PUNCT
fcis-7985	21	17	rank	rank	NOUN
fcis-7985	21	18	preserving	preserve	VERB
fcis-7985	21	19	projection	projection	NOUN
fcis-7985	21	20	(	(	PUNCT
fcis-7985	21	21	nnsdlrpp	nnsdlrpp	NOUN
fcis-7985	21	22	)	)	PUNCT
fcis-7985	21	23	for	for	ADP
fcis-7985	21	24	face	face	NOUN
fcis-7985	21	25	classification	classification	NOUN
fcis-7985	21	26	.	.	PUNCT
fcis-7985	22	1	this	this	DET
fcis-7985	22	2	model	model	NOUN
fcis-7985	22	3	solves	solve	VERB
fcis-7985	22	4	the	the	DET
fcis-7985	22	5	above	above	ADJ
fcis-7985	22	6	limitations	limitation	NOUN
fcis-7985	22	7	by	by	ADP
fcis-7985	22	8	imposing	impose	VERB
fcis-7985	22	9	the	the	DET
fcis-7985	22	10	sparsity	sparsity	NOUN
fcis-7985	22	11	,	,	PUNCT
fcis-7985	22	12	nonnegativity	nonnegativity	NOUN
fcis-7985	22	13	and	and	CCONJ
fcis-7985	22	14	block	block	NOUN
fcis-7985	22	15	-	-	PUNCT
fcis-7985	22	16	diagonal	diagonal	ADJ
fcis-7985	22	17	regularization	regularization	NOUN
fcis-7985	22	18	to	to	PART
fcis-7985	22	19	obtain	obtain	VERB
fcis-7985	22	20	more	more	ADJ
fcis-7985	22	21	discriminative	discriminative	NOUN
fcis-7985	22	22	feature	feature	NOUN
fcis-7985	22	23	representation	representation	NOUN
fcis-7985	22	24	.	.	PUNCT
fcis-7985	23	1	with	with	ADP
fcis-7985	23	2	such	such	ADJ
fcis-7985	23	3	method	method	NOUN
fcis-7985	23	4	,	,	PUNCT
fcis-7985	23	5	a	a	DET
fcis-7985	23	6	potential	potential	ADJ
fcis-7985	23	7	subspace	subspace	NOUN
fcis-7985	23	8	with	with	ADP
fcis-7985	23	9	good	good	ADJ
fcis-7985	23	10	discriminative	discriminative	NOUN
fcis-7985	23	11	information	information	NOUN
fcis-7985	23	12	and	and	CCONJ
fcis-7985	23	13	strong	strong	ADJ
fcis-7985	23	14	interpretability	interpretability	NOUN
fcis-7985	23	15	can	can	AUX
fcis-7985	23	16	be	be	AUX
fcis-7985	23	17	found	find	VERB
fcis-7985	23	18	and	and	CCONJ
fcis-7985	23	19	achieving	achieve	VERB
fcis-7985	23	20	better	well	ADJ
fcis-7985	23	21	classification	classification	NOUN
fcis-7985	23	22	accuracy	accuracy	NOUN
fcis-7985	23	23	.	.	PUNCT
fcis-7985	24	1	2	2	X
fcis-7985	24	2	.	.	X
fcis-7985	24	3	methodology	methodology	NOUN
fcis-7985	24	4	2.1	2.1	NUM
fcis-7985	24	5	.	.	PUNCT
fcis-7985	25	1	model	model	NOUN
fcis-7985	25	2	formulation	formulation	NOUN
fcis-7985	25	3	in	in	ADP
fcis-7985	25	4	order	order	NOUN
fcis-7985	25	5	to	to	PART
fcis-7985	25	6	preserve	preserve	VERB
fcis-7985	25	7	the	the	DET
fcis-7985	25	8	global	global	ADJ
fcis-7985	25	9	and	and	CCONJ
fcis-7985	25	10	local	local	ADJ
fcis-7985	25	11	structure	structure	NOUN
fcis-7985	25	12	of	of	ADP
fcis-7985	25	13	the	the	DET
fcis-7985	25	14	data	datum	NOUN
fcis-7985	25	15	during	during	ADP
fcis-7985	25	16	feature	feature	NOUN
fcis-7985	25	17	extraction	extraction	NOUN
fcis-7985	25	18	without	without	ADP
fcis-7985	25	19	increasing	increase	VERB
fcis-7985	25	20	complexity	complexity	NOUN
fcis-7985	25	21	,	,	PUNCT
fcis-7985	25	22	we	we	PRON
fcis-7985	25	23	adopt	adopt	VERB
fcis-7985	25	24	the	the	DET
fcis-7985	25	25	lrpp_grr	lrpp_grr	ADJ
fcis-7985	25	26	framework	framework	NOUN
fcis-7985	25	27	.	.	PUNCT
fcis-7985	26	1	in	in	ADP
fcis-7985	26	2	addition	addition	NOUN
fcis-7985	26	3	,	,	PUNCT
fcis-7985	26	4	we	we	PRON
fcis-7985	26	5	introduce	introduce	VERB
fcis-7985	26	6	three	three	NUM
fcis-7985	26	7	helpful	helpful	ADJ
fcis-7985	26	8	regularization	regularization	NOUN
fcis-7985	26	9	terms	term	NOUN
fcis-7985	26	10	which	which	PRON
fcis-7985	26	11	are	be	AUX
fcis-7985	26	12	groupwise	groupwise	NOUN
fcis-7985	26	13	constraint	constraint	NOUN
fcis-7985	26	14	,	,	PUNCT
fcis-7985	26	15	sparse	sparse	ADJ
fcis-7985	26	16	constraint	constraint	NOUN
fcis-7985	26	17	and	and	CCONJ
fcis-7985	26	18	non	non	ADJ
fcis-7985	26	19	-	-	ADJ
fcis-7985	26	20	negative	negative	ADJ
fcis-7985	26	21	constraint	constraint	NOUN
fcis-7985	26	22	.	.	PUNCT
fcis-7985	27	1	to	to	PART
fcis-7985	27	2	remedy	remedy	VERB
fcis-7985	27	3	the	the	DET
fcis-7985	27	4	shortcoming	shortcoming	NOUN
fcis-7985	27	5	of	of	ADP
fcis-7985	27	6	low	low	ADJ
fcis-7985	27	7	discrimination	discrimination	NOUN
fcis-7985	27	8	,	,	PUNCT
fcis-7985	27	9	we	we	PRON
fcis-7985	27	10	seek	seek	VERB
fcis-7985	27	11	a	a	DET
fcis-7985	27	12	representation	representation	NOUN
fcis-7985	27	13	matrix	matrix	NOUN
fcis-7985	27	14	z	z	NOUN
fcis-7985	27	15	with	with	ADP
fcis-7985	27	16	a	a	DET
fcis-7985	27	17	block	block	NOUN
fcis-7985	27	18	-	-	PUNCT
fcis-7985	27	19	diagonal	diagonal	ADJ
fcis-7985	27	20	structure	structure	NOUN
fcis-7985	27	21	.	.	PUNCT
fcis-7985	28	1	specifically	specifically	ADV
fcis-7985	28	2	,	,	PUNCT
fcis-7985	28	3	the	the	DET
fcis-7985	28	4	desired	desire	VERB
fcis-7985	28	5	representation	representation	NOUN
fcis-7985	28	6	z	z	NOUN
fcis-7985	28	7	with	with	ADP
fcis-7985	28	8	c	c	PROPN
fcis-7985	28	9	classes	class	NOUN
fcis-7985	28	10	is	be	AUX
fcis-7985	28	11	as	as	SCONJ
fcis-7985	28	12	follows	follow	VERB
fcis-7985	28	13	:	:	PUNCT
fcis-7985	28	14	⋯	⋯	NOUN
fcis-7985	28	15	⋮	⋮	NOUN
fcis-7985	28	16	⋱	⋱	PUNCT
fcis-7985	28	17	⋮	⋮	NOUN
fcis-7985	28	18	⋯	⋯	VERB
fcis-7985	28	19	(	(	PUNCT
fcis-7985	28	20	1	1	NUM
fcis-7985	28	21	)	)	PUNCT
fcis-7985	28	22	to	to	ADP
fcis-7985	28	23	this	this	DET
fcis-7985	28	24	end	end	NOUN
fcis-7985	28	25	,	,	PUNCT
fcis-7985	28	26	a	a	DET
fcis-7985	28	27	group	group	NOUN
fcis-7985	28	28	-	-	PUNCT
fcis-7985	28	29	wise	wise	ADJ
fcis-7985	28	30	regularization	regularization	NOUN
fcis-7985	28	31	∑	∑	PUNCT
fcis-7985	28	32	‖	‖	PROPN
fcis-7985	28	33	‖	‖	PROPN
fcis-7985	28	34	is	be	AUX
fcis-7985	28	35	introduced	introduce	VERB
fcis-7985	28	36	.	.	PUNCT
fcis-7985	29	1	means	mean	VERB
fcis-7985	29	2	the	the	DET
fcis-7985	29	3	representation	representation	NOUN
fcis-7985	29	4	matrix	matrix	NOUN
fcis-7985	29	5	of	of	ADP
fcis-7985	29	6	the	the	DET
fcis-7985	29	7	i	i	PROPN
fcis-7985	29	8	-	-	PUNCT
fcis-7985	29	9	th	th	VERB
fcis-7985	29	10	class	class	NOUN
fcis-7985	29	11	samples	sample	NOUN
fcis-7985	29	12	subject	subject	ADJ
fcis-7985	29	13	to	to	ADP
fcis-7985	29	14	x	x	PUNCT
fcis-7985	29	15	with	with	SCONJ
fcis-7985	29	16	i	i	NOUN
fcis-7985	29	17	-	-	PUNCT
fcis-7985	29	18	class	class	NOUN
fcis-7985	29	19	samples	sample	NOUN
fcis-7985	29	20	removed	remove	VERB
fcis-7985	29	21	.	.	PUNCT
fcis-7985	30	1	the	the	DET
fcis-7985	30	2	label	label	NOUN
fcis-7985	30	3	information	information	NOUN
fcis-7985	30	4	is	be	AUX
fcis-7985	30	5	incorporated	incorporate	VERB
fcis-7985	30	6	into	into	ADP
fcis-7985	30	7	this	this	DET
fcis-7985	30	8	regularization	regularization	NOUN
fcis-7985	30	9	term	term	NOUN
fcis-7985	30	10	,	,	PUNCT
fcis-7985	30	11	which	which	PRON
fcis-7985	30	12	can	can	AUX
fcis-7985	30	13	promote	promote	VERB
fcis-7985	30	14	each	each	DET
fcis-7985	30	15	sample	sample	NOUN
fcis-7985	30	16	to	to	PART
fcis-7985	30	17	be	be	AUX
fcis-7985	30	18	adaptively	adaptively	ADV
fcis-7985	30	19	represented	represent	VERB
fcis-7985	30	20	by	by	ADP
fcis-7985	30	21	the	the	DET
fcis-7985	30	22	corresponding	correspond	VERB
fcis-7985	30	23	samples	sample	NOUN
fcis-7985	30	24	and	and	CCONJ
fcis-7985	30	25	the	the	DET
fcis-7985	30	26	representation	representation	NOUN
fcis-7985	30	27	coefficient	coefficient	NOUN
fcis-7985	30	28	of	of	ADP
fcis-7985	30	29	interclass	interclass	NOUN
fcis-7985	30	30	samples	sample	NOUN
fcis-7985	30	31	to	to	PART
fcis-7985	30	32	be	be	AUX
fcis-7985	30	33	suppressed	suppress	VERB
fcis-7985	30	34	.	.	PUNCT
fcis-7985	31	1	so	so	ADV
fcis-7985	31	2	a	a	DET
fcis-7985	31	3	more	more	ADV
fcis-7985	31	4	ordered	order	VERB
fcis-7985	31	5	and	and	CCONJ
fcis-7985	31	6	discriminative	discriminative	NOUN
fcis-7985	31	7	representation	representation	NOUN
fcis-7985	31	8	matrix	matrix	NOUN
fcis-7985	31	9	can	can	AUX
fcis-7985	31	10	be	be	AUX
fcis-7985	31	11	obtained	obtain	VERB
fcis-7985	31	12	.	.	PUNCT
fcis-7985	32	1	sparsity	sparsity	NOUN
fcis-7985	32	2	helps	help	VERB
fcis-7985	32	3	to	to	PART
fcis-7985	32	4	reduce	reduce	VERB
fcis-7985	32	5	unnecessary	unnecessary	ADJ
fcis-7985	32	6	connections	connection	NOUN
fcis-7985	32	7	between	between	ADP
fcis-7985	32	8	samples	sample	NOUN
fcis-7985	32	9	with	with	ADP
fcis-7985	32	10	smaller	small	ADJ
fcis-7985	32	11	weights	weight	NOUN
fcis-7985	32	12	and	and	CCONJ
fcis-7985	32	13	ensure	ensure	VERB
fcis-7985	32	14	that	that	SCONJ
fcis-7985	32	15	the	the	DET
fcis-7985	32	16	involved	involved	ADJ
fcis-7985	32	17	neighbors	neighbor	NOUN
fcis-7985	32	18	are	be	AUX
fcis-7985	32	19	fewest	few	ADJ
fcis-7985	32	20	possible	possible	ADJ
fcis-7985	32	21	.	.	PUNCT
fcis-7985	33	1	nonnegativity	nonnegativity	NOUN
fcis-7985	33	2	ensures	ensure	VERB
fcis-7985	33	3	that	that	SCONJ
fcis-7985	33	4	every	every	DET
fcis-7985	33	5	data	data	NOUN
fcis-7985	33	6	vector	vector	NOUN
fcis-7985	33	7	is	be	AUX
fcis-7985	33	8	in	in	ADP
fcis-7985	33	9	the	the	DET
fcis-7985	33	10	convex	convex	NOUN
fcis-7985	33	11	hull	hull	NOUN
fcis-7985	33	12	of	of	ADP
fcis-7985	33	13	its	its	PRON
fcis-7985	33	14	neighbors	neighbor	NOUN
fcis-7985	33	15	.	.	PUNCT
fcis-7985	34	1	the	the	DET
fcis-7985	34	2	three	three	NUM
fcis-7985	34	3	constraints	constraint	NOUN
fcis-7985	34	4	introduce	introduce	VERB
fcis-7985	34	5	richer	rich	ADJ
fcis-7985	34	6	information	information	NOUN
fcis-7985	34	7	on	on	ADP
fcis-7985	34	8	affinity	affinity	NOUN
fcis-7985	34	9	representation	representation	NOUN
fcis-7985	34	10	matrix	matrix	NOUN
fcis-7985	34	11	z	z	PROPN
fcis-7985	34	12	,	,	PUNCT
fcis-7985	34	13	which	which	PRON
fcis-7985	34	14	enables	enable	VERB
fcis-7985	34	15	the	the	DET
fcis-7985	34	16	structure	structure	NOUN
fcis-7985	34	17	of	of	ADP
fcis-7985	34	18	z	z	PROPN
fcis-7985	34	19	to	to	PART
fcis-7985	34	20	reflect	reflect	VERB
fcis-7985	34	21	the	the	DET
fcis-7985	34	22	relationships	relationship	NOUN
fcis-7985	34	23	between	between	ADP
fcis-7985	34	24	samples	sample	NOUN
fcis-7985	34	25	more	more	ADV
fcis-7985	34	26	accurately	accurately	ADV
fcis-7985	34	27	.	.	PUNCT
fcis-7985	35	1	based	base	VERB
fcis-7985	35	2	on	on	ADP
fcis-7985	35	3	the	the	DET
fcis-7985	35	4	above	above	ADJ
fcis-7985	35	5	analysis	analysis	NOUN
fcis-7985	35	6	,	,	PUNCT
fcis-7985	35	7	the	the	DET
fcis-7985	35	8	objective	objective	ADJ
fcis-7985	35	9	function	function	NOUN
fcis-7985	35	10	of	of	ADP
fcis-7985	35	11	nnsdlrpp	nnsdlrpp	NOUN
fcis-7985	35	12	can	can	AUX
fcis-7985	35	13	be	be	AUX
fcis-7985	35	14	formulated	formulate	VERB
fcis-7985	35	15	as	as	SCONJ
fcis-7985	35	16	follows	follow	VERB
fcis-7985	35	17	:	:	PUNCT
fcis-7985	35	18	2	2	NUM
fcis-7985	35	19	2	2	NUM
fcis-7985	35	20	1	1	NUM
fcis-7985	35	21	*	*	SYM
fcis-7985	35	22	1	1	NUM
fcis-7985	35	23	2,12	2,12	NUM
fcis-7985	35	24	,	,	PUNCT
fcis-7985	35	25	,	,	PUNCT
fcis-7985	35	26	1	1	NUM
fcis-7985	35	27	1	1	NUM
fcis-7985	35	28	1	1	NUM
fcis-7985	35	29	min	min	NOUN
fcis-7985	35	30	2	2	NUM
fcis-7985	35	31	.	.	PUNCT
fcis-7985	35	32	.	.	PUNCT
fcis-7985	36	1	,	,	PUNCT
fcis-7985	36	2	n	n	CCONJ
fcis-7985	37	1	n	n	PROPN
fcis-7985	37	2	c	c	NOUN
fcis-7985	38	1	t	t	NOUN
fcis-7985	39	1	i	i	PRON
fcis-7985	39	2	j	j	INTJ
fcis-7985	40	1	ij	ij	INTJ
fcis-7985	41	1	i	i	PRON
fcis-7985	41	2	f	f	PROPN
fcis-7985	42	1	i	i	PRON
fcis-7985	42	2	j	j	INTJ
fcis-7985	43	1	i	i	PRON
fcis-7985	43	2	ts	ts	VERB
fcis-7985	43	3	t	t	PROPN
fcis-7985	43	4			X
fcis-7985	43	5			NUM
fcis-7985	43	6			NOUN
fcis-7985	43	7			NUM
fcis-7985	43	8			NUM
fcis-7985	43	9			PROPN
fcis-7985	43	10			PROPN
fcis-7985	43	11			ADV
fcis-7985	43	12			CCONJ
fcis-7985	43	13			PUNCT
fcis-7985	43	14			PUNCT
fcis-7985	43	15			PROPN
fcis-7985	43	16			NUM
fcis-7985	43	17			CCONJ
fcis-7985	43	18			X
fcis-7985	43	19	r	r	NOUN
fcis-7985	43	20	p	p	X
fcis-7985	43	21	z	z	NOUN
fcis-7985	43	22	x	x	SYM
fcis-7985	43	23	rp	rp	PROPN
fcis-7985	43	24	xz	xz	PROPN
fcis-7985	44	1	w	w	PROPN
fcis-7985	44	2	z	z	PROPN
fcis-7985	44	3	z	z	PROPN
fcis-7985	44	4	z	z	NOUN
fcis-7985	45	1	p	p	NOUN
fcis-7985	45	2	r	r	NOUN
fcis-7985	45	3	r	r	NOUN
fcis-7985	45	4	i	i	NOUN
fcis-7985	45	5	z	z	NOUN
fcis-7985	45	6	0	0	NUM
fcis-7985	46	1	(	(	PUNCT
fcis-7985	46	2	2	2	NUM
fcis-7985	46	3	)	)	PUNCT
fcis-7985	46	4	where	where	SCONJ
fcis-7985	46	5	,	,	PUNCT
fcis-7985	46	6	,	,	PUNCT
fcis-7985	46	7	and	and	CCONJ
fcis-7985	46	8	are	be	AUX
fcis-7985	46	9	regularization	regularization	NOUN
fcis-7985	46	10	parameters	parameter	NOUN
fcis-7985	46	11	;	;	PUNCT
fcis-7985	46	12	the	the	DET
fcis-7985	46	13	first	first	ADJ
fcis-7985	46	14	term	term	NOUN
fcis-7985	46	15	imposes	impose	VERB
fcis-7985	46	16	graph	graph	NOUN
fcis-7985	46	17	constraints	constraint	NOUN
fcis-7985	46	18	on	on	ADP
fcis-7985	46	19	data	datum	NOUN
fcis-7985	46	20	reconstruction	reconstruction	NOUN
fcis-7985	46	21	errors	error	NOUN
fcis-7985	46	22	to	to	PART
fcis-7985	46	23	capture	capture	VERB
fcis-7985	46	24	global	global	ADJ
fcis-7985	46	25	and	and	CCONJ
fcis-7985	46	26	local	local	ADJ
fcis-7985	46	27	information	information	NOUN
fcis-7985	46	28	of	of	ADP
fcis-7985	46	29	data	data	PROPN
fcis-7985	46	30	.	.	PUNCT
fcis-7985	47	1	redundant	redundant	ADJ
fcis-7985	47	2	information	information	NOUN
fcis-7985	47	3	can	can	AUX
fcis-7985	47	4	be	be	AUX
fcis-7985	47	5	removed	remove	VERB
fcis-7985	47	6	by	by	ADP
fcis-7985	47	7	introducing	introduce	VERB
fcis-7985	47	8	the	the	DET
fcis-7985	47	9	ℓ	ℓ	NOUN
fcis-7985	47	10	,	,	PUNCT
fcis-7985	47	11	-norm	-norm	NOUN
fcis-7985	47	12	into	into	ADP
fcis-7985	47	13	the	the	DET
fcis-7985	47	14	projection	projection	NOUN
fcis-7985	47	15	matrix	matrix	NOUN
fcis-7985	47	16	p.	p.	NOUN
fcis-7985	47	17	latent	latent	NOUN
fcis-7985	47	18	subspace	subspace	NOUN
fcis-7985	47	19	derived	derive	VERB
fcis-7985	47	20	by	by	ADP
fcis-7985	47	21	the	the	DET
fcis-7985	47	22	proposed	propose	VERB
fcis-7985	47	23	method	method	NOUN
fcis-7985	47	24	is	be	AUX
fcis-7985	47	25	more	more	ADJ
fcis-7985	47	26	discriminant	discriminant	ADJ
fcis-7985	47	27	and	and	CCONJ
fcis-7985	47	28	interpretable	interpretable	ADJ
fcis-7985	47	29	.	.	PUNCT
fcis-7985	48	1	16	16	NUM
fcis-7985	48	2	2.2	2.2	NUM
fcis-7985	48	3	.	.	PUNCT
fcis-7985	49	1	optimization	optimization	NOUN
fcis-7985	49	2	strategy	strategy	NOUN
fcis-7985	49	3	to	to	PART
fcis-7985	49	4	make	make	VERB
fcis-7985	49	5	(	(	PUNCT
fcis-7985	49	6	2	2	NUM
fcis-7985	49	7	)	)	PUNCT
fcis-7985	49	8	separable	separable	NOUN
fcis-7985	49	9	,	,	PUNCT
fcis-7985	49	10	we	we	PRON
fcis-7985	49	11	first	first	ADV
fcis-7985	49	12	introduce	introduce	VERB
fcis-7985	49	13	three	three	NUM
fcis-7985	49	14	auxiliary	auxiliary	ADJ
fcis-7985	49	15	variables	variable	NOUN
fcis-7985	49	16	z	z	NOUN
fcis-7985	49	17	=	=	PROPN
fcis-7985	49	18	a	a	X
fcis-7985	49	19	,	,	PUNCT
fcis-7985	49	20	z	z	PROPN
fcis-7985	49	21	=	=	PROPN
fcis-7985	49	22	j	j	PROPN
fcis-7985	49	23	and	and	CCONJ
fcis-7985	49	24	rtxz	rtxz	PROPN
fcis-7985	49	25	=	=	PROPN
fcis-7985	49	26	y.	y.	NOUN
fcis-7985	49	27	then	then	ADV
fcis-7985	49	28	,	,	PUNCT
fcis-7985	49	29	the	the	DET
fcis-7985	49	30	augmented	augment	VERB
fcis-7985	49	31	lagrangian	lagrangian	ADJ
fcis-7985	49	32	function	function	NOUN
fcis-7985	49	33	of	of	ADP
fcis-7985	49	34	(	(	PUNCT
fcis-7985	49	35	3	3	X
fcis-7985	49	36	)	)	PUNCT
fcis-7985	49	37	is	be	AUX
fcis-7985	49	38	constructed	construct	VERB
fcis-7985	49	39	as	as	SCONJ
fcis-7985	49	40	follows	follow	VERB
fcis-7985	49	41	:	:	PUNCT
fcis-7985	49	42			NOUN
fcis-7985	50	1			SYM
fcis-7985	50	2			NOUN
fcis-7985	50	3			PUNCT
fcis-7985	50	4	2	2	NUM
fcis-7985	50	5	*	*	SYM
fcis-7985	50	6	1	1	NUM
fcis-7985	50	7	2,1	2,1	NUM
fcis-7985	50	8	1	1	NUM
fcis-7985	50	9	2	2	NUM
fcis-7985	50	10	2	2	NUM
fcis-7985	50	11	2	2	NUM
fcis-7985	50	12	1	1	NUM
fcis-7985	50	13	2	2	NUM
fcis-7985	50	14	3	3	NUM
fcis-7985	50	15	(	(	PUNCT
fcis-7985	50	16	)	)	PUNCT
fcis-7985	50	17	2	2	NUM
fcis-7985	50	18	,	,	PUNCT
fcis-7985	50	19	2	2	NUM
fcis-7985	50	20	,	,	PUNCT
fcis-7985	50	21	,	,	PUNCT
fcis-7985	50	22	c	c	PROPN
fcis-7985	50	23	t	t	PROPN
fcis-7985	50	24	t	t	PROPN
fcis-7985	50	25	t	t	PROPN
fcis-7985	50	26	t	t	PROPN
fcis-7985	51	1	i	i	PRON
fcis-7985	51	2	f	f	PROPN
fcis-7985	52	1	i	i	PRON
fcis-7985	52	2	t	t	X
fcis-7985	52	3	t	t	X
fcis-7985	52	4	f	f	PROPN
fcis-7985	52	5	ff	ff	PROPN
fcis-7985	52	6	l	l	NOUN
fcis-7985	52	7	tr	tr	VERB
fcis-7985	52	8	tr	tr	VERB
fcis-7985	52	9			X
fcis-7985	52	10			NUM
fcis-7985	52	11			ADJ
fcis-7985	52	12			NOUN
fcis-7985	52	13			NOUN
fcis-7985	52	14			PROPN
fcis-7985	52	15			PROPN
fcis-7985	52	16			PROPN
fcis-7985	52	17			ADV
fcis-7985	52	18			CCONJ
fcis-7985	52	19			PROPN
fcis-7985	52	20			PROPN
fcis-7985	52	21			PUNCT
fcis-7985	52	22			PUNCT
fcis-7985	52	23			PROPN
fcis-7985	52	24			VERB
fcis-7985	52	25			PROPN
fcis-7985	52	26			VERB
fcis-7985	52	27			PROPN
fcis-7985	52	28			VERB
fcis-7985	52	29			PROPN
fcis-7985	52	30			VERB
fcis-7985	52	31			PROPN
fcis-7985	52	32			VERB
fcis-7985	52	33			VERB
fcis-7985	52	34	xdx	xdx	PUNCT
fcis-7985	52	35	2xwy	2xwy	NUM
fcis-7985	52	36	r	r	NOUN
fcis-7985	52	37	ydy	ydy	NOUN
fcis-7985	53	1	a	a	DET
fcis-7985	53	2	z	z	NOUN
fcis-7985	53	3	j	j	PROPN
fcis-7985	54	1	p	p	PROPN
fcis-7985	54	2	p	p	X
fcis-7985	54	3	xz	xz	PROPN
fcis-7985	54	4	y	y	PROPN
fcis-7985	54	5	z	z	PROPN
fcis-7985	54	6	a	a	PRON
fcis-7985	54	7	z	z	NOUN
fcis-7985	54	8	j	j	PROPN
fcis-7985	55	1	c	c	PROPN
fcis-7985	55	2	p	p	PROPN
fcis-7985	55	3	xz	xz	PROPN
fcis-7985	55	4	y	y	PROPN
fcis-7985	55	5	c	c	PROPN
fcis-7985	55	6	z	z	PROPN
fcis-7985	56	1	a	a	PRON
fcis-7985	56	2	c	c	PROPN
fcis-7985	56	3	z	z	PROPN
fcis-7985	56	4	j	j	PROPN
fcis-7985	56	5	(	(	PUNCT
fcis-7985	56	6	3	3	NUM
fcis-7985	56	7	)	)	PUNCT
fcis-7985	56	8	where	where	SCONJ
fcis-7985	56	9	,	,	PUNCT
fcis-7985	56	10	and	and	CCONJ
fcis-7985	56	11	are	be	AUX
fcis-7985	56	12	the	the	DET
fcis-7985	56	13	lagrangian	lagrangian	ADJ
fcis-7985	56	14	multipliers	multiplier	NOUN
fcis-7985	56	15	,	,	PUNCT
fcis-7985	56	16	is	be	AUX
fcis-7985	56	17	a	a	DET
fcis-7985	56	18	penalty	penalty	NOUN
fcis-7985	56	19	parameter	parameter	NOUN
fcis-7985	56	20	,	,	PUNCT
fcis-7985	56	21	d	d	PROPN
fcis-7985	56	22	is	be	AUX
fcis-7985	56	23	a	a	DET
fcis-7985	56	24	diagonal	diagonal	ADJ
fcis-7985	56	25	matrix	matrix	NOUN
fcis-7985	56	26	with	with	ADP
fcis-7985	56	27	∑	∑	PROPN
fcis-7985	56	28	.	.	PUNCT
fcis-7985	57	1	by	by	ADP
fcis-7985	57	2	alternately	alternately	ADV
fcis-7985	57	3	updating	update	VERB
fcis-7985	57	4	all	all	DET
fcis-7985	57	5	variables	variable	NOUN
fcis-7985	57	6	,	,	PUNCT
fcis-7985	57	7	we	we	PRON
fcis-7985	57	8	can	can	AUX
fcis-7985	57	9	obtain	obtain	VERB
fcis-7985	57	10	the	the	DET
fcis-7985	57	11	solutions	solution	NOUN
fcis-7985	57	12	of	of	ADP
fcis-7985	57	13	(	(	PUNCT
fcis-7985	57	14	2	2	NUM
fcis-7985	57	15	)	)	PUNCT
fcis-7985	57	16	.	.	PUNCT
fcis-7985	58	1	step	step	NOUN
fcis-7985	58	2	1	1	NUM
fcis-7985	58	3	:	:	PUNCT
fcis-7985	58	4	update	update	VERB
fcis-7985	58	5	p	p	NOUN
fcis-7985	58	6	with	with	ADP
fcis-7985	58	7	the	the	DET
fcis-7985	58	8	other	other	ADJ
fcis-7985	58	9	variables	variable	NOUN
fcis-7985	58	10	fixed	fix	VERB
fcis-7985	58	11	.	.	PUNCT
fcis-7985	59	1	2	2	NUM
fcis-7985	59	2	12,1	12,1	NUM
fcis-7985	59	3	(	(	PUNCT
fcis-7985	59	4	)	)	PUNCT
fcis-7985	59	5	2	2	NUM
fcis-7985	59	6	t	t	NOUN
fcis-7985	59	7	f	f	PROPN
fcis-7985	59	8	l	l	NOUN
fcis-7985	59	9			PRON
fcis-7985	59	10			NOUN
fcis-7985	59	11			X
fcis-7985	59	12			NOUN
fcis-7985	59	13	p	p	PROPN
fcis-7985	59	14	p	p	PROPN
fcis-7985	59	15	p	p	X
fcis-7985	59	16	xz	xz	PROPN
fcis-7985	59	17	y	y	PROPN
fcis-7985	59	18	c	c	PROPN
fcis-7985	59	19	(	(	PUNCT
fcis-7985	59	20	4	4	X
fcis-7985	59	21	)	)	PUNCT
fcis-7985	59	22	taking	take	VERB
fcis-7985	59	23	the	the	DET
fcis-7985	59	24	derivative	derivative	NOUN
fcis-7985	59	25	of	of	ADP
fcis-7985	59	26	and	and	CCONJ
fcis-7985	59	27	setting	set	VERB
fcis-7985	59	28	it	it	PRON
fcis-7985	59	29	to	to	ADP
fcis-7985	59	30	zero	zero	NUM
fcis-7985	59	31	,	,	PUNCT
fcis-7985	59	32	then	then	ADV
fcis-7985	59	33	p	p	NOUN
fcis-7985	59	34	is	be	AUX
fcis-7985	59	35	obtained	obtain	VERB
fcis-7985	59	36	:	:	PUNCT
fcis-7985	60	1			NOUN
fcis-7985	60	2			PROPN
fcis-7985	60	3			NOUN
fcis-7985	60	4	1	1	NOUN
fcis-7985	60	5	t	t	PROPN
fcis-7985	60	6	t	t	X
fcis-7985	60	7	t	t	NUM
fcis-7985	60	8			NOUN
fcis-7985	60	9			NOUN
fcis-7985	60	10			NOUN
fcis-7985	60	11			PRON
fcis-7985	61	1	p	p	PROPN
fcis-7985	61	2	g	g	PROPN
fcis-7985	61	3	xzz	xzz	NOUN
fcis-7985	61	4	x	x	X
fcis-7985	61	5	xzh	xzh	PROPN
fcis-7985	61	6	(	(	PUNCT
fcis-7985	61	7	5	5	NUM
fcis-7985	61	8	)	)	PUNCT
fcis-7985	61	9	where	where	SCONJ
fcis-7985	61	10	⁄	⁄	ADJ
fcis-7985	61	11	,	,	PUNCT
fcis-7985	61	12	1	1	NUM
fcis-7985	61	13	‖	‖	PROPN
fcis-7985	61	14	∙‖⁄	∙‖⁄	NOUN
fcis-7985	61	15	,	,	PUNCT
fcis-7985	61	16	∙	∙	PROPN
fcis-7985	61	17	is	be	AUX
fcis-7985	61	18	the	the	DET
fcis-7985	61	19	i	i	PROPN
fcis-7985	61	20	-	-	PUNCT
fcis-7985	61	21	th	th	X
fcis-7985	61	22	row	row	NOUN
fcis-7985	61	23	vector	vector	NOUN
fcis-7985	61	24	of	of	ADP
fcis-7985	61	25	q	q	ADJ
fcis-7985	61	26	step	step	NOUN
fcis-7985	61	27	2	2	NUM
fcis-7985	61	28	:	:	PUNCT
fcis-7985	61	29	update	update	NOUN
fcis-7985	61	30	z	z	PROPN
fcis-7985	61	31	with	with	ADP
fcis-7985	61	32	the	the	DET
fcis-7985	61	33	other	other	ADJ
fcis-7985	61	34	variables	variable	NOUN
fcis-7985	61	35	fixed	fix	VERB
fcis-7985	61	36	.	.	PUNCT
fcis-7985	62	1			NOUN
fcis-7985	62	2	22	22	NOUN
fcis-7985	62	3	2	2	NUM
fcis-7985	62	4	2	2	NUM
fcis-7985	62	5	1	1	NUM
fcis-7985	62	6	2	2	NUM
fcis-7985	62	7	3	3	NUM
fcis-7985	62	8	1	1	NUM
fcis-7985	62	9	(	(	PUNCT
fcis-7985	62	10	)	)	PUNCT
fcis-7985	62	11	2	2	NUM
fcis-7985	62	12	2	2	NUM
fcis-7985	62	13	c	c	NOUN
fcis-7985	62	14	t	t	NOUN
fcis-7985	63	1	i	i	PRON
fcis-7985	63	2	f	f	NOUN
fcis-7985	64	1	f	f	X
fcis-7985	65	1	ff	ff	INTJ
fcis-7985	65	2	i	i	PROPN
fcis-7985	65	3	l	l	NOUN
fcis-7985	65	4			NOUN
fcis-7985	65	5			NOUN
fcis-7985	65	6			NOUN
fcis-7985	65	7			NOUN
fcis-7985	65	8			PUNCT
fcis-7985	66	1			NUM
fcis-7985	66	2			PROPN
fcis-7985	66	3			ADJ
fcis-7985	66	4			PROPN
fcis-7985	66	5			ADV
fcis-7985	66	6			PUNCT
fcis-7985	66	7			PROPN
fcis-7985	66	8			ADV
fcis-7985	66	9			PUNCT
fcis-7985	66	10			PROPN
fcis-7985	66	11	z	z	NOUN
fcis-7985	66	12	z	z	PROPN
fcis-7985	67	1	p	p	X
fcis-7985	67	2	xz	xz	PROPN
fcis-7985	67	3	y	y	PROPN
fcis-7985	67	4	c	c	PROPN
fcis-7985	67	5	z	z	PROPN
fcis-7985	68	1	a	a	DET
fcis-7985	68	2	c	c	PROPN
fcis-7985	68	3	z	z	NOUN
fcis-7985	68	4	j	j	PROPN
fcis-7985	68	5	c	c	X
fcis-7985	68	6	(	(	PUNCT
fcis-7985	68	7	6	6	NUM
fcis-7985	68	8	)	)	PUNCT
fcis-7985	68	9	let	let	VERB
fcis-7985	68	10	⁄	⁄	PROPN
fcis-7985	68	11	0	0	NUM
fcis-7985	68	12	,	,	PUNCT
fcis-7985	68	13	then	then	ADV
fcis-7985	68	14	we	we	PRON
fcis-7985	68	15	get	get	VERB
fcis-7985	68	16	:	:	PUNCT
fcis-7985	68	17			NOUN
fcis-7985	68	18			SYM
fcis-7985	68	19			NOUN
fcis-7985	68	20	1	1	VERB
fcis-7985	68	21	1	1	NUM
fcis-7985	68	22	2	2	NUM
fcis-7985	68	23	32	32	NUM
fcis-7985	68	24	t	t	NOUN
fcis-7985	68	25	t	t	NOUN
fcis-7985	68	26	t	t	PROPN
fcis-7985	68	27			PROPN
fcis-7985	68	28			NOUN
fcis-7985	68	29			NOUN
fcis-7985	68	30			NOUN
fcis-7985	68	31			NOUN
fcis-7985	68	32			NOUN
fcis-7985	68	33			NOUN
fcis-7985	68	34			PRON
fcis-7985	68	35			PROPN
fcis-7985	68	36			PUNCT
fcis-7985	68	37			PUNCT
fcis-7985	68	38			PROPN
fcis-7985	68	39	z	z	PROPN
fcis-7985	68	40	x	x	PUNCT
fcis-7985	69	1	pp	pp	PROPN
fcis-7985	69	2	x	x	X
fcis-7985	70	1	i	i	PRON
fcis-7985	70	2	i	i	PRON
fcis-7985	70	3	q	q	VERB
fcis-7985	70	4	x	x	VERB
fcis-7985	70	5	pm	pm	VERB
fcis-7985	70	6	m	m	VERB
fcis-7985	70	7	m	m	VERB
fcis-7985	70	8	(	(	PUNCT
fcis-7985	70	9	7	7	NUM
fcis-7985	70	10	)	)	PUNCT
fcis-7985	70	11	where	where	SCONJ
fcis-7985	70	12	,	,	PUNCT
fcis-7985	70	13	μ⁄	μ⁄	NUM
fcis-7985	70	14	,	,	PUNCT
fcis-7985	70	15	μ⁄	μ⁄	NUM
fcis-7985	70	16	,	,	PUNCT
fcis-7985	70	17	μ⁄	μ⁄	NUM
fcis-7985	70	18	.	.	PUNCT
fcis-7985	71	1	step	step	NOUN
fcis-7985	71	2	3	3	NUM
fcis-7985	71	3	:	:	PUNCT
fcis-7985	71	4	update	update	VERB
fcis-7985	71	5	a	a	PRON
fcis-7985	71	6	with	with	ADP
fcis-7985	71	7	the	the	DET
fcis-7985	71	8	other	other	ADJ
fcis-7985	71	9	variables	variable	NOUN
fcis-7985	71	10	fixed	fix	VERB
fcis-7985	71	11	.	.	PUNCT
fcis-7985	72	1	2	2	NUM
fcis-7985	72	2	2	2	NUM
fcis-7985	72	3	*	*	SYM
fcis-7985	72	4	min	min	NOUN
fcis-7985	72	5	2	2	NUM
fcis-7985	72	6	f	f	PROPN
fcis-7985	72	7			NOUN
fcis-7985	72	8			PROPN
fcis-7985	72	9			PROPN
fcis-7985	72	10			VERB
fcis-7985	72	11	a	a	DET
fcis-7985	72	12	a	a	DET
fcis-7985	72	13	z	z	NOUN
fcis-7985	72	14	a	a	DET
fcis-7985	72	15	c	c	NOUN
fcis-7985	72	16	(	(	PUNCT
fcis-7985	72	17	8)	8)	NUM
fcis-7985	72	18	the	the	DET
fcis-7985	72	19	solution	solution	NOUN
fcis-7985	72	20	of	of	ADP
fcis-7985	72	21	problem	problem	NOUN
fcis-7985	72	22	(	(	PUNCT
fcis-7985	72	23	8)	8)	NUM
fcis-7985	72	24	can	can	AUX
fcis-7985	72	25	be	be	AUX
fcis-7985	72	26	obtained	obtain	VERB
fcis-7985	72	27	by	by	ADP
fcis-7985	72	28	singular	singular	ADJ
fcis-7985	72	29	value	value	NOUN
fcis-7985	72	30	thresholding	thresholding	NOUN
fcis-7985	72	31	operator	operator	NOUN
fcis-7985	72	32	[	[	X
fcis-7985	72	33	7	7	NUM
fcis-7985	72	34	]	]	X
fcis-7985	72	35	s	s	X
fcis-7985	72	36	as	as	SCONJ
fcis-7985	72	37	follows	follow	VERB
fcis-7985	72	38	:	:	PUNCT
fcis-7985	72	39			NOUN
fcis-7985	72	40	2s	2s	PROPN
fcis-7985	72	41			NOUN
fcis-7985	72	42			NOUN
fcis-7985	72	43	a	a	X
fcis-7985	72	44	z	z	PROPN
fcis-7985	72	45	c	c	NOUN
fcis-7985	72	46	(	(	PUNCT
fcis-7985	72	47	9	9	NUM
fcis-7985	72	48	)	)	PUNCT
fcis-7985	72	49	step	step	NOUN
fcis-7985	72	50	4	4	NUM
fcis-7985	72	51	:	:	PUNCT
fcis-7985	72	52	update	update	VERB
fcis-7985	72	53	j	j	PROPN
fcis-7985	72	54	with	with	ADP
fcis-7985	72	55	the	the	DET
fcis-7985	72	56	other	other	ADJ
fcis-7985	72	57	variables	variable	NOUN
fcis-7985	72	58	fixed	fix	VERB
fcis-7985	72	59	.	.	PUNCT
fcis-7985	73	1	2	2	NUM
fcis-7985	73	2	310	310	NUM
fcis-7985	73	3	min	min	NOUN
fcis-7985	73	4	2	2	NUM
fcis-7985	73	5	f	f	PROPN
fcis-7985	73	6			PROPN
fcis-7985	73	7			NOUN
fcis-7985	73	8			NUM
fcis-7985	73	9			PUNCT
fcis-7985	73	10			PROPN
fcis-7985	73	11			PROPN
fcis-7985	73	12	j	j	PROPN
fcis-7985	73	13	j	j	PROPN
fcis-7985	73	14	z	z	PROPN
fcis-7985	73	15	j	j	PROPN
fcis-7985	73	16	c	c	X
fcis-7985	73	17	(	(	PUNCT
fcis-7985	73	18	10	10	NUM
fcis-7985	73	19	)	)	PUNCT
fcis-7985	73	20	then	then	ADV
fcis-7985	73	21	j	j	PROPN
fcis-7985	73	22	is	be	AUX
fcis-7985	73	23	obtained	obtain	VERB
fcis-7985	73	24	by	by	ADP
fcis-7985	73	25	using	use	VERB
fcis-7985	73	26	the	the	DET
fcis-7985	73	27	soft	soft	ADJ
fcis-7985	73	28	thresholding	thresholding	NOUN
fcis-7985	73	29	operator	operator	NOUN
fcis-7985	73	30	[	[	X
fcis-7985	73	31	8	8	NUM
fcis-7985	73	32	]	]	PUNCT
fcis-7985	73	33	ω	ω	NOUN
fcis-7985	73	34	as	as	SCONJ
fcis-7985	73	35	follows	follow	VERB
fcis-7985	73	36	:	:	PUNCT
fcis-7985	73	37			PROPN
fcis-7985	73	38			ADJ
fcis-7985	73	39	3max	3max	NOUN
fcis-7985	73	40	,	,	PUNCT
fcis-7985	73	41	0	0	NUM
fcis-7985	73	42			NOUN
fcis-7985	73	43			NOUN
fcis-7985	73	44			PROPN
fcis-7985	74	1	j	j	PROPN
fcis-7985	74	2	z	z	PROPN
fcis-7985	74	3	c	c	NOUN
fcis-7985	74	4	(	(	PUNCT
fcis-7985	74	5	11	11	NUM
fcis-7985	74	6	)	)	PUNCT
fcis-7985	74	7	step	step	NOUN
fcis-7985	74	8	5	5	NUM
fcis-7985	74	9	:	:	PUNCT
fcis-7985	74	10	update	update	VERB
fcis-7985	74	11	y	y	PROPN
fcis-7985	74	12	with	with	ADP
fcis-7985	74	13	the	the	DET
fcis-7985	74	14	other	other	ADJ
fcis-7985	74	15	variables	variable	NOUN
fcis-7985	74	16	fixed	fix	VERB
fcis-7985	74	17	.	.	PUNCT
fcis-7985	75	1	2	2	NUM
fcis-7985	75	2	1	1	NUM
fcis-7985	75	3	(	(	PUNCT
fcis-7985	75	4	)	)	PUNCT
fcis-7985	75	5	(	(	PUNCT
fcis-7985	75	6	)	)	PUNCT
fcis-7985	75	7	2	2	NUM
fcis-7985	75	8	t	t	NOUN
fcis-7985	75	9	t	t	NOUN
fcis-7985	75	10	t	t	PROPN
fcis-7985	75	11	t	t	PROPN
fcis-7985	75	12	f	f	PROPN
fcis-7985	75	13	l	l	NOUN
fcis-7985	75	14	tr	tr	NOUN
fcis-7985	75	15			NOUN
fcis-7985	75	16			NOUN
fcis-7985	75	17			NOUN
fcis-7985	75	18			VERB
fcis-7985	75	19			NOUN
fcis-7985	75	20	y	y	PROPN
fcis-7985	75	21	ydy	ydy	VERB
fcis-7985	75	22	2xwy	2xwy	NUM
fcis-7985	75	23	r	r	NOUN
fcis-7985	75	24	p	p	NOUN
fcis-7985	75	25	xz	xz	PROPN
fcis-7985	75	26	y	y	PROPN
fcis-7985	75	27	c	c	PROPN
fcis-7985	75	28	(	(	PUNCT
fcis-7985	75	29	12	12	NUM
fcis-7985	75	30	)	)	PUNCT
fcis-7985	75	31	let	let	VERB
fcis-7985	75	32	μ⁄	μ⁄	NUM
fcis-7985	75	33	and	and	CCONJ
fcis-7985	75	34	⁄	⁄	ADP
fcis-7985	75	35	0	0	NUM
fcis-7985	75	36	,	,	PUNCT
fcis-7985	75	37	then	then	ADV
fcis-7985	75	38	we	we	PRON
fcis-7985	75	39	can	can	AUX
fcis-7985	75	40	get	get	VERB
fcis-7985	75	41	:	:	PUNCT
fcis-7985	75	42			NOUN
fcis-7985	75	43			SYM
fcis-7985	75	44			NOUN
fcis-7985	75	45	1	1	VERB
fcis-7985	75	46	42	42	NUM
fcis-7985	75	47	2	2	NUM
fcis-7985	75	48	t	t	NUM
fcis-7985	75	49			NOUN
fcis-7985	75	50			ADV
fcis-7985	76	1	y	y	PROPN
fcis-7985	77	1	d	d	X
fcis-7985	78	1	i	i	PRON
fcis-7985	78	2	m	m	VERB
fcis-7985	78	3	r	r	NOUN
fcis-7985	78	4	xw	xw	PROPN
fcis-7985	78	5	(	(	PUNCT
fcis-7985	78	6	13	13	NUM
fcis-7985	78	7	)	)	PUNCT
fcis-7985	78	8	step	step	NOUN
fcis-7985	78	9	6	6	NUM
fcis-7985	78	10	:	:	PUNCT
fcis-7985	78	11	update	update	NOUN
fcis-7985	78	12	r	r	NOUN
fcis-7985	78	13	with	with	ADP
fcis-7985	78	14	the	the	DET
fcis-7985	78	15	other	other	ADJ
fcis-7985	78	16	variables	variable	NOUN
fcis-7985	78	17	fixed	fix	VERB
fcis-7985	78	18	.	.	PUNCT
fcis-7985	79	1			NOUN
fcis-7985	80	1	max	max	NOUN
fcis-7985	80	2	t	t	PROPN
fcis-7985	80	3	t	t	X
fcis-7985	80	4	ttr	ttr	PROPN
fcis-7985	80	5	r	r	PROPN
fcis-7985	81	1	r	r	NOUN
fcis-7985	82	1	i	i	PRON
fcis-7985	82	2	xwy	xwy	NOUN
fcis-7985	82	3	r	r	NOUN
fcis-7985	82	4	(	(	PUNCT
fcis-7985	82	5	14	14	NUM
fcis-7985	82	6	)	)	PUNCT
fcis-7985	82	7	problem	problem	NOUN
fcis-7985	82	8	(	(	PUNCT
fcis-7985	82	9	15	15	NUM
fcis-7985	82	10	)	)	PUNCT
fcis-7985	82	11	is	be	AUX
fcis-7985	82	12	an	an	DET
fcis-7985	82	13	orthogonal	orthogonal	ADJ
fcis-7985	82	14	procrustes	procruste	NOUN
fcis-7985	82	15	problem	problem	NOUN
fcis-7985	82	16	which	which	PRON
fcis-7985	82	17	can	can	AUX
fcis-7985	82	18	be	be	AUX
fcis-7985	82	19	simply	simply	ADV
fcis-7985	82	20	solved	solve	VERB
fcis-7985	82	21	by	by	ADP
fcis-7985	82	22	svd	svd	PROPN
fcis-7985	82	23	,	,	PUNCT
fcis-7985	82	24	i.e.	i.e.	X
fcis-7985	83	1	,	,	PUNCT
fcis-7985	83	2			PROPN
fcis-7985	83	3	svd	svd	SYM
fcis-7985	83	4	t	t	PROPN
fcis-7985	83	5	txwy	txwy	PROPN
fcis-7985	83	6	usv	usv	ADV
fcis-7985	83	7	(	(	PUNCT
fcis-7985	83	8	15	15	NUM
fcis-7985	83	9	)	)	PUNCT
fcis-7985	83	10	tr	tr	NOUN
fcis-7985	84	1	uv	uv	NOUN
fcis-7985	84	2	(	(	PUNCT
fcis-7985	84	3	16	16	NUM
fcis-7985	84	4	)	)	PUNCT
fcis-7985	84	5	step	step	NOUN
fcis-7985	84	6	6	6	NUM
fcis-7985	84	7	:	:	PUNCT
fcis-7985	84	8	update	update	VERB
fcis-7985	84	9	lagrangian	lagrangian	ADJ
fcis-7985	84	10	multipliers	multiplier	NOUN
fcis-7985	84	11	and	and	CCONJ
fcis-7985	84	12	penalty	penalty	NOUN
fcis-7985	84	13	parameter	parameter	NOUN
fcis-7985	84	14	.	.	PUNCT
fcis-7985	85	1			NOUN
fcis-7985	85	2			PUNCT
fcis-7985	85	3			NOUN
fcis-7985	85	4			PUNCT
fcis-7985	85	5			NOUN
fcis-7985	85	6			NOUN
fcis-7985	85	7	1	1	NUM
fcis-7985	85	8	1	1	NUM
fcis-7985	85	9	2	2	NUM
fcis-7985	85	10	2	2	NUM
fcis-7985	85	11	3	3	NUM
fcis-7985	85	12	3	3	NUM
fcis-7985	85	13	maxmin	maxmin	NOUN
fcis-7985	85	14	(	(	PUNCT
fcis-7985	85	15	,	,	PUNCT
fcis-7985	85	16	)	)	PUNCT
fcis-7985	85	17	t	t	NUM
fcis-7985	85	18			NOUN
fcis-7985	85	19			NOUN
fcis-7985	85	20			NOUN
fcis-7985	85	21			NOUN
fcis-7985	85	22			NOUN
fcis-7985	85	23			NOUN
fcis-7985	85	24			ADV
fcis-7985	85	25			PROPN
fcis-7985	85	26			PROPN
fcis-7985	85	27			ADV
fcis-7985	85	28			PROPN
fcis-7985	85	29			PROPN
fcis-7985	85	30			ADV
fcis-7985	85	31			NOUN
fcis-7985	85	32			NOUN
fcis-7985	85	33	c	c	NOUN
fcis-7985	85	34	c	c	NOUN
fcis-7985	86	1	p	p	PROPN
fcis-7985	86	2	xz	xz	PROPN
fcis-7985	86	3	y	y	PROPN
fcis-7985	86	4	c	c	PROPN
fcis-7985	86	5	c	c	PROPN
fcis-7985	86	6	z	z	NOUN
fcis-7985	86	7	a	a	DET
fcis-7985	86	8	c	c	NOUN
fcis-7985	86	9	c	c	PROPN
fcis-7985	86	10	z	z	PROPN
fcis-7985	86	11	j	j	PROPN
fcis-7985	86	12	(	(	PUNCT
fcis-7985	86	13	17	17	NUM
fcis-7985	86	14	)	)	PUNCT
fcis-7985	86	15	3	3	NUM
fcis-7985	86	16	.	.	X
fcis-7985	86	17	experimental	experimental	ADJ
fcis-7985	86	18	result	result	NOUN
fcis-7985	86	19	3.1	3.1	NUM
fcis-7985	86	20	.	.	PUNCT
fcis-7985	86	21	experimental	experimental	ADJ
fcis-7985	86	22	setup	setup	NOUN
fcis-7985	86	23	in	in	ADP
fcis-7985	86	24	this	this	DET
fcis-7985	86	25	section	section	NOUN
fcis-7985	86	26	,	,	PUNCT
fcis-7985	86	27	several	several	ADJ
fcis-7985	86	28	experiments	experiment	NOUN
fcis-7985	86	29	are	be	AUX
fcis-7985	86	30	conducted	conduct	VERB
fcis-7985	86	31	to	to	PART
fcis-7985	86	32	evaluate	evaluate	VERB
fcis-7985	86	33	the	the	DET
fcis-7985	86	34	performance	performance	NOUN
fcis-7985	86	35	of	of	ADP
fcis-7985	86	36	the	the	DET
fcis-7985	86	37	proposed	propose	VERB
fcis-7985	86	38	algorithm	algorithm	NOUN
fcis-7985	86	39	.	.	PUNCT
fcis-7985	87	1	for	for	ADP
fcis-7985	87	2	comparison	comparison	NOUN
fcis-7985	87	3	,	,	PUNCT
fcis-7985	87	4	some	some	DET
fcis-7985	87	5	related	related	ADJ
fcis-7985	87	6	methods	method	NOUN
fcis-7985	87	7	are	be	AUX
fcis-7985	87	8	selected	select	VERB
fcis-7985	87	9	including	include	VERB
fcis-7985	87	10	lpp	lpp	PROPN
fcis-7985	87	11	,	,	PUNCT
fcis-7985	87	12	npe	npe	PROPN
fcis-7985	87	13	,	,	PUNCT
fcis-7985	87	14	spp	spp	NOUN
fcis-7985	87	15	,	,	PUNCT
fcis-7985	87	16	lspp	lspp	PROPN
fcis-7985	87	17	,	,	PUNCT
fcis-7985	87	18	lrpp_grr	lrpp_grr	PROPN
fcis-7985	87	19	.	.	PUNCT
fcis-7985	88	1	the	the	DET
fcis-7985	88	2	datasets	dataset	NOUN
fcis-7985	88	3	used	use	VERB
fcis-7985	88	4	in	in	ADP
fcis-7985	88	5	the	the	DET
fcis-7985	88	6	experiments	experiment	NOUN
fcis-7985	88	7	are	be	AUX
fcis-7985	88	8	orl	orl	PRON
fcis-7985	88	9	and	and	CCONJ
fcis-7985	88	10	extended	extend	VERB
fcis-7985	88	11	yale	yale	PROPN
fcis-7985	88	12	b.	b.	PROPN
fcis-7985	89	1	the	the	DET
fcis-7985	89	2	orl	orl	PROPN
fcis-7985	89	3	dataset	dataset	NOUN
fcis-7985	89	4	contains	contain	VERB
fcis-7985	89	5	400	400	NUM
fcis-7985	89	6	images	image	NOUN
fcis-7985	89	7	of	of	ADP
fcis-7985	89	8	40	40	NUM
fcis-7985	89	9	individuals	individual	NOUN
fcis-7985	89	10	and	and	CCONJ
fcis-7985	89	11	the	the	DET
fcis-7985	89	12	image	image	NOUN
fcis-7985	89	13	size	size	NOUN
fcis-7985	89	14	is	be	AUX
fcis-7985	89	15	32	32	NUM
fcis-7985	89	16	32	32	NUM
fcis-7985	89	17	.	.	PUNCT
fcis-7985	90	1	fig	fig	NOUN
fcis-7985	90	2	.	.	PUNCT
fcis-7985	91	1	1(a	1(a	NUM
fcis-7985	91	2	)	)	PUNCT
fcis-7985	91	3	show	show	VERB
fcis-7985	91	4	some	some	DET
fcis-7985	91	5	orl	orl	NOUN
fcis-7985	91	6	images	image	NOUN
fcis-7985	91	7	with	with	ADP
fcis-7985	91	8	different	different	ADJ
fcis-7985	91	9	percentages	percentage	NOUN
fcis-7985	91	10	of	of	ADP
fcis-7985	91	11	pixel	pixel	ADJ
fcis-7985	91	12	corruption	corruption	NOUN
fcis-7985	91	13	.	.	PUNCT
fcis-7985	92	1	the	the	DET
fcis-7985	92	2	extended	extended	ADJ
fcis-7985	92	3	yale	yale	PROPN
fcis-7985	92	4	b	b	PROPN
fcis-7985	92	5	contains	contain	VERB
fcis-7985	92	6	2414	2414	NUM
fcis-7985	92	7	images	image	NOUN
fcis-7985	92	8	with	with	ADP
fcis-7985	92	9	the	the	DET
fcis-7985	92	10	size	size	NOUN
fcis-7985	92	11	of	of	ADP
fcis-7985	92	12	32	32	NUM
fcis-7985	92	13	32	32	NUM
fcis-7985	92	14	.	.	PUNCT
fcis-7985	93	1	fig	fig	NOUN
fcis-7985	93	2	.	.	PUNCT
fcis-7985	94	1	1(b	1(b	X
fcis-7985	94	2	)	)	PUNCT
fcis-7985	94	3	show	show	VERB
fcis-7985	94	4	some	some	DET
fcis-7985	94	5	eyaleb	eyaleb	ADJ
fcis-7985	94	6	images	image	NOUN
fcis-7985	94	7	with	with	ADP
fcis-7985	94	8	different	different	ADJ
fcis-7985	94	9	levels	level	NOUN
fcis-7985	94	10	of	of	ADP
fcis-7985	94	11	block	block	NOUN
fcis-7985	94	12	occlusion	occlusion	NOUN
fcis-7985	94	13	.	.	PUNCT
fcis-7985	95	1	(	(	PUNCT
fcis-7985	95	2	a	a	X
fcis-7985	95	3	)	)	PUNCT
fcis-7985	95	4	the	the	DET
fcis-7985	95	5	orl	orl	PROPN
fcis-7985	95	6	dataset	dataset	VERB
fcis-7985	95	7	with	with	ADP
fcis-7985	95	8	random	random	ADJ
fcis-7985	95	9	pixel	pixel	NOUN
fcis-7985	95	10	corruptions	corruption	NOUN
fcis-7985	95	11	of	of	ADP
fcis-7985	95	12	none	none	NOUN
fcis-7985	95	13	,	,	PUNCT
fcis-7985	95	14	0.05,0.1	0.05,0.1	NOUN
fcis-7985	95	15	(	(	PUNCT
fcis-7985	95	16	b	b	X
fcis-7985	95	17	)	)	PUNCT
fcis-7985	95	18	the	the	DET
fcis-7985	95	19	eyaleb	eyaleb	NOUN
fcis-7985	95	20	dataset	dataset	VERB
fcis-7985	95	21	with	with	ADP
fcis-7985	95	22	block	block	NOUN
fcis-7985	95	23	occlusions	occlusion	NOUN
fcis-7985	95	24	of	of	ADP
fcis-7985	95	25	none	none	NOUN
fcis-7985	95	26	,	,	PUNCT
fcis-7985	95	27	6	6	NUM
fcis-7985	95	28	6	6	NUM
fcis-7985	95	29	,	,	PUNCT
fcis-7985	95	30	8	8	NUM
fcis-7985	95	31	8	8	NUM
fcis-7985	95	32	figure	figure	NOUN
fcis-7985	95	33	1	1	NUM
fcis-7985	95	34	.	.	PUNCT
fcis-7985	96	1	some	some	DET
fcis-7985	96	2	corrupted	corrupted	ADJ
fcis-7985	96	3	samples	sample	NOUN
fcis-7985	96	4	from	from	ADP
fcis-7985	96	5	the	the	DET
fcis-7985	96	6	orl	orl	PROPN
fcis-7985	96	7	and	and	CCONJ
fcis-7985	96	8	eyaleb	eyaleb	ADJ
fcis-7985	96	9	datasets	dataset	NOUN
fcis-7985	96	10	.	.	PUNCT
fcis-7985	97	1	3.2	3.2	NUM
fcis-7985	97	2	.	.	PUNCT
fcis-7985	98	1	experimental	experimental	ADJ
fcis-7985	98	2	results	result	NOUN
fcis-7985	98	3	and	and	CCONJ
fcis-7985	98	4	analysis	analysis	NOUN
fcis-7985	98	5	figure	figure	NOUN
fcis-7985	98	6	2	2	NUM
fcis-7985	98	7	is	be	AUX
fcis-7985	98	8	the	the	DET
fcis-7985	98	9	classification	classification	NOUN
fcis-7985	98	10	accuracy	accuracy	NOUN
fcis-7985	98	11	of	of	ADP
fcis-7985	98	12	different	different	ADJ
fcis-7985	98	13	methods	method	NOUN
fcis-7985	98	14	under	under	ADP
fcis-7985	98	15	different	different	ADJ
fcis-7985	98	16	feature	feature	NOUN
fcis-7985	98	17	dimensions	dimension	NOUN
fcis-7985	98	18	in	in	ADP
fcis-7985	98	19	the	the	DET
fcis-7985	98	20	orl	orl	PROPN
fcis-7985	98	21	dataset	dataset	NOUN
fcis-7985	98	22	and	and	CCONJ
fcis-7985	98	23	the	the	DET
fcis-7985	98	24	eyaleb	eyaleb	ADJ
fcis-7985	98	25	dataset	dataset	NOUN
fcis-7985	98	26	.	.	PUNCT
fcis-7985	99	1	other	other	ADJ
fcis-7985	99	2	comparison	comparison	NOUN
fcis-7985	99	3	methods	method	NOUN
fcis-7985	99	4	use	use	VERB
fcis-7985	99	5	suggested	suggest	VERB
fcis-7985	99	6	parameters	parameter	NOUN
fcis-7985	99	7	.	.	PUNCT
fcis-7985	100	1	in	in	ADP
fcis-7985	100	2	random	random	ADJ
fcis-7985	100	3	pixel	pixel	NOUN
fcis-7985	100	4	corruptions	corruption	NOUN
fcis-7985	100	5	experiment	experiment	NOUN
fcis-7985	100	6	on	on	ADP
fcis-7985	100	7	the	the	DET
fcis-7985	100	8	orl	orl	PROPN
fcis-7985	100	9	dataset	dataset	NOUN
fcis-7985	100	10	,	,	PUNCT
fcis-7985	100	11	our	our	PRON
fcis-7985	100	12	algorithm	algorithm	NOUN
fcis-7985	100	13	outperforms	outperform	VERB
fcis-7985	100	14	other	other	ADJ
fcis-7985	100	15	algorithms	algorithm	NOUN
fcis-7985	100	16	and	and	CCONJ
fcis-7985	100	17	the	the	DET
fcis-7985	100	18	highest	high	ADJ
fcis-7985	100	19	classification	classification	NOUN
fcis-7985	100	20	accuracy	accuracy	NOUN
fcis-7985	100	21	are	be	AUX
fcis-7985	100	22	85.6	85.6	NUM
fcis-7985	100	23	%	%	NOUN
fcis-7985	100	24	and	and	CCONJ
fcis-7985	100	25	78.1	78.1	NUM
fcis-7985	100	26	%	%	NOUN
fcis-7985	100	27	,	,	PUNCT
fcis-7985	100	28	respectively	respectively	ADV
fcis-7985	100	29	.	.	PUNCT
fcis-7985	101	1	in	in	ADP
fcis-7985	101	2	the	the	DET
fcis-7985	101	3	block	block	NOUN
fcis-7985	101	4	occlusion	occlusion	NOUN
fcis-7985	101	5	experiment	experiment	NOUN
fcis-7985	101	6	on	on	ADP
fcis-7985	101	7	the	the	DET
fcis-7985	101	8	eyaleb	eyaleb	ADJ
fcis-7985	101	9	dataset	dataset	NOUN
fcis-7985	101	10	,	,	PUNCT
fcis-7985	101	11	the	the	DET
fcis-7985	101	12	proposed	propose	VERB
fcis-7985	101	13	algorithm	algorithm	NOUN
fcis-7985	101	14	achieved	achieve	VERB
fcis-7985	101	15	the	the	DET
fcis-7985	101	16	best	good	ADJ
fcis-7985	101	17	17	17	NUM
fcis-7985	101	18	performance	performance	NOUN
fcis-7985	101	19	with	with	ADP
fcis-7985	101	20	the	the	DET
fcis-7985	101	21	highest	high	ADJ
fcis-7985	101	22	classification	classification	NOUN
fcis-7985	101	23	accuracy	accuracy	NOUN
fcis-7985	101	24	of	of	ADP
fcis-7985	101	25	84.3	84.3	NUM
fcis-7985	101	26	%	%	NOUN
fcis-7985	101	27	and	and	CCONJ
fcis-7985	101	28	80.5	80.5	NUM
fcis-7985	101	29	%	%	NOUN
fcis-7985	101	30	.	.	PUNCT
fcis-7985	102	1	in	in	ADP
fcis-7985	102	2	general	general	ADJ
fcis-7985	102	3	,	,	PUNCT
fcis-7985	102	4	the	the	DET
fcis-7985	102	5	proposed	propose	VERB
fcis-7985	102	6	method	method	NOUN
fcis-7985	102	7	outperforms	outperform	VERB
fcis-7985	102	8	other	other	ADJ
fcis-7985	102	9	methods	method	NOUN
fcis-7985	102	10	with	with	ADP
fcis-7985	102	11	stronger	strong	ADJ
fcis-7985	102	12	feature	feature	NOUN
fcis-7985	102	13	extraction	extraction	NOUN
fcis-7985	102	14	capability	capability	NOUN
fcis-7985	102	15	and	and	CCONJ
fcis-7985	102	16	robustness	robustness	NOUN
fcis-7985	102	17	.	.	PUNCT
fcis-7985	103	1	(	(	PUNCT
fcis-7985	103	2	a	a	X
fcis-7985	103	3	)	)	PUNCT
fcis-7985	103	4	classification	classification	NOUN
fcis-7985	103	5	accuracy	accuracy	NOUN
fcis-7985	103	6	in	in	ADP
fcis-7985	103	7	orl	orl	PROPN
fcis-7985	103	8	dataset	dataset	PROPN
fcis-7985	103	9	(	(	PUNCT
fcis-7985	103	10	b	b	NOUN
fcis-7985	103	11	)	)	PUNCT
fcis-7985	103	12	classification	classification	NOUN
fcis-7985	103	13	accuracy	accuracy	NOUN
fcis-7985	103	14	in	in	ADP
fcis-7985	103	15	eyaleb	eyaleb	ADJ
fcis-7985	103	16	dataset	dataset	NOUN
fcis-7985	103	17	figure	figure	NOUN
fcis-7985	103	18	2	2	NUM
fcis-7985	103	19	.	.	PUNCT
fcis-7985	104	1	the	the	DET
fcis-7985	104	2	relationship	relationship	NOUN
fcis-7985	104	3	between	between	ADP
fcis-7985	104	4	features	feature	NOUN
fcis-7985	104	5	and	and	CCONJ
fcis-7985	104	6	classification	classification	NOUN
fcis-7985	104	7	accuracy	accuracy	NOUN
fcis-7985	104	8	in	in	ADP
fcis-7985	104	9	two	two	NUM
fcis-7985	104	10	datasets	dataset	NOUN
fcis-7985	104	11	.	.	PUNCT
fcis-7985	105	1	3.3	3.3	NUM
fcis-7985	105	2	.	.	PUNCT
fcis-7985	106	1	parameter	parameter	NOUN
fcis-7985	106	2	sensitiveness	sensitiveness	NOUN
fcis-7985	106	3	there	there	PRON
fcis-7985	106	4	are	be	VERB
fcis-7985	106	5	4	4	NUM
fcis-7985	106	6	parameters	parameter	NOUN
fcis-7985	106	7	to	to	PART
fcis-7985	106	8	be	be	AUX
fcis-7985	106	9	tuned	tune	VERB
fcis-7985	106	10	in	in	ADP
fcis-7985	106	11	our	our	PRON
fcis-7985	106	12	proposed	propose	VERB
fcis-7985	106	13	methods	method	NOUN
fcis-7985	106	14	,	,	PUNCT
fcis-7985	106	15	including	include	VERB
fcis-7985	106	16	,	,	PUNCT
fcis-7985	106	17	,	,	PUNCT
fcis-7985	106	18	and	and	CCONJ
fcis-7985	106	19	.	.	PUNCT
fcis-7985	107	1	in	in	ADP
fcis-7985	107	2	our	our	PRON
fcis-7985	107	3	experiments	experiment	NOUN
fcis-7985	107	4	,	,	PUNCT
fcis-7985	107	5	we	we	PRON
fcis-7985	107	6	choose	choose	VERB
fcis-7985	107	7	them	they	PRON
fcis-7985	107	8	from	from	ADP
fcis-7985	107	9	{	{	PUNCT
fcis-7985	107	10	10	10	NUM
fcis-7985	107	11	-	-	PUNCT
fcis-7985	107	12	4,10	4,10	NUM
fcis-7985	107	13	-	-	PUNCT
fcis-7985	107	14	3,10	3,10	NUM
fcis-7985	107	15	-	-	PUNCT
fcis-7985	107	16	2,10	2,10	NUM
fcis-7985	107	17	-	-	PUNCT
fcis-7985	107	18	1,100,101,102	1,100,101,102	NUM
fcis-7985	107	19	}	}	PUNCT
fcis-7985	107	20	by	by	ADP
fcis-7985	107	21	a	a	DET
fcis-7985	107	22	grid	grid	NOUN
fcis-7985	107	23	search	search	NOUN
fcis-7985	107	24	manner	manner	NOUN
fcis-7985	107	25	.	.	PUNCT
fcis-7985	108	1	the	the	DET
fcis-7985	108	2	performance	performance	NOUN
fcis-7985	108	3	variance	variance	NOUN
fcis-7985	108	4	of	of	ADP
fcis-7985	108	5	each	each	DET
fcis-7985	108	6	combination	combination	NOUN
fcis-7985	108	7	of	of	ADP
fcis-7985	108	8	parameter	parameter	NOUN
fcis-7985	108	9	value	value	NOUN
fcis-7985	108	10	on	on	ADP
fcis-7985	108	11	the	the	DET
fcis-7985	108	12	orl	orl	PROPN
fcis-7985	108	13	dataset	dataset	NOUN
fcis-7985	108	14	is	be	AUX
fcis-7985	108	15	shown	show	VERB
fcis-7985	108	16	in	in	ADP
fcis-7985	108	17	fig	fig	NOUN
fcis-7985	108	18	.	.	PUNCT
fcis-7985	109	1	3	3	X
fcis-7985	109	2	.	.	X
fcis-7985	109	3	for	for	ADP
fcis-7985	109	4	orl	orl	PROPN
fcis-7985	109	5	dataset	dataset	PROPN
fcis-7985	109	6	,	,	PUNCT
fcis-7985	109	7	the	the	DET
fcis-7985	109	8	optimal	optimal	ADJ
fcis-7985	109	9	parameters	parameter	NOUN
fcis-7985	109	10	are	be	AUX
fcis-7985	109	11	=	=	NOUN
fcis-7985	109	12	10	10	NUM
fcis-7985	109	13	-	-	SYM
fcis-7985	109	14	3	3	NUM
fcis-7985	109	15	,	,	PUNCT
fcis-7985	109	16	=	=	NOUN
fcis-7985	109	17	10	10	NUM
fcis-7985	109	18	-	-	SYM
fcis-7985	109	19	4	4	NUM
fcis-7985	109	20	,	,	PUNCT
fcis-7985	109	21	=	=	NOUN
fcis-7985	109	22	10	10	NUM
fcis-7985	109	23	-	-	SYM
fcis-7985	109	24	3	3	NUM
fcis-7985	109	25	and	and	CCONJ
fcis-7985	109	26	=	=	NOUN
fcis-7985	109	27	10	10	NUM
fcis-7985	109	28	-	-	SYM
fcis-7985	109	29	1	1	NUM
fcis-7985	109	30	.	.	PUNCT
fcis-7985	110	1	for	for	ADP
fcis-7985	110	2	eyaleb	eyaleb	ADJ
fcis-7985	110	3	dataset	dataset	NOUN
fcis-7985	110	4	,	,	PUNCT
fcis-7985	110	5	the	the	DET
fcis-7985	110	6	optimal	optimal	ADJ
fcis-7985	110	7	parameters	parameter	NOUN
fcis-7985	110	8	are	be	AUX
fcis-7985	110	9	=	=	NOUN
fcis-7985	110	10	10	10	NUM
fcis-7985	110	11	-	-	SYM
fcis-7985	110	12	2	2	NUM
fcis-7985	110	13	,	,	PUNCT
fcis-7985	110	14	=	=	NOUN
fcis-7985	110	15	10	10	NUM
fcis-7985	110	16	-	-	SYM
fcis-7985	110	17	4	4	NUM
fcis-7985	110	18	,	,	PUNCT
fcis-7985	110	19	=	=	NOUN
fcis-7985	110	20	10	10	NUM
fcis-7985	110	21	-	-	SYM
fcis-7985	110	22	4	4	NUM
fcis-7985	110	23	and	and	CCONJ
fcis-7985	110	24	=	=	NOUN
fcis-7985	110	25	10	10	NUM
fcis-7985	110	26	-	-	SYM
fcis-7985	110	27	4	4	NUM
fcis-7985	110	28	.	.	PUNCT
fcis-7985	110	29	figure	figure	NOUN
fcis-7985	110	30	3	3	NUM
fcis-7985	110	31	.	.	PUNCT
fcis-7985	110	32	classification	classification	NOUN
fcis-7985	110	33	accuracy	accuracy	NOUN
fcis-7985	110	34	under	under	ADP
fcis-7985	110	35	different	different	ADJ
fcis-7985	110	36	parameter	parameter	NOUN
fcis-7985	110	37	combinations	combination	NOUN
fcis-7985	110	38	4	4	NUM
fcis-7985	110	39	.	.	PUNCT
fcis-7985	110	40	conclusion	conclusion	NOUN
fcis-7985	110	41	in	in	ADP
fcis-7985	110	42	this	this	DET
fcis-7985	110	43	paper	paper	NOUN
fcis-7985	110	44	,	,	PUNCT
fcis-7985	110	45	we	we	PRON
fcis-7985	110	46	develop	develop	VERB
fcis-7985	110	47	a	a	DET
fcis-7985	110	48	nnsdlrpp	nnsdlrpp	NOUN
fcis-7985	110	49	model	model	NOUN
fcis-7985	110	50	for	for	ADP
fcis-7985	110	51	face	face	NOUN
fcis-7985	110	52	feature	feature	NOUN
fcis-7985	110	53	extraction	extraction	NOUN
fcis-7985	110	54	and	and	CCONJ
fcis-7985	110	55	classification	classification	NOUN
fcis-7985	110	56	.	.	PUNCT
fcis-7985	111	1	in	in	ADP
fcis-7985	111	2	order	order	NOUN
fcis-7985	111	3	to	to	PART
fcis-7985	111	4	get	get	VERB
fcis-7985	111	5	a	a	DET
fcis-7985	111	6	more	more	ADV
fcis-7985	111	7	discriminant	discriminant	ADJ
fcis-7985	111	8	feature	feature	NOUN
fcis-7985	111	9	distribution	distribution	NOUN
fcis-7985	111	10	,	,	PUNCT
fcis-7985	111	11	a	a	DET
fcis-7985	111	12	group	group	NOUN
fcis-7985	111	13	-	-	PUNCT
fcis-7985	111	14	wise	wise	ADJ
fcis-7985	111	15	regularization	regularization	NOUN
fcis-7985	111	16	is	be	AUX
fcis-7985	111	17	introduced	introduce	VERB
fcis-7985	111	18	.	.	PUNCT
fcis-7985	112	1	the	the	DET
fcis-7985	112	2	non	non	ADJ
fcis-7985	112	3	-	-	ADJ
fcis-7985	112	4	negative	negative	ADJ
fcis-7985	112	5	constraint	constraint	NOUN
fcis-7985	112	6	improves	improve	VERB
fcis-7985	112	7	the	the	DET
fcis-7985	112	8	interpretability	interpretability	NOUN
fcis-7985	112	9	of	of	ADP
fcis-7985	112	10	representation	representation	NOUN
fcis-7985	112	11	while	while	SCONJ
fcis-7985	112	12	the	the	DET
fcis-7985	112	13	sparse	sparse	ADJ
fcis-7985	112	14	constraint	constraint	NOUN
fcis-7985	112	15	help	help	NOUN
fcis-7985	112	16	capture	capture	VERB
fcis-7985	112	17	the	the	DET
fcis-7985	112	18	local	local	ADJ
fcis-7985	112	19	information	information	NOUN
fcis-7985	112	20	.	.	PUNCT
fcis-7985	113	1	we	we	PRON
fcis-7985	113	2	develop	develop	VERB
fcis-7985	113	3	an	an	DET
fcis-7985	113	4	alternate	alternate	ADJ
fcis-7985	113	5	algorithm	algorithm	NOUN
fcis-7985	113	6	to	to	PART
fcis-7985	113	7	solve	solve	VERB
fcis-7985	113	8	the	the	DET
fcis-7985	113	9	model	model	NOUN
fcis-7985	113	10	.	.	PUNCT
fcis-7985	114	1	extensive	extensive	ADJ
fcis-7985	114	2	experiments	experiment	NOUN
fcis-7985	114	3	on	on	ADP
fcis-7985	114	4	two	two	NUM
fcis-7985	114	5	well	well	ADV
fcis-7985	114	6	-	-	PUNCT
fcis-7985	114	7	known	know	VERB
fcis-7985	114	8	databases	database	NOUN
fcis-7985	114	9	have	have	AUX
fcis-7985	114	10	demonstrated	demonstrate	VERB
fcis-7985	114	11	the	the	DET
fcis-7985	114	12	superiority	superiority	NOUN
fcis-7985	114	13	and	and	CCONJ
fcis-7985	114	14	effectiveness	effectiveness	NOUN
fcis-7985	114	15	of	of	ADP
fcis-7985	114	16	our	our	PRON
fcis-7985	114	17	method	method	NOUN
fcis-7985	114	18	.	.	PUNCT
fcis-7985	115	1	references	reference	NOUN
fcis-7985	115	2	[	[	X
fcis-7985	115	3	1	1	X
fcis-7985	115	4	]	]	X
fcis-7985	115	5	lei	lei	PROPN
fcis-7985	115	6	chen	chen	PROPN
fcis-7985	115	7	,	,	PUNCT
fcis-7985	115	8	jianjun	jianjun	PROPN
fcis-7985	115	9	xu	xu	PROPN
fcis-7985	115	10	,	,	PUNCT
fcis-7985	115	11	shuang	shuang	PROPN
fcis-7985	115	12	chen	chen	PROPN
fcis-7985	115	13	,	,	PUNCT
fcis-7985	115	14	hui	hui	PROPN
fcis-7985	115	15	yang	yang	PROPN
fcis-7985	115	16	,	,	PUNCT
fcis-7985	115	17	linhu	linhu	PROPN
fcis-7985	115	18	liu	liu	PROPN
fcis-7985	115	19	.	.	PUNCT
fcis-7985	116	1	power	power	NOUN
fcis-7985	116	2	quality	quality	PROPN
fcis-7985	116	3	disturbance	disturbance	NOUN
fcis-7985	116	4	signal	signal	NOUN
fcis-7985	116	5	denoising	denoising	NOUN
fcis-7985	116	6	based	base	VERB
fcis-7985	116	7	on	on	ADP
fcis-7985	116	8	overcomplete	overcomplete	ADJ
fcis-7985	116	9	representation	representation	NOUN
fcis-7985	116	10	.	.	PUNCT
fcis-7985	117	1	ieej	ieej	NOUN
fcis-7985	117	2	transactions	transaction	NOUN
fcis-7985	117	3	on	on	ADP
fcis-7985	117	4	electrical	electrical	ADJ
fcis-7985	117	5	and	and	CCONJ
fcis-7985	117	6	electronic	electronic	ADJ
fcis-7985	117	7	engineering.2022	engineering.2022	NOUN
fcis-7985	117	8	;	;	PUNCT
fcis-7985	117	9	17	17	NUM
fcis-7985	117	10	:	:	PUNCT
fcis-7985	117	11	544–555	544–555	NUM
fcis-7985	117	12	.	.	PUNCT
fcis-7985	118	1	[	[	X
fcis-7985	118	2	2	2	NUM
fcis-7985	118	3	]	]	PUNCT
fcis-7985	118	4	xu	xu	PROPN
fcis-7985	118	5	jianjun	jianjun	PROPN
fcis-7985	118	6	,	,	PUNCT
fcis-7985	118	7	huang	huang	PROPN
fcis-7985	118	8	lida	lida	PROPN
fcis-7985	118	9	,	,	PUNCT
fcis-7985	118	10	yan	yan	PROPN
fcis-7985	118	11	limei	limei	PROPN
fcis-7985	118	12	,	,	PUNCT
fcis-7985	118	13	yi	yi	PROPN
fcis-7985	118	14	na	na	NOUN
fcis-7985	118	15	.	.	NOUN
fcis-7985	118	16	insulator	insulator	NOUN
fcis-7985	118	17	selfexplosion	selfexplosion	NOUN
fcis-7985	118	18	defect	defect	NOUN
fcis-7985	118	19	detection	detection	NOUN
fcis-7985	118	20	based	base	VERB
fcis-7985	118	21	on	on	ADP
fcis-7985	118	22	hierarchical	hierarchical	ADJ
fcis-7985	118	23	multi	multi	NOUN
fcis-7985	118	24	-	-	NOUN
fcis-7985	118	25	task	task	ADJ
fcis-7985	118	26	18	18	NUM
fcis-7985	118	27	deep	deep	ADJ
fcis-7985	118	28	learning	learning	NOUN
fcis-7985	118	29	.	.	PUNCT
fcis-7985	119	1	transactions	transaction	NOUN
fcis-7985	119	2	of	of	ADP
fcis-7985	119	3	china	china	PROPN
fcis-7985	119	4	electrotechnical	electrotechnical	ADJ
fcis-7985	119	5	society	society	NOUN
fcis-7985	119	6	,	,	PUNCT
fcis-7985	119	7	2021,36(07):1407	2021,36(07):1407	NUM
fcis-7985	119	8	-	-	SYM
fcis-7985	119	9	1415	1415	NUM
fcis-7985	119	10	.	.	PUNCT
fcis-7985	120	1	[	[	X
fcis-7985	120	2	3	3	X
fcis-7985	120	3	]	]	X
fcis-7985	120	4	yan	yan	PROPN
fcis-7985	120	5	limei	limei	PROPN
fcis-7985	120	6	,	,	PUNCT
fcis-7985	120	7	liu	liu	PROPN
fcis-7985	120	8	yongqiang	yongqiang	PROPN
fcis-7985	120	9	,	,	PUNCT
fcis-7985	120	10	xu	xu	PROPN
fcis-7985	120	11	jianjun	jianjun	PROPN
fcis-7985	120	12	,	,	PUNCT
fcis-7985	120	13	et	et	PROPN
fcis-7985	120	14	al.broken	al.broken	VERB
fcis-7985	120	15	string	string	NOUN
fcis-7985	120	16	diagnosis	diagnosis	NOUN
fcis-7985	120	17	of	of	ADP
fcis-7985	120	18	composite	composite	ADJ
fcis-7985	120	19	insulator	insulator	NOUN
fcis-7985	120	20	based	base	VERB
fcis-7985	120	21	on	on	ADP
fcis-7985	120	22	grabcut	grabcut	NOUN
fcis-7985	120	23	segmentation	segmentation	NOUN
fcis-7985	120	24	and	and	CCONJ
fcis-7985	120	25	filler	filler	NOUN
fcis-7985	120	26	area	area	NOUN
fcis-7985	120	27	discrimination.power	discrimination.power	NOUN
fcis-7985	120	28	system	system	NOUN
fcis-7985	120	29	protection	protection	NOUN
fcis-7985	120	30	and	and	CCONJ
fcis-7985	120	31	control,2021,49(22):114	control,2021,49(22):114	NOUN
fcis-7985	120	32	-	-	SYM
fcis-7985	120	33	119	119	NUM
fcis-7985	120	34	.	.	PUNCT
fcis-7985	121	1	[	[	X
fcis-7985	121	2	4	4	NUM
fcis-7985	121	3	]	]	SYM
fcis-7985	121	4	yi	yi	PROPN
fcis-7985	121	5	,	,	PUNCT
fcis-7985	121	6	q.	q.	PROPN
fcis-7985	121	7	wang	wang	PROPN
fcis-7985	121	8	,	,	PUNCT
fcis-7985	121	9	l.	l.	PROPN
fcis-7985	121	10	yan	yan	PROPN
fcis-7985	121	11	,	,	PUNCT
fcis-7985	121	12	et	et	PROPN
fcis-7985	121	13	al	al	PROPN
fcis-7985	121	14	.	.	PROPN
fcis-7985	121	15	,	,	PUNCT
fcis-7985	121	16	a	a	DET
fcis-7985	121	17	multi	multi	ADJ
fcis-7985	121	18	-	-	ADJ
fcis-7985	121	19	stage	stage	ADJ
fcis-7985	121	20	game	game	NOUN
fcis-7985	121	21	model	model	NOUN
fcis-7985	121	22	for	for	ADP
fcis-7985	121	23	the	the	DET
fcis-7985	121	24	false	false	ADJ
fcis-7985	121	25	data	datum	NOUN
fcis-7985	121	26	injection	injection	NOUN
fcis-7985	121	27	attack	attack	NOUN
fcis-7985	121	28	from	from	ADP
fcis-7985	121	29	attacker	attacker	PROPN
fcis-7985	121	30	’s	’s	PART
fcis-7985	121	31	perspective	perspective	NOUN
fcis-7985	121	32	.	.	PUNCT
fcis-7985	122	1	sustainable	sustainable	ADJ
fcis-7985	122	2	energy	energy	NOUN
fcis-7985	122	3	grids	grid	NOUN
fcis-7985	122	4	&	&	CCONJ
fcis-7985	122	5	networks	network	NOUN
fcis-7985	122	6	28	28	NUM
fcis-7985	122	7	(	(	PUNCT
fcis-7985	122	8	2021	2021	NUM
fcis-7985	122	9	)	)	PUNCT
fcis-7985	122	10	.	.	PUNCT
fcis-7985	123	1	[	[	X
fcis-7985	123	2	5	5	X
fcis-7985	123	3	]	]	PUNCT
fcis-7985	123	4	na	na	PART
fcis-7985	123	5	yi	yi	PROPN
fcis-7985	123	6	,	,	PUNCT
fcis-7985	123	7	jianjun	jianjun	PROPN
fcis-7985	123	8	xu	xu	PROPN
fcis-7985	123	9	,	,	PUNCT
fcis-7985	123	10	limei	limei	PROPN
fcis-7985	123	11	yan	yan	PROPN
fcis-7985	123	12	,	,	PUNCT
fcis-7985	123	13	lin	lin	PROPN
fcis-7985	123	14	huang	huang	PROPN
fcis-7985	123	15	.	.	PROPN
fcis-7985	124	1	task	task	PROPN
fcis-7985	124	2	optimization	optimization	NOUN
fcis-7985	124	3	and	and	CCONJ
fcis-7985	124	4	scheduling	scheduling	NOUN
fcis-7985	124	5	of	of	ADP
fcis-7985	124	6	distributed	distribute	VERB
fcis-7985	124	7	cyber	cyber	NOUN
fcis-7985	124	8	-	-	PUNCT
fcis-7985	124	9	physical	physical	ADJ
fcis-7985	124	10	system	system	NOUN
fcis-7985	124	11	based	base	VERB
fcis-7985	124	12	on	on	ADP
fcis-7985	124	13	improved	improve	VERB
fcis-7985	124	14	ant	ant	ADJ
fcis-7985	124	15	colony	colony	NOUN
fcis-7985	124	16	algorithm	algorithm	NOUN
fcis-7985	124	17	.	.	PUNCT
fcis-7985	125	1	future	future	ADJ
fcis-7985	125	2	generation	generation	NOUN
fcis-7985	125	3	computer	computer	NOUN
fcis-7985	125	4	systems	system	NOUN
fcis-7985	125	5	,	,	PUNCT
fcis-7985	125	6	109(aug	109(aug	NUM
fcis-7985	125	7	.	.	PUNCT
fcis-7985	126	1	2020),134	2020),134	NUM
fcis-7985	126	2	-	-	SYM
fcis-7985	126	3	148	148	NUM
fcis-7985	126	4	.	.	PUNCT
fcis-7985	127	1	[	[	X
fcis-7985	127	2	6	6	NUM
fcis-7985	127	3	]	]	PUNCT
fcis-7985	127	4	yang	yang	PROPN
fcis-7985	127	5	zhao	zhao	PROPN
fcis-7985	127	6	,	,	PUNCT
fcis-7985	127	7	jianjun	jianjun	PROPN
fcis-7985	127	8	xu	xu	PROPN
fcis-7985	127	9	,	,	PUNCT
fcis-7985	127	10	jingchun	jingchun	PROPN
fcis-7985	127	11	wu	wu	PROPN
fcis-7985	127	12	.	.	PUNCT
fcis-7985	128	1	a	a	DET
fcis-7985	128	2	new	new	ADJ
fcis-7985	128	3	method	method	NOUN
fcis-7985	128	4	for	for	ADP
fcis-7985	128	5	bad	bad	ADJ
fcis-7985	128	6	data	datum	NOUN
fcis-7985	128	7	identification	identification	NOUN
fcis-7985	128	8	of	of	ADP
fcis-7985	128	9	oilfield	oilfield	NOUN
fcis-7985	128	10	power	power	NOUN
fcis-7985	128	11	system	system	NOUN
fcis-7985	128	12	based	base	VERB
fcis-7985	128	13	on	on	ADP
fcis-7985	128	14	enhanced	enhanced	ADJ
fcis-7985	128	15	gravitational	gravitational	ADJ
fcis-7985	128	16	search	search	NOUN
fcis-7985	128	17	-	-	PUNCT
fcis-7985	128	18	fuzzy	fuzzy	ADJ
fcis-7985	128	19	c	c	NOUN
fcis-7985	128	20	-	-	PUNCT
fcis-7985	128	21	means	means	NOUN
fcis-7985	128	22	algorithm	algorithm	NOUN
fcis-7985	128	23	.	.	PUNCT
fcis-7985	129	1	ieee	ieee	NOUN
fcis-7985	129	2	transactions	transaction	NOUN
fcis-7985	129	3	on	on	ADP
fcis-7985	129	4	industrial	industrial	ADJ
fcis-7985	129	5	informatics	informatic	NOUN
fcis-7985	129	6	.	.	PUNCT
fcis-7985	130	1	vol	vol	NOUN
fcis-7985	130	2	.	.	PROPN
fcis-7985	131	1	15	15	NUM
fcis-7985	131	2	,	,	PUNCT
fcis-7985	131	3	no	no	INTJ
fcis-7985	131	4	.	.	NOUN
fcis-7985	131	5	11	11	NUM
fcis-7985	131	6	,	,	PUNCT
fcis-7985	131	7	november	november	PROPN
fcis-7985	131	8	2019	2019	NUM
fcis-7985	131	9	5963	5963	NUM
fcis-7985	131	10	-	-	SYM
fcis-7985	131	11	5970	5970	NUM
fcis-7985	131	12	.	.	PUNCT
fcis-7985	132	1	[	[	X
fcis-7985	132	2	7	7	NUM
fcis-7985	132	3	]	]	X
fcis-7985	132	4	cai	cai	X
fcis-7985	132	5	,	,	PUNCT
fcis-7985	132	6	j.	j.	PROPN
fcis-7985	132	7	f.	f.	PROPN
fcis-7985	132	8	,	,	PUNCT
fcis-7985	132	9	candès	candès	PROPN
fcis-7985	132	10	,	,	PUNCT
fcis-7985	132	11	e.	e.	PROPN
fcis-7985	132	12	j.	j.	PROPN
fcis-7985	132	13	,	,	PUNCT
fcis-7985	132	14	&	&	CCONJ
fcis-7985	132	15	shen	shen	PROPN
fcis-7985	132	16	,	,	PUNCT
fcis-7985	132	17	z.	z.	PROPN
fcis-7985	132	18	(	(	PUNCT
fcis-7985	132	19	2010	2010	NUM
fcis-7985	132	20	)	)	PUNCT
fcis-7985	132	21	.	.	PUNCT
fcis-7985	133	1	a	a	PRON
fcis-7985	133	2	singular	singular	ADJ
fcis-7985	133	3	value	value	NOUN
fcis-7985	133	4	thresholding	thresholde	VERB
fcis-7985	133	5	algorithm	algorithm	NOUN
fcis-7985	133	6	for	for	ADP
fcis-7985	133	7	matrix	matrix	NOUN
fcis-7985	133	8	completion	completion	NOUN
fcis-7985	133	9	.	.	PUNCT
fcis-7985	134	1	siam	siam	PROPN
fcis-7985	134	2	journal	journal	PROPN
fcis-7985	134	3	on	on	ADP
fcis-7985	134	4	optimization	optimization	NOUN
fcis-7985	134	5	,	,	PUNCT
fcis-7985	134	6	20(4	20(4	NOUN
fcis-7985	134	7	)	)	PUNCT
fcis-7985	134	8	,	,	PUNCT
fcis-7985	134	9	1956	1956	NUM
fcis-7985	134	10	-	-	SYM
fcis-7985	134	11	1982	1982	NUM
fcis-7985	134	12	.	.	PUNCT
fcis-7985	135	1	[	[	X
fcis-7985	135	2	8	8	NUM
fcis-7985	135	3	]	]	X
fcis-7985	135	4	lin	lin	PROPN
fcis-7985	135	5	,	,	PUNCT
fcis-7985	135	6	z.	z.	PROPN
fcis-7985	135	7	,	,	PUNCT
fcis-7985	135	8	chen	chen	PROPN
fcis-7985	135	9	,	,	PUNCT
fcis-7985	135	10	m.	m.	NOUN
fcis-7985	135	11	,	,	PUNCT
fcis-7985	135	12	&	&	CCONJ
fcis-7985	135	13	ma	ma	PROPN
fcis-7985	135	14	,	,	PUNCT
fcis-7985	135	15	y.	y.	PROPN
fcis-7985	135	16	(	(	PUNCT
fcis-7985	135	17	2010	2010	NUM
fcis-7985	135	18	)	)	PUNCT
fcis-7985	135	19	.	.	PUNCT
fcis-7985	136	1	the	the	DET
fcis-7985	136	2	augmented	augment	VERB
fcis-7985	136	3	lagrange	lagrange	NOUN
fcis-7985	136	4	multiplier	multipli	ADJ
fcis-7985	136	5	method	method	NOUN
fcis-7985	136	6	for	for	ADP
fcis-7985	136	7	exact	exact	ADJ
fcis-7985	136	8	recovery	recovery	NOUN
fcis-7985	136	9	of	of	ADP
fcis-7985	136	10	corrupted	corrupted	ADJ
fcis-7985	136	11	low	low	ADJ
fcis-7985	136	12	-	-	PUNCT
fcis-7985	136	13	rank	rank	NOUN
fcis-7985	136	14	matrices	matrix	NOUN
fcis-7985	136	15	.	.	PUNCT
fcis-7985	137	1	arxiv	arxiv	PROPN
fcis-7985	137	2	preprint	preprint	NOUN
fcis-7985	137	3	arxiv:1009.5055	arxiv:1009.5055	NOUN
fcis-7985	137	4	.	.	PUNCT