id	sid	tid	token	lemma	pos
fcis-17883	1	1	frontiers	frontier	NOUN
fcis-17883	1	2	in	in	ADP
fcis-17883	1	3	computing	computing	NOUN
fcis-17883	1	4	and	and	CCONJ
fcis-17883	1	5	intelligent	intelligent	ADJ
fcis-17883	1	6	systems	system	NOUN
fcis-17883	1	7	issn	issn	VERB
fcis-17883	1	8	:	:	PUNCT
fcis-17883	1	9	2832	2832	NUM
fcis-17883	1	10	-	-	SYM
fcis-17883	1	11	6024	6024	NUM
fcis-17883	1	12	|	|	NOUN
fcis-17883	1	13	vol	vol	NOUN
fcis-17883	1	14	.	.	PROPN
fcis-17883	2	1	7	7	NUM
fcis-17883	2	2	,	,	PUNCT
fcis-17883	2	3	no	no	INTJ
fcis-17883	2	4	.	.	NOUN
fcis-17883	2	5	2	2	NUM
fcis-17883	2	6	,	,	PUNCT
fcis-17883	2	7	2024	2024	NUM
fcis-17883	2	8	34	34	NUM
fcis-17883	2	9	a	a	DET
fcis-17883	2	10	sigma‐pi‐sigma	sigma‐pi‐sigma	PROPN
fcis-17883	2	11	neural	neural	ADJ
fcis-17883	2	12	network	network	NOUN
fcis-17883	2	13	model	model	NOUN
fcis-17883	2	14	with	with	ADP
fcis-17883	2	15	graph	graph	NOUN
fcis-17883	2	16	regularity	regularity	NOUN
fcis-17883	2	17	term	term	NOUN
fcis-17883	2	18	qianru	qianru	NOUN
fcis-17883	2	19	huang	huang	PROPN
fcis-17883	2	20	*	*	PROPN
fcis-17883	2	21	,	,	PUNCT
fcis-17883	2	22	qingmei	qingmei	NUM
fcis-17883	2	23	dong	dong	PROPN
fcis-17883	2	24	,	,	PUNCT
fcis-17883	2	25	yunlong	yunlong	PROPN
fcis-17883	2	26	liu	liu	PROPN
fcis-17883	2	27	,	,	PUNCT
fcis-17883	2	28	deqing	deqing	PROPN
fcis-17883	2	29	ji	ji	PROPN
fcis-17883	2	30	,	,	PUNCT
fcis-17883	2	31	qinwei	qinwei	VERB
fcis-17883	2	32	fan	fan	PROPN
fcis-17883	2	33	xi’an	xi’an	PROPN
fcis-17883	2	34	polytechnic	polytechnic	PROPN
fcis-17883	2	35	university	university	PROPN
fcis-17883	2	36	,	,	PUNCT
fcis-17883	2	37	xi’an	xi’an	PROPN
fcis-17883	2	38	,	,	PUNCT
fcis-17883	2	39	shaanxi	shaanxi	PROPN
fcis-17883	2	40	,	,	PUNCT
fcis-17883	2	41	710600	710600	NUM
fcis-17883	2	42	,	,	PUNCT
fcis-17883	2	43	china	china	PROPN
fcis-17883	2	44	*	*	PUNCT
fcis-17883	2	45	corresponding	correspond	VERB
fcis-17883	2	46	author	author	NOUN
fcis-17883	2	47	:	:	PUNCT
fcis-17883	2	48	qianru	qianru	PROPN
fcis-17883	2	49	huang	huang	PROPN
fcis-17883	2	50	abstract	abstract	PROPN
fcis-17883	2	51	:	:	PUNCT
fcis-17883	2	52	in	in	ADP
fcis-17883	2	53	recent	recent	ADJ
fcis-17883	2	54	years	year	NOUN
fcis-17883	2	55	,	,	PUNCT
fcis-17883	2	56	sigma	sigma	PROPN
fcis-17883	2	57	-	-	PUNCT
fcis-17883	2	58	pi	pi	NOUN
fcis-17883	2	59	-	-	PUNCT
fcis-17883	2	60	sigma	sigma	NOUN
fcis-17883	2	61	neural	neural	ADJ
fcis-17883	2	62	network	network	NOUN
fcis-17883	2	63	(	(	PUNCT
fcis-17883	2	64	spsnn	spsnn	NOUN
fcis-17883	2	65	)	)	PUNCT
fcis-17883	2	66	as	as	ADP
fcis-17883	2	67	a	a	DET
fcis-17883	2	68	special	special	ADJ
fcis-17883	2	69	kind	kind	NOUN
fcis-17883	2	70	of	of	ADP
fcis-17883	2	71	higher	high	ADJ
fcis-17883	2	72	-	-	PUNCT
fcis-17883	2	73	order	order	NOUN
fcis-17883	2	74	neural	neural	ADJ
fcis-17883	2	75	network	network	NOUN
fcis-17883	2	76	has	have	AUX
fcis-17883	2	77	attracted	attract	VERB
fcis-17883	2	78	wide	wide	ADJ
fcis-17883	2	79	attention	attention	NOUN
fcis-17883	2	80	for	for	ADP
fcis-17883	2	81	its	its	PRON
fcis-17883	2	82	fast	fast	ADJ
fcis-17883	2	83	convergence	convergence	NOUN
fcis-17883	2	84	speed	speed	NOUN
fcis-17883	2	85	and	and	CCONJ
fcis-17883	2	86	good	good	ADJ
fcis-17883	2	87	approximation	approximation	NOUN
fcis-17883	2	88	ability	ability	NOUN
fcis-17883	2	89	.	.	PUNCT
fcis-17883	3	1	however	however	ADV
fcis-17883	3	2	,	,	PUNCT
fcis-17883	3	3	an	an	DET
fcis-17883	3	4	inappropriate	inappropriate	ADJ
fcis-17883	3	5	number	number	NOUN
fcis-17883	3	6	of	of	ADP
fcis-17883	3	7	hidden	hide	VERB
fcis-17883	3	8	layer	layer	NOUN
fcis-17883	3	9	neurons	neuron	NOUN
fcis-17883	3	10	may	may	AUX
fcis-17883	3	11	also	also	ADV
fcis-17883	3	12	lead	lead	VERB
fcis-17883	3	13	to	to	ADP
fcis-17883	3	14	model	model	NOUN
fcis-17883	3	15	underfitting	underfitting	NOUN
fcis-17883	3	16	or	or	CCONJ
fcis-17883	3	17	overfitting	overfitting	NOUN
fcis-17883	3	18	,	,	PUNCT
fcis-17883	3	19	which	which	PRON
fcis-17883	3	20	affects	affect	VERB
fcis-17883	3	21	the	the	DET
fcis-17883	3	22	performance	performance	NOUN
fcis-17883	3	23	and	and	CCONJ
fcis-17883	3	24	generalization	generalization	NOUN
fcis-17883	3	25	ability	ability	NOUN
fcis-17883	3	26	of	of	ADP
fcis-17883	3	27	the	the	DET
fcis-17883	3	28	model	model	NOUN
fcis-17883	3	29	.	.	PUNCT
fcis-17883	4	1	therefore	therefore	ADV
fcis-17883	4	2	,	,	PUNCT
fcis-17883	4	3	we	we	PRON
fcis-17883	4	4	propose	propose	VERB
fcis-17883	4	5	a	a	DET
fcis-17883	4	6	sigma	sigma	PROPN
fcis-17883	4	7	-	-	PUNCT
fcis-17883	4	8	pi	pi	NOUN
fcis-17883	4	9	-	-	PUNCT
fcis-17883	4	10	sigma	sigma	NOUN
fcis-17883	4	11	neural	neural	ADJ
fcis-17883	4	12	network	network	NOUN
fcis-17883	4	13	with	with	ADP
fcis-17883	4	14	graph	graph	NOUN
fcis-17883	4	15	regularity	regularity	NOUN
fcis-17883	4	16	by	by	ADP
fcis-17883	4	17	adding	add	VERB
fcis-17883	4	18	a	a	DET
fcis-17883	4	19	graph	graph	NOUN
fcis-17883	4	20	regularity	regularity	NOUN
fcis-17883	4	21	term	term	NOUN
fcis-17883	4	22	to	to	ADP
fcis-17883	4	23	the	the	DET
fcis-17883	4	24	network	network	NOUN
fcis-17883	4	25	.	.	PUNCT
fcis-17883	5	1	the	the	DET
fcis-17883	5	2	results	result	NOUN
fcis-17883	5	3	show	show	VERB
fcis-17883	5	4	that	that	SCONJ
fcis-17883	5	5	the	the	DET
fcis-17883	5	6	proposed	propose	VERB
fcis-17883	5	7	algorithm	algorithm	NOUN
fcis-17883	5	8	performs	perform	VERB
fcis-17883	5	9	well	well	ADV
fcis-17883	5	10	in	in	ADP
fcis-17883	5	11	terms	term	NOUN
fcis-17883	5	12	of	of	ADP
fcis-17883	5	13	training	training	NOUN
fcis-17883	5	14	accuracy	accuracy	NOUN
fcis-17883	5	15	,	,	PUNCT
fcis-17883	5	16	testing	testing	NOUN
fcis-17883	5	17	accuracy	accuracy	NOUN
fcis-17883	5	18	and	and	CCONJ
fcis-17883	5	19	efficiency	efficiency	NOUN
fcis-17883	5	20	.	.	PUNCT
fcis-17883	6	1	keywords	keyword	NOUN
fcis-17883	6	2	:	:	PUNCT
fcis-17883	6	3	sigma	sigma	PROPN
fcis-17883	6	4	-	-	PUNCT
fcis-17883	6	5	pi	pi	NOUN
fcis-17883	6	6	-	-	PUNCT
fcis-17883	6	7	sigma	sigma	NOUN
fcis-17883	6	8	neural	neural	ADJ
fcis-17883	6	9	network	network	NOUN
fcis-17883	6	10	;	;	PUNCT
fcis-17883	6	11	graph	graph	NOUN
fcis-17883	6	12	regularization	regularization	NOUN
fcis-17883	6	13	;	;	PUNCT
fcis-17883	6	14	squared	square	VERB
fcis-17883	6	15	error	error	NOUN
fcis-17883	6	16	.	.	PUNCT
fcis-17883	7	1	1	1	X
fcis-17883	7	2	.	.	X
fcis-17883	7	3	introduction	introduction	NOUN
fcis-17883	7	4	in	in	ADP
fcis-17883	7	5	recent	recent	ADJ
fcis-17883	7	6	years	year	NOUN
fcis-17883	7	7	,	,	PUNCT
fcis-17883	7	8	neural	neural	ADJ
fcis-17883	7	9	networks	network	NOUN
fcis-17883	7	10	have	have	AUX
fcis-17883	7	11	become	become	VERB
fcis-17883	7	12	one	one	NUM
fcis-17883	7	13	of	of	ADP
fcis-17883	7	14	the	the	DET
fcis-17883	7	15	focuses	focus	NOUN
fcis-17883	7	16	of	of	ADP
fcis-17883	7	17	today	today	NOUN
fcis-17883	7	18	's	's	PART
fcis-17883	7	19	science	science	NOUN
fcis-17883	7	20	and	and	CCONJ
fcis-17883	7	21	technology	technology	NOUN
fcis-17883	7	22	field	field	NOUN
fcis-17883	7	23	and	and	CCONJ
fcis-17883	7	24	have	have	AUX
fcis-17883	7	25	received	receive	VERB
fcis-17883	7	26	widespread	widespread	ADJ
fcis-17883	7	27	attention	attention	NOUN
fcis-17883	7	28	due	due	ADP
fcis-17883	7	29	to	to	ADP
fcis-17883	7	30	their	their	PRON
fcis-17883	7	31	powerful	powerful	ADJ
fcis-17883	7	32	learning	learning	NOUN
fcis-17883	7	33	ability	ability	NOUN
fcis-17883	7	34	and	and	CCONJ
fcis-17883	7	35	wide	wide	ADJ
fcis-17883	7	36	application	application	NOUN
fcis-17883	7	37	prospects	prospect	NOUN
fcis-17883	7	38	.	.	PUNCT
fcis-17883	8	1	sigma	sigma	PROPN
fcis-17883	8	2	-	-	PUNCT
fcis-17883	8	3	pi	pi	NOUN
fcis-17883	8	4	-	-	PUNCT
fcis-17883	8	5	sigma	sigma	NOUN
fcis-17883	8	6	neural	neural	ADJ
fcis-17883	8	7	network	network	NOUN
fcis-17883	8	8	(	(	PUNCT
fcis-17883	8	9	spsnn	spsnn	NOUN
fcis-17883	8	10	)	)	PUNCT
fcis-17883	8	11	is	be	AUX
fcis-17883	8	12	a	a	DET
fcis-17883	8	13	kind	kind	NOUN
fcis-17883	8	14	of	of	ADP
fcis-17883	8	15	feed	feed	NOUN
fcis-17883	8	16	-	-	PUNCT
fcis-17883	8	17	forward	forward	ADV
fcis-17883	8	18	neural	neural	ADJ
fcis-17883	8	19	network	network	NOUN
fcis-17883	8	20	[	[	X
fcis-17883	8	21	1	1	NUM
fcis-17883	8	22	]	]	PUNCT
fcis-17883	8	23	,	,	PUNCT
fcis-17883	8	24	which	which	PRON
fcis-17883	8	25	consists	consist	VERB
fcis-17883	8	26	of	of	ADP
fcis-17883	8	27	different	different	ADJ
fcis-17883	8	28	orders	order	NOUN
fcis-17883	8	29	of	of	ADP
fcis-17883	8	30	sigma	sigma	PROPN
fcis-17883	8	31	-	-	PUNCT
fcis-17883	8	32	pi	pi	NOUN
fcis-17883	8	33	neural	neural	ADJ
fcis-17883	8	34	networks	network	NOUN
fcis-17883	8	35	(	(	PUNCT
fcis-17883	8	36	psns	psns	NOUN
fcis-17883	8	37	)	)	PUNCT
fcis-17883	8	38	,	,	PUNCT
fcis-17883	8	39	and	and	CCONJ
fcis-17883	8	40	is	be	AUX
fcis-17883	8	41	a	a	DET
fcis-17883	8	42	high	high	ADJ
fcis-17883	8	43	-	-	PUNCT
fcis-17883	8	44	order	order	NOUN
fcis-17883	8	45	neural	neural	ADJ
fcis-17883	8	46	network	network	NOUN
fcis-17883	8	47	with	with	ADP
fcis-17883	8	48	a	a	DET
fcis-17883	8	49	multiplicative	multiplicative	ADJ
fcis-17883	8	50	unit	unit	NOUN
fcis-17883	8	51	and	and	CCONJ
fcis-17883	8	52	stronger	strong	ADJ
fcis-17883	8	53	nonlinear	nonlinear	ADJ
fcis-17883	8	54	mapping	mapping	NOUN
fcis-17883	8	55	ability	ability	NOUN
fcis-17883	8	56	,	,	PUNCT
fcis-17883	8	57	with	with	ADP
fcis-17883	8	58	its	its	PRON
fcis-17883	8	59	fast	fast	ADJ
fcis-17883	8	60	convergence	convergence	NOUN
fcis-17883	8	61	speed	speed	NOUN
fcis-17883	8	62	and	and	CCONJ
fcis-17883	8	63	good	good	ADJ
fcis-17883	8	64	approximation	approximation	NOUN
fcis-17883	8	65	ability	ability	NOUN
fcis-17883	8	66	.	.	PUNCT
fcis-17883	9	1	it	it	PRON
fcis-17883	9	2	is	be	AUX
fcis-17883	9	3	a	a	DET
fcis-17883	9	4	higher	high	ADJ
fcis-17883	9	5	order	order	NOUN
fcis-17883	9	6	neural	neural	ADJ
fcis-17883	9	7	network	network	NOUN
fcis-17883	9	8	with	with	ADP
fcis-17883	9	9	product	product	NOUN
fcis-17883	9	10	unit	unit	NOUN
fcis-17883	9	11	and	and	CCONJ
fcis-17883	9	12	stronger	strong	ADJ
fcis-17883	9	13	nonlinear	nonlinear	ADJ
fcis-17883	9	14	mapping	mapping	NOUN
fcis-17883	9	15	ability	ability	NOUN
fcis-17883	9	16	,	,	PUNCT
fcis-17883	9	17	which	which	PRON
fcis-17883	9	18	has	have	VERB
fcis-17883	9	19	its	its	PRON
fcis-17883	9	20	fast	fast	ADJ
fcis-17883	9	21	convergence	convergence	NOUN
fcis-17883	9	22	speed	speed	NOUN
fcis-17883	9	23	and	and	CCONJ
fcis-17883	9	24	good	good	ADJ
fcis-17883	9	25	approximation	approximation	NOUN
fcis-17883	9	26	ability	ability	NOUN
fcis-17883	9	27	[	[	X
fcis-17883	9	28	2	2	NUM
fcis-17883	9	29	]	]	PUNCT
fcis-17883	9	30	.	.	PUNCT
fcis-17883	10	1	due	due	ADP
fcis-17883	10	2	to	to	ADP
fcis-17883	10	3	its	its	PRON
fcis-17883	10	4	strong	strong	ADJ
fcis-17883	10	5	nonlinear	nonlinear	ADJ
fcis-17883	10	6	mapping	mapping	NOUN
fcis-17883	10	7	ability	ability	NOUN
fcis-17883	10	8	,	,	PUNCT
fcis-17883	10	9	this	this	DET
fcis-17883	10	10	network	network	NOUN
fcis-17883	10	11	has	have	AUX
fcis-17883	10	12	been	be	AUX
fcis-17883	10	13	widely	widely	ADV
fcis-17883	10	14	used	use	VERB
fcis-17883	10	15	in	in	ADP
fcis-17883	10	16	many	many	ADJ
fcis-17883	10	17	fields	field	NOUN
fcis-17883	10	18	such	such	ADJ
fcis-17883	10	19	as	as	ADP
fcis-17883	10	20	medical	medical	ADJ
fcis-17883	10	21	and	and	CCONJ
fcis-17883	10	22	industrial	industrial	ADJ
fcis-17883	10	23	[	[	X
fcis-17883	10	24	3	3	NUM
fcis-17883	10	25	-	-	SYM
fcis-17883	10	26	4	4	NUM
fcis-17883	10	27	]	]	PUNCT
fcis-17883	10	28	.	.	PUNCT
fcis-17883	11	1	generally	generally	ADV
fcis-17883	11	2	,	,	PUNCT
fcis-17883	11	3	we	we	PRON
fcis-17883	11	4	use	use	VERB
fcis-17883	11	5	gradient	gradient	ADJ
fcis-17883	11	6	descent	descent	NOUN
fcis-17883	11	7	method	method	NOUN
fcis-17883	11	8	to	to	PART
fcis-17883	11	9	train	train	VERB
fcis-17883	11	10	the	the	DET
fcis-17883	11	11	network	network	NOUN
fcis-17883	11	12	[	[	X
fcis-17883	11	13	5	5	NUM
fcis-17883	11	14	]	]	PUNCT
fcis-17883	11	15	.	.	PUNCT
fcis-17883	12	1	however	however	ADV
fcis-17883	12	2	,	,	PUNCT
fcis-17883	12	3	when	when	SCONJ
fcis-17883	12	4	we	we	PRON
fcis-17883	12	5	use	use	VERB
fcis-17883	12	6	the	the	DET
fcis-17883	12	7	squared	square	VERB
fcis-17883	12	8	error	error	NOUN
fcis-17883	12	9	sum	sum	NOUN
fcis-17883	12	10	function	function	NOUN
fcis-17883	12	11	to	to	PART
fcis-17883	12	12	train	train	VERB
fcis-17883	12	13	the	the	DET
fcis-17883	12	14	spsnn	spsnn	NOUN
fcis-17883	12	15	,	,	PUNCT
fcis-17883	12	16	the	the	DET
fcis-17883	12	17	inappropriate	inappropriate	ADJ
fcis-17883	12	18	number	number	NOUN
fcis-17883	12	19	of	of	ADP
fcis-17883	12	20	hidden	hide	VERB
fcis-17883	12	21	layer	layer	NOUN
fcis-17883	12	22	neurons	neuron	NOUN
fcis-17883	12	23	and	and	CCONJ
fcis-17883	12	24	the	the	DET
fcis-17883	12	25	magnitude	magnitude	NOUN
fcis-17883	12	26	of	of	ADP
fcis-17883	12	27	the	the	DET
fcis-17883	12	28	weights	weight	NOUN
fcis-17883	12	29	may	may	AUX
fcis-17883	12	30	lead	lead	VERB
fcis-17883	12	31	to	to	ADP
fcis-17883	12	32	underfitting	underfitting	NOUN
fcis-17883	12	33	or	or	CCONJ
fcis-17883	12	34	overfitting	overfitting	NOUN
fcis-17883	12	35	of	of	ADP
fcis-17883	12	36	the	the	DET
fcis-17883	12	37	model	model	NOUN
fcis-17883	12	38	,	,	PUNCT
fcis-17883	12	39	which	which	PRON
fcis-17883	12	40	affects	affect	VERB
fcis-17883	12	41	the	the	DET
fcis-17883	12	42	performance	performance	NOUN
fcis-17883	12	43	and	and	CCONJ
fcis-17883	12	44	generalization	generalization	NOUN
fcis-17883	12	45	ability	ability	NOUN
fcis-17883	12	46	of	of	ADP
fcis-17883	12	47	the	the	DET
fcis-17883	12	48	model	model	NOUN
fcis-17883	12	49	.	.	PUNCT
fcis-17883	13	1	therefore	therefore	ADV
fcis-17883	13	2	,	,	PUNCT
fcis-17883	13	3	the	the	DET
fcis-17883	13	4	structural	structural	ADJ
fcis-17883	13	5	design	design	NOUN
fcis-17883	13	6	of	of	ADP
fcis-17883	13	7	bp	bp	PROPN
fcis-17883	13	8	neural	neural	ADJ
fcis-17883	13	9	networks	network	NOUN
fcis-17883	13	10	and	and	CCONJ
fcis-17883	13	11	the	the	DET
fcis-17883	13	12	acquisition	acquisition	NOUN
fcis-17883	13	13	of	of	ADP
fcis-17883	13	14	optimal	optimal	ADJ
fcis-17883	13	15	parameters	parameter	NOUN
fcis-17883	13	16	remain	remain	VERB
fcis-17883	13	17	a	a	DET
fcis-17883	13	18	great	great	ADJ
fcis-17883	13	19	challenge	challenge	NOUN
fcis-17883	13	20	.	.	PUNCT
fcis-17883	14	1	for	for	ADP
fcis-17883	14	2	the	the	DET
fcis-17883	14	3	time	time	NOUN
fcis-17883	14	4	being	be	AUX
fcis-17883	14	5	,	,	PUNCT
fcis-17883	14	6	adding	add	VERB
fcis-17883	14	7	a	a	DET
fcis-17883	14	8	regularization	regularization	NOUN
fcis-17883	14	9	term	term	NOUN
fcis-17883	14	10	to	to	ADP
fcis-17883	14	11	the	the	DET
fcis-17883	14	12	error	error	NOUN
fcis-17883	14	13	function	function	NOUN
fcis-17883	14	14	is	be	AUX
fcis-17883	14	15	the	the	DET
fcis-17883	14	16	most	most	ADV
fcis-17883	14	17	common	common	ADJ
fcis-17883	14	18	way	way	NOUN
fcis-17883	14	19	to	to	PART
fcis-17883	14	20	solve	solve	VERB
fcis-17883	14	21	this	this	DET
fcis-17883	14	22	problem	problem	NOUN
fcis-17883	14	23	.	.	PUNCT
fcis-17883	15	1	the	the	DET
fcis-17883	15	2	main	main	ADJ
fcis-17883	15	3	common	common	ADJ
fcis-17883	15	4	penalties	penalty	NOUN
fcis-17883	15	5	we	we	PRON
fcis-17883	15	6	use	use	VERB
fcis-17883	15	7	to	to	PART
fcis-17883	15	8	train	train	NOUN
fcis-17883	15	9	networks	network	NOUN
fcis-17883	15	10	are	be	AUX
fcis-17883	15	11	weight	weight	NOUN
fcis-17883	15	12	decay	decay	NOUN
fcis-17883	15	13	,	,	PUNCT
fcis-17883	15	14	weight	weight	NOUN
fcis-17883	15	15	elimination	elimination	NOUN
fcis-17883	15	16	and	and	CCONJ
fcis-17883	15	17	approximate	approximate	ADJ
fcis-17883	15	18	smoothing	smooth	VERB
fcis-17883	15	19	[	[	PUNCT
fcis-17883	15	20	6	6	NUM
fcis-17883	15	21	-	-	SYM
fcis-17883	15	22	8	8	NUM
fcis-17883	15	23	]	]	PUNCT
fcis-17883	15	24	.	.	PUNCT
fcis-17883	16	1	while	while	SCONJ
fcis-17883	16	2	the	the	DET
fcis-17883	16	3	traditional	traditional	ADJ
fcis-17883	16	4	conventional	conventional	ADJ
fcis-17883	16	5	regularization	regularization	NOUN
fcis-17883	16	6	term	term	NOUN
fcis-17883	16	7	is	be	AUX
fcis-17883	16	8	very	very	ADV
fcis-17883	16	9	effective	effective	ADJ
fcis-17883	16	10	in	in	ADP
fcis-17883	16	11	sparsifying	sparsifye	VERB
fcis-17883	16	12	and	and	CCONJ
fcis-17883	16	13	preventing	prevent	VERB
fcis-17883	16	14	overfitting	overfitte	VERB
fcis-17883	16	15	,	,	PUNCT
fcis-17883	16	16	it	it	PRON
fcis-17883	16	17	does	do	AUX
fcis-17883	16	18	not	not	PART
fcis-17883	16	19	take	take	VERB
fcis-17883	16	20	into	into	ADP
fcis-17883	16	21	account	account	NOUN
fcis-17883	16	22	the	the	DET
fcis-17883	16	23	geometric	geometric	ADJ
fcis-17883	16	24	structure	structure	NOUN
fcis-17883	16	25	of	of	ADP
fcis-17883	16	26	the	the	DET
fcis-17883	16	27	original	original	ADJ
fcis-17883	16	28	data	datum	NOUN
fcis-17883	16	29	[	[	X
fcis-17883	16	30	9	9	NUM
fcis-17883	16	31	]	]	PUNCT
fcis-17883	16	32	.	.	PUNCT
fcis-17883	17	1	recently	recently	ADV
fcis-17883	17	2	popular	popular	ADJ
fcis-17883	17	3	graph	graph	NOUN
fcis-17883	17	4	regularization	regularization	NOUN
fcis-17883	17	5	terms	term	NOUN
fcis-17883	17	6	perform	perform	VERB
fcis-17883	17	7	well	well	ADV
fcis-17883	17	8	in	in	ADP
fcis-17883	17	9	this	this	DET
fcis-17883	17	10	regard	regard	NOUN
fcis-17883	17	11	[	[	X
fcis-17883	17	12	10	10	NUM
fcis-17883	17	13	-	-	SYM
fcis-17883	17	14	11	11	NUM
fcis-17883	17	15	]	]	PUNCT
fcis-17883	17	16	.	.	PUNCT
fcis-17883	18	1	therefore	therefore	ADV
fcis-17883	18	2	,	,	PUNCT
fcis-17883	18	3	in	in	ADP
fcis-17883	18	4	this	this	DET
fcis-17883	18	5	paper	paper	NOUN
fcis-17883	18	6	,	,	PUNCT
fcis-17883	18	7	we	we	PRON
fcis-17883	18	8	consider	consider	VERB
fcis-17883	18	9	introducing	introduce	VERB
fcis-17883	18	10	the	the	DET
fcis-17883	18	11	graph	graph	NOUN
fcis-17883	18	12	regularization	regularization	NOUN
fcis-17883	18	13	term	term	NOUN
fcis-17883	18	14	in	in	ADP
fcis-17883	18	15	the	the	DET
fcis-17883	18	16	spsnn	spsnn	NOUN
fcis-17883	18	17	model	model	NOUN
fcis-17883	18	18	to	to	PART
fcis-17883	18	19	obtain	obtain	VERB
fcis-17883	18	20	the	the	DET
fcis-17883	18	21	optimal	optimal	ADJ
fcis-17883	18	22	parameters	parameter	NOUN
fcis-17883	18	23	and	and	CCONJ
fcis-17883	18	24	the	the	DET
fcis-17883	18	25	best	good	ADJ
fcis-17883	18	26	network	network	NOUN
fcis-17883	18	27	structure	structure	NOUN
fcis-17883	18	28	by	by	ADP
fcis-17883	18	29	applying	apply	VERB
fcis-17883	18	30	the	the	DET
fcis-17883	18	31	graph	graph	NOUN
fcis-17883	18	32	regularization	regularization	NOUN
fcis-17883	18	33	term	term	NOUN
fcis-17883	18	34	to	to	ADP
fcis-17883	18	35	the	the	DET
fcis-17883	18	36	error	error	NOUN
fcis-17883	18	37	function	function	NOUN
fcis-17883	18	38	.	.	PUNCT
fcis-17883	19	1	the	the	DET
fcis-17883	19	2	rest	rest	NOUN
fcis-17883	19	3	of	of	ADP
fcis-17883	19	4	the	the	DET
fcis-17883	19	5	paper	paper	NOUN
fcis-17883	19	6	is	be	AUX
fcis-17883	19	7	organized	organize	VERB
fcis-17883	19	8	as	as	SCONJ
fcis-17883	19	9	follows	follow	VERB
fcis-17883	19	10	:	:	PUNCT
fcis-17883	19	11	in	in	ADP
fcis-17883	19	12	section	section	NOUN
fcis-17883	19	13	2	2	NUM
fcis-17883	19	14	,	,	PUNCT
fcis-17883	19	15	the	the	DET
fcis-17883	19	16	sigma	sigma	PROPN
fcis-17883	19	17	-	-	PUNCT
fcis-17883	19	18	pi	pi	NOUN
fcis-17883	19	19	-	-	PUNCT
fcis-17883	19	20	sigma	sigma	NOUN
fcis-17883	19	21	neural	neural	ADJ
fcis-17883	19	22	network	network	NOUN
fcis-17883	19	23	and	and	CCONJ
fcis-17883	19	24	the	the	DET
fcis-17883	19	25	batch	batch	NOUN
fcis-17883	19	26	gradient	gradient	NOUN
fcis-17883	19	27	algorithm	algorithm	NOUN
fcis-17883	19	28	based	base	VERB
fcis-17883	19	29	on	on	ADP
fcis-17883	19	30	graph	graph	NOUN
fcis-17883	19	31	regularization	regularization	NOUN
fcis-17883	19	32	are	be	AUX
fcis-17883	19	33	described	describe	VERB
fcis-17883	19	34	in	in	ADP
fcis-17883	19	35	detail	detail	NOUN
fcis-17883	19	36	.	.	PUNCT
fcis-17883	20	1	section	section	NOUN
fcis-17883	20	2	3	3	NUM
fcis-17883	20	3	analyses	analyse	VERB
fcis-17883	20	4	the	the	DET
fcis-17883	20	5	experimental	experimental	ADJ
fcis-17883	20	6	results	result	NOUN
fcis-17883	20	7	.	.	PUNCT
fcis-17883	21	1	finally	finally	ADV
fcis-17883	21	2	,	,	PUNCT
fcis-17883	21	3	section	section	NOUN
fcis-17883	21	4	4	4	NUM
fcis-17883	21	5	gives	give	VERB
fcis-17883	21	6	some	some	DET
fcis-17883	21	7	conclusions	conclusion	NOUN
fcis-17883	21	8	of	of	ADP
fcis-17883	21	9	this	this	DET
fcis-17883	21	10	paper	paper	NOUN
fcis-17883	21	11	.	.	PUNCT
fcis-17883	22	1	2	2	X
fcis-17883	22	2	.	.	X
fcis-17883	22	3	the	the	DET
fcis-17883	22	4	proposed	propose	VERB
fcis-17883	22	5	approach	approach	NOUN
fcis-17883	22	6	(	(	PUNCT
fcis-17883	22	7	1	1	NUM
fcis-17883	22	8	)	)	PUNCT
fcis-17883	22	9	sigma	sigma	NOUN
fcis-17883	22	10	-	-	PUNCT
fcis-17883	22	11	pi	pi	NOUN
fcis-17883	22	12	-	-	PUNCT
fcis-17883	22	13	sigma	sigma	NOUN
fcis-17883	22	14	neural	neural	ADJ
fcis-17883	22	15	network	network	NOUN
fcis-17883	22	16	(	(	PUNCT
fcis-17883	22	17	spsnn	spsnn	NOUN
fcis-17883	22	18	)	)	PUNCT
fcis-17883	22	19	in	in	ADP
fcis-17883	22	20	this	this	DET
fcis-17883	22	21	section	section	NOUN
fcis-17883	22	22	,	,	PUNCT
fcis-17883	22	23	we	we	PRON
fcis-17883	22	24	introduce	introduce	VERB
fcis-17883	22	25	the	the	DET
fcis-17883	22	26	spsnn	spsnn	NOUN
fcis-17883	22	27	model	model	NOUN
fcis-17883	22	28	.	.	PUNCT
fcis-17883	23	1	p	p	NOUN
fcis-17883	23	2	,	,	PUNCT
fcis-17883	23	3	n	n	PROPN
fcis-17883	23	4	,	,	PUNCT
fcis-17883	23	5	q	q	X
fcis-17883	23	6	,	,	PUNCT
fcis-17883	23	7	and	and	CCONJ
fcis-17883	23	8	1	1	NUM
fcis-17883	23	9	are	be	AUX
fcis-17883	23	10	the	the	DET
fcis-17883	23	11	number	number	NOUN
fcis-17883	23	12	of	of	ADP
fcis-17883	23	13	neurons	neuron	NOUN
fcis-17883	23	14	in	in	ADP
fcis-17883	23	15	the	the	DET
fcis-17883	23	16	input	input	NOUN
fcis-17883	23	17	layer	layer	NOUN
fcis-17883	23	18	,	,	PUNCT
fcis-17883	23	19	1	1	NOUN
fcis-17883	23	20	layer	layer	NOUN
fcis-17883	23	21	,	,	PUNCT
fcis-17883	23	22			NOUN
fcis-17883	23	23	layer	layer	NOUN
fcis-17883	23	24	,	,	PUNCT
fcis-17883	23	25	and	and	CCONJ
fcis-17883	23	26	2	2	NUM
fcis-17883	23	27	layer	layer	NOUN
fcis-17883	23	28	,	,	PUNCT
fcis-17883	23	29	respectively	respectively	ADV
fcis-17883	23	30	(	(	PUNCT
fcis-17883	23	31	see	see	VERB
fcis-17883	23	32	fig	fig	NOUN
fcis-17883	23	33	.	.	PUNCT
fcis-17883	24	1	1	1	NUM
fcis-17883	24	2	)	)	PUNCT
fcis-17883	24	3	.	.	PUNCT
fcis-17883	25	1	let	let	VERB
fcis-17883	25	2			PRON
fcis-17883	25	3			PROPN
fcis-17883	25	4	1	1	NUM
fcis-17883	25	5	,	,	PUNCT
fcis-17883	25	6	j	j	PROPN
fcis-17883	25	7	p	p	PROPN
fcis-17883	25	8	j	j	PROPN
fcis-17883	26	1	j	j	PROPN
fcis-17883	26	2	j	j	PROPN
fcis-17883	27	1	x	x	PUNCT
fcis-17883	27	2	o	o	NOUN
fcis-17883	28	1	r	r	NOUN
fcis-17883	28	2	r	r	NOUN
fcis-17883	28	3			NOUN
fcis-17883	29	1			PROPN
fcis-17883	29	2			AUX
fcis-17883	29	3	be	be	AUX
fcis-17883	29	4	the	the	DET
fcis-17883	29	5	set	set	NOUN
fcis-17883	29	6	of	of	ADP
fcis-17883	29	7	training	training	NOUN
fcis-17883	29	8	samples	sample	NOUN
fcis-17883	29	9	,	,	PUNCT
fcis-17883	29	10	where	where	SCONJ
fcis-17883	29	11	jo	jo	PROPN
fcis-17883	29	12	is	be	AUX
fcis-17883	29	13	the	the	DET
fcis-17883	29	14	desired	desire	VERB
fcis-17883	29	15	output	output	NOUN
fcis-17883	29	16	of	of	ADP
fcis-17883	29	17	input	input	NOUN
fcis-17883	29	18	sample	sample	NOUN
fcis-17883	29	19	jx	jx	PROPN
fcis-17883	29	20	and	and	CCONJ
fcis-17883	29	21	j	j	PROPN
fcis-17883	29	22	is	be	AUX
fcis-17883	29	23	the	the	DET
fcis-17883	29	24	number	number	NOUN
fcis-17883	29	25	of	of	ADP
fcis-17883	29	26	training	training	NOUN
fcis-17883	29	27	samples	sample	NOUN
fcis-17883	29	28	.	.	PUNCT
fcis-17883	30	1	1	1	NUM
fcis-17883	30	2	1	1	NUM
fcis-17883	30	3	1	1	NUM
fcis-17883	30	4	1	1	NUM
fcis-17883	30	5	output	output	NOUN
fcis-17883	30	6	input	input	NOUN
fcis-17883	30	7	  	  	SPACE
fcis-17883	30	8	layer	layer	NOUN
fcis-17883	30	9	nw	nw	PROPN
fcis-17883	30	10	0w	0w	PROPN
fcis-17883	30	11	figure	figure	NOUN
fcis-17883	30	12	1	1	NUM
fcis-17883	30	13	.	.	PUNCT
fcis-17883	30	14	topology	topology	NOUN
fcis-17883	30	15	of	of	ADP
fcis-17883	30	16	an	an	DET
fcis-17883	30	17	spsnn	spsnn	NOUN
fcis-17883	30	18	.	.	PUNCT
fcis-17883	31	1	let	let	VERB
fcis-17883	31	2	1	1	NUM
fcis-17883	31	3	2	2	NUM
fcis-17883	31	4	(	(	PUNCT
fcis-17883	31	5	,	,	PUNCT
fcis-17883	31	6	,	,	PUNCT
fcis-17883	31	7	,	,	PUNCT
fcis-17883	31	8	)	)	PUNCT
fcis-17883	32	1	p	p	X
fcis-17883	33	1	i	i	PRON
fcis-17883	33	2	i	i	PRON
fcis-17883	34	1	i	i	PRON
fcis-17883	34	2	ipw	ipw	VERB
fcis-17883	34	3	w	w	PROPN
fcis-17883	34	4	w	w	PROPN
fcis-17883	34	5	w	w	NOUN
fcis-17883	34	6			NOUN
fcis-17883	34	7			PROPN
fcis-17883	34	8	be	be	AUX
fcis-17883	34	9	the	the	DET
fcis-17883	34	10	weight	weight	NOUN
fcis-17883	34	11	vector	vector	NOUN
fcis-17883	34	12	between	between	ADP
fcis-17883	34	13	the	the	DET
fcis-17883	34	14	input	input	NOUN
fcis-17883	34	15	layer	layer	NOUN
fcis-17883	34	16	and	and	CCONJ
fcis-17883	34	17	the	the	DET
fcis-17883	34	18	1	1	PROPN
fcis-17883	34	19	layer	layer	NOUN
fcis-17883	34	20	at	at	ADP
fcis-17883	34	21	1,2	1,2	NUM
fcis-17883	34	22	,	,	PUNCT
fcis-17883	34	23	,	,	PUNCT
fcis-17883	34	24	i	i	PROPN
fcis-17883	34	25	n	n	NOUN
fcis-17883	34	26			NOUN
fcis-17883	34	27	let	let	VERB
fcis-17883	34	28	0	0	NUM
fcis-17883	34	29	01	01	NUM
fcis-17883	34	30	02	02	NUM
fcis-17883	34	31	0	0	NUM
fcis-17883	34	32	(	(	PUNCT
fcis-17883	34	33	,	,	PUNCT
fcis-17883	34	34	,	,	PUNCT
fcis-17883	34	35	,	,	PUNCT
fcis-17883	34	36	)	)	PUNCT
fcis-17883	34	37	q	q	PROPN
fcis-17883	35	1	qw	qw	INTJ
fcis-17883	35	2	w	w	PROPN
fcis-17883	35	3	w	w	PROPN
fcis-17883	35	4	w	w	NOUN
fcis-17883	35	5			NOUN
fcis-17883	35	6			PROPN
fcis-17883	35	7	be	be	AUX
fcis-17883	35	8	the	the	DET
fcis-17883	35	9	weight	weight	NOUN
fcis-17883	35	10	vector	vector	NOUN
fcis-17883	35	11	connecting	connect	VERB
fcis-17883	35	12	the	the	DET
fcis-17883	35	13	2	2	NUM
fcis-17883	35	14	layer	layer	NOUN
fcis-17883	35	15	to	to	ADP
fcis-17883	35	16	the	the	DET
fcis-17883	35	17			ADJ
fcis-17883	35	18	layer	layer	NOUN
fcis-17883	35	19	.	.	PUNCT
fcis-17883	36	1	0	0	NUM
fcis-17883	37	1	1	1	NUM
fcis-17883	37	2	(	(	PUNCT
fcis-17883	37	3	,	,	PUNCT
fcis-17883	37	4	,	,	PUNCT
fcis-17883	37	5	,	,	PUNCT
fcis-17883	37	6	)	)	PUNCT
fcis-17883	37	7	t	t	PROPN
fcis-17883	37	8	t	t	PROPN
fcis-17883	37	9	t	t	PROPN
fcis-17883	37	10	t	t	PROPN
fcis-17883	37	11	q	q	PROPN
fcis-17883	37	12	np	np	INTJ
fcis-17883	37	13	nw	nw	PROPN
fcis-17883	37	14	w	w	PROPN
fcis-17883	37	15	w	w	PROPN
fcis-17883	37	16	w	w	PROPN
fcis-17883	37	17			ADJ
fcis-17883	37	18			ADJ
fcis-17883	37	19			PROPN
fcis-17883	37	20	we	we	PRON
fcis-17883	37	21	write	write	VERB
fcis-17883	37	22	all	all	DET
fcis-17883	37	23	the	the	DET
fcis-17883	37	24	ownership	ownership	NOUN
fcis-17883	37	25	values	value	NOUN
fcis-17883	37	26	in	in	ADP
fcis-17883	37	27	a	a	DET
fcis-17883	37	28	tightened	tighten	VERB
fcis-17883	37	29	form	form	NOUN
fcis-17883	37	30	.	.	PUNCT
fcis-17883	38	1	let	let	VERB
fcis-17883	38	2	j	j	PROPN
fcis-17883	38	3	0	0	PROPN
fcis-17883	38	4	(	(	PUNCT
fcis-17883	38	5	)	)	PUNCT
fcis-17883	38	6	jy	jy	PROPN
fcis-17883	38	7	g	g	PROPN
fcis-17883	38	8	w	w	PROPN
fcis-17883	38	9			PROPN
fcis-17883	38	10			PROPN
fcis-17883	38	11	as	as	ADP
fcis-17883	38	12	the	the	DET
fcis-17883	38	13	real	real	ADJ
fcis-17883	38	14	output	output	NOUN
fcis-17883	38	15	,	,	PUNCT
fcis-17883	38	16	i.e.	i.e.	X
fcis-17883	38	17	,	,	PUNCT
fcis-17883	38	18	35	35	NUM
fcis-17883	38	19	0	0	NUM
fcis-17883	38	20	1	1	NUM
fcis-17883	38	21	(	(	PUNCT
fcis-17883	38	22	,	,	PUNCT
fcis-17883	38	23	,	,	PUNCT
fcis-17883	38	24	,	,	PUNCT
fcis-17883	38	25	)	)	PUNCT
fcis-17883	38	26	t	t	PROPN
fcis-17883	38	27	t	t	PROPN
fcis-17883	38	28	t	t	PROPN
fcis-17883	38	29	t	t	PROPN
fcis-17883	38	30	q	q	PROPN
fcis-17883	38	31	np	np	INTJ
fcis-17883	38	32	nw	nw	PROPN
fcis-17883	38	33	w	w	PROPN
fcis-17883	38	34	w	w	PROPN
fcis-17883	38	35	w	w	PROPN
fcis-17883	38	36			ADJ
fcis-17883	38	37			NOUN
fcis-17883	38	38			PUNCT
fcis-17883	38	39	.	.	PUNCT
fcis-17883	39	1	define	define	VERB
fcis-17883	39	2	:	:	PUNCT
fcis-17883	39	3	g	g	PROPN
fcis-17883	39	4	r	r	NOUN
fcis-17883	39	5	r	r	NOUN
fcis-17883	39	6	to	to	PART
fcis-17883	39	7	be	be	AUX
fcis-17883	39	8	the	the	DET
fcis-17883	39	9	activation	activation	NOUN
fcis-17883	39	10	function	function	NOUN
fcis-17883	39	11	.	.	PUNCT
fcis-17883	40	1	(	(	PUNCT
fcis-17883	40	2	)	)	PUNCT
fcis-17883	40	3	e	e	X
fcis-17883	40	4	w	w	ADP
fcis-17883	40	5			PROPN
fcis-17883	40	6	to	to	PART
fcis-17883	40	7	be	be	AUX
fcis-17883	40	8	the	the	DET
fcis-17883	40	9	conventional	conventional	ADJ
fcis-17883	40	10	squared	square	VERB
fcis-17883	40	11	error	error	NOUN
fcis-17883	40	12	function	function	NOUN
fcis-17883	40	13	:	:	PUNCT
fcis-17883	40	14			NOUN
fcis-17883	40	15			PROPN
fcis-17883	40	16	1	1	NUM
fcis-17883	40	17	0	0	NUM
fcis-17883	40	18	1	1	NUM
fcis-17883	40	19	1	1	NUM
fcis-17883	40	20	(	(	PUNCT
fcis-17883	40	21	)	)	PUNCT
fcis-17883	40	22	2	2	NUM
fcis-17883	40	23	(	(	PUNCT
fcis-17883	40	24	)	)	PUNCT
fcis-17883	41	1	j	j	PROPN
fcis-17883	42	1	j	j	PROPN
fcis-17883	42	2	j	j	PROPN
fcis-17883	42	3	j	j	PROPN
fcis-17883	42	4	j	j	PROPN
fcis-17883	42	5	j	j	PROPN
fcis-17883	42	6	j	j	PROPN
fcis-17883	42	7	j	j	PROPN
fcis-17883	43	1	e	e	PROPN
fcis-17883	43	2	w	w	PROPN
fcis-17883	43	3	y	y	NOUN
fcis-17883	43	4	o	o	X
fcis-17883	43	5	g	g	PROPN
fcis-17883	43	6	w	w	PROPN
fcis-17883	43	7			PROPN
fcis-17883	44	1			NUM
fcis-17883	44	2			NUM
fcis-17883	44	3			PROPN
fcis-17883	44	4			PROPN
fcis-17883	44	5			PROPN
fcis-17883	44	6			PROPN
fcis-17883	44	7			X
fcis-17883	44	8			X
fcis-17883	44	9			NUM
fcis-17883	44	10	where	where	SCONJ
fcis-17883	44	11	1	1	NUM
fcis-17883	44	12	(	(	PUNCT
fcis-17883	44	13	)	)	PUNCT
fcis-17883	44	14	(	(	PUNCT
fcis-17883	44	15	(	(	PUNCT
fcis-17883	44	16	)	)	PUNCT
fcis-17883	44	17	)	)	PUNCT
fcis-17883	44	18	,	,	PUNCT
fcis-17883	44	19	,	,	PUNCT
fcis-17883	44	20	1	1	NUM
fcis-17883	44	21	j	j	PROPN
fcis-17883	44	22	j	j	PROPN
fcis-17883	44	23	2	2	NUM
fcis-17883	44	24	j	j	PROPN
fcis-17883	44	25	jg	jg	PROPN
fcis-17883	44	26	r	r	NOUN
fcis-17883	45	1	g	g	PROPN
fcis-17883	45	2	r	r	NOUN
fcis-17883	45	3	o	o	NOUN
fcis-17883	45	4	r	r	NOUN
fcis-17883	45	5	r	r	NOUN
fcis-17883	45	6			PROPN
fcis-17883	45	7			NOUN
fcis-17883	45	8			NOUN
fcis-17883	45	9			NOUN
fcis-17883	45	10	.	.	PUNCT
fcis-17883	46	1	at	at	ADP
fcis-17883	46	2	the	the	DET
fcis-17883	46	3	same	same	ADJ
fcis-17883	46	4	time	time	NOUN
fcis-17883	46	5	,	,	PUNCT
fcis-17883	46	6	we	we	PRON
fcis-17883	46	7	take	take	VERB
fcis-17883	46	8	the	the	DET
fcis-17883	46	9	following	follow	VERB
fcis-17883	46	10	forms	form	NOUN
fcis-17883	46	11	.	.	PUNCT
fcis-17883	47	1	1	1	NUM
fcis-17883	47	2	2	2	NUM
fcis-17883	47	3	1	1	NUM
fcis-17883	47	4	2	2	NUM
fcis-17883	47	5	1	1	NUM
fcis-17883	47	6	2	2	NUM
fcis-17883	47	7	(	(	PUNCT
fcis-17883	47	8	,	,	PUNCT
fcis-17883	47	9	,	,	PUNCT
fcis-17883	47	10	)	)	PUNCT
fcis-17883	47	11	(	(	PUNCT
fcis-17883	47	12	,	,	PUNCT
fcis-17883	47	13	,	,	PUNCT
fcis-17883	47	14	)	)	PUNCT
fcis-17883	47	15	(	(	PUNCT
fcis-17883	47	16	(	(	PUNCT
fcis-17883	47	17	)	)	PUNCT
fcis-17883	47	18	,	,	PUNCT
fcis-17883	47	19	(	(	PUNCT
fcis-17883	47	20	)	)	PUNCT
fcis-17883	47	21	,	,	PUNCT
fcis-17883	47	22	(	(	PUNCT
fcis-17883	47	23	)	)	PUNCT
fcis-17883	47	24	)	)	PUNCT
fcis-17883	48	1	q	q	NOUN
fcis-17883	49	1	q	q	X
fcis-17883	49	2	j	j	PROPN
fcis-17883	49	3	j	j	PROPN
fcis-17883	49	4	j	j	PROPN
fcis-17883	50	1	j	j	PROPN
fcis-17883	50	2	q	q	X
fcis-17883	51	1	i	i	PRON
fcis-17883	51	2	i	i	PRON
fcis-17883	52	1	i	i	PRON
fcis-17883	52	2	i	i	PRON
fcis-17883	53	1	i	i	PRON
fcis-17883	54	1	i	i	PRON
fcis-17883	54	2	j	j	PROPN
fcis-17883	55	1	j	j	PROPN
fcis-17883	55	2	j	j	PROPN
fcis-17883	56	1	i	i	PRON
fcis-17883	56	2	i	i	PRON
fcis-17883	57	1	i	i	PRON
fcis-17883	57	2	i	i	PRON
fcis-17883	58	1	i	i	PRON
fcis-17883	58	2	i	i	PRON
fcis-17883	59	1	g	g	VERB
fcis-17883	59	2	w	w	PROPN
fcis-17883	59	3	x	x	X
fcis-17883	59	4	g	g	PROPN
fcis-17883	59	5	w	w	PROPN
fcis-17883	59	6	x	x	X
fcis-17883	59	7	g	g	PROPN
fcis-17883	59	8	w	w	NOUN
fcis-17883	59	9	x	x	VERB
fcis-17883	59	10			NOUN
fcis-17883	60	1			X
fcis-17883	60	2			X
fcis-17883	61	1			X
fcis-17883	61	2			X
fcis-17883	61	3			X
fcis-17883	61	4			X
fcis-17883	61	5			NOUN
fcis-17883	61	6			NOUN
fcis-17883	61	7			ADJ
fcis-17883	61	8			ADJ
fcis-17883	61	9			NOUN
fcis-17883	61	10			NOUN
fcis-17883	62	1			NUM
fcis-17883	62	2			NUM
fcis-17883	62	3			PROPN
fcis-17883	62	4			PROPN
fcis-17883	62	5			PROPN
fcis-17883	62	6			PROPN
fcis-17883	62	7			NUM
fcis-17883	62	8			NUM
fcis-17883	62	9			NUM
fcis-17883	62	10			NUM
fcis-17883	62	11			NUM
fcis-17883	62	12			NUM
fcis-17883	62	13			NOUN
fcis-17883	62	14			NOUN
fcis-17883	62	15			NOUN
fcis-17883	63	1	so	so	ADV
fcis-17883	63	2	,	,	PUNCT
fcis-17883	63	3	~	~	PUNCT
fcis-17883	63	4	(	(	PUNCT
fcis-17883	63	5	)	)	PUNCT
fcis-17883	63	6	e	e	X
fcis-17883	63	7	w	w	NOUN
fcis-17883	63	8	can	can	AUX
fcis-17883	63	9	be	be	AUX
fcis-17883	63	10	represented	represent	VERB
fcis-17883	63	11	as	as	SCONJ
fcis-17883	63	12	follows	follow	VERB
fcis-17883	63	13	:	:	PUNCT
fcis-17883	63	14	~	~	PUNCT
fcis-17883	63	15	0	0	NUM
fcis-17883	63	16	1	1	NUM
fcis-17883	63	17	1	1	NUM
fcis-17883	63	18	(	(	PUNCT
fcis-17883	63	19	)	)	PUNCT
fcis-17883	63	20	(	(	PUNCT
fcis-17883	63	21	(	(	PUNCT
fcis-17883	63	22	)	)	PUNCT
fcis-17883	63	23	)	)	PUNCT
fcis-17883	64	1	qj	qj	PROPN
fcis-17883	65	1	j	j	PROPN
fcis-17883	66	1	j	j	PROPN
fcis-17883	66	2	q	q	X
fcis-17883	67	1	i	i	PRON
fcis-17883	67	2	j	j	INTJ
fcis-17883	68	1	q	q	X
fcis-17883	69	1	i	i	PRON
fcis-17883	69	2	q	q	X
fcis-17883	70	1	e	e	NOUN
fcis-17883	70	2	w	w	PROPN
fcis-17883	70	3	g	g	PROPN
fcis-17883	70	4	w	w	PROPN
fcis-17883	70	5	g	g	PROPN
fcis-17883	70	6	w	w	PROPN
fcis-17883	70	7	x	x	PROPN
fcis-17883	70	8			NUM
fcis-17883	70	9			PROPN
fcis-17883	70	10			PROPN
fcis-17883	70	11			NUM
fcis-17883	70	12			PROPN
fcis-17883	70	13			SYM
fcis-17883	70	14			X
fcis-17883	70	15			NUM
fcis-17883	70	16	and	and	CCONJ
fcis-17883	70	17	we	we	PRON
fcis-17883	70	18	can	can	AUX
fcis-17883	70	19	get	get	VERB
fcis-17883	70	20	the	the	DET
fcis-17883	70	21	partial	partial	ADJ
fcis-17883	70	22	derivative	derivative	NOUN
fcis-17883	70	23	as	as	SCONJ
fcis-17883	70	24	follows	follow	VERB
fcis-17883	70	25	:	:	PUNCT
fcis-17883	70	26	0	0	NUM
fcis-17883	70	27	~	~	PUNCT
fcis-17883	70	28	'	'	PUNCT
fcis-17883	70	29	0	0	NUM
fcis-17883	70	30	1	1	NUM
fcis-17883	70	31	(	(	PUNCT
fcis-17883	70	32	)	)	PUNCT
fcis-17883	70	33	(	(	PUNCT
fcis-17883	70	34	)	)	PUNCT
fcis-17883	71	1	j	j	PROPN
fcis-17883	72	1	j	j	PROPN
fcis-17883	72	2	j	j	PROPN
fcis-17883	72	3	w	w	PROPN
fcis-17883	72	4	j	j	PROPN
fcis-17883	72	5	j	j	PROPN
fcis-17883	72	6	e	e	PROPN
fcis-17883	72	7	w	w	PROPN
fcis-17883	72	8	g	g	PROPN
fcis-17883	72	9	w	w	PROPN
fcis-17883	72	10			NOUN
fcis-17883	73	1			X
fcis-17883	74	1			NUM
fcis-17883	74	2			PROPN
fcis-17883	75	1			X
fcis-17883	75	2	in	in	ADP
fcis-17883	75	3	addition	addition	NOUN
fcis-17883	75	4	,	,	PUNCT
fcis-17883	75	5	according	accord	VERB
fcis-17883	75	6	to	to	ADP
fcis-17883	75	7	(	(	PUNCT
fcis-17883	75	8	)	)	PUNCT
fcis-17883	75	9	q	q	PROPN
fcis-17883	76	1	j	j	PROPN
fcis-17883	77	1	j	j	X
fcis-17883	77	2	q	q	X
fcis-17883	78	1	i	i	PRON
fcis-17883	78	2	i	i	PRON
fcis-17883	78	3	g	g	VERB
fcis-17883	78	4	w	w	PROPN
fcis-17883	78	5	x	x	X
fcis-17883	78	6			NOUN
fcis-17883	78	7			NUM
fcis-17883	78	8			PROPN
fcis-17883	78	9	,	,	PUNCT
fcis-17883	78	10	we	we	PRON
fcis-17883	78	11	get	get	VERB
fcis-17883	78	12	(	(	PUNCT
fcis-17883	78	13	)	)	PUNCT
fcis-17883	78	14	(	(	PUNCT
fcis-17883	78	15	)	)	PUNCT
fcis-17883	78	16	,	,	PUNCT
fcis-17883	78	17	?	?	PUNCT
fcis-17883	79	1	1	1	NUM
fcis-17883	79	2	?	?	SYM
fcis-17883	79	3	0	0	NUM
fcis-17883	79	4	,	,	PUNCT
fcis-17883	79	5	?	?	PUNCT
fcis-17883	80	1	1	1	X
fcis-17883	80	2	?	?	PUNCT
fcis-17883	80	3	q	q	PUNCT
fcis-17883	81	1	j	j	PROPN
fcis-17883	82	1	j	j	PROPN
fcis-17883	82	2	j	j	PROPN
fcis-17883	83	1	i	i	PRON
fcis-17883	83	2	i	i	PRON
fcis-17883	83	3	q	q	VERB
fcis-17883	83	4	q	q	X
fcis-17883	83	5	i	i	VERB
fcis-17883	83	6	n	n	VERB
fcis-17883	83	7	n	n	NOUN
fcis-17883	83	8	q	q	NOUN
fcis-17883	83	9	g	g	PROPN
fcis-17883	83	10	w	w	NOUN
fcis-17883	83	11	x	x	PUNCT
fcis-17883	83	12	x	x	X
fcis-17883	83	13	if	if	SCONJ
fcis-17883	83	14	q	q	PROPN
fcis-17883	83	15	and	and	CCONJ
fcis-17883	83	16	n	n	ADV
fcis-17883	83	17	w	w	NOUN
fcis-17883	83	18	if	if	SCONJ
fcis-17883	83	19	q	q	PROPN
fcis-17883	83	20	or	or	CCONJ
fcis-17883	83	21	n	n	PRON
fcis-17883	83	22			PROPN
fcis-17883	83	23			ADJ
fcis-17883	83	24			ADV
fcis-17883	83	25			NOUN
fcis-17883	83	26			PROPN
fcis-17883	83	27			NOUN
fcis-17883	83	28			PROPN
fcis-17883	83	29			PROPN
fcis-17883	84	1			NUM
fcis-17883	84	2			ADJ
fcis-17883	84	3			NUM
fcis-17883	84	4			PROPN
fcis-17883	84	5			PRON
fcis-17883	84	6			NOUN
fcis-17883	84	7			PUNCT
fcis-17883	84	8	then	then	ADV
fcis-17883	84	9	,	,	PUNCT
fcis-17883	84	10	in	in	ADP
fcis-17883	84	11	light	light	NOUN
fcis-17883	84	12	of	of	ADP
fcis-17883	84	13	the	the	DET
fcis-17883	84	14	above	above	ADJ
fcis-17883	84	15	definitions	definition	NOUN
fcis-17883	84	16	and	and	CCONJ
fcis-17883	84	17	(	(	PUNCT
fcis-17883	84	18	ii.4	ii.4	PROPN
fcis-17883	84	19	)	)	PUNCT
fcis-17883	84	20	,	,	PUNCT
fcis-17883	84	21	we	we	PRON
fcis-17883	84	22	have	have	VERB
fcis-17883	84	23	~	~	PUNCT
fcis-17883	84	24	'	'	PUNCT
fcis-17883	84	25	0	0	NUM
fcis-17883	84	26	1	1	NUM
fcis-17883	84	27	'	'	NUM
fcis-17883	84	28	0	0	NUM
fcis-17883	84	29	1	1	NUM
fcis-17883	84	30	\	\	NOUN
fcis-17883	84	31	(	(	PUNCT
fcis-17883	84	32	)	)	PUNCT
fcis-17883	84	33	(	(	PUNCT
fcis-17883	84	34	)	)	PUNCT
fcis-17883	84	35	(	(	PUNCT
fcis-17883	84	36	)	)	PUNCT
fcis-17883	84	37	(	(	PUNCT
fcis-17883	84	38	)	)	PUNCT
fcis-17883	84	39	(	(	PUNCT
fcis-17883	84	40	)	)	PUNCT
fcis-17883	84	41	n	n	CCONJ
fcis-17883	84	42	jj	jj	PROPN
fcis-17883	84	43	qj	qj	PROPN
fcis-17883	84	44	w	w	PROPN
fcis-17883	84	45	j	j	PROPN
fcis-17883	84	46	oq	oq	PROPN
fcis-17883	84	47	j	j	PROPN
fcis-17883	84	48	q	q	PROPN
fcis-17883	84	49	n	n	PROPN
fcis-17883	84	50	n	n	PRON
fcis-17883	84	51	j	j	PROPN
fcis-17883	85	1	j	j	PROPN
fcis-17883	85	2	j	j	PROPN
fcis-17883	85	3	j	j	PROPN
fcis-17883	86	1	j	j	PROPN
fcis-17883	86	2	oq	oq	INTJ
fcis-17883	86	3	i	i	INTJ
fcis-17883	86	4	j	j	PROPN
fcis-17883	86	5	q	q	VERB
fcis-17883	87	1	n	n	INTJ
fcis-17883	87	2	i	i	PRON
fcis-17883	87	3	q	q	PROPN
fcis-17883	88	1	n	n	INTJ
fcis-17883	88	2	j	j	NOUN
fcis-17883	89	1	i	i	PRON
fcis-17883	89	2	i	i	PRON
fcis-17883	89	3	q	q	X
fcis-17883	90	1	e	e	NOUN
fcis-17883	90	2	w	w	PROPN
fcis-17883	90	3	g	g	PROPN
fcis-17883	90	4	w	w	PROPN
fcis-17883	90	5	w	w	PROPN
fcis-17883	90	6	w	w	PROPN
fcis-17883	90	7	g	g	PROPN
fcis-17883	90	8	w	w	PROPN
fcis-17883	90	9	w	w	PROPN
fcis-17883	90	10	x	x	SYM
fcis-17883	90	11	g	g	PROPN
fcis-17883	90	12	w	w	NOUN
fcis-17883	90	13	x	x	VERB
fcis-17883	90	14			NOUN
fcis-17883	91	1			X
fcis-17883	91	2			X
fcis-17883	91	3			X
fcis-17883	91	4			PROPN
fcis-17883	91	5			NUM
fcis-17883	91	6			NUM
fcis-17883	91	7			NUM
fcis-17883	91	8			NUM
fcis-17883	91	9			PROPN
fcis-17883	91	10			PROPN
fcis-17883	91	11			NOUN
fcis-17883	91	12			PROPN
fcis-17883	91	13			PROPN
fcis-17883	91	14			ADJ
fcis-17883	91	15			ADJ
fcis-17883	91	16			VERB
fcis-17883	91	17			ADJ
fcis-17883	91	18			ADJ
fcis-17883	91	19			PROPN
fcis-17883	91	20			ADJ
fcis-17883	91	21			PROPN
fcis-17883	91	22			VERB
fcis-17883	91	23			X
fcis-17883	91	24			X
fcis-17883	91	25			X
fcis-17883	92	1			X
fcis-17883	92	2			X
fcis-17883	93	1			NUM
fcis-17883	93	2	a	a	DET
fcis-17883	93	3	sigma	sigma	PROPN
fcis-17883	93	4	-	-	PUNCT
fcis-17883	93	5	pi	pi	NOUN
fcis-17883	93	6	-	-	PUNCT
fcis-17883	93	7	sigma	sigma	NOUN
fcis-17883	93	8	neural	neural	ADJ
fcis-17883	93	9	network	network	NOUN
fcis-17883	93	10	with	with	ADP
fcis-17883	93	11	graph	graph	NOUN
fcis-17883	93	12	regularity	regularity	NOUN
fcis-17883	93	13	terms	term	NOUN
fcis-17883	93	14	(	(	PUNCT
fcis-17883	93	15	spsnngr	spsnngr	NOUN
fcis-17883	93	16	)	)	PUNCT
fcis-17883	93	17	in	in	ADP
fcis-17883	93	18	this	this	DET
fcis-17883	93	19	section	section	NOUN
fcis-17883	93	20	,	,	PUNCT
fcis-17883	93	21	we	we	PRON
fcis-17883	93	22	describe	describe	VERB
fcis-17883	93	23	our	our	PRON
fcis-17883	93	24	spsnn	spsnn	NOUN
fcis-17883	93	25	model	model	NOUN
fcis-17883	93	26	with	with	ADP
fcis-17883	93	27	graph	graph	NOUN
fcis-17883	93	28	regular	regular	ADJ
fcis-17883	93	29	terms	term	NOUN
fcis-17883	93	30	,	,	PUNCT
fcis-17883	93	31	firstly	firstly	ADV
fcis-17883	93	32	we	we	PRON
fcis-17883	93	33	construct	construct	VERB
fcis-17883	93	34	the	the	DET
fcis-17883	93	35	weighted	weight	VERB
fcis-17883	93	36	graph	graph	NOUN
fcis-17883	93	37	from	from	ADP
fcis-17883	93	38	the	the	DET
fcis-17883	93	39	input	input	NOUN
fcis-17883	93	40	data	datum	NOUN
fcis-17883	93	41	,	,	PUNCT
fcis-17883	93	42	secondly	secondly	ADV
fcis-17883	93	43	,	,	PUNCT
fcis-17883	93	44	we	we	PRON
fcis-17883	93	45	compute	compute	VERB
fcis-17883	93	46	its	its	PRON
fcis-17883	93	47	laplace	laplace	NOUN
fcis-17883	93	48	matrix	matrix	NOUN
fcis-17883	93	49	l	l	NOUN
fcis-17883	93	50	based	base	VERB
fcis-17883	93	51	on	on	ADP
fcis-17883	93	52	the	the	DET
fcis-17883	93	53	information	information	NOUN
fcis-17883	93	54	given	give	VERB
fcis-17883	93	55	by	by	ADP
fcis-17883	93	56	the	the	DET
fcis-17883	93	57	weighted	weight	VERB
fcis-17883	93	58	graph	graph	NOUN
fcis-17883	93	59	,	,	PUNCT
fcis-17883	93	60	and	and	CCONJ
fcis-17883	93	61	finally	finally	ADV
fcis-17883	93	62	we	we	PRON
fcis-17883	93	63	compute	compute	VERB
fcis-17883	93	64	the	the	DET
fcis-17883	93	65	graph	graph	NOUN
fcis-17883	93	66	regular	regular	ADJ
fcis-17883	93	67	terms	term	NOUN
fcis-17883	93	68	based	base	VERB
fcis-17883	93	69	on	on	ADP
fcis-17883	93	70	the	the	DET
fcis-17883	93	71	above	above	ADJ
fcis-17883	93	72	information	information	NOUN
fcis-17883	93	73	.	.	PUNCT
fcis-17883	94	1	in	in	ADP
fcis-17883	94	2	this	this	DET
fcis-17883	94	3	paper	paper	NOUN
fcis-17883	94	4	,	,	PUNCT
fcis-17883	94	5	we	we	PRON
fcis-17883	94	6	use	use	VERB
fcis-17883	94	7	a	a	DET
fcis-17883	94	8	gaussian	gaussian	ADJ
fcis-17883	94	9	kernel	kernel	NOUN
fcis-17883	94	10	to	to	PART
fcis-17883	94	11	measure	measure	VERB
fcis-17883	94	12	the	the	DET
fcis-17883	94	13	similarity	similarity	NOUN
fcis-17883	94	14	between	between	ADP
fcis-17883	94	15	any	any	DET
fcis-17883	94	16	two	two	NUM
fcis-17883	94	17	instances	instance	NOUN
fcis-17883	94	18	xi	xi	X
fcis-17883	94	19	and	and	CCONJ
fcis-17883	94	20	xj	xj	PROPN
fcis-17883	94	21	to	to	PART
fcis-17883	94	22	generate	generate	VERB
fcis-17883	94	23	a	a	DET
fcis-17883	94	24	weighted	weight	VERB
fcis-17883	94	25	graph	graph	NOUN
fcis-17883	94	26	g	g	PROPN
fcis-17883	94	27	.	.	PUNCT
fcis-17883	95	1	let	let	VERB
fcis-17883	95	2	1	1	NUM
fcis-17883	95	3	{	{	PUNCT
fcis-17883	95	4	,	,	PUNCT
fcis-17883	95	5	}	}	PUNCT
fcis-17883	96	1	j	j	PROPN
fcis-17883	96	2	j	j	PROPN
fcis-17883	96	3	j	j	PROPN
fcis-17883	97	1	p	p	PROPN
fcis-17883	97	2	jx	jx	PROPN
fcis-17883	98	1	o	o	NOUN
fcis-17883	98	2			PROPN
fcis-17883	99	1			PROPN
fcis-17883	99	2			PUNCT
fcis-17883	99	3			PUNCT
fcis-17883	99	4	be	be	AUX
fcis-17883	99	5	a	a	DET
fcis-17883	99	6	given	give	VERB
fcis-17883	99	7	set	set	NOUN
fcis-17883	99	8	of	of	ADP
fcis-17883	99	9	training	training	NOUN
fcis-17883	99	10	samples	sample	NOUN
fcis-17883	99	11	,	,	PUNCT
fcis-17883	99	12	we	we	PRON
fcis-17883	99	13	have	have	VERB
fcis-17883	99	14	2	2	NUM
fcis-17883	99	15	2||	2||	NUM
fcis-17883	99	16	||	||	NOUN
fcis-17883	99	17	/	/	SYM
fcis-17883	99	18	0	0	NUM
fcis-17883	99	19	,	,	PUNCT
fcis-17883	99	20	?	?	PUNCT
fcis-17883	100	1	i	i	PRON
fcis-17883	100	2	jx	jx	NOUN
fcis-17883	100	3	x	x	PUNCT
fcis-17883	101	1	ij	ij	INTJ
fcis-17883	101	2	e	e	NOUN
fcis-17883	101	3	if	if	SCONJ
fcis-17883	101	4	i	i	PRON
fcis-17883	101	5	j	j	PROPN
fcis-17883	102	1	w	w	INTJ
fcis-17883	102	2	if	if	SCONJ
fcis-17883	102	3	otherwise	otherwise	ADV
fcis-17883	102	4			PROPN
fcis-17883	102	5			PROPN
fcis-17883	102	6			PROPN
fcis-17883	102	7			NUM
fcis-17883	102	8			NOUN
fcis-17883	102	9	where	where	SCONJ
fcis-17883	102	10	2	2	NUM
fcis-17883	102	11	2	2	NUM
fcis-17883	102	12	1	1	NUM
fcis-17883	102	13	1	1	NUM
fcis-17883	102	14	j	j	PROPN
fcis-17883	102	15	j	j	PROPN
fcis-17883	102	16	j	j	PROPN
fcis-17883	102	17	x	x	X
fcis-17883	103	1	j	j	PROPN
fcis-17883	103	2			PROPN
fcis-17883	104	1			NUM
fcis-17883	104	2			NOUN
fcis-17883	104	3			INTJ
fcis-17883	104	4			PROPN
fcis-17883	104	5	.	.	PUNCT
fcis-17883	105	1	therefore	therefore	ADV
fcis-17883	105	2	,	,	PUNCT
fcis-17883	105	3	the	the	DET
fcis-17883	105	4	normalized	normalize	VERB
fcis-17883	105	5	laplace	laplace	NOUN
fcis-17883	105	6	matrix	matrix	NOUN
fcis-17883	105	7	l	l	NOUN
fcis-17883	105	8	of	of	ADP
fcis-17883	105	9	the	the	DET
fcis-17883	105	10	weighted	weight	VERB
fcis-17883	105	11	graph	graph	NOUN
fcis-17883	105	12	g	g	PROPN
fcis-17883	105	13	is	be	AUX
fcis-17883	105	14	defined	define	VERB
fcis-17883	105	15	as	as	ADP
fcis-17883	105	16	l	l	NOUN
fcis-17883	105	17	d	d	NOUN
fcis-17883	105	18	w	w	VERB
fcis-17883	105	19			PROPN
fcis-17883	106	1	where	where	SCONJ
fcis-17883	106	2	ii	ii	PROPN
fcis-17883	106	3	ij	ij	INTJ
fcis-17883	106	4	j	j	PROPN
fcis-17883	107	1	i	i	PRON
fcis-17883	107	2	d	d	PROPN
fcis-17883	107	3	w	w	PROPN
fcis-17883	107	4			NOUN
fcis-17883	107	5			X
fcis-17883	107	6	.	.	PUNCT
fcis-17883	108	1	in	in	ADP
fcis-17883	108	2	order	order	NOUN
fcis-17883	108	3	to	to	PART
fcis-17883	108	4	make	make	VERB
fcis-17883	108	5	the	the	DET
fcis-17883	108	6	outputs	output	NOUN
fcis-17883	108	7	of	of	ADP
fcis-17883	108	8	hidden	hidden	ADJ
fcis-17883	108	9	layers	layer	NOUN
fcis-17883	108	10	from	from	ADP
fcis-17883	108	11	the	the	DET
fcis-17883	108	12	same	same	ADJ
fcis-17883	108	13	class	class	NOUN
fcis-17883	108	14	as	as	ADV
fcis-17883	108	15	close	close	ADJ
fcis-17883	108	16	as	as	ADP
fcis-17883	108	17	possible	possible	ADJ
fcis-17883	108	18	and	and	CCONJ
fcis-17883	108	19	thus	thus	ADV
fcis-17883	108	20	improve	improve	VERB
fcis-17883	108	21	the	the	DET
fcis-17883	108	22	network	network	NOUN
fcis-17883	108	23	accuracy	accuracy	NOUN
fcis-17883	108	24	,	,	PUNCT
fcis-17883	108	25	we	we	PRON
fcis-17883	108	26	define	define	VERB
fcis-17883	108	27	the	the	DET
fcis-17883	108	28	penalty	penalty	NOUN
fcis-17883	108	29	term	term	NOUN
fcis-17883	108	30	function	function	NOUN
fcis-17883	108	31	r	r	NOUN
fcis-17883	108	32	as	as	ADP
fcis-17883	108	33	2	2	NUM
fcis-17883	108	34	,	,	PUNCT
fcis-17883	108	35	1	1	NUM
fcis-17883	108	36	2	2	NUM
fcis-17883	108	37	1	1	NUM
fcis-17883	108	38	,	,	PUNCT
fcis-17883	108	39	1	1	NUM
fcis-17883	108	40	1	1	NUM
fcis-17883	108	41	(	(	PUNCT
fcis-17883	108	42	(	(	PUNCT
fcis-17883	108	43	)	)	PUNCT
fcis-17883	108	44	(	(	PUNCT
fcis-17883	108	45	)	)	PUNCT
fcis-17883	108	46	)	)	PUNCT
fcis-17883	108	47	2	2	NUM
fcis-17883	108	48	(	(	PUNCT
fcis-17883	108	49	)	)	PUNCT
fcis-17883	108	50	(	(	PUNCT
fcis-17883	108	51	)	)	PUNCT
fcis-17883	108	52	(	(	PUNCT
fcis-17883	108	53	)	)	PUNCT
fcis-17883	108	54	(	(	PUNCT
fcis-17883	108	55	)	)	PUNCT
fcis-17883	108	56	(	(	PUNCT
fcis-17883	108	57	)	)	PUNCT
fcis-17883	108	58	(	(	PUNCT
fcis-17883	108	59	)	)	PUNCT
fcis-17883	109	1	i	i	PRON
fcis-17883	109	2	j	j	INTJ
fcis-17883	110	1	ij	ij	INTJ
fcis-17883	111	1	i	i	INTJ
fcis-17883	111	2	j	j	PROPN
fcis-17883	112	1	i	i	PRON
fcis-17883	112	2	ii	ii	VERB
fcis-17883	113	1	i	i	PRON
fcis-17883	113	2	j	j	INTJ
fcis-17883	114	1	ij	ij	INTJ
fcis-17883	114	2	i	i	PRON
fcis-17883	115	1	i	i	INTJ
fcis-17883	116	1	j	j	PROPN
fcis-17883	117	1	t	t	PROPN
fcis-17883	117	2	t	t	PROPN
fcis-17883	117	3	j	j	PROPN
fcis-17883	118	1	j	j	PROPN
fcis-17883	118	2	t	t	PROPN
fcis-17883	118	3	j	j	PROPN
fcis-17883	119	1	r	r	NOUN
fcis-17883	119	2	f	f	PROPN
fcis-17883	119	3	x	x	X
fcis-17883	119	4	f	f	PROPN
fcis-17883	119	5	x	x	X
fcis-17883	119	6	w	w	NOUN
fcis-17883	119	7	f	f	NOUN
fcis-17883	119	8	x	x	PUNCT
fcis-17883	120	1	d	d	NOUN
fcis-17883	120	2	f	f	X
fcis-17883	120	3	x	x	X
fcis-17883	120	4	f	f	PROPN
fcis-17883	120	5	x	x	X
fcis-17883	120	6	w	w	PROPN
fcis-17883	120	7	tr	tr	VERB
fcis-17883	120	8	vdv	vdv	PROPN
fcis-17883	120	9	tr	tr	PROPN
fcis-17883	120	10	vwv	vwv	NOUN
fcis-17883	120	11	tr	tr	VERB
fcis-17883	120	12	vlv	vlv	PROPN
fcis-17883	120	13			PROPN
fcis-17883	121	1			NUM
fcis-17883	122	1			NUM
fcis-17883	122	2			PROPN
fcis-17883	122	3			PROPN
fcis-17883	122	4			PROPN
fcis-17883	122	5			PROPN
fcis-17883	122	6			PROPN
fcis-17883	122	7			PROPN
fcis-17883	122	8			NOUN
fcis-17883	122	9			X
fcis-17883	122	10			X
fcis-17883	122	11			X
fcis-17883	122	12	(	(	PUNCT
fcis-17883	122	13	)	)	PUNCT
fcis-17883	122	14	tr	tr	PROPN
fcis-17883	122	15	x	x	PROPN
fcis-17883	122	16	in	in	ADP
fcis-17883	122	17	the	the	DET
fcis-17883	122	18	above	above	ADJ
fcis-17883	122	19	equation	equation	NOUN
fcis-17883	122	20	is	be	AUX
fcis-17883	122	21	the	the	DET
fcis-17883	122	22	trace	trace	NOUN
fcis-17883	122	23	of	of	ADP
fcis-17883	122	24	the	the	DET
fcis-17883	122	25	matrix	matrix	NOUN
fcis-17883	122	26	x	x	SYM
fcis-17883	122	27	,	,	PUNCT
fcis-17883	122	28	(	(	PUNCT
fcis-17883	122	29	)	)	PUNCT
fcis-17883	122	30	if	if	SCONJ
fcis-17883	122	31	x	x	PRON
fcis-17883	122	32	is	be	AUX
fcis-17883	122	33	a	a	DET
fcis-17883	122	34	function	function	NOUN
fcis-17883	122	35	of	of	ADP
fcis-17883	122	36	the	the	DET
fcis-17883	122	37	mapping	mapping	NOUN
fcis-17883	122	38	of	of	ADP
fcis-17883	122	39	sample	sample	NOUN
fcis-17883	122	40	points	point	NOUN
fcis-17883	122	41	ix	ix	ADV
fcis-17883	122	42	in	in	ADP
fcis-17883	122	43	high	high	ADJ
fcis-17883	122	44	-	-	PUNCT
fcis-17883	122	45	dimensional	dimensional	ADJ
fcis-17883	122	46	space	space	NOUN
fcis-17883	122	47	to	to	ADP
fcis-17883	122	48	low	low	ADV
fcis-17883	122	49	-	-	PUNCT
fcis-17883	122	50	dimensional	dimensional	ADJ
fcis-17883	122	51	,	,	PUNCT
fcis-17883	122	52	and	and	CCONJ
fcis-17883	122	53	v	v	NOUN
fcis-17883	122	54	is	be	AUX
fcis-17883	122	55	the	the	DET
fcis-17883	122	56	low	low	ADJ
fcis-17883	122	57	-	-	PUNCT
fcis-17883	122	58	dimensional	dimensional	ADJ
fcis-17883	122	59	representation	representation	NOUN
fcis-17883	122	60	matrix	matrix	NOUN
fcis-17883	122	61	.	.	PUNCT
fcis-17883	123	1	therefore	therefore	ADV
fcis-17883	123	2	,	,	PUNCT
fcis-17883	123	3	the	the	DET
fcis-17883	123	4	error	error	NOUN
fcis-17883	123	5	function	function	NOUN
fcis-17883	123	6	can	can	AUX
fcis-17883	123	7	be	be	AUX
fcis-17883	123	8	defined	define	VERB
fcis-17883	123	9	as	as	ADP
fcis-17883	123	10	~	~	PUNCT
fcis-17883	123	11	0	0	PUNCT
fcis-17883	123	12	~	~	PUNCT
fcis-17883	123	13	(	(	PUNCT
fcis-17883	123	14	)	)	PUNCT
fcis-17883	123	15	(	(	PUNCT
fcis-17883	123	16	)	)	PUNCT
fcis-17883	123	17	(	(	PUNCT
fcis-17883	123	18	,	,	PUNCT
fcis-17883	123	19	)	)	PUNCT
fcis-17883	123	20	2	2	NUM
fcis-17883	123	21	(	(	PUNCT
fcis-17883	123	22	)	)	PUNCT
fcis-17883	123	23	(	(	PUNCT
fcis-17883	123	24	(	(	PUNCT
fcis-17883	123	25	)	)	PUNCT
fcis-17883	123	26	(	(	PUNCT
fcis-17883	123	27	)	)	PUNCT
fcis-17883	123	28	)	)	PUNCT
fcis-17883	123	29	2	2	NUM
fcis-17883	123	30	n	n	ADP
fcis-17883	123	31	t	t	NOUN
fcis-17883	123	32	n	n	CCONJ
fcis-17883	123	33	n	n	PROPN
fcis-17883	123	34	e	e	PROPN
fcis-17883	123	35	w	w	PROPN
fcis-17883	123	36	e	e	PROPN
fcis-17883	123	37	w	w	PROPN
fcis-17883	123	38	r	r	NOUN
fcis-17883	123	39	w	w	PROPN
fcis-17883	123	40	w	w	PROPN
fcis-17883	123	41	e	e	PROPN
fcis-17883	123	42	w	w	PROPN
fcis-17883	123	43	tr	tr	PROPN
fcis-17883	123	44	f	f	PROPN
fcis-17883	123	45	w	w	PROPN
fcis-17883	123	46	x	x	VERB
fcis-17883	123	47	lf	lf	ADP
fcis-17883	123	48	w	w	PROPN
fcis-17883	123	49	x	x	PROPN
fcis-17883	123	50			X
fcis-17883	123	51			ADJ
fcis-17883	123	52			PROPN
fcis-17883	123	53			ADV
fcis-17883	123	54			NOUN
fcis-17883	123	55			ADJ
fcis-17883	123	56	for	for	ADP
fcis-17883	123	57	the	the	DET
fcis-17883	123	58	sake	sake	NOUN
fcis-17883	123	59	of	of	ADP
fcis-17883	123	60	simpler	simple	ADJ
fcis-17883	123	61	calculations	calculation	NOUN
fcis-17883	123	62	,	,	PUNCT
fcis-17883	123	63	we	we	PRON
fcis-17883	123	64	consider	consider	VERB
fcis-17883	123	65	~	~	PUNCT
fcis-17883	123	66	(	(	PUNCT
fcis-17883	123	67	)	)	PUNCT
fcis-17883	123	68	(	(	PUNCT
fcis-17883	123	69	)	)	PUNCT
fcis-17883	123	70	(	(	PUNCT
fcis-17883	123	71	(	(	PUNCT
fcis-17883	123	72	)	)	PUNCT
fcis-17883	123	73	(	(	PUNCT
fcis-17883	123	74	)	)	PUNCT
fcis-17883	123	75	)	)	PUNCT
fcis-17883	123	76	2	2	NUM
fcis-17883	123	77	t	t	NOUN
fcis-17883	123	78	n	n	X
fcis-17883	123	79	ne	ne	PROPN
fcis-17883	123	80	w	w	PROPN
fcis-17883	123	81	e	e	PROPN
fcis-17883	123	82	w	w	ADP
fcis-17883	123	83	tr	tr	NOUN
fcis-17883	123	84	w	w	NOUN
fcis-17883	123	85	x	x	NOUN
fcis-17883	123	86	l	l	NOUN
fcis-17883	123	87	w	w	NOUN
fcis-17883	123	88	x	x	X
fcis-17883	123	89			X
fcis-17883	123	90			NOUN
fcis-17883	123	91			PUNCT
fcis-17883	123	92	so	so	ADV
fcis-17883	123	93	,	,	PUNCT
fcis-17883	123	94	we	we	PRON
fcis-17883	123	95	can	can	AUX
fcis-17883	123	96	have	have	VERB
fcis-17883	123	97	0	0	NUM
fcis-17883	123	98	0	0	NUM
fcis-17883	123	99	~	~	PUNCT
fcis-17883	123	100	'	'	PUNCT
fcis-17883	123	101	0	0	NUM
fcis-17883	123	102	1	1	NUM
fcis-17883	123	103	(	(	PUNCT
fcis-17883	123	104	)	)	PUNCT
fcis-17883	123	105	(	(	PUNCT
fcis-17883	123	106	)	)	PUNCT
fcis-17883	123	107	(	(	PUNCT
fcis-17883	123	108	)	)	PUNCT
fcis-17883	124	1	j	j	PROPN
fcis-17883	125	1	j	j	PROPN
fcis-17883	125	2	j	j	PROPN
fcis-17883	125	3	w	w	PROPN
fcis-17883	125	4	w	w	PROPN
fcis-17883	125	5	j	j	PROPN
fcis-17883	125	6	j	j	PROPN
fcis-17883	125	7	e	e	PROPN
fcis-17883	125	8	w	w	PROPN
fcis-17883	125	9	e	e	PROPN
fcis-17883	125	10	w	w	PROPN
fcis-17883	125	11	g	g	PROPN
fcis-17883	125	12	w	w	PROPN
fcis-17883	125	13			NOUN
fcis-17883	126	1			X
fcis-17883	127	1			NOUN
fcis-17883	127	2			PROPN
fcis-17883	127	3			PROPN
fcis-17883	127	4			X
fcis-17883	127	5	(	(	PUNCT
fcis-17883	127	6	)	)	PUNCT
fcis-17883	127	7	(	(	PUNCT
fcis-17883	127	8	)	)	PUNCT
fcis-17883	127	9	(	(	PUNCT
fcis-17883	127	10	(	(	PUNCT
fcis-17883	127	11	)	)	PUNCT
fcis-17883	127	12	(	(	PUNCT
fcis-17883	127	13	'	'	PUNCT
fcis-17883	127	14	(	(	PUNCT
fcis-17883	127	15	)	)	PUNCT
fcis-17883	127	16	)	)	PUNCT
fcis-17883	127	17	)	)	PUNCT
fcis-17883	128	1	2	2	NUM
fcis-17883	128	2	n	n	NUM
fcis-17883	128	3	n	n	NOUN
fcis-17883	128	4	t	t	PROPN
fcis-17883	128	5	w	w	PROPN
fcis-17883	128	6	w	w	PROPN
fcis-17883	128	7	n	n	PROPN
fcis-17883	128	8	ne	ne	PROPN
fcis-17883	128	9	w	w	PROPN
fcis-17883	128	10	e	e	PROPN
fcis-17883	128	11	w	w	PROPN
fcis-17883	128	12	tr	tr	PROPN
fcis-17883	128	13	f	f	PROPN
fcis-17883	128	14	w	w	PROPN
fcis-17883	128	15	x	x	PUNCT
fcis-17883	128	16	l	l	NOUN
fcis-17883	128	17	f	f	X
fcis-17883	128	18	w	w	NOUN
fcis-17883	128	19	x	x	PUNCT
fcis-17883	128	20	x	x	X
fcis-17883	128	21			ADJ
fcis-17883	128	22			PROPN
fcis-17883	128	23			ADV
fcis-17883	128	24			PROPN
fcis-17883	128	25	the	the	DET
fcis-17883	128	26	initial	initial	ADJ
fcis-17883	128	27	weight	weight	NOUN
fcis-17883	128	28	0w	0w	PROPN
fcis-17883	128	29	is	be	AUX
fcis-17883	128	30	given	give	VERB
fcis-17883	128	31	,	,	PUNCT
fcis-17883	128	32	and	and	CCONJ
fcis-17883	128	33	the	the	DET
fcis-17883	128	34	iterative	iterative	NOUN
fcis-17883	128	35	updates	update	NOUN
fcis-17883	128	36	of	of	ADP
fcis-17883	128	37	the	the	DET
fcis-17883	128	38	weights	weight	NOUN
fcis-17883	128	39	of	of	ADP
fcis-17883	128	40	the	the	DET
fcis-17883	128	41	spsnngr	spsnngr	ADJ
fcis-17883	128	42	algorithm	algorithm	NOUN
fcis-17883	128	43	are	be	AUX
fcis-17883	128	44	as	as	SCONJ
fcis-17883	128	45	follows	follow	VERB
fcis-17883	128	46	:	:	PUNCT
fcis-17883	128	47	1	1	NUM
fcis-17883	128	48	1	1	NUM
fcis-17883	128	49	,	,	PUNCT
fcis-17883	128	50	2	2	NUM
fcis-17883	128	51	,	,	PUNCT
fcis-17883	128	52	m	m	VERB
fcis-17883	128	53	m	m	VERB
fcis-17883	128	54	mw	mw	ADJ
fcis-17883	128	55	w	w	PROPN
fcis-17883	128	56	w	w	PROPN
fcis-17883	128	57	m	m	PROPN
fcis-17883	128	58			PROPN
fcis-17883	128	59			ADV
fcis-17883	128	60			NUM
fcis-17883	128	61			NOUN
fcis-17883	128	62	and	and	CCONJ
fcis-17883	128	63	(	(	PUNCT
fcis-17883	128	64	)	)	PUNCT
fcis-17883	128	65	m	m	VERB
fcis-17883	128	66	m	m	VERB
fcis-17883	128	67	ww	ww	PROPN
fcis-17883	128	68	e	e	PROPN
fcis-17883	128	69	w	w	PROPN
fcis-17883	128	70			NOUN
fcis-17883	128	71	.	.	PUNCT
fcis-17883	129	1	where	where	SCONJ
fcis-17883	129	2			NUM
fcis-17883	129	3	is	be	AUX
fcis-17883	129	4	the	the	DET
fcis-17883	129	5	learning	learning	NOUN
fcis-17883	129	6	rate	rate	NOUN
fcis-17883	129	7	.	.	PUNCT
fcis-17883	130	1	3	3	X
fcis-17883	130	2	.	.	X
fcis-17883	130	3	simulation	simulation	NOUN
fcis-17883	130	4	results	result	NOUN
fcis-17883	130	5	(	(	PUNCT
fcis-17883	130	6	1	1	X
fcis-17883	130	7	)	)	PUNCT
fcis-17883	130	8	function	function	NOUN
fcis-17883	130	9	approximation	approximation	NOUN
fcis-17883	130	10	problems	problem	NOUN
fcis-17883	130	11	to	to	PART
fcis-17883	130	12	illustrate	illustrate	VERB
fcis-17883	130	13	the	the	DET
fcis-17883	130	14	effectiveness	effectiveness	NOUN
fcis-17883	130	15	of	of	ADP
fcis-17883	130	16	the	the	DET
fcis-17883	130	17	spsnngr	spsnngr	ADJ
fcis-17883	130	18	algorithm	algorithm	NOUN
fcis-17883	130	19	,	,	PUNCT
fcis-17883	130	20	the	the	DET
fcis-17883	130	21	following	follow	VERB
fcis-17883	130	22	is	be	AUX
fcis-17883	130	23	an	an	DET
fcis-17883	130	24	example	example	NOUN
fcis-17883	130	25	to	to	PART
fcis-17883	130	26	compare	compare	VERB
fcis-17883	130	27	the	the	DET
fcis-17883	130	28	spsnn	spsnn	VERB
fcis-17883	130	29	algorithm	algorithm	NOUN
fcis-17883	130	30	and	and	CCONJ
fcis-17883	130	31	the	the	DET
fcis-17883	130	32	spssnngr	spssnngr	NOUN
fcis-17883	130	33	algorithm	algorithm	NOUN
fcis-17883	130	34	with	with	ADP
fcis-17883	130	35	a	a	DET
fcis-17883	130	36	nonlinear	nonlinear	ADJ
fcis-17883	130	37	function	function	NOUN
fcis-17883	130	38	that	that	SCONJ
fcis-17883	130	39	36	36	NUM
fcis-17883	130	40	is	be	AUX
fcis-17883	130	41	formulated	formulate	VERB
fcis-17883	130	42	as	as	ADP
fcis-17883	130	43	0.2	0.2	NUM
fcis-17883	130	44	(	(	PUNCT
fcis-17883	130	45	sin	sin	NOUN
fcis-17883	130	46	(	(	PUNCT
fcis-17883	130	47	)	)	PUNCT
fcis-17883	130	48	1.2)y	1.2)y	PROPN
fcis-17883	130	49	x	x	PUNCT
fcis-17883	130	50			NUM
fcis-17883	130	51			VERB
fcis-17883	130	52	for	for	ADP
fcis-17883	130	53	the	the	DET
fcis-17883	130	54	function	function	NOUN
fcis-17883	130	55	approximation	approximation	NOUN
fcis-17883	130	56	problem	problem	NOUN
fcis-17883	130	57	,	,	PUNCT
fcis-17883	130	58	we	we	PRON
fcis-17883	130	59	used	use	VERB
fcis-17883	130	60	the	the	DET
fcis-17883	130	61	network	network	NOUN
fcis-17883	130	62	utilizing	utilize	VERB
fcis-17883	130	63	this	this	DET
fcis-17883	130	64	structure	structure	NOUN
fcis-17883	130	65	to	to	PART
fcis-17883	130	66	simulate	simulate	VERB
fcis-17883	130	67	all	all	DET
fcis-17883	130	68	the	the	DET
fcis-17883	130	69	algorithms(input	algorithms(input	PROPN
fcis-17883	130	70	2p	2p	NUM
fcis-17883	130	71			PROPN
fcis-17883	130	72	,	,	PUNCT
fcis-17883	130	73	1	1	NUM
fcis-17883	130	74	layer	layer	NOUN
fcis-17883	130	75	4n	4n	NOUN
fcis-17883	130	76			PRON
fcis-17883	130	77	,	,	PUNCT
fcis-17883	130	78			ADP
fcis-17883	130	79	layer	layer	NOUN
fcis-17883	130	80	16q	16q	NOUN
fcis-17883	130	81			NUM
fcis-17883	130	82	and	and	CCONJ
fcis-17883	130	83	2	2	NUM
fcis-17883	130	84	layer	layer	NOUN
fcis-17883	130	85	)	)	PUNCT
fcis-17883	130	86	.	.	PUNCT
fcis-17883	131	1	in	in	ADP
fcis-17883	131	2	this	this	DET
fcis-17883	131	3	experiment	experiment	NOUN
fcis-17883	131	4	,	,	PUNCT
fcis-17883	131	5	151	151	NUM
fcis-17883	131	6	input	input	NOUN
fcis-17883	131	7	variables	variable	NOUN
fcis-17883	131	8	were	be	AUX
fcis-17883	131	9	selected	select	VERB
fcis-17883	131	10	in	in	ADP
fcis-17883	131	11	the	the	DET
fcis-17883	131	12	interval	interval	NOUN
fcis-17883	131	13	[	[	PUNCT
fcis-17883	131	14	3,3]	3,3]	NUM
fcis-17883	131	15	.	.	PUNCT
fcis-17883	132	1	the	the	DET
fcis-17883	132	2	initial	initial	ADJ
fcis-17883	132	3	weight	weight	NOUN
fcis-17883	132	4	w	w	NOUN
fcis-17883	132	5	is	be	AUX
fcis-17883	132	6	randomly	randomly	ADV
fcis-17883	132	7	selected	select	VERB
fcis-17883	132	8	in	in	ADP
fcis-17883	132	9	the	the	DET
fcis-17883	132	10	interval	interval	NOUN
fcis-17883	132	11	[	[	PUNCT
fcis-17883	132	12	0.5,0.5]	0.5,0.5]	PROPN
fcis-17883	132	13	,	,	PUNCT
fcis-17883	132	14	the	the	DET
fcis-17883	132	15	learning	learning	NOUN
fcis-17883	132	16	rate	rate	NOUN
fcis-17883	132	17	0.01	0.01	PROPN
fcis-17883	133	1			ADJ
fcis-17883	133	2	and	and	CCONJ
fcis-17883	133	3	the	the	DET
fcis-17883	133	4	penalty	penalty	NOUN
fcis-17883	133	5	term	term	NOUN
fcis-17883	133	6	coefficient	coefficient	NOUN
fcis-17883	133	7	0.001	0.001	PUNCT
fcis-17883	134	1			NOUN
fcis-17883	134	2	.	.	PUNCT
fcis-17883	135	1	the	the	DET
fcis-17883	135	2	experimental	experimental	ADJ
fcis-17883	135	3	termination	termination	NOUN
fcis-17883	135	4	condition	condition	NOUN
fcis-17883	135	5	is	be	AUX
fcis-17883	135	6	that	that	SCONJ
fcis-17883	135	7	the	the	DET
fcis-17883	135	8	error	error	NOUN
fcis-17883	135	9	is	be	AUX
fcis-17883	135	10	less	less	ADJ
fcis-17883	135	11	than	than	ADP
fcis-17883	135	12	51	51	NUM
fcis-17883	135	13	10	10	NUM
fcis-17883	135	14	or	or	CCONJ
fcis-17883	135	15	the	the	DET
fcis-17883	135	16	number	number	NOUN
fcis-17883	135	17	of	of	ADP
fcis-17883	135	18	iterations	iteration	NOUN
fcis-17883	135	19	reaches	reach	VERB
fcis-17883	135	20	5000	5000	NUM
fcis-17883	135	21	times	time	NOUN
fcis-17883	135	22	.	.	PUNCT
fcis-17883	136	1	figure	figure	NOUN
fcis-17883	136	2	2	2	NUM
fcis-17883	136	3	shows	show	VERB
fcis-17883	136	4	the	the	DET
fcis-17883	136	5	network	network	NOUN
fcis-17883	136	6	approximation	approximation	NOUN
fcis-17883	136	7	graph	graph	NOUN
fcis-17883	136	8	of	of	ADP
fcis-17883	136	9	spsnngr	spsnngr	NOUN
fcis-17883	136	10	,	,	PUNCT
fcis-17883	136	11	and	and	CCONJ
fcis-17883	136	12	figure	figure	VERB
fcis-17883	136	13	3	3	NUM
fcis-17883	136	14	shows	show	VERB
fcis-17883	136	15	the	the	DET
fcis-17883	136	16	error	error	NOUN
fcis-17883	136	17	function	function	NOUN
fcis-17883	136	18	graphs	graph	NOUN
fcis-17883	136	19	of	of	ADP
fcis-17883	136	20	spsnn	spsnn	VERB
fcis-17883	136	21	algorithm	algorithm	NOUN
fcis-17883	136	22	and	and	CCONJ
fcis-17883	136	23	spsnngr	spsnngr	ADJ
fcis-17883	136	24	algorithm	algorithm	NOUN
fcis-17883	136	25	,	,	PUNCT
fcis-17883	136	26	it	it	PRON
fcis-17883	136	27	can	can	AUX
fcis-17883	136	28	be	be	AUX
fcis-17883	136	29	seen	see	VERB
fcis-17883	136	30	that	that	SCONJ
fcis-17883	136	31	the	the	DET
fcis-17883	136	32	spsnngr	spsnngr	ADJ
fcis-17883	136	33	algorithm	algorithm	NOUN
fcis-17883	136	34	has	have	VERB
fcis-17883	136	35	good	good	ADJ
fcis-17883	136	36	approximation	approximation	NOUN
fcis-17883	136	37	ability	ability	NOUN
fcis-17883	136	38	and	and	CCONJ
fcis-17883	136	39	the	the	DET
fcis-17883	136	40	error	error	NOUN
fcis-17883	136	41	of	of	ADP
fcis-17883	136	42	spsnngr	spsnngr	ADJ
fcis-17883	136	43	algorithm	algorithm	NOUN
fcis-17883	136	44	decreases	decrease	VERB
fcis-17883	136	45	faster	fast	ADV
fcis-17883	136	46	,	,	PUNCT
fcis-17883	136	47	which	which	PRON
fcis-17883	136	48	indicates	indicate	VERB
fcis-17883	136	49	that	that	SCONJ
fcis-17883	136	50	our	our	PRON
fcis-17883	136	51	proposed	propose	VERB
fcis-17883	136	52	algorithm	algorithm	NOUN
fcis-17883	136	53	performs	perform	VERB
fcis-17883	136	54	better	well	ADJ
fcis-17883	136	55	.	.	PUNCT
fcis-17883	137	1	figure	figure	NOUN
fcis-17883	137	2	2	2	NUM
fcis-17883	137	3	.	.	PUNCT
fcis-17883	137	4	approximation	approximation	NOUN
fcis-17883	137	5	result	result	NOUN
fcis-17883	137	6	of	of	ADP
fcis-17883	137	7	spsnngr	spsnngr	ADJ
fcis-17883	137	8	figure	figure	NOUN
fcis-17883	137	9	3	3	NUM
fcis-17883	137	10	.	.	PUNCT
fcis-17883	137	11	errors	error	NOUN
fcis-17883	137	12	of	of	ADP
fcis-17883	137	13	spsnn	spsnn	NOUN
fcis-17883	137	14	and	and	CCONJ
fcis-17883	137	15	spsnngr	spsnngr	NOUN
fcis-17883	137	16	.	.	PUNCT
fcis-17883	138	1	(	(	PUNCT
fcis-17883	138	2	2	2	X
fcis-17883	138	3	)	)	PUNCT
fcis-17883	138	4	real	real	ADJ
fcis-17883	138	5	-	-	PUNCT
fcis-17883	138	6	world	world	NOUN
fcis-17883	138	7	classification	classification	NOUN
fcis-17883	138	8	problems	problem	NOUN
fcis-17883	138	9	to	to	PART
fcis-17883	138	10	further	far	ADV
fcis-17883	138	11	validate	validate	VERB
fcis-17883	138	12	the	the	DET
fcis-17883	138	13	algorithmic	algorithmic	ADJ
fcis-17883	138	14	performance	performance	NOUN
fcis-17883	138	15	of	of	ADP
fcis-17883	138	16	spsnngr	spsnngr	NOUN
fcis-17883	138	17	,	,	PUNCT
fcis-17883	138	18	we	we	PRON
fcis-17883	138	19	conducted	conduct	VERB
fcis-17883	138	20	numerical	numerical	ADJ
fcis-17883	138	21	experiments	experiment	NOUN
fcis-17883	138	22	on	on	ADP
fcis-17883	138	23	4	4	NUM
fcis-17883	138	24	real	real	ADJ
fcis-17883	138	25	datasets	dataset	NOUN
fcis-17883	138	26	,	,	PUNCT
fcis-17883	138	27	all	all	PRON
fcis-17883	138	28	selected	select	VERB
fcis-17883	138	29	from	from	ADP
fcis-17883	138	30	the	the	DET
fcis-17883	138	31	uci	uci	PROPN
fcis-17883	138	32	database	database	NOUN
fcis-17883	138	33	.	.	PUNCT
fcis-17883	139	1	in	in	ADP
fcis-17883	139	2	this	this	DET
fcis-17883	139	3	experiment	experiment	NOUN
fcis-17883	139	4	,	,	PUNCT
fcis-17883	139	5	we	we	PRON
fcis-17883	139	6	compare	compare	VERB
fcis-17883	139	7	the	the	DET
fcis-17883	139	8	performance	performance	NOUN
fcis-17883	139	9	of	of	ADP
fcis-17883	139	10	the	the	DET
fcis-17883	139	11	spsnn	spsnn	VERB
fcis-17883	139	12	algorithm	algorithm	NOUN
fcis-17883	139	13	and	and	CCONJ
fcis-17883	139	14	the	the	DET
fcis-17883	139	15	spsnnger	spsnng	ADJ
fcis-17883	139	16	algorithm	algorithm	NOUN
fcis-17883	139	17	.	.	PUNCT
fcis-17883	140	1	we	we	PRON
fcis-17883	140	2	set	set	VERB
fcis-17883	140	3	the	the	DET
fcis-17883	140	4	initial	initial	ADJ
fcis-17883	140	5	network	network	NOUN
fcis-17883	140	6	architecture	architecture	NOUN
fcis-17883	140	7	of	of	ADP
fcis-17883	140	8	spsnn	spsnn	VERB
fcis-17883	140	9	as(input	as(input	PROPN
fcis-17883	140	10	2p	2p	NOUN
fcis-17883	140	11			PROPN
fcis-17883	140	12	,	,	PUNCT
fcis-17883	140	13	1	1	NUM
fcis-17883	140	14	layer	layer	NOUN
fcis-17883	140	15	4n	4n	NOUN
fcis-17883	140	16			PRON
fcis-17883	140	17	,	,	PUNCT
fcis-17883	140	18			ADP
fcis-17883	140	19	layer	layer	NOUN
fcis-17883	140	20	16q	16q	NOUN
fcis-17883	140	21			NUM
fcis-17883	140	22	and	and	CCONJ
fcis-17883	140	23	2	2	NUM
fcis-17883	140	24	layer	layer	NOUN
fcis-17883	140	25	)	)	PUNCT
fcis-17883	140	26	,	,	PUNCT
fcis-17883	140	27	where	where	SCONJ
fcis-17883	140	28	the	the	DET
fcis-17883	140	29	learning	learning	NOUN
fcis-17883	140	30	rate	rate	NOUN
fcis-17883	140	31	0.001	0.001	PROPN
fcis-17883	140	32			ADJ
fcis-17883	140	33	and	and	CCONJ
fcis-17883	140	34	the	the	DET
fcis-17883	140	35	penalty	penalty	NOUN
fcis-17883	140	36	term	term	NOUN
fcis-17883	140	37	coefficient	coefficient	NOUN
fcis-17883	140	38	0.001	0.001	PUNCT
fcis-17883	141	1			INTJ
fcis-17883	141	2	.	.	PUNCT
fcis-17883	142	1	in	in	ADP
fcis-17883	142	2	addition	addition	NOUN
fcis-17883	142	3	,	,	PUNCT
fcis-17883	142	4	our	our	PRON
fcis-17883	142	5	initial	initial	ADJ
fcis-17883	142	6	weights	weight	NOUN
fcis-17883	142	7	w	w	NOUN
fcis-17883	142	8	are	be	AUX
fcis-17883	142	9	randomly	randomly	ADV
fcis-17883	142	10	generated	generate	VERB
fcis-17883	142	11	from	from	ADP
fcis-17883	142	12	the	the	DET
fcis-17883	142	13	interval	interval	NOUN
fcis-17883	142	14	,	,	PUNCT
fcis-17883	142	15	and	and	CCONJ
fcis-17883	142	16	we	we	PRON
fcis-17883	142	17	set	set	VERB
fcis-17883	142	18	the	the	DET
fcis-17883	142	19	training	training	NOUN
fcis-17883	142	20	to	to	PART
fcis-17883	142	21	stop	stop	VERB
fcis-17883	142	22	when	when	SCONJ
fcis-17883	142	23	the	the	DET
fcis-17883	142	24	error	error	NOUN
fcis-17883	142	25	is	be	AUX
fcis-17883	142	26	less	less	ADJ
fcis-17883	142	27	than	than	ADP
fcis-17883	142	28	51	51	NUM
fcis-17883	142	29	10	10	NUM
fcis-17883	142	30	or	or	CCONJ
fcis-17883	142	31	the	the	DET
fcis-17883	142	32	number	number	NOUN
fcis-17883	142	33	of	of	ADP
fcis-17883	142	34	iterations	iteration	NOUN
fcis-17883	142	35	reach	reach	VERB
fcis-17883	142	36	5000	5000	NUM
fcis-17883	142	37	.	.	PUNCT
fcis-17883	143	1	we	we	PRON
fcis-17883	143	2	used	use	VERB
fcis-17883	143	3	a	a	DET
fcis-17883	143	4	cross	cross	ADJ
fcis-17883	143	5	-	-	ADJ
fcis-17883	143	6	validation	validation	ADJ
fcis-17883	143	7	approach	approach	NOUN
fcis-17883	143	8	that	that	PRON
fcis-17883	143	9	,	,	PUNCT
fcis-17883	143	10	70	70	NUM
fcis-17883	143	11	%	%	NOUN
fcis-17883	143	12	of	of	ADP
fcis-17883	143	13	each	each	DET
fcis-17883	143	14	dataset	dataset	NOUN
fcis-17883	143	15	is	be	AUX
fcis-17883	143	16	selected	select	VERB
fcis-17883	143	17	as	as	ADP
fcis-17883	143	18	the	the	DET
fcis-17883	143	19	training	training	NOUN
fcis-17883	143	20	set	set	NOUN
fcis-17883	143	21	,	,	PUNCT
fcis-17883	143	22	and	and	CCONJ
fcis-17883	143	23	the	the	DET
fcis-17883	143	24	rest	rest	NOUN
fcis-17883	143	25	is	be	AUX
fcis-17883	143	26	used	use	VERB
fcis-17883	143	27	as	as	ADP
fcis-17883	143	28	the	the	DET
fcis-17883	143	29	test	test	NOUN
fcis-17883	143	30	set	set	NOUN
fcis-17883	143	31	,	,	PUNCT
fcis-17883	143	32	to	to	PART
fcis-17883	143	33	obtain	obtain	VERB
fcis-17883	143	34	the	the	DET
fcis-17883	143	35	average	average	ADJ
fcis-17883	143	36	results	result	NOUN
fcis-17883	143	37	of	of	ADP
fcis-17883	143	38	each	each	DET
fcis-17883	143	39	algorithm	algorithm	NOUN
fcis-17883	143	40	under	under	ADP
fcis-17883	143	41	10	10	NUM
fcis-17883	143	42	simulation	simulation	NOUN
fcis-17883	143	43	experiments	experiment	NOUN
fcis-17883	143	44	,	,	PUNCT
fcis-17883	143	45	and	and	CCONJ
fcis-17883	143	46	to	to	PART
fcis-17883	143	47	provide	provide	VERB
fcis-17883	143	48	a	a	DET
fcis-17883	143	49	detailed	detailed	ADJ
fcis-17883	143	50	analysis	analysis	NOUN
fcis-17883	143	51	and	and	CCONJ
fcis-17883	143	52	comparison	comparison	NOUN
fcis-17883	143	53	of	of	ADP
fcis-17883	143	54	their	their	PRON
fcis-17883	143	55	performance	performance	NOUN
fcis-17883	143	56	.	.	PUNCT
fcis-17883	144	1	table	table	NOUN
fcis-17883	144	2	1	1	NUM
fcis-17883	144	3	.	.	PUNCT
fcis-17883	145	1	performance	performance	NOUN
fcis-17883	145	2	comparison	comparison	NOUN
fcis-17883	145	3	of	of	ADP
fcis-17883	145	4	classification	classification	NOUN
fcis-17883	145	5	problems	problem	NOUN
fcis-17883	145	6	data	datum	NOUN
fcis-17883	145	7	set	set	VERB
fcis-17883	145	8	accuracy	accuracy	NOUN
fcis-17883	145	9	rate	rate	NOUN
fcis-17883	145	10	spsnn	spsnn	VERB
fcis-17883	145	11	spsnngr	spsnngr	ADJ
fcis-17883	145	12	tic	tic	ADJ
fcis-17883	145	13	training	training	NOUN
fcis-17883	145	14	0.9584	0.9584	NUM
fcis-17883	145	15	0.9642	0.9642	NUM
fcis-17883	145	16	testing	test	VERB
fcis-17883	145	17	0.9061	0.9061	NUM
fcis-17883	145	18	0.9229	0.9229	NUM
fcis-17883	145	19	heart	heart	NOUN
fcis-17883	145	20	training	train	VERB
fcis-17883	145	21	0.9670	0.9670	NUM
fcis-17883	145	22	0.9952	0.9952	NUM
fcis-17883	145	23	testing	test	VERB
fcis-17883	145	24	0.9187	0.9187	NUM
fcis-17883	145	25	0.9890	0.9890	NUM
fcis-17883	145	26	wireless	wireless	ADJ
fcis-17883	145	27	training	train	VERB
fcis-17883	145	28	0.8350	0.8350	NUM
fcis-17883	145	29	0.8486	0.8486	NUM
fcis-17883	145	30	testing	testing	NOUN
fcis-17883	145	31	0.8171	0.8171	NUM
fcis-17883	145	32	0.8404	0.8404	NUM
fcis-17883	145	33	wholesale	wholesale	NOUN
fcis-17883	145	34	training	train	VERB
fcis-17883	145	35	0.8539	0.8539	NUM
fcis-17883	145	36	0.8571	0.8571	NUM
fcis-17883	145	37	testing	test	VERB
fcis-17883	145	38	0.8045	0.8045	NUM
fcis-17883	145	39	0.8271	0.8271	NUM
fcis-17883	145	40	the	the	DET
fcis-17883	145	41	performance	performance	NOUN
fcis-17883	145	42	comparison	comparison	NOUN
fcis-17883	145	43	results	result	NOUN
fcis-17883	145	44	are	be	AUX
fcis-17883	145	45	shown	show	VERB
fcis-17883	145	46	in	in	ADP
fcis-17883	145	47	the	the	DET
fcis-17883	145	48	table	table	NOUN
fcis-17883	145	49	1	1	NUM
fcis-17883	145	50	,	,	PUNCT
fcis-17883	145	51	we	we	PRON
fcis-17883	145	52	can	can	AUX
fcis-17883	145	53	clearly	clearly	ADV
fcis-17883	145	54	see	see	VERB
fcis-17883	145	55	that	that	PRON
fcis-17883	145	56	for	for	ADP
fcis-17883	145	57	most	most	ADJ
fcis-17883	145	58	of	of	ADP
fcis-17883	145	59	the	the	DET
fcis-17883	145	60	datasets	dataset	NOUN
fcis-17883	145	61	,	,	PUNCT
fcis-17883	145	62	the	the	DET
fcis-17883	145	63	spsnngr	spsnngr	ADJ
fcis-17883	145	64	algorithm	algorithm	NOUN
fcis-17883	145	65	has	have	VERB
fcis-17883	145	66	higher	high	ADJ
fcis-17883	145	67	training	training	NOUN
fcis-17883	145	68	and	and	CCONJ
fcis-17883	145	69	testing	testing	NOUN
fcis-17883	145	70	accuracy	accuracy	NOUN
fcis-17883	145	71	compared	compare	VERB
fcis-17883	145	72	to	to	ADP
fcis-17883	145	73	the	the	DET
fcis-17883	145	74	spsnn	spsnn	VERB
fcis-17883	145	75	algorithm	algorithm	NOUN
fcis-17883	145	76	,	,	PUNCT
fcis-17883	145	77	which	which	PRON
fcis-17883	145	78	indicates	indicate	VERB
fcis-17883	145	79	that	that	SCONJ
fcis-17883	145	80	the	the	DET
fcis-17883	145	81	spsnngr	spsnngr	ADJ
fcis-17883	145	82	algorithm	algorithm	NOUN
fcis-17883	145	83	is	be	AUX
fcis-17883	145	84	more	more	ADV
fcis-17883	145	85	generalized	generalized	ADJ
fcis-17883	145	86	and	and	CCONJ
fcis-17883	145	87	stable	stable	ADJ
fcis-17883	145	88	.	.	PUNCT
fcis-17883	146	1	4	4	X
fcis-17883	146	2	.	.	X
fcis-17883	146	3	conclusion	conclusion	NOUN
fcis-17883	146	4	in	in	ADP
fcis-17883	146	5	his	his	PRON
fcis-17883	146	6	paper	paper	NOUN
fcis-17883	146	7	,	,	PUNCT
fcis-17883	146	8	a	a	DET
fcis-17883	146	9	new	new	ADJ
fcis-17883	146	10	sigma	sigma	ADJ
fcis-17883	146	11	-	-	PUNCT
fcis-17883	146	12	pi	pi	NOUN
fcis-17883	146	13	-	-	PUNCT
fcis-17883	146	14	sigma	sigma	NOUN
fcis-17883	146	15	neural	neural	ADJ
fcis-17883	146	16	network	network	NOUN
fcis-17883	146	17	model	model	NOUN
fcis-17883	146	18	with	with	ADP
fcis-17883	146	19	graph	graph	NOUN
fcis-17883	146	20	regularity	regularity	NOUN
fcis-17883	146	21	is	be	AUX
fcis-17883	146	22	established	establish	VERB
fcis-17883	146	23	by	by	ADP
fcis-17883	146	24	introducing	introduce	VERB
fcis-17883	146	25	the	the	DET
fcis-17883	146	26	graph	graph	NOUN
fcis-17883	146	27	regularity	regularity	NOUN
fcis-17883	146	28	term	term	NOUN
fcis-17883	146	29	,	,	PUNCT
fcis-17883	146	30	and	and	CCONJ
fcis-17883	146	31	the	the	DET
fcis-17883	146	32	algorithm	algorithm	NOUN
fcis-17883	146	33	not	not	PART
fcis-17883	146	34	only	only	ADV
fcis-17883	146	35	improves	improve	VERB
fcis-17883	146	36	the	the	DET
fcis-17883	146	37	generalization	generalization	NOUN
fcis-17883	146	38	ability	ability	NOUN
fcis-17883	146	39	of	of	ADP
fcis-17883	146	40	the	the	DET
fcis-17883	146	41	network	network	NOUN
fcis-17883	146	42	,	,	PUNCT
fcis-17883	146	43	but	but	CCONJ
fcis-17883	146	44	also	also	ADV
fcis-17883	146	45	improves	improve	VERB
fcis-17883	146	46	the	the	DET
fcis-17883	146	47	convergence	convergence	NOUN
fcis-17883	146	48	speed	speed	NOUN
fcis-17883	146	49	.	.	PUNCT
fcis-17883	147	1	in	in	ADP
fcis-17883	147	2	addition	addition	NOUN
fcis-17883	147	3	,	,	PUNCT
fcis-17883	147	4	we	we	PRON
fcis-17883	147	5	conducted	conduct	VERB
fcis-17883	147	6	numerical	numerical	ADJ
fcis-17883	147	7	experiments	experiment	NOUN
fcis-17883	147	8	to	to	PART
fcis-17883	147	9	verify	verify	VERB
fcis-17883	147	10	the	the	DET
fcis-17883	147	11	effectiveness	effectiveness	NOUN
fcis-17883	147	12	of	of	ADP
fcis-17883	147	13	the	the	DET
fcis-17883	147	14	new	new	ADJ
fcis-17883	147	15	algorithm	algorithm	NOUN
fcis-17883	147	16	,	,	PUNCT
fcis-17883	147	17	and	and	CCONJ
fcis-17883	147	18	it	it	PRON
fcis-17883	147	19	can	can	AUX
fcis-17883	147	20	be	be	AUX
fcis-17883	147	21	seen	see	VERB
fcis-17883	147	22	that	that	SCONJ
fcis-17883	147	23	,	,	PUNCT
fcis-17883	147	24	compared	compare	VERB
fcis-17883	147	25	with	with	ADP
fcis-17883	147	26	the	the	DET
fcis-17883	147	27	spsnn	spsnn	VERB
fcis-17883	147	28	algorithm	algorithm	NOUN
fcis-17883	147	29	,	,	PUNCT
fcis-17883	147	30	the	the	DET
fcis-17883	147	31	new	new	ADJ
fcis-17883	147	32	algorithm	algorithm	NOUN
fcis-17883	147	33	not	not	PART
fcis-17883	147	34	only	only	ADV
fcis-17883	147	35	has	have	VERB
fcis-17883	147	36	good	good	ADJ
fcis-17883	147	37	performance	performance	NOUN
fcis-17883	147	38	in	in	ADP
fcis-17883	147	39	the	the	DET
fcis-17883	147	40	function	function	NOUN
fcis-17883	147	41	approximation	approximation	NOUN
fcis-17883	147	42	experiments	experiment	NOUN
fcis-17883	147	43	,	,	PUNCT
fcis-17883	147	44	but	but	CCONJ
fcis-17883	147	45	also	also	ADV
fcis-17883	147	46	has	have	VERB
fcis-17883	147	47	higher	high	ADJ
fcis-17883	147	48	accuracy	accuracy	NOUN
fcis-17883	147	49	and	and	CCONJ
fcis-17883	147	50	better	well	ADJ
fcis-17883	147	51	generalization	generalization	NOUN
fcis-17883	147	52	performance	performance	NOUN
fcis-17883	147	53	in	in	ADP
fcis-17883	147	54	the	the	DET
fcis-17883	147	55	classification	classification	NOUN
fcis-17883	147	56	experiments	experiment	NOUN
fcis-17883	147	57	references	reference	NOUN
fcis-17883	147	58	[	[	X
fcis-17883	147	59	1	1	NUM
fcis-17883	147	60	]	]	X
fcis-17883	147	61	c	c	PROPN
fcis-17883	147	62	k	k	PROPN
fcis-17883	147	63	li	li	PROPN
fcis-17883	147	64	.	.	PUNCT
fcis-17883	148	1	a	a	DET
fcis-17883	148	2	sigma	sigma	PROPN
fcis-17883	148	3	-	-	PUNCT
fcis-17883	148	4	pi	pi	NOUN
fcis-17883	148	5	-	-	PUNCT
fcis-17883	148	6	sigma	sigma	NOUN
fcis-17883	148	7	neural	neural	ADJ
fcis-17883	148	8	network	network	NOUN
fcis-17883	148	9	(	(	PUNCT
fcis-17883	148	10	spsnn)[j	spsnn)[j	PROPN
fcis-17883	148	11	]	]	PUNCT
fcis-17883	148	12	.	.	PUNCT
fcis-17883	149	1	neural	neural	ADJ
fcis-17883	149	2	processing	processing	NOUN
fcis-17883	149	3	letters	letter	NOUN
fcis-17883	149	4	,	,	PUNCT
fcis-17883	149	5	2003(17	2003(17	NUM
fcis-17883	149	6	):	):	PUNCT
fcis-17883	149	7	1	1	NUM
fcis-17883	149	8	-	-	SYM
fcis-17883	149	9	19	19	NUM
fcis-17883	149	10	.	.	PUNCT
fcis-17883	150	1	[	[	X
fcis-17883	150	2	2	2	X
fcis-17883	150	3	]	]	X
fcis-17883	150	4	q	q	PROPN
fcis-17883	150	5	w	w	PROPN
fcis-17883	150	6	fan	fan	PROPN
fcis-17883	150	7	,	,	PUNCT
fcis-17883	150	8	q	q	PROPN
fcis-17883	150	9	kang	kang	PROPN
fcis-17883	150	10	,	,	PUNCT
fcis-17883	150	11	j	j	PROPN
fcis-17883	150	12	m	m	PROPN
fcis-17883	150	13	zurada	zurada	PROPN
fcis-17883	150	14	.	.	PUNCT
fcis-17883	151	1	convergence	convergence	NOUN
fcis-17883	151	2	analysis	analysis	NOUN
fcis-17883	151	3	for	for	ADP
fcis-17883	151	4	sigma	sigma	ADJ
fcis-17883	151	5	-	-	PUNCT
fcis-17883	151	6	pi	pi	NOUN
fcis-17883	151	7	-	-	PUNCT
fcis-17883	151	8	sigma	sigma	NOUN
fcis-17883	151	9	neural	neural	ADJ
fcis-17883	151	10	network	network	NOUN
fcis-17883	151	11	based	base	VERB
fcis-17883	151	12	on	on	ADP
fcis-17883	151	13	some	some	DET
fcis-17883	151	14	relaxed	relaxed	ADJ
fcis-17883	151	15	conditions[j	conditions[j	PROPN
fcis-17883	151	16	]	]	PUNCT
fcis-17883	151	17	.	.	PUNCT
fcis-17883	152	1	information	information	NOUN
fcis-17883	152	2	sciences	sciences	PROPN
fcis-17883	152	3	,	,	PUNCT
fcis-17883	152	4	2022(585	2022(585	NUM
fcis-17883	152	5	):	):	PUNCT
fcis-17883	152	6	70	70	NUM
fcis-17883	152	7	-	-	SYM
fcis-17883	152	8	88	88	NUM
fcis-17883	152	9	.	.	PUNCT
fcis-17883	153	1	[	[	X
fcis-17883	153	2	3	3	NUM
fcis-17883	153	3	]	]	SYM
fcis-17883	153	4	s	s	NOUN
fcis-17883	153	5	n	n	PRON
fcis-17883	153	6	arsla	arsla	ADJ
fcis-17883	153	7	.	.	PUNCT
fcis-17883	154	1	a	a	DET
fcis-17883	154	2	hybrid	hybrid	ADJ
fcis-17883	154	3	sigma	sigma	ADJ
fcis-17883	154	4	-	-	PUNCT
fcis-17883	154	5	pi	pi	NOUN
fcis-17883	154	6	neural	neural	ADJ
fcis-17883	154	7	network	network	NOUN
fcis-17883	154	8	for	for	ADP
fcis-17883	154	9	combined	combined	ADJ
fcis-17883	154	10	intuitionistic	intuitionistic	ADJ
fcis-17883	154	11	fuzzy	fuzzy	ADJ
fcis-17883	154	12	time	time	NOUN
fcis-17883	154	13	series	series	PROPN
fcis-17883	154	14	prediction	prediction	NOUN
fcis-17883	154	15	model[j	model[j	PROPN
fcis-17883	154	16	]	]	X
fcis-17883	154	17	.	.	PUNCT
fcis-17883	155	1	neural	neural	ADJ
fcis-17883	155	2	comput	comput	NOUN
fcis-17883	155	3	and	and	CCONJ
fcis-17883	155	4	applic,2022(34	applic,2022(34	PROPN
fcis-17883	155	5	):	):	PUNCT
fcis-17883	155	6	12895	12895	NUM
fcis-17883	155	7	-	-	SYM
fcis-17883	155	8	12917	12917	NUM
fcis-17883	155	9	.	.	PUNCT
fcis-17883	156	1	[	[	X
fcis-17883	156	2	4	4	NUM
fcis-17883	156	3	]	]	X
fcis-17883	156	4	r	r	NOUN
fcis-17883	156	5	s	s	PROPN
fcis-17883	156	6	neville	neville	NOUN
fcis-17883	156	7	.	.	PUNCT
fcis-17883	157	1	reuse	reuse	NOUN
fcis-17883	157	2	of	of	ADP
fcis-17883	157	3	information	information	NOUN
fcis-17883	157	4	in	in	ADP
fcis-17883	157	5	multi	multi	ADJ
fcis-17883	157	6	-	-	NOUN
fcis-17883	157	7	layer	layer	ADJ
fcis-17883	157	8	sigma	sigma	PROPN
fcis-17883	157	9	-	-	PUNCT
fcis-17883	157	10	pi	pi	NOUN
fcis-17883	157	11	neural	neural	ADJ
fcis-17883	157	12	networks[j	networks[j	PROPN
fcis-17883	157	13	]	]	X
fcis-17883	157	14	.	.	PUNCT
fcis-17883	158	1	connection	connection	NOUN
fcis-17883	158	2	science	science	NOUN
fcis-17883	158	3	,	,	PUNCT
fcis-17883	158	4	2006(18	2006(18	NUM
fcis-17883	158	5	):	):	PUNCT
fcis-17883	158	6	43	43	NUM
fcis-17883	158	7	-	-	SYM
fcis-17883	158	8	59	59	NUM
fcis-17883	158	9	.	.	PUNCT
fcis-17883	159	1	[	[	X
fcis-17883	159	2	5	5	NUM
fcis-17883	159	3	]	]	SYM
fcis-17883	159	4	g	g	PROPN
fcis-17883	159	5	welper	welper	NOUN
fcis-17883	159	6	.	.	PUNCT
fcis-17883	160	1	universality	universality	NOUN
fcis-17883	160	2	of	of	ADP
fcis-17883	160	3	gradient	gradient	ADJ
fcis-17883	160	4	descent	descent	NOUN
fcis-17883	160	5	neural	neural	ADJ
fcis-17883	160	6	network	network	NOUN
fcis-17883	160	7	training[j	training[j	NOUN
fcis-17883	160	8	]	]	PUNCT
fcis-17883	160	9	.	.	PUNCT
fcis-17883	161	1	neural	neural	ADJ
fcis-17883	161	2	networks	network	NOUN
fcis-17883	161	3	,	,	PUNCT
fcis-17883	161	4	2022(150	2022(150	NUM
fcis-17883	161	5	):	):	PUNCT
fcis-17883	161	6	259	259	NUM
fcis-17883	161	7	-	-	SYM
fcis-17883	161	8	273	273	NUM
fcis-17883	161	9	.	.	PUNCT
fcis-17883	162	1	[	[	X
fcis-17883	162	2	6	6	NUM
fcis-17883	162	3	]	]	X
fcis-17883	162	4	y	y	PROPN
fcis-17883	162	5	liu	liu	PROPN
fcis-17883	162	6	,	,	PUNCT
fcis-17883	162	7	d	d	PROPN
fcis-17883	162	8	k	k	PROPN
fcis-17883	162	9	yang	yang	PROPN
fcis-17883	162	10	,	,	PUNCT
fcis-17883	162	11	f	f	PROPN
fcis-17883	162	12	li	li	PROPN
fcis-17883	162	13	.	.	PROPN
fcis-17883	162	14	smoothed	smooth	VERB
fcis-17883	162	15	regularizer	regularizer	NOUN
fcis-17883	162	16	learning	learn	VERB
fcis-17883	162	17	for	for	ADP
fcis-17883	162	18	splitcomplex	splitcomplex	NOUN
fcis-17883	162	19	valued	value	VERB
fcis-17883	162	20	neuro	neuro	NOUN
fcis-17883	162	21	-	-	PUNCT
fcis-17883	162	22	fuzzy	fuzzy	ADJ
fcis-17883	162	23	algorithm	algorithm	NOUN
fcis-17883	162	24	for	for	ADP
fcis-17883	162	25	tsk	tsk	NOUN
fcis-17883	162	26	system	system	NOUN
fcis-17883	162	27	and	and	CCONJ
fcis-17883	162	28	its	its	PRON
fcis-17883	162	29	convergence	convergence	NOUN
fcis-17883	162	30	results[j	results[j	PROPN
fcis-17883	162	31	]	]	PUNCT
fcis-17883	162	32	.	.	PUNCT
fcis-17883	163	1	journal	journal	PROPN
fcis-17883	163	2	of	of	ADP
fcis-17883	163	3	the	the	DET
fcis-17883	163	4	franklin	franklin	PROPN
fcis-17883	163	5	institute	institute	PROPN
fcis-17883	163	6	,	,	PUNCT
fcis-17883	163	7	2018	2018	NUM
fcis-17883	163	8	(	(	PUNCT
fcis-17883	163	9	355	355	NUM
fcis-17883	163	10	):	):	PUNCT
fcis-17883	163	11	6132	6132	NUM
fcis-17883	163	12	-	-	SYM
fcis-17883	163	13	6151	6151	NUM
fcis-17883	163	14	.	.	PUNCT
fcis-17883	164	1	[	[	X
fcis-17883	164	2	7	7	X
fcis-17883	164	3	]	]	X
fcis-17883	164	4	p	p	X
fcis-17883	164	5	may	may	AUX
fcis-17883	164	6	,	,	PUNCT
fcis-17883	164	7	e	e	PROPN
fcis-17883	164	8	zhou	zhou	PROPN
fcis-17883	164	9	,	,	PUNCT
fcis-17883	164	10	c	c	PROPN
fcis-17883	164	11	w	w	PROPN
fcis-17883	164	12	lee	lee	PROPN
fcis-17883	164	13	.	.	PUNCT
fcis-17883	165	1	a	a	DET
fcis-17883	165	2	comprehensive	comprehensive	ADJ
fcis-17883	165	3	evaluation	evaluation	NOUN
fcis-17883	165	4	of	of	ADP
fcis-17883	165	5	weight	weight	NOUN
fcis-17883	165	6	growth	growth	NOUN
fcis-17883	165	7	and	and	CCONJ
fcis-17883	165	8	weight	weight	NOUN
fcis-17883	165	9	elimination	elimination	NOUN
fcis-17883	165	10	methods	method	NOUN
fcis-17883	165	11	using	use	VERB
fcis-17883	165	12	the	the	DET
fcis-17883	165	13	37	37	NUM
fcis-17883	165	14	tangent	tangent	NOUN
fcis-17883	165	15	plane	plane	NOUN
fcis-17883	165	16	algorithm[j	algorithm[j	PROPN
fcis-17883	165	17	]	]	PUNCT
fcis-17883	165	18	.	.	PUNCT
fcis-17883	166	1	international	international	ADJ
fcis-17883	166	2	journal	journal	PROPN
fcis-17883	166	3	of	of	ADP
fcis-17883	166	4	advanced	advanced	ADJ
fcis-17883	166	5	computer	computer	NOUN
fcis-17883	166	6	science	science	NOUN
fcis-17883	166	7	and	and	CCONJ
fcis-17883	166	8	applications	application	NOUN
fcis-17883	166	9	,	,	PUNCT
fcis-17883	166	10	2013(6):149	2013(6):149	NUM
fcis-17883	166	11	-	-	SYM
fcis-17883	166	12	156	156	NUM
fcis-17883	166	13	.	.	PUNCT
fcis-17883	167	1	[	[	X
fcis-17883	167	2	8	8	NUM
fcis-17883	167	3	]	]	X
fcis-17883	167	4	e	e	X
fcis-17883	167	5	geoffrey	geoffrey	PROPN
fcis-17883	167	6	.	.	PUNCT
fcis-17883	167	7	connectionist	connectionist	NOUN
fcis-17883	167	8	learning	learn	VERB
fcis-17883	167	9	procedures[j	procedures[j	PROPN
fcis-17883	167	10	]	]	PUNCT
fcis-17883	167	11	.	.	PUNCT
fcis-17883	168	1	artificial	artificial	ADJ
fcis-17883	168	2	intelligence	intelligence	NOUN
fcis-17883	168	3	,	,	PUNCT
fcis-17883	168	4	1989(40):185	1989(40):185	NUM
fcis-17883	168	5	-	-	SYM
fcis-17883	168	6	234	234	NUM
fcis-17883	168	7	.	.	PUNCT
fcis-17883	169	1	[	[	X
fcis-17883	169	2	9	9	NUM
fcis-17883	169	3	]	]	PUNCT
fcis-17883	169	4	p	p	X
fcis-17883	169	5	l	l	NOUN
fcis-17883	169	6	zhai	zhai	PROPN
fcis-17883	169	7	,	,	PUNCT
fcis-17883	169	8	s	s	PROPN
fcis-17883	170	1	h	h	PROPN
fcis-17883	170	2	zhang	zhang	PROPN
fcis-17883	170	3	.	.	PUNCT
fcis-17883	170	4	learnable	learnable	ADJ
fcis-17883	170	5	graph	graph	NOUN
fcis-17883	170	6	-	-	PUNCT
fcis-17883	170	7	regularization	regularization	NOUN
fcis-17883	170	8	for	for	ADP
fcis-17883	170	9	matrix	matrix	NOUN
fcis-17883	170	10	decomposition[j	decomposition[j	PROPN
fcis-17883	170	11	]	]	PUNCT
fcis-17883	170	12	.	.	PUNCT
fcis-17883	171	1	association	association	NOUN
fcis-17883	171	2	for	for	ADP
fcis-17883	171	3	computing	compute	VERB
fcis-17883	171	4	machinery	machinery	NOUN
fcis-17883	171	5	,	,	PUNCT
fcis-17883	171	6	2023(17):1	2023(17):1	NUM
fcis-17883	171	7	-	-	SYM
fcis-17883	171	8	20	20	NUM
fcis-17883	171	9	.	.	PUNCT
fcis-17883	172	1	[	[	X
fcis-17883	172	2	10	10	NUM
fcis-17883	172	3	]	]	X
fcis-17883	172	4	r	r	NOUN
fcis-17883	172	5	nassif	nassif	NOUN
fcis-17883	172	6	,	,	PUNCT
fcis-17883	172	7	s	s	PART
fcis-17883	172	8	vlaski	vlaski	NOUN
fcis-17883	172	9	.	.	PUNCT
fcis-17883	173	1	a	a	DET
fcis-17883	173	2	regularization	regularization	NOUN
fcis-17883	173	3	framework	framework	NOUN
fcis-17883	173	4	for	for	ADP
fcis-17883	173	5	learning	learn	VERB
fcis-17883	173	6	over	over	ADP
fcis-17883	173	7	multitask	multitask	PROPN
fcis-17883	173	8	graphs[j	graphs[j	PROPN
fcis-17883	173	9	]	]	PUNCT
fcis-17883	173	10	.	.	PUNCT
fcis-17883	174	1	in	in	ADP
fcis-17883	174	2	ieee	ieee	NOUN
fcis-17883	174	3	signal	signal	NOUN
fcis-17883	174	4	processing	processing	NOUN
fcis-17883	174	5	letters	letter	NOUN
fcis-17883	174	6	,	,	PUNCT
fcis-17883	174	7	2019(26):297	2019(26):297	PROPN
fcis-17883	174	8	-	-	SYM
fcis-17883	174	9	301	301	NUM
fcis-17883	174	10	.	.	PUNCT
fcis-17883	175	1	[	[	X
fcis-17883	175	2	11	11	NUM
fcis-17883	175	3	]	]	X
fcis-17883	175	4	y	y	PROPN
fcis-17883	175	5	t	t	PROPN
fcis-17883	175	6	fan	fan	PROPN
fcis-17883	175	7	,	,	PUNCT
fcis-17883	175	8	w	w	PROPN
fcis-17883	175	9	y	y	PROPN
fcis-17883	175	10	yang	yang	PROPN
fcis-17883	175	11	.	.	PUNCT
fcis-17883	176	1	a	a	DET
fcis-17883	176	2	backpropagation	backpropagation	NOUN
fcis-17883	176	3	learning	learn	VERB
fcis-17883	176	4	algorithm	algorithm	NOUN
fcis-17883	176	5	with	with	ADP
fcis-17883	176	6	graph	graph	NOUN
fcis-17883	176	7	regularization	regularization	NOUN
fcis-17883	176	8	for	for	ADP
fcis-17883	176	9	feedforward	feedforward	NOUN
fcis-17883	176	10	neural	neural	ADJ
fcis-17883	176	11	networks[j	networks[j	PROPN
fcis-17883	176	12	]	]	X
fcis-17883	176	13	.	.	PUNCT
fcis-17883	177	1	information	information	NOUN
fcis-17883	177	2	sciences	sciences	PROPN
fcis-17883	177	3	,	,	PUNCT
fcis-17883	177	4	2022(607	2022(607	NUM
fcis-17883	177	5	):	):	PUNCT
fcis-17883	177	6	263	263	NUM
fcis-17883	177	7	-	-	SYM
fcis-17883	177	8	277	277	NUM
fcis-17883	177	9	.	.	PUNCT