id	sid	tid	token	lemma	pos
work_ezwbhq2nrvadncyjpwjlluuona	1	1	Connection	Connection	NNP
work_ezwbhq2nrvadncyjpwjlluuona	1	2	Science	Science	NNP
work_ezwbhq2nrvadncyjpwjlluuona	1	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	1	4	1991	1991	CD
work_ezwbhq2nrvadncyjpwjlluuona	1	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	1	6	3(2	3(2	NNP
work_ezwbhq2nrvadncyjpwjlluuona	1	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	1	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	1	9	163–178	163–178	CD
work_ezwbhq2nrvadncyjpwjlluuona	1	10	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	2	1	Simulating	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	2	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	3	“	"	``
work_ezwbhq2nrvadncyjpwjlluuona	2	4	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	2	5	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	2	6	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	7	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	8	”	"	''
work_ezwbhq2nrvadncyjpwjlluuona	2	9	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	11	problem	problem	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	12	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	13	perceptual	perceptual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	2	14	learning	learn	VBG
work_ezwbhq2nrvadncyjpwjlluuona	2	15	Alice	Alice	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	16	O’Toole	O’Toole	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	17	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	18	University	University	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	20	Texas	Texas	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	21	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	22	Dallas	Dallas	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	23	Ken	Ken	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	24	Deffenbacher	Deffenbacher	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	25	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	26	University	University	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	27	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	28	Nebraska	Nebraska	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	29	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	30	Omaha	Omaha	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	31	Hervé	Hervé	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	32	Abdi	Abdi	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	33	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	34	University	University	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	35	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	36	Texas	Texas	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	37	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	38	Dallas	Dallas	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	39	Jim	Jim	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	40	Bartlett	Bartlett	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	41	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	42	University	University	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	43	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	44	Texas	Texas	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	45	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	46	Dallas	Dallas	NNP
work_ezwbhq2nrvadncyjpwjlluuona	2	47	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	2	48	report	report	VBP
work_ezwbhq2nrvadncyjpwjlluuona	2	49	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	50	series	series	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	51	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	52	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	2	53	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	54	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	55	well	well	RB
work_ezwbhq2nrvadncyjpwjlluuona	2	56	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	2	57	known	know	VBN
work_ezwbhq2nrvadncyjpwjlluuona	2	58	“	"	``
work_ezwbhq2nrvadncyjpwjlluuona	2	59	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	2	60	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	2	61	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	62	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	63	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	2	64	”	"	''
work_ezwbhq2nrvadncyjpwjlluuona	2	65	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	2	66	trained	train	VBD
work_ezwbhq2nrvadncyjpwjlluuona	2	67	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	68	autoassociative	autoassociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	2	69	network	network	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	70	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	71	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	72	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	73	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	2	74	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	75	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	76	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	77	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	78	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	2	79	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	2	80	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	2	81	tested	test	VBD
work_ezwbhq2nrvadncyjpwjlluuona	2	82	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	83	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	84	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	2	85	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	2	86	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	2	87	process	process	VB
work_ezwbhq2nrvadncyjpwjlluuona	2	88	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	2	89	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	90	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	2	91	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	2	92	races	race	NNS
work_ezwbhq2nrvadncyjpwjlluuona	2	93	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	2	94	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	2	95	ways	way	NNS
work_ezwbhq2nrvadncyjpwjlluuona	2	96	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	3	1	First	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	3	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	3	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	3	4	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	3	5	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	3	6	better	well	RBR
work_ezwbhq2nrvadncyjpwjlluuona	3	7	able	able	JJ
work_ezwbhq2nrvadncyjpwjlluuona	3	8	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	3	9	reconstruct	reconstruct	VB
work_ezwbhq2nrvadncyjpwjlluuona	3	10	unlearned	unlearned	JJ
work_ezwbhq2nrvadncyjpwjlluuona	3	11	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	3	12	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	3	13	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	3	14	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	3	15	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	3	16	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	4	1	Secondly	secondly	RB
work_ezwbhq2nrvadncyjpwjlluuona	4	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	4	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	4	4	average	average	JJ
work_ezwbhq2nrvadncyjpwjlluuona	4	5	inter	inter	JJ
work_ezwbhq2nrvadncyjpwjlluuona	4	6	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	4	7	face	face	JJ
work_ezwbhq2nrvadncyjpwjlluuona	4	8	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	4	9	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	4	10	higher	high	JJR
work_ezwbhq2nrvadncyjpwjlluuona	4	11	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	4	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	4	13	reconstructed	reconstructed	JJ
work_ezwbhq2nrvadncyjpwjlluuona	4	14	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	4	15	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	4	16	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	4	17	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	4	18	reconstructed	reconstructed	JJ
work_ezwbhq2nrvadncyjpwjlluuona	4	19	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	4	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	4	21	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	4	22	indi-	indi-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	4	23	cating	cating	NN
work_ezwbhq2nrvadncyjpwjlluuona	4	24	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	4	25	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	4	26	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	4	27	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	4	28	coding	code	VBG
work_ezwbhq2nrvadncyjpwjlluuona	4	29	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	4	30	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	4	31	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	4	32	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	4	33	distinctively	distinctively	RB
work_ezwbhq2nrvadncyjpwjlluuona	4	34	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	4	35	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	4	36	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	4	37	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	4	38	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	5	1	These	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	5	2	results	result	NNS
work_ezwbhq2nrvadncyjpwjlluuona	5	3	held	hold	VBN
work_ezwbhq2nrvadncyjpwjlluuona	5	4	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	5	5	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	5	6	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	5	7	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	5	8	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	5	9	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	5	10	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	5	11	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	5	12	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	5	13	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	5	14	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	5	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	5	16	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	5	17	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	5	18	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	5	19	vice	vice	NN
work_ezwbhq2nrvadncyjpwjlluuona	5	20	versa	versa	RB
work_ezwbhq2nrvadncyjpwjlluuona	5	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	6	1	Thirdly	thirdly	RB
work_ezwbhq2nrvadncyjpwjlluuona	6	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	6	3	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	6	4	simulated	simulate	VBD
work_ezwbhq2nrvadncyjpwjlluuona	6	5	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	6	6	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	7	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	8	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	6	9	same-	same-	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	10	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	6	11	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	6	12	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	6	13	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	14	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	6	15	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	6	16	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	6	17	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	6	18	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	19	history	history	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	20	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	21	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	6	22	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	6	23	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	24	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	25	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	26	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	6	27	equal	equal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	6	28	numbers	number	NNS
work_ezwbhq2nrvadncyjpwjlluuona	6	29	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	6	30	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	6	31	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	6	32	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	6	33	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	6	34	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	6	35	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	6	36	reconstructing	reconstruct	VBG
work_ezwbhq2nrvadncyjpwjlluuona	6	37	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	6	38	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	6	39	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	6	40	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	6	41	weighted	weighted	JJ
work_ezwbhq2nrvadncyjpwjlluuona	6	42	combination	combination	NN
work_ezwbhq2nrvadncyjpwjlluuona	6	43	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	6	44	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	6	45	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	6	46	matrices	matrix	NNS
work_ezwbhq2nrvadncyjpwjlluuona	6	47	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	7	1	Using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	7	2	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	7	3	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	7	4	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	7	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	7	6	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	7	7	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	7	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	7	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	7	10	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	7	11	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	7	12	better	well	RBR
work_ezwbhq2nrvadncyjpwjlluuona	7	13	able	able	JJ
work_ezwbhq2nrvadncyjpwjlluuona	7	14	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	7	15	discriminate	discriminate	VB
work_ezwbhq2nrvadncyjpwjlluuona	7	16	learned	learn	VBN
work_ezwbhq2nrvadncyjpwjlluuona	7	17	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	7	18	new	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	7	19	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	7	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	7	21	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	7	22	learned	learn	VBN
work_ezwbhq2nrvadncyjpwjlluuona	7	23	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	7	24	new	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	7	25	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	7	26	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	7	27	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	8	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	8	2	discuss	discuss	VBP
work_ezwbhq2nrvadncyjpwjlluuona	8	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	8	4	results	result	NNS
work_ezwbhq2nrvadncyjpwjlluuona	8	5	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	8	6	terms	term	NNS
work_ezwbhq2nrvadncyjpwjlluuona	8	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	8	8	perceptual	perceptual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	8	9	tuning	tuning	NN
work_ezwbhq2nrvadncyjpwjlluuona	8	10	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	8	11	information	information	NN
work_ezwbhq2nrvadncyjpwjlluuona	8	12	useful	useful	JJ
work_ezwbhq2nrvadncyjpwjlluuona	8	13	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	8	14	processing	processing	NN
work_ezwbhq2nrvadncyjpwjlluuona	8	15	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	8	16	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	8	17	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	8	18	single	single	JJ
work_ezwbhq2nrvadncyjpwjlluuona	8	19	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	8	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	9	1	Keywords	keyword	NNS
work_ezwbhq2nrvadncyjpwjlluuona	9	2	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	9	3	Face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	9	4	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	9	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	9	6	autoassociative	autoassociative	VBP
work_ezwbhq2nrvadncyjpwjlluuona	9	7	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	9	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	9	9	neural	neural	JJ
work_ezwbhq2nrvadncyjpwjlluuona	9	10	network	network	NN
work_ezwbhq2nrvadncyjpwjlluuona	9	11	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	9	12	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	9	13	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	9	14	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	9	15	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	9	16	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	10	1	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	11	1	Introduction	introduction	NN
work_ezwbhq2nrvadncyjpwjlluuona	11	2	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	3	many	many	JJ
work_ezwbhq2nrvadncyjpwjlluuona	11	4	years	year	NNS
work_ezwbhq2nrvadncyjpwjlluuona	11	5	scientist	scientist	NN
work_ezwbhq2nrvadncyjpwjlluuona	11	6	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	11	7	layperson	layperson	NN
work_ezwbhq2nrvadncyjpwjlluuona	11	8	alike	alike	RB
work_ezwbhq2nrvadncyjpwjlluuona	11	9	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	11	10	suspected	suspect	VBN
work_ezwbhq2nrvadncyjpwjlluuona	11	11	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	12	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	11	13	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	14	one	one	PRP
work_ezwbhq2nrvadncyjpwjlluuona	11	15	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	11	16	own	own	JJ
work_ezwbhq2nrvadncyjpwjlluuona	11	17	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	11	18	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	11	19	Thanks	thank	NNS
work_ezwbhq2nrvadncyjpwjlluuona	11	20	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	11	21	due	due	JJ
work_ezwbhq2nrvadncyjpwjlluuona	11	22	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	23	June	June	NNP
work_ezwbhq2nrvadncyjpwjlluuona	11	24	Chance	Chance	NNP
work_ezwbhq2nrvadncyjpwjlluuona	11	25	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	11	26	Al	Al	NNP
work_ezwbhq2nrvadncyjpwjlluuona	11	27	Goldstein	Goldstein	NNP
work_ezwbhq2nrvadncyjpwjlluuona	11	28	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	29	providing	provide	VBG
work_ezwbhq2nrvadncyjpwjlluuona	11	30	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	11	31	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	11	32	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	11	33	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	11	34	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	11	35	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	11	36	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	37	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	11	38	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	11	39	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	11	40	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	11	41	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	42	Peter	Peter	NNP
work_ezwbhq2nrvadncyjpwjlluuona	11	43	Assmann	Assmann	NNP
work_ezwbhq2nrvadncyjpwjlluuona	11	44	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	11	45	Barbara	Barbara	NNP
work_ezwbhq2nrvadncyjpwjlluuona	11	46	Edwards	Edwards	NNP
work_ezwbhq2nrvadncyjpwjlluuona	11	47	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	11	48	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	11	49	anonymous	anonymous	JJ
work_ezwbhq2nrvadncyjpwjlluuona	11	50	reviewers	reviewer	NNS
work_ezwbhq2nrvadncyjpwjlluuona	11	51	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	52	helpful	helpful	JJ
work_ezwbhq2nrvadncyjpwjlluuona	11	53	comments	comment	NNS
work_ezwbhq2nrvadncyjpwjlluuona	11	54	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	55	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	11	56	earlier	early	JJR
work_ezwbhq2nrvadncyjpwjlluuona	11	57	version	version	NN
work_ezwbhq2nrvadncyjpwjlluuona	11	58	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	11	59	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	11	60	manuscript	manuscript	NN
work_ezwbhq2nrvadncyjpwjlluuona	11	61	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	12	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	12	2	project	project	NN
work_ezwbhq2nrvadncyjpwjlluuona	12	3	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	12	4	supported	support	VBN
work_ezwbhq2nrvadncyjpwjlluuona	12	5	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	12	6	part	part	NN
work_ezwbhq2nrvadncyjpwjlluuona	12	7	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	12	8	National	National	NNP
work_ezwbhq2nrvadncyjpwjlluuona	12	9	Institute	Institute	NNP
work_ezwbhq2nrvadncyjpwjlluuona	12	10	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	12	11	Aging	Aging	NNP
work_ezwbhq2nrvadncyjpwjlluuona	12	12	Grant	Grant	NNP
work_ezwbhq2nrvadncyjpwjlluuona	12	13	RO1-AG07798	RO1-AG07798	NNP
work_ezwbhq2nrvadncyjpwjlluuona	12	14	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	12	15	J.C.B.	J.C.B.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	13	1	Send	send	VB
work_ezwbhq2nrvadncyjpwjlluuona	13	2	correspondence	correspondence	NN
work_ezwbhq2nrvadncyjpwjlluuona	13	3	about	about	IN
work_ezwbhq2nrvadncyjpwjlluuona	13	4	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	13	5	paper	paper	NN
work_ezwbhq2nrvadncyjpwjlluuona	13	6	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	13	7	Hervé	Hervé	NNP
work_ezwbhq2nrvadncyjpwjlluuona	13	8	Abdi	Abdi	NNPS
work_ezwbhq2nrvadncyjpwjlluuona	13	9	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	13	10	herve@utdallas.edu	herve@utdallas.edu	ADD
work_ezwbhq2nrvadncyjpwjlluuona	13	11	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	14	1	Home	home	NN
work_ezwbhq2nrvadncyjpwjlluuona	14	2	page	page	NN
work_ezwbhq2nrvadncyjpwjlluuona	14	3	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	14	4	http://www.utdallas.edu/∼herve	http://www.utdallas.edu/∼herve	NNP
work_ezwbhq2nrvadncyjpwjlluuona	14	5	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	15	1	recognized	recognize	VBN
work_ezwbhq2nrvadncyjpwjlluuona	15	2	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	15	3	accurately	accurately	RB
work_ezwbhq2nrvadncyjpwjlluuona	15	4	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	15	5	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	15	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	15	7	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	15	8	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	15	9	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	16	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	16	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	16	3	early	early	JJ
work_ezwbhq2nrvadncyjpwjlluuona	16	4	part	part	NN
work_ezwbhq2nrvadncyjpwjlluuona	16	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	16	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	16	7	century	century	NN
work_ezwbhq2nrvadncyjpwjlluuona	16	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	16	9	Feingold	Feingold	NNP
work_ezwbhq2nrvadncyjpwjlluuona	16	10	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	16	11	1914,p	1914,p	NNP
work_ezwbhq2nrvadncyjpwjlluuona	16	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	17	1	50	50	CD
work_ezwbhq2nrvadncyjpwjlluuona	17	2	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	17	3	stated	state	VBD
work_ezwbhq2nrvadncyjpwjlluuona	17	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	17	5	supposition	supposition	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	6	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	17	7	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	17	8	plausible	plausible	JJ
work_ezwbhq2nrvadncyjpwjlluuona	17	9	reason	reason	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	10	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	17	11	its	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	17	12	existence	existence	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	13	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	17	14	way	way	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	15	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	17	16	things	thing	NNS
work_ezwbhq2nrvadncyjpwjlluuona	17	17	being	be	VBG
work_ezwbhq2nrvadncyjpwjlluuona	17	18	equal	equal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	17	19	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	17	20	individuals	individual	NNS
work_ezwbhq2nrvadncyjpwjlluuona	17	21	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	17	22	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	17	23	given	give	VBN
work_ezwbhq2nrvadncyjpwjlluuona	17	24	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	25	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	17	26	distinguishable	distinguishable	JJ
work_ezwbhq2nrvadncyjpwjlluuona	17	27	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	17	28	each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	17	29	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	17	30	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	17	31	proportion	proportion	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	32	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	17	33	our	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	17	34	familiarity	familiarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	35	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	17	36	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	17	37	our	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	17	38	contact	contact	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	39	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	17	40	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	17	41	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	42	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	17	43	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	17	44	whole	whole	NN
work_ezwbhq2nrvadncyjpwjlluuona	17	45	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	18	1	Thus	thus	RB
work_ezwbhq2nrvadncyjpwjlluuona	18	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	18	3	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	18	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	18	5	uninitiated	uninitiated	JJ
work_ezwbhq2nrvadncyjpwjlluuona	18	6	American	American	NNP
work_ezwbhq2nrvadncyjpwjlluuona	18	7	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	18	8	Asi-	Asi-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	18	9	atics	atic	NNS
work_ezwbhq2nrvadncyjpwjlluuona	18	10	look	look	VBP
work_ezwbhq2nrvadncyjpwjlluuona	18	11	alike	alike	RB
work_ezwbhq2nrvadncyjpwjlluuona	18	12	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	18	13	while	while	IN
work_ezwbhq2nrvadncyjpwjlluuona	18	14	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	18	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	18	16	Asiatic	Asiatic	NNP
work_ezwbhq2nrvadncyjpwjlluuona	18	17	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	18	18	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	18	19	White	white	JJ
work_ezwbhq2nrvadncyjpwjlluuona	18	20	men	man	NNS
work_ezwbhq2nrvadncyjpwjlluuona	18	21	look	look	VBP
work_ezwbhq2nrvadncyjpwjlluuona	18	22	alike	alike	RB
work_ezwbhq2nrvadncyjpwjlluuona	18	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	19	1	More	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	19	2	recently	recently	RB
work_ezwbhq2nrvadncyjpwjlluuona	19	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	19	4	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	19	5	has	have	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	19	6	been	be	VBN
work_ezwbhq2nrvadncyjpwjlluuona	19	7	shown	show	VBN
work_ezwbhq2nrvadncyjpwjlluuona	19	8	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	19	9	approxi-	approxi-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	19	10	mately	mately	RB
work_ezwbhq2nrvadncyjpwjlluuona	19	11	half	half	NN
work_ezwbhq2nrvadncyjpwjlluuona	19	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	19	13	potential	potential	JJ
work_ezwbhq2nrvadncyjpwjlluuona	19	14	jurors	juror	NNS
work_ezwbhq2nrvadncyjpwjlluuona	19	15	believe	believe	VBP
work_ezwbhq2nrvadncyjpwjlluuona	19	16	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	19	17	such	such	PDT
work_ezwbhq2nrvadncyjpwjlluuona	19	18	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	19	19	bias	bias	NN
work_ezwbhq2nrvadncyjpwjlluuona	19	20	exists	exist	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	19	21	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	19	22	Deffenbacher	Deffenbacher	NNP
work_ezwbhq2nrvadncyjpwjlluuona	19	23	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	19	24	Loftus	Loftus	NNP
work_ezwbhq2nrvadncyjpwjlluuona	19	25	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	19	26	1982	1982	CD
work_ezwbhq2nrvadncyjpwjlluuona	19	27	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	19	28	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	20	1	Indeed	indeed	RB
work_ezwbhq2nrvadncyjpwjlluuona	20	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	20	3	abundant	abundant	JJ
work_ezwbhq2nrvadncyjpwjlluuona	20	4	empirical	empirical	JJ
work_ezwbhq2nrvadncyjpwjlluuona	20	5	support	support	NN
work_ezwbhq2nrvadncyjpwjlluuona	20	6	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	20	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	20	8	own	own	JJ
work_ezwbhq2nrvadncyjpwjlluuona	20	9	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	20	10	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	20	11	bias	bias	NN
work_ezwbhq2nrvadncyjpwjlluuona	20	12	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	20	13	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	20	14	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	20	15	accuracy	accuracy	NN
work_ezwbhq2nrvadncyjpwjlluuona	20	16	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	20	17	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	20	18	found	find	VBN
work_ezwbhq2nrvadncyjpwjlluuona	20	19	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	20	20	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	20	21	recent	recent	JJ
work_ezwbhq2nrvadncyjpwjlluuona	20	22	meta	meta	NN
work_ezwbhq2nrvadncyjpwjlluuona	20	23	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	20	24	analyses	analysis	NNS
work_ezwbhq2nrvadncyjpwjlluuona	20	25	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	20	26	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	20	27	large	large	JJ
work_ezwbhq2nrvadncyjpwjlluuona	20	28	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	20	29	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	20	30	studies	study	NNS
work_ezwbhq2nrvadncyjpwjlluuona	20	31	163	163	CD
work_ezwbhq2nrvadncyjpwjlluuona	20	32	164	164	CD
work_ezwbhq2nrvadncyjpwjlluuona	20	33	O’TOOLE	o’toole	CD
work_ezwbhq2nrvadncyjpwjlluuona	20	34	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	20	35	DEFFENBACHER	DEFFENBACHER	NNP
work_ezwbhq2nrvadncyjpwjlluuona	20	36	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	20	37	ABDI	ABDI	NNP
work_ezwbhq2nrvadncyjpwjlluuona	20	38	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	20	39	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	20	40	BARTLETT	BARTLETT	NNP
work_ezwbhq2nrvadncyjpwjlluuona	20	41	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	20	42	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	20	43	topic	topic	NN
work_ezwbhq2nrvadncyjpwjlluuona	20	44	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	20	45	Shapiro	Shapiro	NNP
work_ezwbhq2nrvadncyjpwjlluuona	20	46	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	20	47	Period	Period	NNP
work_ezwbhq2nrvadncyjpwjlluuona	20	48	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	20	49	1986	1986	CD
work_ezwbhq2nrvadncyjpwjlluuona	20	50	;	;	:
work_ezwbhq2nrvadncyjpwjlluuona	20	51	Bothwell	Bothwell	NNP
work_ezwbhq2nrvadncyjpwjlluuona	20	52	et	et	FW
work_ezwbhq2nrvadncyjpwjlluuona	20	53	al	al	NNP
work_ezwbhq2nrvadncyjpwjlluuona	20	54	.	.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	20	55	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	20	56	1989	1989	CD
work_ezwbhq2nrvadncyjpwjlluuona	20	57	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	20	58	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	21	1	Several	several	JJ
work_ezwbhq2nrvadncyjpwjlluuona	21	2	hypotheses	hypothesis	NNS
work_ezwbhq2nrvadncyjpwjlluuona	21	3	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	21	4	been	be	VBN
work_ezwbhq2nrvadncyjpwjlluuona	21	5	advanced	advance	VBN
work_ezwbhq2nrvadncyjpwjlluuona	21	6	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	21	7	ac-	ac-	RB
work_ezwbhq2nrvadncyjpwjlluuona	21	8	count	count	VB
work_ezwbhq2nrvadncyjpwjlluuona	21	9	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	21	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	21	11	cross	cross	JJ
work_ezwbhq2nrvadncyjpwjlluuona	21	12	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	21	13	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	21	14	phenomenon	phenomenon	NN
work_ezwbhq2nrvadncyjpwjlluuona	21	15	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	21	16	Faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	21	17	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	21	18	some	some	DT
work_ezwbhq2nrvadncyjpwjlluuona	21	19	races	race	NNS
work_ezwbhq2nrvadncyjpwjlluuona	21	20	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	21	21	inherently	inherently	RB
work_ezwbhq2nrvadncyjpwjlluuona	21	22	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	21	23	difficult	difficult	JJ
work_ezwbhq2nrvadncyjpwjlluuona	21	24	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	21	25	identify	identify	VB
work_ezwbhq2nrvadncyjpwjlluuona	21	26	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	21	27	others	other	NNS
work_ezwbhq2nrvadncyjpwjlluuona	21	28	;	;	:
work_ezwbhq2nrvadncyjpwjlluuona	21	29	prejudicial	prejudicial	JJ
work_ezwbhq2nrvadncyjpwjlluuona	21	30	attitudes	attitude	NNS
work_ezwbhq2nrvadncyjpwjlluuona	21	31	lead	lead	VBP
work_ezwbhq2nrvadncyjpwjlluuona	21	32	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	21	33	less	less	RBR
work_ezwbhq2nrvadncyjpwjlluuona	21	34	ac-	ac-	RB
work_ezwbhq2nrvadncyjpwjlluuona	21	35	curate	curate	JJ
work_ezwbhq2nrvadncyjpwjlluuona	21	36	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	21	37	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	21	38	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	21	39	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	21	40	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	21	41	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	21	42	;	;	:
work_ezwbhq2nrvadncyjpwjlluuona	21	43	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	21	44	other-	other-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	21	45	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	21	46	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	21	47	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	21	48	processed	process	VBN
work_ezwbhq2nrvadncyjpwjlluuona	21	49	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	21	50	superficially	superficially	RB
work_ezwbhq2nrvadncyjpwjlluuona	21	51	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	21	52	same	same	JJ
work_ezwbhq2nrvadncyjpwjlluuona	21	53	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	21	54	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	21	55	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	21	56	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	22	1	There	there	EX
work_ezwbhq2nrvadncyjpwjlluuona	22	2	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	22	3	little	little	JJ
work_ezwbhq2nrvadncyjpwjlluuona	22	4	support	support	NN
work_ezwbhq2nrvadncyjpwjlluuona	22	5	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	22	6	any	any	DT
work_ezwbhq2nrvadncyjpwjlluuona	22	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	22	8	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	22	9	accounts	account	NNS
work_ezwbhq2nrvadncyjpwjlluuona	22	10	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	22	11	Brigham	Brigham	NNP
work_ezwbhq2nrvadncyjpwjlluuona	22	12	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	22	13	1986	1986	CD
work_ezwbhq2nrvadncyjpwjlluuona	22	14	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	22	15	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	23	1	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	23	2	fourth	fourth	JJ
work_ezwbhq2nrvadncyjpwjlluuona	23	3	possi-	possi-	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	4	bility	bility	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	5	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	23	6	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	23	7	implied	imply	VBN
work_ezwbhq2nrvadncyjpwjlluuona	23	8	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	23	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	23	10	quote	quote	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	11	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	23	12	Feingold	Feingold	NNP
work_ezwbhq2nrvadncyjpwjlluuona	23	13	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	23	14	1914	1914	CD
work_ezwbhq2nrvadncyjpwjlluuona	23	15	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	23	16	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	23	17	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	23	18	own	own	JJ
work_ezwbhq2nrvadncyjpwjlluuona	23	19	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	23	20	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	21	bias	bias	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	22	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	23	23	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	23	24	direct	direct	JJ
work_ezwbhq2nrvadncyjpwjlluuona	23	25	proportion	proportion	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	26	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	23	27	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	23	28	difference	difference	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	29	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	23	30	amount	amount	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	31	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	23	32	contact	contact	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	33	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	23	34	persons	person	NNS
work_ezwbhq2nrvadncyjpwjlluuona	23	35	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	23	36	one	one	PRP
work_ezwbhq2nrvadncyjpwjlluuona	23	37	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	23	38	own	own	JJ
work_ezwbhq2nrvadncyjpwjlluuona	23	39	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	23	40	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	23	41	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	23	42	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	24	1	Several	several	JJ
work_ezwbhq2nrvadncyjpwjlluuona	24	2	studies	study	NNS
work_ezwbhq2nrvadncyjpwjlluuona	24	3	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	24	4	found	find	VBN
work_ezwbhq2nrvadncyjpwjlluuona	24	5	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	24	6	smaller	small	JJR
work_ezwbhq2nrvadncyjpwjlluuona	24	7	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	24	8	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	24	9	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	24	10	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	24	11	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	24	12	persons	person	NNS
work_ezwbhq2nrvadncyjpwjlluuona	24	13	living	live	VBG
work_ezwbhq2nrvadncyjpwjlluuona	24	14	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	24	15	more	more	JJR
work_ezwbhq2nrvadncyjpwjlluuona	24	16	racially	racially	RB
work_ezwbhq2nrvadncyjpwjlluuona	24	17	integrated	integrate	VBN
work_ezwbhq2nrvadncyjpwjlluuona	24	18	circumstances	circumstance	NNS
work_ezwbhq2nrvadncyjpwjlluuona	24	19	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	24	20	Cross	Cross	NNP
work_ezwbhq2nrvadncyjpwjlluuona	24	21	et	et	NNP
work_ezwbhq2nrvadncyjpwjlluuona	24	22	al	al	NNP
work_ezwbhq2nrvadncyjpwjlluuona	24	23	.	.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	24	24	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	24	25	1971	1971	CD
work_ezwbhq2nrvadncyjpwjlluuona	24	26	;	;	:
work_ezwbhq2nrvadncyjpwjlluuona	24	27	Feinman	Feinman	NNP
work_ezwbhq2nrvadncyjpwjlluuona	24	28	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	24	29	Entwistle	Entwistle	NNP
work_ezwbhq2nrvadncyjpwjlluuona	24	30	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	24	31	1976	1976	CD
work_ezwbhq2nrvadncyjpwjlluuona	24	32	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	24	33	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	25	1	However	however	RB
work_ezwbhq2nrvadncyjpwjlluuona	25	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	25	3	while	while	IN
work_ezwbhq2nrvadncyjpwjlluuona	25	4	at	at	RB
work_ezwbhq2nrvadncyjpwjlluuona	25	5	least	least	RBS
work_ezwbhq2nrvadncyjpwjlluuona	25	6	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	25	7	study	study	NN
work_ezwbhq2nrvadncyjpwjlluuona	25	8	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	25	9	Brigham	Brigham	NNP
work_ezwbhq2nrvadncyjpwjlluuona	25	10	et	et	FW
work_ezwbhq2nrvadncyjpwjlluuona	25	11	al	al	NNP
work_ezwbhq2nrvadncyjpwjlluuona	25	12	.	.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	25	13	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	25	14	1982	1982	CD
work_ezwbhq2nrvadncyjpwjlluuona	25	15	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	25	16	yielded	yield	VBD
work_ezwbhq2nrvadncyjpwjlluuona	25	17	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	25	18	small	small	JJ
work_ezwbhq2nrvadncyjpwjlluuona	25	19	but	but	CC
work_ezwbhq2nrvadncyjpwjlluuona	25	20	significant	significant	JJ
work_ezwbhq2nrvadncyjpwjlluuona	25	21	correlation	correlation	NN
work_ezwbhq2nrvadncyjpwjlluuona	25	22	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	25	23	self	self	NN
work_ezwbhq2nrvadncyjpwjlluuona	25	24	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	25	25	reported	report	VBN
work_ezwbhq2nrvadncyjpwjlluuona	25	26	degree	degree	NN
work_ezwbhq2nrvadncyjpwjlluuona	25	27	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	25	28	cross	cross	JJ
work_ezwbhq2nrvadncyjpwjlluuona	25	29	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	25	30	race	race	JJ
work_ezwbhq2nrvadncyjpwjlluuona	25	31	contact	contact	NN
work_ezwbhq2nrvadncyjpwjlluuona	25	32	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	25	33	cross	cross	JJ
work_ezwbhq2nrvadncyjpwjlluuona	25	34	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	25	35	race	race	JJ
work_ezwbhq2nrvadncyjpwjlluuona	25	36	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	25	37	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	25	38	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	25	39	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	25	40	studies	study	NNS
work_ezwbhq2nrvadncyjpwjlluuona	25	41	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	25	42	e.g.	e.g.	RB
work_ezwbhq2nrvadncyjpwjlluuona	26	1	Brigham	Brigham	NNP
work_ezwbhq2nrvadncyjpwjlluuona	26	2	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	26	3	Barkowitz	Barkowitz	NNP
work_ezwbhq2nrvadncyjpwjlluuona	26	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	26	5	1978	1978	CD
work_ezwbhq2nrvadncyjpwjlluuona	26	6	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	26	7	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	26	8	found	find	VBN
work_ezwbhq2nrvadncyjpwjlluuona	26	9	no	no	DT
work_ezwbhq2nrvadncyjpwjlluuona	26	10	rela-	rela-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	26	11	tionship	tionship	NN
work_ezwbhq2nrvadncyjpwjlluuona	26	12	at	at	RB
work_ezwbhq2nrvadncyjpwjlluuona	26	13	all	all	RB
work_ezwbhq2nrvadncyjpwjlluuona	26	14	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	27	1	It	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	27	2	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	27	3	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	27	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	27	5	however	however	RB
work_ezwbhq2nrvadncyjpwjlluuona	27	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	27	7	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	27	8	current	current	JJ
work_ezwbhq2nrvadncyjpwjlluuona	27	9	techniques	technique	NNS
work_ezwbhq2nrvadncyjpwjlluuona	27	10	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	27	11	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	27	12	sensitive	sensitive	JJ
work_ezwbhq2nrvadncyjpwjlluuona	27	13	enough	enough	RB
work_ezwbhq2nrvadncyjpwjlluuona	27	14	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	27	15	adequately	adequately	RB
work_ezwbhq2nrvadncyjpwjlluuona	27	16	assess	assess	VB
work_ezwbhq2nrvadncyjpwjlluuona	27	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	27	18	quan-	quan-	NN
work_ezwbhq2nrvadncyjpwjlluuona	27	19	tity	tity	NN
work_ezwbhq2nrvadncyjpwjlluuona	27	20	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	27	21	quality	quality	NN
work_ezwbhq2nrvadncyjpwjlluuona	27	22	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	27	23	contact	contact	NN
work_ezwbhq2nrvadncyjpwjlluuona	27	24	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	27	25	persons	person	NNS
work_ezwbhq2nrvadncyjpwjlluuona	27	26	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	27	27	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	27	28	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	27	29	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	28	1	There	there	EX
work_ezwbhq2nrvadncyjpwjlluuona	28	2	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	28	3	data	datum	NNS
work_ezwbhq2nrvadncyjpwjlluuona	28	4	which	which	WDT
work_ezwbhq2nrvadncyjpwjlluuona	28	5	suggest	suggest	VBP
work_ezwbhq2nrvadncyjpwjlluuona	28	6	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	28	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	28	8	cross-	cross-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	28	9	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	28	10	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	28	11	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	28	12	indeed	indeed	RB
work_ezwbhq2nrvadncyjpwjlluuona	28	13	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	28	14	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	28	15	matter	matter	NN
work_ezwbhq2nrvadncyjpwjlluuona	28	16	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	28	17	differen-	differen-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	28	18	tial	tial	JJ
work_ezwbhq2nrvadncyjpwjlluuona	28	19	exposure	exposure	NN
work_ezwbhq2nrvadncyjpwjlluuona	28	20	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	28	21	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	28	22	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	28	23	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	28	24	races	race	NNS
work_ezwbhq2nrvadncyjpwjlluuona	28	25	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	29	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	29	2	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	29	3	thing	thing	NN
work_ezwbhq2nrvadncyjpwjlluuona	29	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	29	5	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	29	6	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	29	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	29	8	attempts	attempt	NNS
work_ezwbhq2nrvadncyjpwjlluuona	29	9	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	29	10	improve	improve	VB
work_ezwbhq2nrvadncyjpwjlluuona	29	11	same	same	JJ
work_ezwbhq2nrvadncyjpwjlluuona	29	12	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	29	13	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	29	14	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	29	15	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	29	16	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	29	17	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	29	18	failed	fail	VBN
work_ezwbhq2nrvadncyjpwjlluuona	29	19	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	29	20	Malpass	Malpass	NNP
work_ezwbhq2nrvadncyjpwjlluuona	29	21	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	29	22	1981	1981	CD
work_ezwbhq2nrvadncyjpwjlluuona	29	23	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	29	24	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	30	1	However	however	RB
work_ezwbhq2nrvadncyjpwjlluuona	30	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	30	3	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	30	4	training	training	NN
work_ezwbhq2nrvadncyjpwjlluuona	30	5	efforts	effort	NNS
work_ezwbhq2nrvadncyjpwjlluuona	30	6	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	30	7	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	30	8	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	30	9	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	30	10	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	30	11	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	30	12	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	30	13	yielded	yield	VBN
work_ezwbhq2nrvadncyjpwjlluuona	30	14	improvements	improvement	NNS
work_ezwbhq2nrvadncyjpwjlluuona	30	15	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	30	16	e.g.	e.g.	RB
work_ezwbhq2nrvadncyjpwjlluuona	31	1	Goldstein	Goldstein	NNP
work_ezwbhq2nrvadncyjpwjlluuona	31	2	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	31	3	Chance	Chance	NNP
work_ezwbhq2nrvadncyjpwjlluuona	31	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	31	5	1985	1985	CD
work_ezwbhq2nrvadncyjpwjlluuona	31	6	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	31	7	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	32	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	32	2	think	think	VBP
work_ezwbhq2nrvadncyjpwjlluuona	32	3	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	32	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	32	5	differential	differential	JJ
work_ezwbhq2nrvadncyjpwjlluuona	32	6	effects	effect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	32	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	32	8	further	further	JJ
work_ezwbhq2nrvadncyjpwjlluuona	32	9	training	training	NN
work_ezwbhq2nrvadncyjpwjlluuona	32	10	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	32	11	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	32	12	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	32	13	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	32	14	due	due	JJ
work_ezwbhq2nrvadncyjpwjlluuona	32	15	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	32	16	differen-	differen-	NN
work_ezwbhq2nrvadncyjpwjlluuona	32	17	tial	tial	JJ
work_ezwbhq2nrvadncyjpwjlluuona	32	18	amounts	amount	NNS
work_ezwbhq2nrvadncyjpwjlluuona	32	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	32	20	perceptual	perceptual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	32	21	learning	learning	NN
work_ezwbhq2nrvadncyjpwjlluuona	32	22	associated	associate	VBN
work_ezwbhq2nrvadncyjpwjlluuona	32	23	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	32	24	same	same	JJ
work_ezwbhq2nrvadncyjpwjlluuona	32	25	versus	versus	IN
work_ezwbhq2nrvadncyjpwjlluuona	32	26	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	32	27	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	32	28	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	32	29	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	32	30	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	33	1	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	33	2	cross	cross	JJ
work_ezwbhq2nrvadncyjpwjlluuona	33	3	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	33	4	sectional	sectional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	33	5	study	study	NN
work_ezwbhq2nrvadncyjpwjlluuona	33	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	33	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	33	8	development	development	NN
work_ezwbhq2nrvadncyjpwjlluuona	33	9	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	33	10	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	33	11	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	33	12	abil-	abil-	XX
work_ezwbhq2nrvadncyjpwjlluuona	33	13	ity	ity	NN
work_ezwbhq2nrvadncyjpwjlluuona	33	14	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	33	15	Chance	Chance	NNP
work_ezwbhq2nrvadncyjpwjlluuona	33	16	et	et	NNP
work_ezwbhq2nrvadncyjpwjlluuona	33	17	al	al	NNP
work_ezwbhq2nrvadncyjpwjlluuona	33	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	34	1	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	34	2	1982	1982	CD
work_ezwbhq2nrvadncyjpwjlluuona	34	3	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	34	4	found	find	VBD
work_ezwbhq2nrvadncyjpwjlluuona	34	5	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	34	6	6	6	CD
work_ezwbhq2nrvadncyjpwjlluuona	34	7	year	year	NN
work_ezwbhq2nrvadncyjpwjlluuona	34	8	old	old	JJ
work_ezwbhq2nrvadncyjpwjlluuona	34	9	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	34	10	children	child	NNS
work_ezwbhq2nrvadncyjpwjlluuona	34	11	show	show	VBP
work_ezwbhq2nrvadncyjpwjlluuona	34	12	only	only	RB
work_ezwbhq2nrvadncyjpwjlluuona	34	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	34	14	small	small	JJ
work_ezwbhq2nrvadncyjpwjlluuona	34	15	cross	cross	JJ
work_ezwbhq2nrvadncyjpwjlluuona	34	16	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	34	17	race	race	JJ
work_ezwbhq2nrvadncyjpwjlluuona	34	18	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	34	19	recognizing	recognize	VBG
work_ezwbhq2nrvadncyjpwjlluuona	34	20	Japanese	Japanese	NNP
work_ezwbhq2nrvadncyjpwjlluuona	34	21	compared	compare	VBN
work_ezwbhq2nrvadncyjpwjlluuona	34	22	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	34	23	Cau-	Cau-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	34	24	casian	casian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	34	25	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	34	26	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	35	1	At	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	35	2	successively	successively	RB
work_ezwbhq2nrvadncyjpwjlluuona	35	3	older	old	JJR
work_ezwbhq2nrvadncyjpwjlluuona	35	4	ages	age	NNS
work_ezwbhq2nrvadncyjpwjlluuona	35	5	up	up	IN
work_ezwbhq2nrvadncyjpwjlluuona	35	6	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	35	7	early	early	JJ
work_ezwbhq2nrvadncyjpwjlluuona	35	8	adulthood	adulthood	NN
work_ezwbhq2nrvadncyjpwjlluuona	35	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	35	10	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	35	11	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	35	12	recognize	recognize	VB
work_ezwbhq2nrvadncyjpwjlluuona	35	13	both	both	DT
work_ezwbhq2nrvadncyjpwjlluuona	35	14	races	race	NNS
work_ezwbhq2nrvadncyjpwjlluuona	35	15	in-	in-	RB
work_ezwbhq2nrvadncyjpwjlluuona	35	16	creased	crease	VBD
work_ezwbhq2nrvadncyjpwjlluuona	35	17	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	35	18	but	but	CC
work_ezwbhq2nrvadncyjpwjlluuona	35	19	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	35	20	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	35	21	recognize	recognize	VB
work_ezwbhq2nrvadncyjpwjlluuona	35	22	same	same	JJ
work_ezwbhq2nrvadncyjpwjlluuona	35	23	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	35	24	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	35	25	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	35	26	Cau-	Cau-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	35	27	casian	casian	NNP
work_ezwbhq2nrvadncyjpwjlluuona	35	28	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	35	29	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	35	30	increased	increase	VBD
work_ezwbhq2nrvadncyjpwjlluuona	35	31	much	much	RB
work_ezwbhq2nrvadncyjpwjlluuona	35	32	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	35	33	rapidly	rapidly	RB
work_ezwbhq2nrvadncyjpwjlluuona	35	34	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	36	1	Hence	hence	RB
work_ezwbhq2nrvadncyjpwjlluuona	36	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	36	3	any	any	DT
work_ezwbhq2nrvadncyjpwjlluuona	36	4	attempt	attempt	NN
work_ezwbhq2nrvadncyjpwjlluuona	36	5	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	36	6	improve	improve	VB
work_ezwbhq2nrvadncyjpwjlluuona	36	7	same	same	JJ
work_ezwbhq2nrvadncyjpwjlluuona	36	8	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	36	9	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	36	10	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	36	11	recogni-	recogni-	XX
work_ezwbhq2nrvadncyjpwjlluuona	36	12	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	36	13	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	36	14	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	36	15	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	36	16	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	36	17	training	training	NN
work_ezwbhq2nrvadncyjpwjlluuona	36	18	programs	program	NNS
work_ezwbhq2nrvadncyjpwjlluuona	36	19	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	36	20	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	36	21	inade-	inade-	VBN
work_ezwbhq2nrvadncyjpwjlluuona	36	22	quate	quate	NN
work_ezwbhq2nrvadncyjpwjlluuona	36	23	compared	compare	VBN
work_ezwbhq2nrvadncyjpwjlluuona	36	24	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	36	25	years	year	NNS
work_ezwbhq2nrvadncyjpwjlluuona	36	26	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	36	27	extensive	extensive	JJ
work_ezwbhq2nrvadncyjpwjlluuona	36	28	processing	processing	NN
work_ezwbhq2nrvadncyjpwjlluuona	36	29	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	36	30	same	same	JJ
work_ezwbhq2nrvadncyjpwjlluuona	36	31	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	36	32	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	36	33	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	36	34	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	37	1	Studies	study	NNS
work_ezwbhq2nrvadncyjpwjlluuona	37	2	examining	examine	VBG
work_ezwbhq2nrvadncyjpwjlluuona	37	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	37	4	role	role	NN
work_ezwbhq2nrvadncyjpwjlluuona	37	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	37	6	perceptual	perceptual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	37	7	learn-	learn-	FW
work_ezwbhq2nrvadncyjpwjlluuona	37	8	ing1	ing1	FW
work_ezwbhq2nrvadncyjpwjlluuona	37	9	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	37	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	37	11	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	37	12	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	37	13	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	37	14	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	37	15	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	37	16	difficult	difficult	JJ
work_ezwbhq2nrvadncyjpwjlluuona	37	17	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	37	18	carry	carry	VB
work_ezwbhq2nrvadncyjpwjlluuona	37	19	out	out	RP
work_ezwbhq2nrvadncyjpwjlluuona	37	20	empirically	empirically	RB
work_ezwbhq2nrvadncyjpwjlluuona	37	21	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	37	22	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	37	23	reasons	reason	NNS
work_ezwbhq2nrvadncyjpwjlluuona	37	24	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	38	1	First	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	38	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	38	3	while	while	IN
work_ezwbhq2nrvadncyjpwjlluuona	38	4	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	38	5	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	38	6	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	38	7	possible	possible	JJ
work_ezwbhq2nrvadncyjpwjlluuona	38	8	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	38	9	find	find	VB
work_ezwbhq2nrvadncyjpwjlluuona	38	10	subject	subject	JJ
work_ezwbhq2nrvadncyjpwjlluuona	38	11	populations	population	NNS
work_ezwbhq2nrvadncyjpwjlluuona	38	12	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	38	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	38	14	relatively	relatively	RB
work_ezwbhq2nrvadncyjpwjlluuona	38	15	controlled	control	VBN
work_ezwbhq2nrvadncyjpwjlluuona	38	16	“	"	``
work_ezwbhq2nrvadncyjpwjlluuona	38	17	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	38	18	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	38	19	learning	learn	VBG
work_ezwbhq2nrvadncyjpwjlluuona	38	20	”	"	''
work_ezwbhq2nrvadncyjpwjlluuona	38	21	history	history	NN
work_ezwbhq2nrvadncyjpwjlluuona	38	22	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	38	23	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	38	24	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	38	25	generally	generally	RB
work_ezwbhq2nrvadncyjpwjlluuona	38	26	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	38	27	possible	possible	JJ
work_ezwbhq2nrvadncyjpwjlluuona	38	28	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	38	29	equate	equate	VB
work_ezwbhq2nrvadncyjpwjlluuona	38	30	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	38	31	populations	population	NNS
work_ezwbhq2nrvadncyjpwjlluuona	38	32	along	along	IN
work_ezwbhq2nrvadncyjpwjlluuona	38	33	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	38	34	important	important	JJ
work_ezwbhq2nrvadncyjpwjlluuona	38	35	cultural	cultural	JJ
work_ezwbhq2nrvadncyjpwjlluuona	38	36	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	38	37	social	social	JJ
work_ezwbhq2nrvadncyjpwjlluuona	38	38	dimen-	dimen-	NN
work_ezwbhq2nrvadncyjpwjlluuona	38	39	sions	sion	NNS
work_ezwbhq2nrvadncyjpwjlluuona	38	40	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	38	41	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	38	42	affect	affect	VB
work_ezwbhq2nrvadncyjpwjlluuona	38	43	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	38	44	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	38	45	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	38	46	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	38	47	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	39	1	Sec-	sec-	VB
work_ezwbhq2nrvadncyjpwjlluuona	39	2	ondly	ondly	RB
work_ezwbhq2nrvadncyjpwjlluuona	39	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	39	4	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	39	5	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	39	6	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	39	7	already	already	RB
work_ezwbhq2nrvadncyjpwjlluuona	39	8	noted	note	VBN
work_ezwbhq2nrvadncyjpwjlluuona	39	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	39	10	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	39	11	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	39	12	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	39	13	per-	per-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	39	14	ceptual	ceptual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	39	15	learning	learning	NN
work_ezwbhq2nrvadncyjpwjlluuona	39	16	studies	study	NNS
work_ezwbhq2nrvadncyjpwjlluuona	39	17	involving	involve	VBG
work_ezwbhq2nrvadncyjpwjlluuona	39	18	practice	practice	NN
work_ezwbhq2nrvadncyjpwjlluuona	39	19	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	39	20	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	39	21	single	single	JJ
work_ezwbhq2nrvadncyjpwjlluuona	39	22	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	39	23	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	39	24	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	39	25	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	39	26	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	39	27	necessarily	necessarily	RB
work_ezwbhq2nrvadncyjpwjlluuona	39	28	adequate	adequate	JJ
work_ezwbhq2nrvadncyjpwjlluuona	39	29	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	39	30	control	control	VB
work_ezwbhq2nrvadncyjpwjlluuona	39	31	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	39	32	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	39	33	lifetime	lifetime	NN
work_ezwbhq2nrvadncyjpwjlluuona	39	34	experience	experience	NN
work_ezwbhq2nrvadncyjpwjlluuona	39	35	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	39	36	observers	observer	NNS
work_ezwbhq2nrvadncyjpwjlluuona	39	37	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	39	38	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	39	39	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	39	40	their	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	39	41	own	own	JJ
work_ezwbhq2nrvadncyjpwjlluuona	39	42	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	39	43	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	40	1	These	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	40	2	methodological	methodological	JJ
work_ezwbhq2nrvadncyjpwjlluuona	40	3	difficulties	difficulty	NNS
work_ezwbhq2nrvadncyjpwjlluuona	40	4	make	make	VBP
work_ezwbhq2nrvadncyjpwjlluuona	40	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	40	6	cross	cross	JJ
work_ezwbhq2nrvadncyjpwjlluuona	40	7	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	40	8	race	race	JJ
work_ezwbhq2nrvadncyjpwjlluuona	40	9	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	40	10	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	40	11	ideal	ideal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	40	12	can-	can-	NN
work_ezwbhq2nrvadncyjpwjlluuona	40	13	didate	didate	NN
work_ezwbhq2nrvadncyjpwjlluuona	40	14	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	40	15	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	40	16	approaches	approach	NNS
work_ezwbhq2nrvadncyjpwjlluuona	40	17	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	40	18	understanding	understand	VBG
work_ezwbhq2nrvadncyjpwjlluuona	40	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	40	20	psychological	psychological	JJ
work_ezwbhq2nrvadncyjpwjlluuona	40	21	data	datum	NNS
work_ezwbhq2nrvadncyjpwjlluuona	40	22	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	41	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	41	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	41	3	present	present	JJ
work_ezwbhq2nrvadncyjpwjlluuona	41	4	study	study	NN
work_ezwbhq2nrvadncyjpwjlluuona	41	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	41	6	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	41	7	present	present	VBP
work_ezwbhq2nrvadncyjpwjlluuona	41	8	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	41	9	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	41	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	41	11	perceptual	perceptual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	41	12	learning	learning	NN
work_ezwbhq2nrvadncyjpwjlluuona	41	13	account	account	NN
work_ezwbhq2nrvadncyjpwjlluuona	41	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	41	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	41	16	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	41	17	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	41	18	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	41	19	ef-	ef-	XX
work_ezwbhq2nrvadncyjpwjlluuona	41	20	fect	fect	JJ
work_ezwbhq2nrvadncyjpwjlluuona	41	21	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	41	22	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	41	23	based	base	VBN
work_ezwbhq2nrvadncyjpwjlluuona	41	24	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	41	25	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	41	26	following	follow	VBG
work_ezwbhq2nrvadncyjpwjlluuona	41	27	principles	principle	NNS
work_ezwbhq2nrvadncyjpwjlluuona	41	28	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	42	1	First	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	42	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	42	3	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	42	4	assume	assume	VBP
work_ezwbhq2nrvadncyjpwjlluuona	42	5	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	42	6	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	42	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	42	8	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	42	9	races	race	NNS
work_ezwbhq2nrvadncyjpwjlluuona	42	10	comprise	comprise	VBP
work_ezwbhq2nrvadncyjpwjlluuona	42	11	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	42	12	statistical	statistical	JJ
work_ezwbhq2nrvadncyjpwjlluuona	42	13	categories	category	NNS
work_ezwbhq2nrvadncyjpwjlluuona	42	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	42	15	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	42	16	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	43	1	Secondly	secondly	RB
work_ezwbhq2nrvadncyjpwjlluuona	43	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	43	3	within	within	IN
work_ezwbhq2nrvadncyjpwjlluuona	43	4	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	43	5	given	give	VBN
work_ezwbhq2nrvadncyjpwjlluuona	43	6	category	category	NN
work_ezwbhq2nrvadncyjpwjlluuona	43	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	43	8	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	43	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	43	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	43	11	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	43	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	43	13	differen-	differen-	NN
work_ezwbhq2nrvadncyjpwjlluuona	43	14	tially	tially	NNP
work_ezwbhq2nrvadncyjpwjlluuona	43	15	weighted	weight	VBD
work_ezwbhq2nrvadncyjpwjlluuona	43	16	“	"	``
work_ezwbhq2nrvadncyjpwjlluuona	43	17	features2	features2	NN
work_ezwbhq2nrvadncyjpwjlluuona	43	18	”	"	''
work_ezwbhq2nrvadncyjpwjlluuona	43	19	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	43	20	optimal	optimal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	43	21	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	43	22	encoding	encode	VBG
work_ezwbhq2nrvadncyjpwjlluuona	43	23	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	43	24	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	43	25	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	43	26	manner	manner	NN
work_ezwbhq2nrvadncyjpwjlluuona	43	27	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	43	28	makes	make	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	43	29	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	43	30	within	within	IN
work_ezwbhq2nrvadncyjpwjlluuona	43	31	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	43	32	cate-	cate-	NN
work_ezwbhq2nrvadncyjpwjlluuona	43	33	gory	gory	NN
work_ezwbhq2nrvadncyjpwjlluuona	43	34	most	most	RBS
work_ezwbhq2nrvadncyjpwjlluuona	43	35	discriminable	discriminable	JJ
work_ezwbhq2nrvadncyjpwjlluuona	43	36	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	44	1	Different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	44	2	feature	feature	NN
work_ezwbhq2nrvadncyjpwjlluuona	44	3	sets	set	NNS
work_ezwbhq2nrvadncyjpwjlluuona	44	4	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	44	5	weightings	weighting	NNS
work_ezwbhq2nrvadncyjpwjlluuona	44	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	44	7	however	however	RB
work_ezwbhq2nrvadncyjpwjlluuona	44	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	44	9	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	44	10	optimal	optimal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	44	11	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	44	12	processing	processing	NN
work_ezwbhq2nrvadncyjpwjlluuona	44	13	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	44	14	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	44	15	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	44	16	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	44	17	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	44	18	categories	category	NNS
work_ezwbhq2nrvadncyjpwjlluuona	44	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	44	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	44	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	45	1	Thirdly	thirdly	RB
work_ezwbhq2nrvadncyjpwjlluuona	45	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	45	3	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	4	exposure	exposure	NN
work_ezwbhq2nrvadncyjpwjlluuona	45	5	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	6	many	many	JJ
work_ezwbhq2nrvadncyjpwjlluuona	45	7	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	45	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	9	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	45	10	given	give	VBN
work_ezwbhq2nrvadncyjpwjlluuona	45	11	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	45	12	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	45	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	45	14	smaller	small	JJR
work_ezwbhq2nrvadncyjpwjlluuona	45	15	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	45	16	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	17	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	45	18	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	19	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	45	20	races	race	NNS
work_ezwbhq2nrvadncyjpwjlluuona	45	21	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	45	22	perceptual	perceptual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	45	23	learning	learning	NN
work_ezwbhq2nrvadncyjpwjlluuona	45	24	enables	enable	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	45	25	observers	observer	NNS
work_ezwbhq2nrvadncyjpwjlluuona	45	26	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	45	27	make	make	VB
work_ezwbhq2nrvadncyjpwjlluuona	45	28	optimal	optimal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	45	29	use	use	NN
work_ezwbhq2nrvadncyjpwjlluuona	45	30	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	31	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	45	32	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	45	33	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	45	34	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	45	35	best	good	JJS
work_ezwbhq2nrvadncyjpwjlluuona	45	36	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	37	processing	process	VBG
work_ezwbhq2nrvadncyjpwjlluuona	45	38	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	45	39	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	40	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	45	41	category	category	NN
work_ezwbhq2nrvadncyjpwjlluuona	45	42	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	43	which	which	WDT
work_ezwbhq2nrvadncyjpwjlluuona	45	44	they	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	45	45	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	45	46	had	have	VBN
work_ezwbhq2nrvadncyjpwjlluuona	45	47	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	45	48	most	most	JJS
work_ezwbhq2nrvadncyjpwjlluuona	45	49	experience	experience	NN
work_ezwbhq2nrvadncyjpwjlluuona	45	50	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	45	51	topically	topically	RB
work_ezwbhq2nrvadncyjpwjlluuona	45	52	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	45	53	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	45	54	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	45	55	their	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	45	56	own	own	JJ
work_ezwbhq2nrvadncyjpwjlluuona	45	57	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	45	58	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	46	1	By	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	46	2	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	46	3	account	account	NN
work_ezwbhq2nrvadncyjpwjlluuona	46	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	46	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	46	6	difficulties	difficulty	NNS
work_ezwbhq2nrvadncyjpwjlluuona	46	7	experienced	experience	VBD
work_ezwbhq2nrvadncyjpwjlluuona	46	8	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	46	9	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	46	10	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	46	11	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	46	12	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	46	13	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	46	14	due	due	JJ
work_ezwbhq2nrvadncyjpwjlluuona	46	15	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	46	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	46	17	fact	fact	NN
work_ezwbhq2nrvadncyjpwjlluuona	46	18	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	46	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	46	20	optimal	optimal	NN
work_ezwbhq2nrvadncyjpwjlluuona	46	21	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	46	22	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	46	23	distinguishing	distinguish	VBG
work_ezwbhq2nrvadncyjpwjlluuona	46	24	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	46	25	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	46	26	one	one	PRP
work_ezwbhq2nrvadncyjpwjlluuona	46	27	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	46	28	own	own	JJ
work_ezwbhq2nrvadncyjpwjlluuona	46	29	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	46	30	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	46	31	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	46	32	optimal	optimal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	46	33	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	46	34	processing	process	VBG
work_ezwbhq2nrvadncyjpwjlluuona	46	35	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	46	36	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	46	37	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	46	38	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	46	39	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	46	40	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	47	1	One	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	47	2	way	way	NN
work_ezwbhq2nrvadncyjpwjlluuona	47	3	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	47	4	simulate	simulate	VB
work_ezwbhq2nrvadncyjpwjlluuona	47	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	47	6	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	47	7	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	47	8	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	47	9	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	47	10	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	47	11	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	47	12	train	train	VB
work_ezwbhq2nrvadncyjpwjlluuona	47	13	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	47	14	autoassociative	autoassociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	47	15	network	network	NN
work_ezwbhq2nrvadncyjpwjlluuona	47	16	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	47	17	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	47	18	pro-	pro-	NN
work_ezwbhq2nrvadncyjpwjlluuona	47	19	portions	portion	NNS
work_ezwbhq2nrvadncyjpwjlluuona	47	20	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	47	21	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	47	22	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	47	23	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	47	24	’	'	``
work_ezwbhq2nrvadncyjpwjlluuona	47	25	own	own	JJ
work_ezwbhq2nrvadncyjpwjlluuona	47	26	’	'	''
work_ezwbhq2nrvadncyjpwjlluuona	47	27	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	47	28	’	'	''
work_ezwbhq2nrvadncyjpwjlluuona	47	29	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	47	30	’	'	''
work_ezwbhq2nrvadncyjpwjlluuona	47	31	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	47	32	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	48	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	48	2	1	1	CD
work_ezwbhq2nrvadncyjpwjlluuona	48	3	Referred	refer	VBN
work_ezwbhq2nrvadncyjpwjlluuona	48	4	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	48	5	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	48	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	48	7	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	48	8	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	48	9	literature	literature	NN
work_ezwbhq2nrvadncyjpwjlluuona	48	10	simply	simply	RB
work_ezwbhq2nrvadncyjpwjlluuona	48	11	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	48	12	“	"	``
work_ezwbhq2nrvadncyjpwjlluuona	48	13	experience	experience	NN
work_ezwbhq2nrvadncyjpwjlluuona	48	14	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	48	15	”	"	''
work_ezwbhq2nrvadncyjpwjlluuona	48	16	2	2	CD
work_ezwbhq2nrvadncyjpwjlluuona	48	17	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	48	18	word	word	NN
work_ezwbhq2nrvadncyjpwjlluuona	48	19	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	48	20	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	48	21	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	48	22	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	48	23	its	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	48	24	most	most	RBS
work_ezwbhq2nrvadncyjpwjlluuona	48	25	general	general	JJ
work_ezwbhq2nrvadncyjpwjlluuona	48	26	sense	sense	NN
work_ezwbhq2nrvadncyjpwjlluuona	48	27	without	without	IN
work_ezwbhq2nrvadncyjpwjlluuona	48	28	commitment	commitment	NN
work_ezwbhq2nrvadncyjpwjlluuona	48	29	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	48	30	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	48	31	specific	specific	JJ
work_ezwbhq2nrvadncyjpwjlluuona	48	32	definition	definition	NN
work_ezwbhq2nrvadncyjpwjlluuona	48	33	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	49	1	SIMULATING	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	49	2	THE	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	49	3	OTHER	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	49	4	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	49	5	RACE	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	6	EFFECT	EFFECT	NNP
work_ezwbhq2nrvadncyjpwjlluuona	49	7	165	165	CD
work_ezwbhq2nrvadncyjpwjlluuona	49	8	trained	train	VBD
work_ezwbhq2nrvadncyjpwjlluuona	49	9	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	49	10	autoassociative	autoassociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	49	11	system	system	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	12	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	49	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	49	14	large	large	JJ
work_ezwbhq2nrvadncyjpwjlluuona	49	15	num-	num-	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	16	ber	ber	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	17	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	49	18	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	49	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	49	20	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	49	21	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	22	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	23	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	49	24	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	49	25	smaller	small	JJR
work_ezwbhq2nrvadncyjpwjlluuona	49	26	num-	num-	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	27	ber	ber	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	28	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	49	29	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	49	30	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	49	31	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	49	32	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	33	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	34	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	49	35	mimic	mimic	VB
work_ezwbhq2nrvadncyjpwjlluuona	49	36	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	49	37	other-	other-	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	38	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	39	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	49	40	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	50	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	50	2	advantages	advantage	NNS
work_ezwbhq2nrvadncyjpwjlluuona	50	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	50	4	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	50	5	autoassociative	autoassociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	50	6	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	50	7	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	50	8	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	50	9	Widrow	Widrow	NNP
work_ezwbhq2nrvadncyjpwjlluuona	50	10	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	50	11	Hoff	Hoff	NNP
work_ezwbhq2nrvadncyjpwjlluuona	50	12	error	error	NN
work_ezwbhq2nrvadncyjpwjlluuona	50	13	correction	correction	NN
work_ezwbhq2nrvadncyjpwjlluuona	50	14	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	50	15	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	50	16	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	50	17	will	will	MD
work_ezwbhq2nrvadncyjpwjlluuona	50	18	develop	develop	VB
work_ezwbhq2nrvadncyjpwjlluuona	50	19	connection	connection	NN
work_ezwbhq2nrvadncyjpwjlluuona	50	20	weights	weight	NNS
work_ezwbhq2nrvadncyjpwjlluuona	50	21	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	50	22	such	such	PDT
work_ezwbhq2nrvadncyjpwjlluuona	50	23	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	50	24	way	way	NN
work_ezwbhq2nrvadncyjpwjlluuona	50	25	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	50	26	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	50	27	optimize	optimize	VB
work_ezwbhq2nrvadncyjpwjlluuona	50	28	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	50	29	storage	storage	NN
work_ezwbhq2nrvadncyjpwjlluuona	50	30	capacity	capacity	NN
work_ezwbhq2nrvadncyjpwjlluuona	50	31	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	50	32	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	50	33	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	50	34	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	51	1	Thus	thus	RB
work_ezwbhq2nrvadncyjpwjlluuona	51	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	51	3	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	51	4	very	very	RB
work_ezwbhq2nrvadncyjpwjlluuona	51	5	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	51	6	stimuli	stimulus	NNS
work_ezwbhq2nrvadncyjpwjlluuona	51	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	51	8	such	such	JJ
work_ezwbhq2nrvadncyjpwjlluuona	51	9	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	51	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	51	11	single	single	JJ
work_ezwbhq2nrvadncyjpwjlluuona	51	12	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	51	13	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	51	14	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	51	15	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	51	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	51	17	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	51	18	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	51	19	tune	tune	VB
work_ezwbhq2nrvadncyjpwjlluuona	51	20	itself	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	51	21	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	51	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	51	23	information	information	NN
work_ezwbhq2nrvadncyjpwjlluuona	51	24	important	important	JJ
work_ezwbhq2nrvadncyjpwjlluuona	51	25	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	51	26	processing	processing	NN
work_ezwbhq2nrvadncyjpwjlluuona	51	27	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	51	28	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	51	29	within	within	IN
work_ezwbhq2nrvadncyjpwjlluuona	51	30	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	51	31	class	class	NN
work_ezwbhq2nrvadncyjpwjlluuona	51	32	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	52	1	Due	due	IN
work_ezwbhq2nrvadncyjpwjlluuona	52	2	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	52	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	52	4	distributed	distribute	VBN
work_ezwbhq2nrvadncyjpwjlluuona	52	5	nature	nature	NN
work_ezwbhq2nrvadncyjpwjlluuona	52	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	52	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	52	8	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	52	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	52	10	when	when	WRB
work_ezwbhq2nrvadncyjpwjlluuona	52	11	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	52	12	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	52	13	retrieved	retrieve	VBN
work_ezwbhq2nrvadncyjpwjlluuona	52	14	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	52	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	52	16	system	system	NN
work_ezwbhq2nrvadncyjpwjlluuona	52	17	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	52	18	they	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	52	19	will	will	MD
work_ezwbhq2nrvadncyjpwjlluuona	52	20	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	52	21	filtered	filter	VBN
work_ezwbhq2nrvadncyjpwjlluuona	52	22	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	52	23	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	52	24	learning	learn	VBG
work_ezwbhq2nrvadncyjpwjlluuona	52	25	history	history	NN
work_ezwbhq2nrvadncyjpwjlluuona	52	26	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	52	27	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	52	28	system	system	NN
work_ezwbhq2nrvadncyjpwjlluuona	52	29	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	53	1	Several	several	JJ
work_ezwbhq2nrvadncyjpwjlluuona	53	2	predictions	prediction	NNS
work_ezwbhq2nrvadncyjpwjlluuona	53	3	about	about	IN
work_ezwbhq2nrvadncyjpwjlluuona	53	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	53	5	system	system	NN
work_ezwbhq2nrvadncyjpwjlluuona	53	6	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	53	7	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	53	8	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	53	9	process	process	VB
work_ezwbhq2nrvadncyjpwjlluuona	53	10	same-	same-	NN
work_ezwbhq2nrvadncyjpwjlluuona	53	11	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	53	12	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	53	13	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	53	14	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	53	15	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	53	16	follow	follow	VBP
work_ezwbhq2nrvadncyjpwjlluuona	53	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	54	1	First	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	54	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	54	3	when	when	WRB
work_ezwbhq2nrvadncyjpwjlluuona	54	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	54	5	network	network	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	6	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	54	7	trained	train	VBN
work_ezwbhq2nrvadncyjpwjlluuona	54	8	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	54	9	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	54	10	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	11	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	54	12	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	54	13	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	54	14	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	54	15	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	16	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	54	17	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	54	18	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	54	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	54	21	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	54	22	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	54	23	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	24	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	54	25	its	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	54	26	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	27	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	54	28	represent	represent	VB
work_ezwbhq2nrvadncyjpwjlluuona	54	29	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	54	30	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	54	31	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	54	32	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	33	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	34	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	54	35	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	54	36	better	well	JJR
work_ezwbhq2nrvadncyjpwjlluuona	54	37	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	54	38	its	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	54	39	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	40	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	54	41	represent	represent	VB
work_ezwbhq2nrvadncyjpwjlluuona	54	42	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	54	43	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	54	44	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	54	45	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	46	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	54	47	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	55	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	55	2	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	55	3	due	due	JJ
work_ezwbhq2nrvadncyjpwjlluuona	55	4	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	55	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	55	6	fact	fact	NN
work_ezwbhq2nrvadncyjpwjlluuona	55	7	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	55	8	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	55	9	will	will	MD
work_ezwbhq2nrvadncyjpwjlluuona	55	10	have	have	VB
work_ezwbhq2nrvadncyjpwjlluuona	55	11	developed	develop	VBN
work_ezwbhq2nrvadncyjpwjlluuona	55	12	“	"	``
work_ezwbhq2nrvadncyjpwjlluuona	55	13	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	55	14	”	"	''
work_ezwbhq2nrvadncyjpwjlluuona	55	15	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	55	16	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	55	17	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	55	18	appropriate	appropriate	JJ
work_ezwbhq2nrvadncyjpwjlluuona	55	19	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	55	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	55	21	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	55	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	55	23	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	55	24	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	55	25	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	56	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	56	2	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	56	3	assess	assess	VB
work_ezwbhq2nrvadncyjpwjlluuona	56	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	56	5	validity	validity	NN
work_ezwbhq2nrvadncyjpwjlluuona	56	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	56	7	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	56	8	prediction	prediction	NN
work_ezwbhq2nrvadncyjpwjlluuona	56	9	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	56	10	looking	look	VBG
work_ezwbhq2nrvadncyjpwjlluuona	56	11	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	56	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	56	13	quality	quality	NN
work_ezwbhq2nrvadncyjpwjlluuona	56	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	56	15	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	56	16	reconstructions	reconstruction	NNS
work_ezwbhq2nrvadncyjpwjlluuona	56	17	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	56	18	new	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	56	19	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	56	20	pre-	pre-	RB
work_ezwbhq2nrvadncyjpwjlluuona	56	21	viously	viously	RB
work_ezwbhq2nrvadncyjpwjlluuona	56	22	unencountered	unencountered	JJ
work_ezwbhq2nrvadncyjpwjlluuona	56	23	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	56	24	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	56	25	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	56	26	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	56	27	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	56	28	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	56	29	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	56	30	races	race	NNS
work_ezwbhq2nrvadncyjpwjlluuona	56	31	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	57	1	Secondly	secondly	RB
work_ezwbhq2nrvadncyjpwjlluuona	57	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	57	3	reconstructed	reconstruct	VBD
work_ezwbhq2nrvadncyjpwjlluuona	57	4	new	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	57	5	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	57	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	57	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	57	8	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	57	9	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	57	10	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	57	11	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	57	12	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	57	13	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	57	14	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	57	15	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	57	16	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	57	17	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	57	18	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	57	19	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	57	20	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	57	21	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	57	22	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	57	23	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	57	24	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	58	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	58	2	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	58	3	words	word	NNS
work_ezwbhq2nrvadncyjpwjlluuona	58	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	58	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	58	6	average	average	JJ
work_ezwbhq2nrvadncyjpwjlluuona	58	7	inter	inter	JJ
work_ezwbhq2nrvadncyjpwjlluuona	58	8	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	58	9	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	58	10	sim-	sim-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	58	11	ilarity	ilarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	58	12	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	58	13	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	58	14	greater	great	JJR
work_ezwbhq2nrvadncyjpwjlluuona	58	15	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	58	16	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	58	17	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	58	18	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	58	19	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	58	20	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	58	21	reconstructed	reconstructed	JJ
work_ezwbhq2nrvadncyjpwjlluuona	58	22	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	58	23	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	58	24	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	59	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	59	2	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	59	3	because	because	IN
work_ezwbhq2nrvadncyjpwjlluuona	59	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	59	5	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	59	6	will	will	MD
work_ezwbhq2nrvadncyjpwjlluuona	59	7	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	59	8	develop	develop	VB
work_ezwbhq2nrvadncyjpwjlluuona	59	9	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	59	10	coding	coding	NN
work_ezwbhq2nrvadncyjpwjlluuona	59	11	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	59	12	makes	make	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	59	13	optimal	optimal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	59	14	use	use	NN
work_ezwbhq2nrvadncyjpwjlluuona	59	15	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	59	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	59	17	distinguishing	distinguishing	NN
work_ezwbhq2nrvadncyjpwjlluuona	59	18	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	59	19	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	59	20	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	59	21	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	59	22	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	59	23	;	;	:
work_ezwbhq2nrvadncyjpwjlluuona	59	24	hence	hence	RB
work_ezwbhq2nrvadncyjpwjlluuona	59	25	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	59	26	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	59	27	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	59	28	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	59	29	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	59	30	“	"	``
work_ezwbhq2nrvadncyjpwjlluuona	59	31	perceptually	perceptually	RB
work_ezwbhq2nrvadncyjpwjlluuona	59	32	”	"	''
work_ezwbhq2nrvadncyjpwjlluuona	59	33	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	59	34	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	59	35	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	59	36	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	59	37	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	59	38	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	60	1	Fi-	fi-	IN
work_ezwbhq2nrvadncyjpwjlluuona	60	2	nally	nally	RB
work_ezwbhq2nrvadncyjpwjlluuona	60	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	60	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	60	5	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	60	6	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	60	7	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	60	8	better	well	RBR
work_ezwbhq2nrvadncyjpwjlluuona	60	9	able	able	JJ
work_ezwbhq2nrvadncyjpwjlluuona	60	10	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	60	11	recognize	recognize	VB
work_ezwbhq2nrvadncyjpwjlluuona	60	12	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	60	13	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	60	14	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	60	15	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	60	16	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	60	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	61	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	61	2	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	61	3	serve	serve	VBP
work_ezwbhq2nrvadncyjpwjlluuona	61	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	61	5	first	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	61	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	61	7	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	61	8	test	test	VB
work_ezwbhq2nrvadncyjpwjlluuona	61	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	61	10	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	61	11	qualitatively	qualitatively	RB
work_ezwbhq2nrvadncyjpwjlluuona	61	12	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	61	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	61	14	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	61	15	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	61	16	tool	tool	NN
work_ezwbhq2nrvadncyjpwjlluuona	61	17	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	61	18	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	61	19	much	much	RB
work_ezwbhq2nrvadncyjpwjlluuona	61	20	larger	large	JJR
work_ezwbhq2nrvadncyjpwjlluuona	61	21	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	61	22	higher	high	JJR
work_ezwbhq2nrvadncyjpwjlluuona	61	23	quality	quality	NN
work_ezwbhq2nrvadncyjpwjlluuona	61	24	stimulus	stimulus	NN
work_ezwbhq2nrvadncyjpwjlluuona	61	25	set	set	VBN
work_ezwbhq2nrvadncyjpwjlluuona	61	26	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	61	27	that	that	DT
work_ezwbhq2nrvadncyjpwjlluuona	61	28	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	61	29	previously	previously	RB
work_ezwbhq2nrvadncyjpwjlluuona	61	30	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	61	31	Kohonen	Kohonen	NNP
work_ezwbhq2nrvadncyjpwjlluuona	61	32	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	61	33	1984	1984	CD
work_ezwbhq2nrvadncyjpwjlluuona	61	34	;	;	:
work_ezwbhq2nrvadncyjpwjlluuona	61	35	O’Toole	O’Toole	NNP
work_ezwbhq2nrvadncyjpwjlluuona	61	36	et	et	NNP
work_ezwbhq2nrvadncyjpwjlluuona	61	37	al	al	NNP
work_ezwbhq2nrvadncyjpwjlluuona	61	38	.	.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	61	39	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	61	40	1988	1988	CD
work_ezwbhq2nrvadncyjpwjlluuona	61	41	;	;	:
work_ezwbhq2nrvadncyjpwjlluuona	61	42	O’Toole	O’Toole	NNP
work_ezwbhq2nrvadncyjpwjlluuona	61	43	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	61	44	Abdi	Abdi	NNP
work_ezwbhq2nrvadncyjpwjlluuona	61	45	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	61	46	1989	1989	CD
work_ezwbhq2nrvadncyjpwjlluuona	61	47	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	61	48	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	62	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	62	2	will	will	MD
work_ezwbhq2nrvadncyjpwjlluuona	62	3	look	look	VB
work_ezwbhq2nrvadncyjpwjlluuona	62	4	specifically	specifically	RB
work_ezwbhq2nrvadncyjpwjlluuona	62	5	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	62	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	62	7	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	62	8	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	62	9	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	62	10	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	62	11	re-	re-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	62	12	spect	spect	NN
work_ezwbhq2nrvadncyjpwjlluuona	62	13	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	62	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	62	15	predictions	prediction	NNS
work_ezwbhq2nrvadncyjpwjlluuona	62	16	stated	state	VBN
work_ezwbhq2nrvadncyjpwjlluuona	62	17	above	above	RB
work_ezwbhq2nrvadncyjpwjlluuona	62	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	63	1	Secondly	secondly	RB
work_ezwbhq2nrvadncyjpwjlluuona	63	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	63	3	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	63	4	type	type	NN
work_ezwbhq2nrvadncyjpwjlluuona	63	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	63	6	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	63	7	suggests	suggest	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	63	8	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	63	9	differ-	differ-	NN
work_ezwbhq2nrvadncyjpwjlluuona	63	10	ent	ent	NN
work_ezwbhq2nrvadncyjpwjlluuona	63	11	definition	definition	NN
work_ezwbhq2nrvadncyjpwjlluuona	63	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	63	13	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	63	14	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	63	15	has	have	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	63	16	previously	previously	RB
work_ezwbhq2nrvadncyjpwjlluuona	63	17	been	be	VBN
work_ezwbhq2nrvadncyjpwjlluuona	63	18	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	63	19	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	63	20	characterize	characterize	VB
work_ezwbhq2nrvadncyjpwjlluuona	63	21	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	63	22	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	64	1	Since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	64	3	autoasso-	autoasso-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	64	4	ciative	ciative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	64	5	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	64	6	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	64	7	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	64	8	decomposed	decompose	VBN
work_ezwbhq2nrvadncyjpwjlluuona	64	9	into	into	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	64	11	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	64	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	13	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	64	14	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	64	15	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	64	16	since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	17	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	64	18	learned	learn	VBN
work_ezwbhq2nrvadncyjpwjlluuona	64	19	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	20	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	64	21	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	64	22	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	64	23	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	64	24	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	64	25	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	26	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	64	27	weighted	weighted	NNP
work_ezwbhq2nrvadncyjpwjlluuona	64	28	combina-	combina-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	64	29	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	64	30	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	31	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	64	32	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	64	33	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	64	34	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	64	35	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	64	36	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	64	37	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	64	38	thought	think	VBN
work_ezwbhq2nrvadncyjpwjlluuona	64	39	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	40	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	41	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	64	42	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	64	43	characterizing	characterize	VBG
work_ezwbhq2nrvadncyjpwjlluuona	64	44	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	64	45	stimulus	stimulus	NN
work_ezwbhq2nrvadncyjpwjlluuona	64	46	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	64	47	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	65	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	65	2	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	65	3	expect	expect	VB
work_ezwbhq2nrvadncyjpwjlluuona	65	4	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	65	5	see	see	VB
work_ezwbhq2nrvadncyjpwjlluuona	65	6	differences	difference	NNS
work_ezwbhq2nrvadncyjpwjlluuona	65	7	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	65	8	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	65	9	based	base	VBN
work_ezwbhq2nrvadncyjpwjlluuona	65	10	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	65	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	65	12	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	65	13	history	history	NN
work_ezwbhq2nrvadncyjpwjlluuona	65	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	65	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	65	16	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	65	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	66	1	Furthermore	furthermore	RB
work_ezwbhq2nrvadncyjpwjlluuona	66	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	66	3	since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	66	4	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	66	5	used	use	VBD
work_ezwbhq2nrvadncyjpwjlluuona	66	6	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	66	7	simple	simple	JJ
work_ezwbhq2nrvadncyjpwjlluuona	66	8	visual	visual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	66	9	code	code	NN
work_ezwbhq2nrvadncyjpwjlluuona	66	10	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	66	11	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	66	12	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	66	13	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	66	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	66	15	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	66	16	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	66	17	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	66	18	displayed	display	VBN
work_ezwbhq2nrvadncyjpwjlluuona	66	19	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	66	20	images	image	NNS
work_ezwbhq2nrvadncyjpwjlluuona	66	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	67	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	67	2	shall	shall	MD
work_ezwbhq2nrvadncyjpwjlluuona	67	3	discuss	discuss	VB
work_ezwbhq2nrvadncyjpwjlluuona	67	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	67	5	potential	potential	JJ
work_ezwbhq2nrvadncyjpwjlluuona	67	6	role	role	NN
work_ezwbhq2nrvadncyjpwjlluuona	67	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	67	8	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	67	9	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	67	10	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	67	11	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	67	12	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	67	13	characteriz-	characteriz-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	67	14	ing	ing	JJ
work_ezwbhq2nrvadncyjpwjlluuona	67	15	same	same	JJ
work_ezwbhq2nrvadncyjpwjlluuona	67	16	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	67	17	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	67	18	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	67	19	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	67	20	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	67	21	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	67	22	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	68	1	Simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	68	2	1	1	CD
work_ezwbhq2nrvadncyjpwjlluuona	68	3	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	68	4	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	68	5	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	68	6	defined	define	VBN
work_ezwbhq2nrvadncyjpwjlluuona	68	7	first	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	68	8	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	68	9	then	then	RB
work_ezwbhq2nrvadncyjpwjlluuona	68	10	its	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	68	11	applica-	applica-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	68	12	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	68	13	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	68	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	68	15	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	68	16	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	68	17	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	68	18	problem	problem	NN
work_ezwbhq2nrvadncyjpwjlluuona	68	19	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	68	20	presented	present	VBN
work_ezwbhq2nrvadncyjpwjlluuona	68	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	69	1	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	69	2	dig-	dig-	RB
work_ezwbhq2nrvadncyjpwjlluuona	69	3	itized	itize	VBN
work_ezwbhq2nrvadncyjpwjlluuona	69	4	image	image	NN
work_ezwbhq2nrvadncyjpwjlluuona	69	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	69	6	each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	69	7	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	69	8	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	69	9	coded	code	VBN
work_ezwbhq2nrvadncyjpwjlluuona	69	10	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	69	11	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	69	12	vector	vector	NN
work_ezwbhq2nrvadncyjpwjlluuona	69	13	comprised	comprise	VBN
work_ezwbhq2nrvadncyjpwjlluuona	69	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	69	15	pixel	pixel	NNP
work_ezwbhq2nrvadncyjpwjlluuona	69	16	elements	element	NNS
work_ezwbhq2nrvadncyjpwjlluuona	69	17	concatenated	concatenate	VBD
work_ezwbhq2nrvadncyjpwjlluuona	69	18	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	69	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	69	20	rows	row	NNS
work_ezwbhq2nrvadncyjpwjlluuona	69	21	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	69	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	69	23	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	69	24	image	image	NN
work_ezwbhq2nrvadncyjpwjlluuona	69	25	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	70	1	Thus	thus	RB
work_ezwbhq2nrvadncyjpwjlluuona	70	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	70	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	70	4	ith	ith	NNP
work_ezwbhq2nrvadncyjpwjlluuona	70	5	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	70	6	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	70	7	rep-	rep-	RB
work_ezwbhq2nrvadncyjpwjlluuona	70	8	resented	resent	VBN
work_ezwbhq2nrvadncyjpwjlluuona	70	9	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	70	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	70	11	J	J	NNP
work_ezwbhq2nrvadncyjpwjlluuona	70	12	×	×	NN
work_ezwbhq2nrvadncyjpwjlluuona	70	13	1	1	CD
work_ezwbhq2nrvadncyjpwjlluuona	70	14	vector	vector	NN
work_ezwbhq2nrvadncyjpwjlluuona	70	15	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	70	16	where	where	WRB
work_ezwbhq2nrvadncyjpwjlluuona	70	17	J	J	NNP
work_ezwbhq2nrvadncyjpwjlluuona	70	18	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	70	19	equal	equal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	70	20	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	70	21	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	70	22	width	width	NN
work_ezwbhq2nrvadncyjpwjlluuona	70	23	times	time	NNS
work_ezwbhq2nrvadncyjpwjlluuona	70	24	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	70	25	height	height	NN
work_ezwbhq2nrvadncyjpwjlluuona	70	26	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	70	27	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	70	28	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	70	29	image	image	NN
work_ezwbhq2nrvadncyjpwjlluuona	70	30	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	70	31	pixels	pixel	NNS
work_ezwbhq2nrvadncyjpwjlluuona	70	32	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	70	33	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	70	34	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	70	35	denoted	denote	VBN
work_ezwbhq2nrvadncyjpwjlluuona	70	36	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	70	37	fi	fi	NN
work_ezwbhq2nrvadncyjpwjlluuona	70	38	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	71	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	71	2	convenience	convenience	NN
work_ezwbhq2nrvadncyjpwjlluuona	71	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	71	4	normalized	normalized	JJ
work_ezwbhq2nrvadncyjpwjlluuona	71	5	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	71	6	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	71	7	assumed	assume	VBN
work_ezwbhq2nrvadncyjpwjlluuona	71	8	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	71	9	i.e.	i.e.	FW
work_ezwbhq2nrvadncyjpwjlluuona	72	1	fTi	fTi	NNS
work_ezwbhq2nrvadncyjpwjlluuona	72	2	fi	fi	IN
work_ezwbhq2nrvadncyjpwjlluuona	72	3	=	=	SYM
work_ezwbhq2nrvadncyjpwjlluuona	72	4	1	1	CD
work_ezwbhq2nrvadncyjpwjlluuona	72	5	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	72	6	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	73	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	73	2	autoasso-	autoasso-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	73	3	ciative	ciative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	73	4	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	5	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	73	6	constructed	construct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	73	7	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	73	8	A	a	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	9	=	=	SYM
work_ezwbhq2nrvadncyjpwjlluuona	73	10	∑	∑	.
work_ezwbhq2nrvadncyjpwjlluuona	73	11	i	i	PRP
work_ezwbhq2nrvadncyjpwjlluuona	73	12	fif	fif	VBP
work_ezwbhq2nrvadncyjpwjlluuona	73	13	T	t	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	14	i	i	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	15	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	73	16	1	1	LS
work_ezwbhq2nrvadncyjpwjlluuona	73	17	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	73	18	Recall	recall	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	73	20	individual	individual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	73	21	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	73	22	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	73	23	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	73	24	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	25	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	73	26	done	do	VBN
work_ezwbhq2nrvadncyjpwjlluuona	73	27	according	accord	VBG
work_ezwbhq2nrvadncyjpwjlluuona	73	28	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	73	29	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	73	30	rule	rule	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	31	f̂i	f̂i	NNP
work_ezwbhq2nrvadncyjpwjlluuona	73	32	=	=	SYM
work_ezwbhq2nrvadncyjpwjlluuona	73	33	Afi	Afi	NNP
work_ezwbhq2nrvadncyjpwjlluuona	73	34	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	73	35	2	2	CD
work_ezwbhq2nrvadncyjpwjlluuona	73	36	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	73	37	where	where	WRB
work_ezwbhq2nrvadncyjpwjlluuona	73	38	f̂i	f̂i	NNP
work_ezwbhq2nrvadncyjpwjlluuona	73	39	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	73	40	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	73	41	system	system	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	42	estimate	estimate	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	43	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	73	44	fi	fi	NN
work_ezwbhq2nrvadncyjpwjlluuona	73	45	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	74	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	74	2	qual-	qual-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	74	3	ity	ity	NN
work_ezwbhq2nrvadncyjpwjlluuona	74	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	74	5	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	74	6	estimate	estimate	NN
work_ezwbhq2nrvadncyjpwjlluuona	74	7	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	74	8	measured	measure	VBN
work_ezwbhq2nrvadncyjpwjlluuona	74	9	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	74	10	comparing	compare	VBG
work_ezwbhq2nrvadncyjpwjlluuona	74	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	74	12	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	74	13	image	image	NN
work_ezwbhq2nrvadncyjpwjlluuona	74	14	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	74	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	74	16	original	original	JJ
work_ezwbhq2nrvadncyjpwjlluuona	74	17	image	image	NN
work_ezwbhq2nrvadncyjpwjlluuona	74	18	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	74	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	74	20	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	74	21	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	74	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	74	23	angle	angle	NN
work_ezwbhq2nrvadncyjpwjlluuona	74	24	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	74	25	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	74	26	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	74	27	f̂i	f̂i	NNP
work_ezwbhq2nrvadncyjpwjlluuona	74	28	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	74	29	fi	fi	NN
work_ezwbhq2nrvadncyjpwjlluuona	74	30	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	75	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	75	2	Widrow	Widrow	NNP
work_ezwbhq2nrvadncyjpwjlluuona	75	3	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	75	4	Hoff	Hoff	NNP
work_ezwbhq2nrvadncyjpwjlluuona	75	5	error	error	NN
work_ezwbhq2nrvadncyjpwjlluuona	75	6	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	75	7	correction	correction	NN
work_ezwbhq2nrvadncyjpwjlluuona	75	8	rule	rule	NN
work_ezwbhq2nrvadncyjpwjlluuona	75	9	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	75	10	ap-	ap-	RB
work_ezwbhq2nrvadncyjpwjlluuona	75	11	plied	ply	VBN
work_ezwbhq2nrvadncyjpwjlluuona	75	12	iteratively	iteratively	RB
work_ezwbhq2nrvadncyjpwjlluuona	75	13	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	75	14	optimize	optimize	VB
work_ezwbhq2nrvadncyjpwjlluuona	75	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	75	16	quality	quality	NN
work_ezwbhq2nrvadncyjpwjlluuona	75	17	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	75	18	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	75	19	recall	recall	NN
work_ezwbhq2nrvadncyjpwjlluuona	75	20	across	across	IN
work_ezwbhq2nrvadncyjpwjlluuona	75	21	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	75	22	stimulus	stimulus	NN
work_ezwbhq2nrvadncyjpwjlluuona	75	23	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	75	24	A[t+1	A[t+1	NNP
work_ezwbhq2nrvadncyjpwjlluuona	75	25	]	]	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	75	26	=	=	NFP
work_ezwbhq2nrvadncyjpwjlluuona	75	27	A[t	a[t	ADD
work_ezwbhq2nrvadncyjpwjlluuona	75	28	]	]	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	75	29	−	−	NNP
work_ezwbhq2nrvadncyjpwjlluuona	75	30	γ	γ	NNP
work_ezwbhq2nrvadncyjpwjlluuona	75	31	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	75	32	fi	fi	IN
work_ezwbhq2nrvadncyjpwjlluuona	75	33	−	−	NNP
work_ezwbhq2nrvadncyjpwjlluuona	75	34	A[t]fi	A[t]fi	NNP
work_ezwbhq2nrvadncyjpwjlluuona	75	35	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	75	36	fTi	fTi	NNP
work_ezwbhq2nrvadncyjpwjlluuona	75	37	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	75	38	3	3	LS
work_ezwbhq2nrvadncyjpwjlluuona	75	39	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	75	40	where	where	WRB
work_ezwbhq2nrvadncyjpwjlluuona	75	41	i	i	PRP
work_ezwbhq2nrvadncyjpwjlluuona	75	42	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	75	43	randomly	randomly	RB
work_ezwbhq2nrvadncyjpwjlluuona	75	44	chosen	choose	VBN
work_ezwbhq2nrvadncyjpwjlluuona	75	45	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	75	46	γ	γ	DT
work_ezwbhq2nrvadncyjpwjlluuona	75	47	decreases	decrease	NNS
work_ezwbhq2nrvadncyjpwjlluuona	75	48	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	75	49	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	75	50	reciprocal	reciprocal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	75	51	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	75	52	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	75	53	iteration	iteration	NN
work_ezwbhq2nrvadncyjpwjlluuona	75	54	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	75	55	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	76	1	Since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	76	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	76	3	eigen	eigen	NN
work_ezwbhq2nrvadncyjpwjlluuona	76	4	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	76	5	decomposition	decomposition	NN
work_ezwbhq2nrvadncyjpwjlluuona	76	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	76	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	76	8	autoasso-	autoasso-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	76	9	ciative	ciative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	76	10	matrix	matrix	NNP
work_ezwbhq2nrvadncyjpwjlluuona	76	11	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	76	12	equivalent	equivalent	JJ
work_ezwbhq2nrvadncyjpwjlluuona	76	13	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	76	14	principal	principal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	76	15	component	component	NN
work_ezwbhq2nrvadncyjpwjlluuona	76	16	analysis	analysis	NN
work_ezwbhq2nrvadncyjpwjlluuona	76	17	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	76	18	Abdi	Abdi	NNP
work_ezwbhq2nrvadncyjpwjlluuona	76	19	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	76	20	1988	1988	CD
work_ezwbhq2nrvadncyjpwjlluuona	76	21	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	76	22	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	76	23	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	76	24	autoassociative	autoassociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	76	25	matrix	matrix	NNP
work_ezwbhq2nrvadncyjpwjlluuona	76	26	166	166	CD
work_ezwbhq2nrvadncyjpwjlluuona	76	27	O’TOOLE	O’TOOLE	NNP
work_ezwbhq2nrvadncyjpwjlluuona	76	28	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	76	29	DEFFENBACHER	DEFFENBACHER	NNP
work_ezwbhq2nrvadncyjpwjlluuona	76	30	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	76	31	ABDI	ABDI	NNP
work_ezwbhq2nrvadncyjpwjlluuona	76	32	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	76	33	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	76	34	BARTLETT	BARTLETT	NNP
work_ezwbhq2nrvadncyjpwjlluuona	76	35	Figure	Figure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	76	36	1	1	CD
work_ezwbhq2nrvadncyjpwjlluuona	76	37	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	77	1	Mean	mean	JJ
work_ezwbhq2nrvadncyjpwjlluuona	77	2	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	77	3	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	77	4	original	original	JJ
work_ezwbhq2nrvadncyjpwjlluuona	77	5	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	77	6	reconstructed	reconstructed	JJ
work_ezwbhq2nrvadncyjpwjlluuona	77	7	images	image	NNS
work_ezwbhq2nrvadncyjpwjlluuona	77	8	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	77	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	77	10	OLD	OLD	NNP
work_ezwbhq2nrvadncyjpwjlluuona	77	11	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	77	12	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	77	13	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	77	14	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	77	15	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	77	16	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	77	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	78	1	SIMULATING	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	78	2	THE	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	78	3	OTHER	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	78	4	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	78	5	RACE	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	78	6	EFFECT	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	78	7	167	167	CD
work_ezwbhq2nrvadncyjpwjlluuona	78	8	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	78	9	give	give	VB
work_ezwbhq2nrvadncyjpwjlluuona	78	10	indications	indication	NNS
work_ezwbhq2nrvadncyjpwjlluuona	78	11	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	78	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	78	13	statistical	statistical	JJ
work_ezwbhq2nrvadncyjpwjlluuona	78	14	structure	structure	NN
work_ezwbhq2nrvadncyjpwjlluuona	78	15	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	78	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	78	17	stimulus	stimulus	NN
work_ezwbhq2nrvadncyjpwjlluuona	78	18	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	78	19	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	79	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	79	2	storage	storage	NN
work_ezwbhq2nrvadncyjpwjlluuona	79	3	capacity	capacity	NN
work_ezwbhq2nrvadncyjpwjlluuona	79	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	79	5	such	such	PDT
work_ezwbhq2nrvadncyjpwjlluuona	79	6	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	79	7	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	79	8	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	79	9	approximately	approximately	RB
work_ezwbhq2nrvadncyjpwjlluuona	79	10	15	15	CD
work_ezwbhq2nrvadncyjpwjlluuona	79	11	%	%	NN
work_ezwbhq2nrvadncyjpwjlluuona	79	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	79	13	its	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	79	14	dimensionality	dimensionality	NN
work_ezwbhq2nrvadncyjpwjlluuona	79	15	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	79	16	random	random	JJ
work_ezwbhq2nrvadncyjpwjlluuona	79	17	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	79	18	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	79	19	Hopfield	Hopfield	NNP
work_ezwbhq2nrvadncyjpwjlluuona	79	20	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	79	21	1984	1984	CD
work_ezwbhq2nrvadncyjpwjlluuona	79	22	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	79	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	80	1	Since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	80	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	80	3	di-	di-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	80	4	mension	mension	NN
work_ezwbhq2nrvadncyjpwjlluuona	80	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	80	6	our	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	80	7	images	image	NNS
work_ezwbhq2nrvadncyjpwjlluuona	80	8	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	80	9	very	very	RB
work_ezwbhq2nrvadncyjpwjlluuona	80	10	large	large	JJ
work_ezwbhq2nrvadncyjpwjlluuona	80	11	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	80	12	compar-	compar-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	80	13	ison	ison	NNP
work_ezwbhq2nrvadncyjpwjlluuona	80	14	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	80	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	80	16	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	80	17	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	80	18	stimuli	stimulus	NNS
work_ezwbhq2nrvadncyjpwjlluuona	80	19	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	80	20	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	80	21	limit	limit	NN
work_ezwbhq2nrvadncyjpwjlluuona	80	22	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	80	23	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	80	24	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	80	25	problem	problem	NN
work_ezwbhq2nrvadncyjpwjlluuona	80	26	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	80	27	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	80	28	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	80	29	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	81	1	Method	Method	NNP
work_ezwbhq2nrvadncyjpwjlluuona	81	2	Stimuli	Stimuli	NNP
work_ezwbhq2nrvadncyjpwjlluuona	81	3	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	82	1	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	82	2	total	total	NN
work_ezwbhq2nrvadncyjpwjlluuona	82	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	82	4	319	319	CD
work_ezwbhq2nrvadncyjpwjlluuona	82	5	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	82	6	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	82	7	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	82	8	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	82	9	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	82	10	digitized	digitize	VBN
work_ezwbhq2nrvadncyjpwjlluuona	82	11	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	82	12	slides	slide	NNS
work_ezwbhq2nrvadncyjpwjlluuona	82	13	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	82	14	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	82	15	resolu-	resolu-	NNS
work_ezwbhq2nrvadncyjpwjlluuona	82	16	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	82	17	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	82	18	16	16	CD
work_ezwbhq2nrvadncyjpwjlluuona	82	19	grey	grey	NN
work_ezwbhq2nrvadncyjpwjlluuona	82	20	levels	level	NNS
work_ezwbhq2nrvadncyjpwjlluuona	82	21	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	82	22	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	82	23	Fotovix	Fotovix	NNP
work_ezwbhq2nrvadncyjpwjlluuona	82	24	digitizer	digitizer	NN
work_ezwbhq2nrvadncyjpwjlluuona	82	25	attached	attach	VBN
work_ezwbhq2nrvadncyjpwjlluuona	82	26	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	82	27	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	82	28	286-based	286-based	CD
work_ezwbhq2nrvadncyjpwjlluuona	82	29	computer	computer	NN
work_ezwbhq2nrvadncyjpwjlluuona	82	30	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	82	31	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	82	32	16-bit	16-bit	CD
work_ezwbhq2nrvadncyjpwjlluuona	82	33	TARGA	TARGA	NNP
work_ezwbhq2nrvadncyjpwjlluuona	82	34	board	board	NN
work_ezwbhq2nrvadncyjpwjlluuona	82	35	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	82	36	True	True	NNP
work_ezwbhq2nrvadncyjpwjlluuona	82	37	Vision	Vision	NNP
work_ezwbhq2nrvadncyjpwjlluuona	82	38	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	82	39	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	83	1	Faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	83	2	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	83	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	83	4	young	young	JJ
work_ezwbhq2nrvadncyjpwjlluuona	83	5	adults	adult	NNS
work_ezwbhq2nrvadncyjpwjlluuona	83	6	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	83	7	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	83	8	roughly	roughly	RB
work_ezwbhq2nrvadncyjpwjlluuona	83	9	half	half	RB
work_ezwbhq2nrvadncyjpwjlluuona	83	10	male	male	JJ
work_ezwbhq2nrvadncyjpwjlluuona	83	11	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	83	12	half	half	NN
work_ezwbhq2nrvadncyjpwjlluuona	83	13	female	female	NN
work_ezwbhq2nrvadncyjpwjlluuona	83	14	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	84	1	None	none	NN
work_ezwbhq2nrvadncyjpwjlluuona	84	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	84	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	84	4	slides	slide	NNS
work_ezwbhq2nrvadncyjpwjlluuona	84	5	pictured	picture	VBD
work_ezwbhq2nrvadncyjpwjlluuona	84	6	people	people	NNS
work_ezwbhq2nrvadncyjpwjlluuona	84	7	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	84	8	facial	facial	JJ
work_ezwbhq2nrvadncyjpwjlluuona	84	9	hair	hair	NN
work_ezwbhq2nrvadncyjpwjlluuona	84	10	or	or	CC
work_ezwbhq2nrvadncyjpwjlluuona	84	11	glasses	glass	NNS
work_ezwbhq2nrvadncyjpwjlluuona	84	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	85	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	85	2	images	image	NNS
work_ezwbhq2nrvadncyjpwjlluuona	85	3	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	85	4	aligned	align	VBN
work_ezwbhq2nrvadncyjpwjlluuona	85	5	so	so	IN
work_ezwbhq2nrvadncyjpwjlluuona	85	6	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	85	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	85	8	eyes	eye	NNS
work_ezwbhq2nrvadncyjpwjlluuona	85	9	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	85	10	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	85	11	about	about	IN
work_ezwbhq2nrvadncyjpwjlluuona	85	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	85	13	same	same	JJ
work_ezwbhq2nrvadncyjpwjlluuona	85	14	height	height	NN
work_ezwbhq2nrvadncyjpwjlluuona	85	15	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	86	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	86	2	images	image	NNS
work_ezwbhq2nrvadncyjpwjlluuona	86	3	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	86	4	cropped	crop	VBN
work_ezwbhq2nrvadncyjpwjlluuona	86	5	around	around	IN
work_ezwbhq2nrvadncyjpwjlluuona	86	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	86	7	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	86	8	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	86	9	eliminate	eliminate	VB
work_ezwbhq2nrvadncyjpwjlluuona	86	10	cloth-	cloth-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	86	11	ing	ing	NN
work_ezwbhq2nrvadncyjpwjlluuona	86	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	87	1	Each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	87	2	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	87	3	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	87	4	151	151	CD
work_ezwbhq2nrvadncyjpwjlluuona	87	5	pixels	pixel	NNS
work_ezwbhq2nrvadncyjpwjlluuona	87	6	wide	wide	JJ
work_ezwbhq2nrvadncyjpwjlluuona	87	7	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	87	8	225	225	CD
work_ezwbhq2nrvadncyjpwjlluuona	87	9	pix-	pix-	NNS
work_ezwbhq2nrvadncyjpwjlluuona	87	10	els	els	NNP
work_ezwbhq2nrvadncyjpwjlluuona	87	11	long	long	JJ
work_ezwbhq2nrvadncyjpwjlluuona	87	12	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	87	13	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	87	14	so	so	RB
work_ezwbhq2nrvadncyjpwjlluuona	87	15	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	87	16	represented	represent	VBN
work_ezwbhq2nrvadncyjpwjlluuona	87	17	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	87	18	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	87	19	33	33	CD
work_ezwbhq2nrvadncyjpwjlluuona	87	20	975-pixel	975-pixel	CD
work_ezwbhq2nrvadncyjpwjlluuona	87	21	vector	vector	NN
work_ezwbhq2nrvadncyjpwjlluuona	87	22	consisting	consist	VBG
work_ezwbhq2nrvadncyjpwjlluuona	87	23	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	87	24	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	87	25	concatenation	concatenation	NN
work_ezwbhq2nrvadncyjpwjlluuona	87	26	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	87	27	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	87	28	pixels	pixel	NNS
work_ezwbhq2nrvadncyjpwjlluuona	87	29	rows	row	NNS
work_ezwbhq2nrvadncyjpwjlluuona	87	30	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	88	1	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	88	2	spatial	spatial	JJ
work_ezwbhq2nrvadncyjpwjlluuona	88	3	differentiation	differentiation	NN
work_ezwbhq2nrvadncyjpwjlluuona	88	4	encoding	encoding	NN
work_ezwbhq2nrvadncyjpwjlluuona	88	5	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	88	6	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	88	7	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	88	8	enhance	enhance	VB
work_ezwbhq2nrvadncyjpwjlluuona	88	9	lines	line	NNS
work_ezwbhq2nrvadncyjpwjlluuona	88	10	prior	prior	RB
work_ezwbhq2nrvadncyjpwjlluuona	88	11	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	88	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	88	13	extraction	extraction	NN
work_ezwbhq2nrvadncyjpwjlluuona	88	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	88	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	88	16	pixel	pixel	NN
work_ezwbhq2nrvadncyjpwjlluuona	88	17	vector	vector	NN
work_ezwbhq2nrvadncyjpwjlluuona	88	18	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	88	19	cf	cf	NNP
work_ezwbhq2nrvadncyjpwjlluuona	88	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	89	1	O’Toole	O’Toole	NNP
work_ezwbhq2nrvadncyjpwjlluuona	89	2	et	et	FW
work_ezwbhq2nrvadncyjpwjlluuona	89	3	al	al	NNP
work_ezwbhq2nrvadncyjpwjlluuona	89	4	.	.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	89	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	89	6	1988	1988	CD
work_ezwbhq2nrvadncyjpwjlluuona	89	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	89	8	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	90	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	90	2	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	90	3	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	90	4	carried	carry	VBN
work_ezwbhq2nrvadncyjpwjlluuona	90	5	out	out	RP
work_ezwbhq2nrvadncyjpwjlluuona	90	6	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	90	7	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	90	8	Sun	Sun	NNP
work_ezwbhq2nrvadncyjpwjlluuona	90	9	MicroSystems	MicroSystems	NNP
work_ezwbhq2nrvadncyjpwjlluuona	90	10	SparcSta-	SparcSta-	:
work_ezwbhq2nrvadncyjpwjlluuona	90	11	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	90	12	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	90	13	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	90	14	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	90	15	Convex	Convex	NNP
work_ezwbhq2nrvadncyjpwjlluuona	90	16	C-1	C-1	NNP
work_ezwbhq2nrvadncyjpwjlluuona	90	17	Vector	Vector	NNP
work_ezwbhq2nrvadncyjpwjlluuona	90	18	computer	computer	NN
work_ezwbhq2nrvadncyjpwjlluuona	90	19	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	91	1	Procedure	procedure	NN
work_ezwbhq2nrvadncyjpwjlluuona	91	2	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	92	1	Two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	92	2	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	92	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	92	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	92	5	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	92	6	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	92	7	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	92	8	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	92	9	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	92	10	performed	perform	VBN
work_ezwbhq2nrvadncyjpwjlluuona	92	11	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	92	12	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	92	13	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	92	14	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	92	15	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	92	16	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	92	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	92	18	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	92	19	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	92	20	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	92	21	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	92	22	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	92	23	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	92	24	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	92	25	mi-	mi-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	92	26	nority	nority	NN
work_ezwbhq2nrvadncyjpwjlluuona	92	27	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	92	28	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	92	29	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	92	30	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	92	31	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	92	32	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	92	33	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	92	34	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	92	35	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	92	36	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	92	37	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	92	38	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	92	39	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	92	40	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	92	41	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	92	42	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	92	43	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	92	44	group	group	NN
work_ezwbhq2nrvadncyjpwjlluuona	92	45	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	93	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	93	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	93	3	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	93	4	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	93	5	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	93	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	93	7	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	93	8	associative	associative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	93	9	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	93	10	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	93	11	trained	train	VBN
work_ezwbhq2nrvadncyjpwjlluuona	93	12	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	93	13	error	error	NN
work_ezwbhq2nrvadncyjpwjlluuona	93	14	correc-	correc-	FW
work_ezwbhq2nrvadncyjpwjlluuona	93	15	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	93	16	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	93	17	95	95	CD
work_ezwbhq2nrvadncyjpwjlluuona	93	18	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	93	19	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	93	20	five	five	CD
work_ezwbhq2nrvadncyjpwjlluuona	93	21	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	93	22	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	93	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	94	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	94	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	94	3	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	94	4	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	94	5	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	94	6	95	95	CD
work_ezwbhq2nrvadncyjpwjlluuona	94	7	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	94	8	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	94	9	five	five	CD
work_ezwbhq2nrvadncyjpwjlluuona	94	10	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	94	11	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	94	12	served	serve	VBD
work_ezwbhq2nrvadncyjpwjlluuona	94	13	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	94	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	94	15	training	training	NN
work_ezwbhq2nrvadncyjpwjlluuona	94	16	set3	set3	NN
work_ezwbhq2nrvadncyjpwjlluuona	94	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	95	1	Results	result	NNS
work_ezwbhq2nrvadncyjpwjlluuona	95	2	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	95	3	Discussion	Discussion	NNP
work_ezwbhq2nrvadncyjpwjlluuona	95	4	Representations	Representations	NNPS
work_ezwbhq2nrvadncyjpwjlluuona	95	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	95	6	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	95	7	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	95	8	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	95	9	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	95	10	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	95	11	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	96	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	96	2	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	96	3	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	96	4	tested	test	VBN
work_ezwbhq2nrvadncyjpwjlluuona	96	5	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	96	6	reconstructing	reconstruct	VBG
work_ezwbhq2nrvadncyjpwjlluuona	96	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	96	8	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	96	9	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	96	10	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	96	11	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	96	12	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	96	13	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	96	14	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	96	15	learned	learn	VBN
work_ezwbhq2nrvadncyjpwjlluuona	96	16	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	96	17	OLD	OLD	NNP
work_ezwbhq2nrvadncyjpwjlluuona	96	18	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	96	19	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	96	20	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	96	21	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	96	22	reconstructing	reconstruct	VBG
work_ezwbhq2nrvadncyjpwjlluuona	96	23	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	96	24	sample	sample	NN
work_ezwbhq2nrvadncyjpwjlluuona	96	25	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	96	26	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	96	27	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	96	28	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	96	29	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	96	30	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	96	31	learned	learn	VBN
work_ezwbhq2nrvadncyjpwjlluuona	96	32	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	96	33	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	96	34	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	96	35	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	96	36	NEW	NEW	NNP
work_ezwbhq2nrvadncyjpwjlluuona	96	37	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	96	38	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	97	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	97	2	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	97	3	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	97	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	97	5	origi-	origi-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	97	6	nal	nal	NNP
work_ezwbhq2nrvadncyjpwjlluuona	97	7	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	97	8	reconstructed	reconstructed	JJ
work_ezwbhq2nrvadncyjpwjlluuona	97	9	image	image	NN
work_ezwbhq2nrvadncyjpwjlluuona	97	10	indicates	indicate	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	97	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	97	12	quality	quality	NN
work_ezwbhq2nrvadncyjpwjlluuona	97	13	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	97	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	97	15	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	97	16	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	97	17	representation	representation	NN
work_ezwbhq2nrvadncyjpwjlluuona	97	18	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	97	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	97	20	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	97	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	98	1	Fig-	Fig-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	98	2	ure	ure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	98	3	1	1	CD
work_ezwbhq2nrvadncyjpwjlluuona	98	4	shows	show	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	98	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	98	6	mean	mean	JJ
work_ezwbhq2nrvadncyjpwjlluuona	98	7	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	98	8	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	98	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	98	10	OLD	OLD	NNP
work_ezwbhq2nrvadncyjpwjlluuona	98	11	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	98	12	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	98	13	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	98	14	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	98	15	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	98	16	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	98	17	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	98	18	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	98	19	simu-	simu-	NN
work_ezwbhq2nrvadncyjpwjlluuona	98	20	lations	lation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	98	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	99	1	Three	three	CD
work_ezwbhq2nrvadncyjpwjlluuona	99	2	points	point	NNS
work_ezwbhq2nrvadncyjpwjlluuona	99	3	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	99	4	worth	worth	JJ
work_ezwbhq2nrvadncyjpwjlluuona	99	5	noting	note	VBG
work_ezwbhq2nrvadncyjpwjlluuona	99	6	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	100	1	First	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	100	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	100	3	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	100	4	both	both	DT
work_ezwbhq2nrvadncyjpwjlluuona	100	5	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	100	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	100	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	100	8	OLD	old	JJ
work_ezwbhq2nrvadncyjpwjlluuona	100	9	stimuli	stimulus	NNS
work_ezwbhq2nrvadncyjpwjlluuona	100	10	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	100	11	both	both	DT
work_ezwbhq2nrvadncyjpwjlluuona	100	12	Cau-	Cau-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	100	13	casian	casian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	100	14	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	100	15	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	100	16	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	100	17	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	100	18	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	100	19	nearly	nearly	RB
work_ezwbhq2nrvadncyjpwjlluuona	100	20	perfectly	perfectly	RB
work_ezwbhq2nrvadncyjpwjlluuona	100	21	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	100	22	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	100	23	mean	mean	VB
work_ezwbhq2nrvadncyjpwjlluuona	100	24	cosine=0.98	cosine=0.98	ADD
work_ezwbhq2nrvadncyjpwjlluuona	100	25	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	100	26	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	101	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	101	2	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	101	3	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	101	4	con-	con-	NN
work_ezwbhq2nrvadncyjpwjlluuona	101	5	sequence	sequence	NN
work_ezwbhq2nrvadncyjpwjlluuona	101	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	101	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	101	8	fact	fact	NN
work_ezwbhq2nrvadncyjpwjlluuona	101	9	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	101	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	101	11	capacity	capacity	NN
work_ezwbhq2nrvadncyjpwjlluuona	101	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	101	13	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	101	14	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	101	15	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	101	16	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	101	17	challenged	challenge	VBN
work_ezwbhq2nrvadncyjpwjlluuona	101	18	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	101	19	cf	cf	NN
work_ezwbhq2nrvadncyjpwjlluuona	101	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	102	1	Hopfield	Hopfield	NNP
work_ezwbhq2nrvadncyjpwjlluuona	102	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	102	3	1984	1984	CD
work_ezwbhq2nrvadncyjpwjlluuona	102	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	102	5	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	102	6	be-	be-	RB
work_ezwbhq2nrvadncyjpwjlluuona	102	7	low	low	RB
work_ezwbhq2nrvadncyjpwjlluuona	102	8	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	102	9	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	103	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	103	2	discuss	discuss	VBP
work_ezwbhq2nrvadncyjpwjlluuona	103	3	below	below	IN
work_ezwbhq2nrvadncyjpwjlluuona	103	4	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	103	5	method	method	NN
work_ezwbhq2nrvadncyjpwjlluuona	103	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	103	7	degrading	degrade	VBG
work_ezwbhq2nrvadncyjpwjlluuona	103	8	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	103	9	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	103	10	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	103	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	103	12	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	103	13	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	103	14	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	103	15	psychologically	psychologically	RB
work_ezwbhq2nrvadncyjpwjlluuona	103	16	interesting	interesting	JJ
work_ezwbhq2nrvadncyjpwjlluuona	103	17	way	way	NN
work_ezwbhq2nrvadncyjpwjlluuona	103	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	104	1	Secondly	secondly	RB
work_ezwbhq2nrvadncyjpwjlluuona	104	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	104	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	104	4	average	average	JJ
work_ezwbhq2nrvadncyjpwjlluuona	104	5	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	104	6	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	104	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	104	8	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	104	9	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	104	10	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	104	11	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	104	12	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	104	13	greater	great	JJR
work_ezwbhq2nrvadncyjpwjlluuona	104	14	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	104	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	104	16	average	average	JJ
work_ezwbhq2nrvadncyjpwjlluuona	104	17	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	104	18	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	104	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	104	20	reconstructed	reconstructed	JJ
work_ezwbhq2nrvadncyjpwjlluuona	104	21	NEW	NEW	NNP
work_ezwbhq2nrvadncyjpwjlluuona	104	22	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	104	23	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	104	24	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	105	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	105	2	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	105	3	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	105	4	seen	see	VBN
work_ezwbhq2nrvadncyjpwjlluuona	105	5	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	105	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	105	7	interaction	interaction	NN
work_ezwbhq2nrvadncyjpwjlluuona	105	8	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	105	9	Figure	Figure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	105	10	1	1	CD
work_ezwbhq2nrvadncyjpwjlluuona	105	11	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	106	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	106	2	differences	difference	NNS
work_ezwbhq2nrvadncyjpwjlluuona	106	3	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	106	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	106	5	quality	quality	NN
work_ezwbhq2nrvadncyjpwjlluuona	106	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	106	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	106	8	reconstructions	reconstruction	NNS
work_ezwbhq2nrvadncyjpwjlluuona	106	9	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	106	10	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	106	11	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	106	12	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	106	13	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	106	14	reflects	reflect	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	106	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	106	16	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	106	17	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	106	18	greater	great	JJR
work_ezwbhq2nrvadncyjpwjlluuona	106	19	success	success	NN
work_ezwbhq2nrvadncyjpwjlluuona	106	20	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	106	21	coding	code	VBG
work_ezwbhq2nrvadncyjpwjlluuona	106	22	or	or	CC
work_ezwbhq2nrvadncyjpwjlluuona	106	23	representing	represent	VBG
work_ezwbhq2nrvadncyjpwjlluuona	106	24	novel	novel	JJ
work_ezwbhq2nrvadncyjpwjlluuona	106	25	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	106	26	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	106	27	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	106	28	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	106	29	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	106	30	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	106	31	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	106	32	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	106	33	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	106	34	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	106	35	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	107	1	Finally	finally	RB
work_ezwbhq2nrvadncyjpwjlluuona	107	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	107	3	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	107	4	both	both	DT
work_ezwbhq2nrvadncyjpwjlluuona	107	5	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	107	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	107	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	107	8	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	107	9	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	107	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	107	11	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	107	12	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	107	13	did	do	VBD
work_ezwbhq2nrvadncyjpwjlluuona	107	14	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	107	15	reflect	reflect	VB
work_ezwbhq2nrvadncyjpwjlluuona	107	16	random	random	JJ
work_ezwbhq2nrvadncyjpwjlluuona	107	17	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	107	18	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	107	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	107	20	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	107	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	108	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	108	2	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	108	3	words	word	NNS
work_ezwbhq2nrvadncyjpwjlluuona	108	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	108	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	108	6	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	108	7	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	108	8	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	108	9	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	108	10	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	108	11	completely	completely	RB
work_ezwbhq2nrvadncyjpwjlluuona	108	12	unfamiliar	unfamiliar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	108	13	stimuli	stimulus	NNS
work_ezwbhq2nrvadncyjpwjlluuona	108	14	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	108	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	108	16	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	108	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	109	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	109	2	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	109	3	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	109	4	consequence	consequence	NN
work_ezwbhq2nrvadncyjpwjlluuona	109	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	109	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	109	7	fact	fact	NN
work_ezwbhq2nrvadncyjpwjlluuona	109	8	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	109	9	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	109	10	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	109	11	share	share	VBP
work_ezwbhq2nrvadncyjpwjlluuona	109	12	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	109	13	general	general	JJ
work_ezwbhq2nrvadncyjpwjlluuona	109	14	schema	schema	NN
work_ezwbhq2nrvadncyjpwjlluuona	109	15	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	109	16	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	109	17	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	109	18	so	so	RB
work_ezwbhq2nrvadncyjpwjlluuona	109	19	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	109	20	given	give	VBN
work_ezwbhq2nrvadncyjpwjlluuona	109	21	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	109	22	might	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	109	23	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	109	24	best	well	RBS
work_ezwbhq2nrvadncyjpwjlluuona	109	25	thought	think	VBN
work_ezwbhq2nrvadncyjpwjlluuona	109	26	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	109	27	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	109	28	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	109	29	subcate-	subcate-	NN
work_ezwbhq2nrvadncyjpwjlluuona	109	30	gory	gory	NN
work_ezwbhq2nrvadncyjpwjlluuona	109	31	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	109	32	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	109	33	general	general	JJ
work_ezwbhq2nrvadncyjpwjlluuona	109	34	class	class	NN
work_ezwbhq2nrvadncyjpwjlluuona	109	35	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	109	36	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	109	37	stimuli	stimulus	NNS
work_ezwbhq2nrvadncyjpwjlluuona	109	38	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	110	1	Similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	110	2	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	111	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	111	2	tested	test	VBD
work_ezwbhq2nrvadncyjpwjlluuona	111	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	111	4	prediction	prediction	NN
work_ezwbhq2nrvadncyjpwjlluuona	111	5	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	111	6	novel	novel	JJ
work_ezwbhq2nrvadncyjpwjlluuona	111	7	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	111	8	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	111	9	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	111	10	perceived	perceive	VBN
work_ezwbhq2nrvadncyjpwjlluuona	111	11	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	111	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	111	13	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	111	14	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	111	15	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	111	16	less	less	RBR
work_ezwbhq2nrvadncyjpwjlluuona	111	17	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	111	18	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	111	19	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	111	20	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	111	21	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	111	22	i.e.	i.e.	FW
work_ezwbhq2nrvadncyjpwjlluuona	112	1	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	112	2	distinctive	distinctive	JJ
work_ezwbhq2nrvadncyjpwjlluuona	112	3	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	112	4	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	112	5	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	112	6	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	112	7	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	113	1	An	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	113	2	analysis	analysis	NN
work_ezwbhq2nrvadncyjpwjlluuona	113	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	113	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	113	5	similar-	similar-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	113	6	ity	ity	NN
work_ezwbhq2nrvadncyjpwjlluuona	113	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	113	8	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	113	9	reconstructed	reconstructed	JJ
work_ezwbhq2nrvadncyjpwjlluuona	113	10	NEW	NEW	NNP
work_ezwbhq2nrvadncyjpwjlluuona	113	11	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	113	12	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	113	13	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	113	14	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	113	15	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	113	16	carried	carry	VBN
work_ezwbhq2nrvadncyjpwjlluuona	113	17	out	out	RP
work_ezwbhq2nrvadncyjpwjlluuona	113	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	114	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	114	2	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	114	3	analysis	analysis	NN
work_ezwbhq2nrvadncyjpwjlluuona	114	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	114	5	50	50	CD
work_ezwbhq2nrvadncyjpwjlluuona	114	6	randomly	randomly	RB
work_ezwbhq2nrvadncyjpwjlluuona	114	7	cho-	cho-	RB
work_ezwbhq2nrvadncyjpwjlluuona	114	8	sen	sen	NNP
work_ezwbhq2nrvadncyjpwjlluuona	114	9	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	114	10	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	114	11	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	114	12	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	114	13	50	50	CD
work_ezwbhq2nrvadncyjpwjlluuona	114	14	randomly	randomly	RB
work_ezwbhq2nrvadncyjpwjlluuona	114	15	cho-	cho-	RB
work_ezwbhq2nrvadncyjpwjlluuona	114	16	sen	sen	NNP
work_ezwbhq2nrvadncyjpwjlluuona	114	17	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	114	18	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	114	19	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	114	20	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	114	21	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	114	22	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	114	23	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	114	24	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	114	25	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	114	26	estimate	estimate	NN
work_ezwbhq2nrvadncyjpwjlluuona	114	27	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	114	28	each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	114	29	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	114	30	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	114	31	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	114	32	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	114	33	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	114	34	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	114	35	analysis	analysis	NN
work_ezwbhq2nrvadncyjpwjlluuona	114	36	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	115	1	These	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	115	2	stimuli	stimulus	NNS
work_ezwbhq2nrvadncyjpwjlluuona	115	3	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	115	4	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	115	5	thought	think	VBN
work_ezwbhq2nrvadncyjpwjlluuona	115	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	115	7	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	115	8	filtered	filter	VBN
work_ezwbhq2nrvadncyjpwjlluuona	115	9	or	or	CC
work_ezwbhq2nrvadncyjpwjlluuona	115	10	“	"	``
work_ezwbhq2nrvadncyjpwjlluuona	115	11	perceived	perceive	VBN
work_ezwbhq2nrvadncyjpwjlluuona	115	12	”	"	''
work_ezwbhq2nrvadncyjpwjlluuona	115	13	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	115	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	115	15	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	115	16	trained	train	VBN
work_ezwbhq2nrvadncyjpwjlluuona	115	17	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	115	18	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	115	19	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	115	20	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	115	21	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	115	22	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	115	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	116	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	116	2	each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	116	3	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	116	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	116	5	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	116	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	116	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	116	8	inter	inter	JJ
work_ezwbhq2nrvadncyjpwjlluuona	116	9	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	116	10	face	face	JJ
work_ezwbhq2nrvadncyjpwjlluuona	116	11	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	116	12	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	116	13	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	116	14	recalled	recall	VBN
work_ezwbhq2nrvadncyjpwjlluuona	116	15	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	116	16	3	3	CD
work_ezwbhq2nrvadncyjpwjlluuona	116	17	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	116	18	choice	choice	NN
work_ezwbhq2nrvadncyjpwjlluuona	116	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	116	20	95	95	CD
work_ezwbhq2nrvadncyjpwjlluuona	116	21	%	%	NN
work_ezwbhq2nrvadncyjpwjlluuona	116	22	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	116	23	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	116	24	5	5	CD
work_ezwbhq2nrvadncyjpwjlluuona	116	25	%	%	NN
work_ezwbhq2nrvadncyjpwjlluuona	116	26	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	116	27	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	116	28	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	116	29	arbitrary	arbitrary	JJ
work_ezwbhq2nrvadncyjpwjlluuona	116	30	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	117	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	117	2	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	117	3	carried	carry	VBN
work_ezwbhq2nrvadncyjpwjlluuona	117	4	out	out	RP
work_ezwbhq2nrvadncyjpwjlluuona	117	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	117	6	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	117	7	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	117	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	117	9	sim-	sim-	DT
work_ezwbhq2nrvadncyjpwjlluuona	117	10	ulations	ulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	117	11	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	117	12	75	75	CD
work_ezwbhq2nrvadncyjpwjlluuona	117	13	%	%	NN
work_ezwbhq2nrvadncyjpwjlluuona	117	14	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	117	15	25	25	CD
work_ezwbhq2nrvadncyjpwjlluuona	117	16	%	%	NN
work_ezwbhq2nrvadncyjpwjlluuona	117	17	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	117	18	as	as	RB
work_ezwbhq2nrvadncyjpwjlluuona	117	19	well	well	RB
work_ezwbhq2nrvadncyjpwjlluuona	117	20	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	117	21	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	117	22	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	117	23	found	find	VBN
work_ezwbhq2nrvadncyjpwjlluuona	117	24	qualitatively	qualitatively	RB
work_ezwbhq2nrvadncyjpwjlluuona	117	25	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	117	26	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	117	27	though	though	IN
work_ezwbhq2nrvadncyjpwjlluuona	117	28	less	less	RBR
work_ezwbhq2nrvadncyjpwjlluuona	117	29	extreme	extreme	JJ
work_ezwbhq2nrvadncyjpwjlluuona	117	30	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	117	31	results	result	NNS
work_ezwbhq2nrvadncyjpwjlluuona	117	32	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	117	33	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	117	34	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	117	35	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	117	36	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	117	37	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	117	38	represent	represent	VB
work_ezwbhq2nrvadncyjpwjlluuona	117	39	new	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	117	40	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	117	41	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	117	42	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	117	43	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	117	44	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	117	45	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	118	1	168	168	CD
work_ezwbhq2nrvadncyjpwjlluuona	118	2	O’TOOLE	O’TOOLE	NNP
work_ezwbhq2nrvadncyjpwjlluuona	118	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	118	4	DEFFENBACHER	DEFFENBACHER	NNP
work_ezwbhq2nrvadncyjpwjlluuona	118	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	118	6	ABDI	ABDI	NNP
work_ezwbhq2nrvadncyjpwjlluuona	118	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	118	8	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	118	9	BARTLETT	BARTLETT	NNP
work_ezwbhq2nrvadncyjpwjlluuona	118	10	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	118	11	calculated	calculate	VBN
work_ezwbhq2nrvadncyjpwjlluuona	118	12	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	118	13	taking	take	VBG
work_ezwbhq2nrvadncyjpwjlluuona	118	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	118	15	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	118	16	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	118	17	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	118	18	possible	possible	JJ
work_ezwbhq2nrvadncyjpwjlluuona	118	19	pairs	pair	NNS
work_ezwbhq2nrvadncyjpwjlluuona	118	20	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	118	21	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	118	22	‘	'	``
work_ezwbhq2nrvadncyjpwjlluuona	118	23	perceived	perceive	VBN
work_ezwbhq2nrvadncyjpwjlluuona	118	24	’	'	''
work_ezwbhq2nrvadncyjpwjlluuona	118	25	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	118	26	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	119	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	119	2	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	119	3	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	119	4	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	119	5	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	119	6	indicates	indicate	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	119	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	119	8	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	119	9	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	119	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	119	11	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	119	12	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	119	13	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	119	14	identical	identical	JJ
work_ezwbhq2nrvadncyjpwjlluuona	119	15	or	or	CC
work_ezwbhq2nrvadncyjpwjlluuona	119	16	scaled	scaled	JJ
work_ezwbhq2nrvadncyjpwjlluuona	119	17	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	119	18	yielding	yield	VBG
work_ezwbhq2nrvadncyjpwjlluuona	119	19	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	119	20	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	119	21	1.0	1.0	CD
work_ezwbhq2nrvadncyjpwjlluuona	119	22	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	120	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	120	2	average	average	JJ
work_ezwbhq2nrvadncyjpwjlluuona	120	3	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	120	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	120	5	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	120	6	possible	possible	JJ
work_ezwbhq2nrvadncyjpwjlluuona	120	7	pairs	pair	NNS
work_ezwbhq2nrvadncyjpwjlluuona	120	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	120	9	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	120	10	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	120	11	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	120	12	taken	take	VBN
work_ezwbhq2nrvadncyjpwjlluuona	120	13	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	120	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	120	15	average	average	JJ
work_ezwbhq2nrvadncyjpwjlluuona	120	16	inter	inter	JJ
work_ezwbhq2nrvadncyjpwjlluuona	120	17	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	120	18	face	face	JJ
work_ezwbhq2nrvadncyjpwjlluuona	120	19	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	120	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	121	1	Several	several	JJ
work_ezwbhq2nrvadncyjpwjlluuona	121	2	conditions	condition	NNS
work_ezwbhq2nrvadncyjpwjlluuona	121	3	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	121	4	analyzed	analyze	VBN
work_ezwbhq2nrvadncyjpwjlluuona	121	5	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	122	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	122	2	each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	122	3	ma-	ma-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	122	4	jority	jority	NN
work_ezwbhq2nrvadncyjpwjlluuona	122	5	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	122	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	122	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	122	8	average	average	JJ
work_ezwbhq2nrvadncyjpwjlluuona	122	9	inter	inter	JJ
work_ezwbhq2nrvadncyjpwjlluuona	122	10	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	122	11	face	face	JJ
work_ezwbhq2nrvadncyjpwjlluuona	122	12	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	122	13	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	122	14	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	122	15	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	122	16	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	122	17	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	122	18	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	122	19	computed	compute	VBN
work_ezwbhq2nrvadncyjpwjlluuona	122	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	123	1	Furthermore	furthermore	RB
work_ezwbhq2nrvadncyjpwjlluuona	123	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	123	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	123	4	reconstructions	reconstruction	NNS
work_ezwbhq2nrvadncyjpwjlluuona	123	5	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	123	6	carried	carry	VBN
work_ezwbhq2nrvadncyjpwjlluuona	123	7	out	out	RP
work_ezwbhq2nrvadncyjpwjlluuona	123	8	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	123	9	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	123	10	numbers	number	NNS
work_ezwbhq2nrvadncyjpwjlluuona	123	11	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	123	12	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	123	13	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	123	14	look	look	VB
work_ezwbhq2nrvadncyjpwjlluuona	123	15	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	123	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	123	17	consistency	consistency	NN
work_ezwbhq2nrvadncyjpwjlluuona	123	18	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	123	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	123	20	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	123	21	effects	effect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	123	22	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	124	1	To	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	124	2	ex-	ex-	RB
work_ezwbhq2nrvadncyjpwjlluuona	124	3	plain	plain	JJ
work_ezwbhq2nrvadncyjpwjlluuona	124	4	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	124	5	latter	latter	JJ
work_ezwbhq2nrvadncyjpwjlluuona	124	6	analysis	analysis	NN
work_ezwbhq2nrvadncyjpwjlluuona	124	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	124	8	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	124	9	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	124	10	digression	digression	NN
work_ezwbhq2nrvadncyjpwjlluuona	124	11	into	into	IN
work_ezwbhq2nrvadncyjpwjlluuona	124	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	124	13	properties	property	NNS
work_ezwbhq2nrvadncyjpwjlluuona	124	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	124	15	associative	associative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	124	16	matrices	matrix	NNS
work_ezwbhq2nrvadncyjpwjlluuona	124	17	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	124	18	necessary	necessary	JJ
work_ezwbhq2nrvadncyjpwjlluuona	124	19	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	125	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	2	reconstruction	reconstruction	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	4	any	any	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	5	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	6	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	8	autoas-	autoas-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	125	9	sociative	sociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	125	10	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	11	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	125	12	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	125	13	achieved	achieve	VBN
work_ezwbhq2nrvadncyjpwjlluuona	125	14	either	either	RB
work_ezwbhq2nrvadncyjpwjlluuona	125	15	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	16	Equa-	Equa-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	125	17	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	18	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	125	19	2	2	CD
work_ezwbhq2nrvadncyjpwjlluuona	125	20	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	125	21	or	or	CC
work_ezwbhq2nrvadncyjpwjlluuona	125	22	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	125	23	equivalently	equivalently	RB
work_ezwbhq2nrvadncyjpwjlluuona	125	24	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	125	25	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	26	taking	take	VBG
work_ezwbhq2nrvadncyjpwjlluuona	125	27	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	28	weighted	weighted	JJ
work_ezwbhq2nrvadncyjpwjlluuona	125	29	sum	sum	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	30	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	31	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	32	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	125	33	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	34	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	35	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	36	A	A	NNP
work_ezwbhq2nrvadncyjpwjlluuona	125	37	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	125	38	where	where	WRB
work_ezwbhq2nrvadncyjpwjlluuona	125	39	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	40	weights	weight	NNS
work_ezwbhq2nrvadncyjpwjlluuona	125	41	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	42	each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	43	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	44	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	45	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	46	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	47	fi	fi	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	48	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	125	49	equal	equal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	125	50	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	51	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	52	dot	dot	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	53	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	125	54	product	product	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	55	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	56	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	57	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	58	vector	vector	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	59	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	125	60	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	61	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	62	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	125	63	multiplied	multiply	VBN
work_ezwbhq2nrvadncyjpwjlluuona	125	64	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	65	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	66	eigenvalue	eigenvalue	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	67	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	68	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	69	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	70	—	—	:
work_ezwbhq2nrvadncyjpwjlluuona	125	71	when	when	WRB
work_ezwbhq2nrvadncyjpwjlluuona	125	72	error	error	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	73	correction	correction	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	74	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	125	75	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	125	76	being	be	VBG
work_ezwbhq2nrvadncyjpwjlluuona	125	77	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	125	78	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	125	79	since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	80	error	error	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	81	correction	correction	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	82	has	have	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	125	83	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	84	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	85	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	125	86	equal-	equal-	NN
work_ezwbhq2nrvadncyjpwjlluuona	125	87	izing	ize	VBG
work_ezwbhq2nrvadncyjpwjlluuona	125	88	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	125	89	eigenvalues	eigenvalue	NNS
work_ezwbhq2nrvadncyjpwjlluuona	125	90	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	125	91	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	126	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	126	2	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	126	3	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	126	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	126	5	error	error	NN
work_ezwbhq2nrvadncyjpwjlluuona	126	6	correction	correction	NN
work_ezwbhq2nrvadncyjpwjlluuona	126	7	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	126	8	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	126	9	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	126	10	so	so	RB
work_ezwbhq2nrvadncyjpwjlluuona	126	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	126	12	eigenvalue	eigenvalue	NN
work_ezwbhq2nrvadncyjpwjlluuona	126	13	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	126	14	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	126	15	included	include	VBN
work_ezwbhq2nrvadncyjpwjlluuona	126	16	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	126	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	126	18	weights	weight	NNS
work_ezwbhq2nrvadncyjpwjlluuona	126	19	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	127	1	Thus	thus	RB
work_ezwbhq2nrvadncyjpwjlluuona	127	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	127	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	127	4	reconstruction	reconstruction	NN
work_ezwbhq2nrvadncyjpwjlluuona	127	5	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	127	6	given	give	VBN
work_ezwbhq2nrvadncyjpwjlluuona	127	7	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	127	8	f̂i	f̂i	NNP
work_ezwbhq2nrvadncyjpwjlluuona	127	9	=	=	NFP
work_ezwbhq2nrvadncyjpwjlluuona	127	10	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	127	11	fi	fi	NN
work_ezwbhq2nrvadncyjpwjlluuona	127	12	·	·	NFP
work_ezwbhq2nrvadncyjpwjlluuona	127	13	e1)e1	e1)e1	NNP
work_ezwbhq2nrvadncyjpwjlluuona	127	14	+	+	NFP
work_ezwbhq2nrvadncyjpwjlluuona	127	15	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	127	16	fi	fi	NN
work_ezwbhq2nrvadncyjpwjlluuona	127	17	·	·	NFP
work_ezwbhq2nrvadncyjpwjlluuona	127	18	e2)e2	e2)e2	NN
work_ezwbhq2nrvadncyjpwjlluuona	127	19	+	+	SYM
work_ezwbhq2nrvadncyjpwjlluuona	127	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	128	1	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	129	1	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	130	1	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	131	1	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	132	1	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	133	1	+	+	NFP
work_ezwbhq2nrvadncyjpwjlluuona	133	2	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	133	3	fi	fi	NN
work_ezwbhq2nrvadncyjpwjlluuona	133	4	·	·	NFP
work_ezwbhq2nrvadncyjpwjlluuona	133	5	e`)e	e`)e	XX
work_ezwbhq2nrvadncyjpwjlluuona	133	6	`	`	''
work_ezwbhq2nrvadncyjpwjlluuona	133	7	+	+	SYM
work_ezwbhq2nrvadncyjpwjlluuona	133	8	·	·	NFP
work_ezwbhq2nrvadncyjpwjlluuona	133	9	·	·	NFP
work_ezwbhq2nrvadncyjpwjlluuona	133	10	·	·	NFP
work_ezwbhq2nrvadncyjpwjlluuona	133	11	+	+	NFP
work_ezwbhq2nrvadncyjpwjlluuona	133	12	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	133	13	fi	fi	NN
work_ezwbhq2nrvadncyjpwjlluuona	133	14	·	·	NFP
work_ezwbhq2nrvadncyjpwjlluuona	133	15	en)en	en)en	NNP
work_ezwbhq2nrvadncyjpwjlluuona	133	16	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	133	17	4	4	CD
work_ezwbhq2nrvadncyjpwjlluuona	133	18	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	133	19	where	where	WRB
work_ezwbhq2nrvadncyjpwjlluuona	133	20	e	e	NNP
work_ezwbhq2nrvadncyjpwjlluuona	133	21	`	`	''
work_ezwbhq2nrvadncyjpwjlluuona	133	22	indicates	indicate	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	133	23	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	133	24	`	`	''
work_ezwbhq2nrvadncyjpwjlluuona	133	25	-th	-th	:
work_ezwbhq2nrvadncyjpwjlluuona	133	26	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	133	27	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	134	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	134	2	reconstruction	reconstruction	NN
work_ezwbhq2nrvadncyjpwjlluuona	134	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	134	4	each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	134	5	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	134	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	134	7	then	then	RB
work_ezwbhq2nrvadncyjpwjlluuona	134	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	134	9	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	134	10	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	134	11	quantified	quantify	VBN
work_ezwbhq2nrvadncyjpwjlluuona	134	12	precisely	precisely	RB
work_ezwbhq2nrvadncyjpwjlluuona	134	13	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	134	14	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	134	15	list	list	NN
work_ezwbhq2nrvadncyjpwjlluuona	134	16	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	134	17	coefficients	coefficient	NNS
work_ezwbhq2nrvadncyjpwjlluuona	134	18	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	134	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	134	20	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	134	21	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	134	22	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	134	23	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	134	24	A4	A4	NNP
work_ezwbhq2nrvadncyjpwjlluuona	134	25	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	135	1	Returning	return	VBG
work_ezwbhq2nrvadncyjpwjlluuona	135	2	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	135	3	our	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	135	4	digression	digression	NN
work_ezwbhq2nrvadncyjpwjlluuona	135	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	135	6	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	135	7	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	135	8	dear	dear	JJ
work_ezwbhq2nrvadncyjpwjlluuona	135	9	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	135	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	135	11	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	135	12	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	135	13	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	135	14	recalled	recall	VBN
work_ezwbhq2nrvadncyjpwjlluuona	135	15	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	135	16	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	135	17	second	second	JJ
work_ezwbhq2nrvadncyjpwjlluuona	135	18	procedure	procedure	NN
work_ezwbhq2nrvadncyjpwjlluuona	135	19	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	135	20	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	135	21	or	or	CC
work_ezwbhq2nrvadncyjpwjlluuona	135	22	any	any	DT
work_ezwbhq2nrvadncyjpwjlluuona	135	23	subset	subset	NN
work_ezwbhq2nrvadncyjpwjlluuona	135	24	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	135	25	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	135	26	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	136	1	Since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	136	2	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	136	3	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	136	4	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	136	5	ordered	order	VBN
work_ezwbhq2nrvadncyjpwjlluuona	136	6	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	136	7	importance	importance	NN
work_ezwbhq2nrvadncyjpwjlluuona	136	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	136	9	contribution	contribution	NN
work_ezwbhq2nrvadncyjpwjlluuona	136	10	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	136	11	their	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	136	12	associ-	associ-	VBP
work_ezwbhq2nrvadncyjpwjlluuona	136	13	ated	ated	JJ
work_ezwbhq2nrvadncyjpwjlluuona	136	14	eigenvalues	eigenvalue	NNS
work_ezwbhq2nrvadncyjpwjlluuona	136	15	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	136	16	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	136	17	recalled	recall	VBD
work_ezwbhq2nrvadncyjpwjlluuona	136	18	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	136	19	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	136	20	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	136	21	numbers	number	NNS
work_ezwbhq2nrvadncyjpwjlluuona	136	22	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	136	23	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	136	24	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	136	25	test	test	VB
work_ezwbhq2nrvadncyjpwjlluuona	136	26	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	136	27	consistency	consistency	NN
work_ezwbhq2nrvadncyjpwjlluuona	136	28	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	136	29	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	136	30	results	result	NNS
work_ezwbhq2nrvadncyjpwjlluuona	136	31	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	136	32	more	more	JJR
work_ezwbhq2nrvadncyjpwjlluuona	136	33	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	136	34	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	136	35	included	include	VBN
work_ezwbhq2nrvadncyjpwjlluuona	136	36	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	137	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	137	2	results	result	NNS
work_ezwbhq2nrvadncyjpwjlluuona	137	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	137	4	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	137	5	analysis	analysis	NN
work_ezwbhq2nrvadncyjpwjlluuona	137	6	appear	appear	VBP
work_ezwbhq2nrvadncyjpwjlluuona	137	7	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	137	8	Figure	Figure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	137	9	2	2	CD
work_ezwbhq2nrvadncyjpwjlluuona	137	10	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	137	11	a	a	NN
work_ezwbhq2nrvadncyjpwjlluuona	137	12	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	137	13	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	137	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	137	15	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	137	16	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	137	17	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	137	18	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	137	19	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	137	20	Figure	Figure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	137	21	2	2	CD
work_ezwbhq2nrvadncyjpwjlluuona	137	22	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	137	23	b	b	NN
work_ezwbhq2nrvadncyjpwjlluuona	137	24	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	137	25	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	137	26	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	137	27	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	137	28	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	137	29	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	137	30	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	138	1	Average	average	JJ
work_ezwbhq2nrvadncyjpwjlluuona	138	2	interface	interface	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	3	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	138	5	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	6	defined	define	VBN
work_ezwbhq2nrvadncyjpwjlluuona	138	7	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	8	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	138	9	av-	av-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	138	10	erage	erage	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	11	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	12	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	13	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	138	14	possible	possible	JJ
work_ezwbhq2nrvadncyjpwjlluuona	138	15	pairs	pair	NNS
work_ezwbhq2nrvadncyjpwjlluuona	138	16	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	17	recon-	recon-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	138	18	structed	structe	VBD
work_ezwbhq2nrvadncyjpwjlluuona	138	19	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	138	20	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	138	21	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	22	coded	code	VBN
work_ezwbhq2nrvadncyjpwjlluuona	138	23	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	24	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	138	25	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	26	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	27	coefficients	coefficient	NNS
work_ezwbhq2nrvadncyjpwjlluuona	138	28	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	138	29	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	138	30	reconstruct	reconstruct	VB
work_ezwbhq2nrvadncyjpwjlluuona	138	31	them	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	138	32	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	138	33	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	138	34	appears	appear	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	138	35	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	36	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	138	37	y	y	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	38	axis	axis	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	39	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	138	40	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	41	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	42	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	138	43	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	138	44	plotted	plot	VBN
work_ezwbhq2nrvadncyjpwjlluuona	138	45	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	138	46	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	138	47	x	x	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	48	axis	axis	NN
work_ezwbhq2nrvadncyjpwjlluuona	138	49	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	139	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	139	2	both	both	DT
work_ezwbhq2nrvadncyjpwjlluuona	139	3	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	139	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	139	5	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	139	6	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	139	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	139	8	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	139	9	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	139	10	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	139	11	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	139	12	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	139	13	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	139	14	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	139	15	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	139	16	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	139	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	139	18	average	average	NN
work_ezwbhq2nrvadncyjpwjlluuona	139	19	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	139	20	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	139	21	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	139	22	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	139	23	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	139	24	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	139	25	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	139	26	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	140	1	Thus	thus	RB
work_ezwbhq2nrvadncyjpwjlluuona	140	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	140	3	when	when	WRB
work_ezwbhq2nrvadncyjpwjlluuona	140	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	140	5	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	140	6	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	140	7	trained	train	VBN
work_ezwbhq2nrvadncyjpwjlluuona	140	8	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	140	9	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	140	10	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	140	11	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	140	12	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	140	13	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	140	14	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	140	15	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	140	16	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	140	17	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	140	18	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	140	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	140	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	140	21	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	140	22	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	140	23	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	140	24	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	140	25	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	140	26	cre-	cre-	XX
work_ezwbhq2nrvadncyjpwjlluuona	140	27	ates	ate	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	140	28	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	140	29	distinct	distinct	JJ
work_ezwbhq2nrvadncyjpwjlluuona	140	30	codings	coding	NNS
work_ezwbhq2nrvadncyjpwjlluuona	140	31	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	140	32	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	140	33	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	140	34	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	140	35	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	141	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	141	2	finding	finding	NN
work_ezwbhq2nrvadncyjpwjlluuona	141	3	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	141	4	reminiscent	reminiscent	JJ
work_ezwbhq2nrvadncyjpwjlluuona	141	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	141	6	Feingold	Feingold	NNP
work_ezwbhq2nrvadncyjpwjlluuona	141	7	’s	’s	''
work_ezwbhq2nrvadncyjpwjlluuona	141	8	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	141	9	1914	1914	CD
work_ezwbhq2nrvadncyjpwjlluuona	141	10	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	141	11	quote	quote	UH
work_ezwbhq2nrvadncyjpwjlluuona	141	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	142	1	Simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	142	2	2	2	CD
work_ezwbhq2nrvadncyjpwjlluuona	142	3	As	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	142	4	previously	previously	RB
work_ezwbhq2nrvadncyjpwjlluuona	142	5	noted	note	VBN
work_ezwbhq2nrvadncyjpwjlluuona	142	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	142	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	142	8	capacity	capacity	NN
work_ezwbhq2nrvadncyjpwjlluuona	142	9	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	142	10	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	142	11	autoas-	autoas-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	142	12	sociative	sociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	142	13	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	142	14	without	without	IN
work_ezwbhq2nrvadncyjpwjlluuona	142	15	error	error	NN
work_ezwbhq2nrvadncyjpwjlluuona	142	16	correction	correction	NN
work_ezwbhq2nrvadncyjpwjlluuona	142	17	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	142	18	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	142	19	estimated	estimate	VBN
work_ezwbhq2nrvadncyjpwjlluuona	142	20	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	142	21	approximately	approximately	RB
work_ezwbhq2nrvadncyjpwjlluuona	142	22	15	15	CD
work_ezwbhq2nrvadncyjpwjlluuona	142	23	%	%	NN
work_ezwbhq2nrvadncyjpwjlluuona	142	24	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	142	25	its	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	142	26	dimension-	dimension-	NN
work_ezwbhq2nrvadncyjpwjlluuona	142	27	ality	ality	NN
work_ezwbhq2nrvadncyjpwjlluuona	142	28	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	143	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	143	2	estimate	estimate	NN
work_ezwbhq2nrvadncyjpwjlluuona	143	3	assumes	assume	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	143	4	random	random	JJ
work_ezwbhq2nrvadncyjpwjlluuona	143	5	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	143	6	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	144	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	144	2	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	144	3	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	144	4	used	use	VBD
work_ezwbhq2nrvadncyjpwjlluuona	144	5	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	144	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	144	7	dimensionality	dimensionality	NN
work_ezwbhq2nrvadncyjpwjlluuona	144	8	33	33	CD
work_ezwbhq2nrvadncyjpwjlluuona	144	9	975	975	CD
work_ezwbhq2nrvadncyjpwjlluuona	144	10	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	144	11	so	so	RB
work_ezwbhq2nrvadncyjpwjlluuona	144	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	144	13	capacity	capacity	NN
work_ezwbhq2nrvadncyjpwjlluuona	144	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	144	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	144	16	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	144	17	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	144	18	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	144	19	roughly	roughly	RB
work_ezwbhq2nrvadncyjpwjlluuona	144	20	5096	5096	CD
work_ezwbhq2nrvadncyjpwjlluuona	144	21	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	144	22	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	145	1	While	while	IN
work_ezwbhq2nrvadncyjpwjlluuona	145	2	there	there	EX
work_ezwbhq2nrvadncyjpwjlluuona	145	3	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	145	4	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	145	5	differences	difference	NNS
work_ezwbhq2nrvadncyjpwjlluuona	145	6	be-	be-	VBN
work_ezwbhq2nrvadncyjpwjlluuona	145	7	tween	tween	NN
work_ezwbhq2nrvadncyjpwjlluuona	145	8	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	145	9	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	145	10	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	145	11	those	those	DT
work_ezwbhq2nrvadncyjpwjlluuona	145	12	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	145	13	which	which	WDT
work_ezwbhq2nrvadncyjpwjlluuona	145	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	145	15	capacity	capacity	NN
work_ezwbhq2nrvadncyjpwjlluuona	145	16	estimates	estimate	NNS
work_ezwbhq2nrvadncyjpwjlluuona	145	17	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	145	18	derived	derive	VBN
work_ezwbhq2nrvadncyjpwjlluuona	145	19	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	145	20	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	145	21	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	145	22	inverse	inverse	JJ
work_ezwbhq2nrvadncyjpwjlluuona	145	23	effects	effect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	145	24	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	145	25	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	145	26	capacity	capacity	NN
work_ezwbhq2nrvadncyjpwjlluuona	145	27	estimates	estimate	NNS
work_ezwbhq2nrvadncyjpwjlluuona	145	28	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	146	1	First	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	146	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	146	3	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	146	4	used	use	VBD
work_ezwbhq2nrvadncyjpwjlluuona	146	5	error	error	NN
work_ezwbhq2nrvadncyjpwjlluuona	146	6	correction	correction	NN
work_ezwbhq2nrvadncyjpwjlluuona	146	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	146	8	which	which	WDT
work_ezwbhq2nrvadncyjpwjlluuona	146	9	improved	improve	VBD
work_ezwbhq2nrvadncyjpwjlluuona	146	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	146	11	capacity	capacity	NN
work_ezwbhq2nrvadncyjpwjlluuona	146	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	146	13	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	146	14	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	146	15	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	147	1	Secondly	secondly	RB
work_ezwbhq2nrvadncyjpwjlluuona	147	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	147	3	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	147	4	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	147	5	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	147	6	random	random	JJ
work_ezwbhq2nrvadncyjpwjlluuona	147	7	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	147	8	but	but	CC
work_ezwbhq2nrvadncyjpwjlluuona	147	9	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	147	10	highly	highly	RB
work_ezwbhq2nrvadncyjpwjlluuona	147	11	correlated	correlate	VBN
work_ezwbhq2nrvadncyjpwjlluuona	147	12	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	147	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	147	14	factor	factor	NN
work_ezwbhq2nrvadncyjpwjlluuona	147	15	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	147	16	lessens	lessen	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	147	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	147	18	capacity	capacity	NN
work_ezwbhq2nrvadncyjpwjlluuona	147	19	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	148	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	148	2	any	any	DT
work_ezwbhq2nrvadncyjpwjlluuona	148	3	ease	ease	NN
work_ezwbhq2nrvadncyjpwjlluuona	148	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	148	5	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	148	6	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	148	7	clear	clear	JJ
work_ezwbhq2nrvadncyjpwjlluuona	148	8	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	148	9	Simulation	Simulation	NNP
work_ezwbhq2nrvadncyjpwjlluuona	148	10	1	1	CD
work_ezwbhq2nrvadncyjpwjlluuona	148	11	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	148	12	100	100	CD
work_ezwbhq2nrvadncyjpwjlluuona	148	13	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	148	14	did	do	VBD
work_ezwbhq2nrvadncyjpwjlluuona	148	15	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	148	16	challenge	challenge	VB
work_ezwbhq2nrvadncyjpwjlluuona	148	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	148	18	capacities	capacity	NNS
work_ezwbhq2nrvadncyjpwjlluuona	148	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	148	20	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	148	21	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	148	22	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	149	1	Since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	149	2	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	149	3	had	have	VBD
work_ezwbhq2nrvadncyjpwjlluuona	149	4	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	149	5	limited	limit	VBN
work_ezwbhq2nrvadncyjpwjlluuona	149	6	data	data	NN
work_ezwbhq2nrvadncyjpwjlluuona	149	7	base	base	NN
work_ezwbhq2nrvadncyjpwjlluuona	149	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	149	9	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	149	10	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	149	11	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	149	12	explored	explore	VBD
work_ezwbhq2nrvadncyjpwjlluuona	149	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	149	14	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	149	15	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	149	16	methods	method	NNS
work_ezwbhq2nrvadncyjpwjlluuona	149	17	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	149	18	de-	de-	RB
work_ezwbhq2nrvadncyjpwjlluuona	149	19	grading	grade	VBG
work_ezwbhq2nrvadncyjpwjlluuona	149	20	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	149	21	system	system	NN
work_ezwbhq2nrvadncyjpwjlluuona	149	22	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	149	23	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	149	24	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	150	1	At	at	RB
work_ezwbhq2nrvadncyjpwjlluuona	150	2	least	least	RBS
work_ezwbhq2nrvadncyjpwjlluuona	150	3	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	150	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	150	5	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	150	6	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	150	7	interesting	interesting	JJ
work_ezwbhq2nrvadncyjpwjlluuona	150	8	psychologically	psychologically	RB
work_ezwbhq2nrvadncyjpwjlluuona	150	9	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	150	10	would	would	MD
work_ezwbhq2nrvadncyjpwjlluuona	150	11	merit	merit	VB
work_ezwbhq2nrvadncyjpwjlluuona	150	12	attention	attention	NN
work_ezwbhq2nrvadncyjpwjlluuona	150	13	regardless	regardless	RB
work_ezwbhq2nrvadncyjpwjlluuona	150	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	150	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	150	16	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	150	17	con-	con-	NN
work_ezwbhq2nrvadncyjpwjlluuona	150	18	straints	straint	NNS
work_ezwbhq2nrvadncyjpwjlluuona	150	19	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	151	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	151	2	method	method	NN
work_ezwbhq2nrvadncyjpwjlluuona	151	3	draws	draw	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	151	4	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	151	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	151	6	metric	metric	JJ
work_ezwbhq2nrvadncyjpwjlluuona	151	7	multidi-	multidi-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	151	8	mensional	mensional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	151	9	sealing	sealing	NNP
work_ezwbhq2nrvadncyjpwjlluuona	151	10	analogy	analogy	NN
work_ezwbhq2nrvadncyjpwjlluuona	151	11	cited	cite	VBD
work_ezwbhq2nrvadncyjpwjlluuona	151	12	previously	previously	RB
work_ezwbhq2nrvadncyjpwjlluuona	151	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	152	1	Multidimensional	multidimensional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	152	2	scaling	scaling	NN
work_ezwbhq2nrvadncyjpwjlluuona	152	3	tries	try	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	152	4	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	152	5	represent	represent	VB
work_ezwbhq2nrvadncyjpwjlluuona	152	6	space	space	NN
work_ezwbhq2nrvadncyjpwjlluuona	152	7	relations	relation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	152	8	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	152	9	entities	entity	NNS
work_ezwbhq2nrvadncyjpwjlluuona	152	10	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	152	11	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	152	12	stimulus	stimulus	JJ
work_ezwbhq2nrvadncyjpwjlluuona	152	13	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	152	14	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	152	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	152	16	smallest	small	JJS
work_ezwbhq2nrvadncyjpwjlluuona	152	17	dimensional	dimensional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	152	18	space	space	NN
work_ezwbhq2nrvadncyjpwjlluuona	152	19	possible	possible	JJ
work_ezwbhq2nrvadncyjpwjlluuona	152	20	while	while	IN
work_ezwbhq2nrvadncyjpwjlluuona	152	21	accounting	account	VBG
work_ezwbhq2nrvadncyjpwjlluuona	152	22	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	152	23	some	some	DT
work_ezwbhq2nrvadncyjpwjlluuona	152	24	experimenter	experimenter	NN
work_ezwbhq2nrvadncyjpwjlluuona	152	25	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	152	26	set	set	VBN
work_ezwbhq2nrvadncyjpwjlluuona	152	27	criterion	criterion	NN
work_ezwbhq2nrvadncyjpwjlluuona	152	28	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	152	29	variance	variance	NN
work_ezwbhq2nrvadncyjpwjlluuona	152	30	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	153	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	153	2	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	153	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	153	4	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	153	5	associative	associative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	153	6	mem-	mem-	NN
work_ezwbhq2nrvadncyjpwjlluuona	153	7	4	4	CD
work_ezwbhq2nrvadncyjpwjlluuona	153	8	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	153	9	strong	strong	JJ
work_ezwbhq2nrvadncyjpwjlluuona	153	10	analogy	analogy	NN
work_ezwbhq2nrvadncyjpwjlluuona	153	11	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	153	12	metric	metric	JJ
work_ezwbhq2nrvadncyjpwjlluuona	153	13	multidimensional	multidimensional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	153	14	scaling	scaling	NN
work_ezwbhq2nrvadncyjpwjlluuona	153	15	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	153	16	present	present	JJ
work_ezwbhq2nrvadncyjpwjlluuona	153	17	here	here	RB
work_ezwbhq2nrvadncyjpwjlluuona	153	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	154	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	154	2	axes	axis	NNS
work_ezwbhq2nrvadncyjpwjlluuona	154	3	or	or	CC
work_ezwbhq2nrvadncyjpwjlluuona	154	4	dimensions	dimension	NNS
work_ezwbhq2nrvadncyjpwjlluuona	154	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	154	6	met-	met-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	154	7	ric	ric	NNP
work_ezwbhq2nrvadncyjpwjlluuona	154	8	multidimensional	multidimensional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	154	9	scaling	scaling	NN
work_ezwbhq2nrvadncyjpwjlluuona	154	10	solutions	solution	NNS
work_ezwbhq2nrvadncyjpwjlluuona	154	11	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	154	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	154	13	eigen-	eigen-	NN
work_ezwbhq2nrvadncyjpwjlluuona	154	14	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	154	15	ordered	order	VBN
work_ezwbhq2nrvadncyjpwjlluuona	154	16	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	154	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	154	18	magnitude	magnitude	NN
work_ezwbhq2nrvadncyjpwjlluuona	154	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	154	20	their	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	154	21	associated	associate	VBN
work_ezwbhq2nrvadncyjpwjlluuona	154	22	eigenvalues	eigenvalue	NNS
work_ezwbhq2nrvadncyjpwjlluuona	154	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	155	1	Thus	thus	RB
work_ezwbhq2nrvadncyjpwjlluuona	155	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	155	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	155	4	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	155	5	axis	axis	NN
work_ezwbhq2nrvadncyjpwjlluuona	155	6	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	155	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	155	8	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	155	9	eigenvec-	eigenvec-	NN
work_ezwbhq2nrvadncyjpwjlluuona	155	10	tor	tor	NN
work_ezwbhq2nrvadncyjpwjlluuona	155	11	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	155	12	etc	etc	FW
work_ezwbhq2nrvadncyjpwjlluuona	155	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	156	1	Typically	typically	RB
work_ezwbhq2nrvadncyjpwjlluuona	156	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	156	3	multidimensional	multidimensional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	156	4	scaling	scale	VBG
work_ezwbhq2nrvadncyjpwjlluuona	156	5	solutions	solution	NNS
work_ezwbhq2nrvadncyjpwjlluuona	156	6	use	use	VBP
work_ezwbhq2nrvadncyjpwjlluuona	156	7	as	as	RB
work_ezwbhq2nrvadncyjpwjlluuona	156	8	many	many	JJ
work_ezwbhq2nrvadncyjpwjlluuona	156	9	axes	axis	NNS
work_ezwbhq2nrvadncyjpwjlluuona	156	10	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	156	11	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	156	12	needed	need	VBN
work_ezwbhq2nrvadncyjpwjlluuona	156	13	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	156	14	account	account	VB
work_ezwbhq2nrvadncyjpwjlluuona	156	15	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	156	16	some	some	DT
work_ezwbhq2nrvadncyjpwjlluuona	156	17	experimenter	experimenter	NN
work_ezwbhq2nrvadncyjpwjlluuona	156	18	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	156	19	specified	specify	VBN
work_ezwbhq2nrvadncyjpwjlluuona	156	20	proportion	proportion	NN
work_ezwbhq2nrvadncyjpwjlluuona	156	21	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	156	22	variance	variance	NN
work_ezwbhq2nrvadncyjpwjlluuona	156	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	157	1	SIMULATING	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	157	2	THE	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	157	3	OTHER	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	157	4	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	157	5	RACE	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	157	6	EFFECT	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	157	7	169	169	CD
work_ezwbhq2nrvadncyjpwjlluuona	157	8	Figure	figure	NN
work_ezwbhq2nrvadncyjpwjlluuona	157	9	2	2	CD
work_ezwbhq2nrvadncyjpwjlluuona	157	10	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	158	1	Average	average	JJ
work_ezwbhq2nrvadncyjpwjlluuona	158	2	inter	inter	JJ
work_ezwbhq2nrvadncyjpwjlluuona	158	3	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	158	4	face	face	JJ
work_ezwbhq2nrvadncyjpwjlluuona	158	5	similarity	similarity	NN
work_ezwbhq2nrvadncyjpwjlluuona	158	6	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	158	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	158	8	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	158	9	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	158	10	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	158	11	Caucasian	Caucasian	NNP
work_ezwbhq2nrvadncyjpwjlluuona	158	12	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	158	13	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	158	14	b	b	LS
work_ezwbhq2nrvadncyjpwjlluuona	158	15	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	158	16	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	158	17	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	158	18	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	158	19	95	95	CD
work_ezwbhq2nrvadncyjpwjlluuona	158	20	%	%	NN
work_ezwbhq2nrvadncyjpwjlluuona	158	21	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	158	22	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	158	23	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	158	24	plotted	plot	VBD
work_ezwbhq2nrvadncyjpwjlluuona	158	25	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	158	26	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	158	27	function	function	NN
work_ezwbhq2nrvadncyjpwjlluuona	158	28	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	158	29	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	158	30	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	158	31	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	158	32	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	158	33	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	158	34	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	158	35	reconstruct	reconstruct	VB
work_ezwbhq2nrvadncyjpwjlluuona	158	36	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	158	37	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	158	38	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	159	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	159	2	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	159	3	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	159	4	5	5	CD
work_ezwbhq2nrvadncyjpwjlluuona	159	5	%	%	NN
work_ezwbhq2nrvadncyjpwjlluuona	159	6	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	159	7	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	159	8	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	159	9	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	159	10	both	both	DT
work_ezwbhq2nrvadncyjpwjlluuona	159	11	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	159	12	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	159	13	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	159	14	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	159	15	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	159	16	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	159	17	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	159	18	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	159	19	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	159	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	159	21	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	159	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	159	23	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	159	24	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	159	25	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	160	1	170	170	CD
work_ezwbhq2nrvadncyjpwjlluuona	160	2	O’TOOLE	o’toole	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	160	4	DEFFENBACHER	DEFFENBACHER	NNP
work_ezwbhq2nrvadncyjpwjlluuona	160	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	160	6	ABDI	ABDI	NNP
work_ezwbhq2nrvadncyjpwjlluuona	160	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	160	8	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	160	9	BARTLETT	BARTLETT	NNP
work_ezwbhq2nrvadncyjpwjlluuona	160	10	ory	ory	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	11	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	160	12	equivalent	equivalent	JJ
work_ezwbhq2nrvadncyjpwjlluuona	160	13	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	160	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	160	15	axes	axis	NNS
work_ezwbhq2nrvadncyjpwjlluuona	160	16	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	160	17	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	160	18	multidimen-	multidimen-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	160	19	sional	sional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	160	20	scaling	scaling	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	21	solution	solution	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	22	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	160	23	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	160	24	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	160	25	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	26	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	160	27	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	160	28	largest	large	JJS
work_ezwbhq2nrvadncyjpwjlluuona	160	29	positive	positive	JJ
work_ezwbhq2nrvadncyjpwjlluuona	160	30	eigenvalue	eigenvalue	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	31	accounting	account	VBG
work_ezwbhq2nrvadncyjpwjlluuona	160	32	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	160	33	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	160	34	largest	large	JJS
work_ezwbhq2nrvadncyjpwjlluuona	160	35	proportion	proportion	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	36	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	160	37	variance	variance	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	38	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	160	39	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	160	40	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	160	41	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	42	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	160	43	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	160	44	second	second	RB
work_ezwbhq2nrvadncyjpwjlluuona	160	45	largest	large	JJS
work_ezwbhq2nrvadncyjpwjlluuona	160	46	eigenvalue	eigenvalue	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	47	accounting	account	VBG
work_ezwbhq2nrvadncyjpwjlluuona	160	48	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	160	49	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	160	50	second	second	RB
work_ezwbhq2nrvadncyjpwjlluuona	160	51	largest	large	JJS
work_ezwbhq2nrvadncyjpwjlluuona	160	52	proportion	proportion	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	53	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	160	54	variance	variance	NN
work_ezwbhq2nrvadncyjpwjlluuona	160	55	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	160	56	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	160	57	so	so	RB
work_ezwbhq2nrvadncyjpwjlluuona	160	58	on	on	RB
work_ezwbhq2nrvadncyjpwjlluuona	160	59	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	161	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	161	2	maximum	maximum	JJ
work_ezwbhq2nrvadncyjpwjlluuona	161	3	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	161	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	161	5	dimensions	dimension	NNS
work_ezwbhq2nrvadncyjpwjlluuona	161	6	needed	need	VBN
work_ezwbhq2nrvadncyjpwjlluuona	161	7	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	161	8	account	account	VB
work_ezwbhq2nrvadncyjpwjlluuona	161	9	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	161	10	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	161	11	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	161	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	161	13	variance	variance	NN
work_ezwbhq2nrvadncyjpwjlluuona	161	14	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	161	15	equal	equal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	161	16	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	161	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	161	18	rank	rank	NN
work_ezwbhq2nrvadncyjpwjlluuona	161	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	161	20	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	161	21	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	161	22	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	161	23	i.e.	i.e.	FW
work_ezwbhq2nrvadncyjpwjlluuona	162	1	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	162	2	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	162	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	162	4	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	162	5	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	162	6	non	non	JJ
work_ezwbhq2nrvadncyjpwjlluuona	162	7	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	162	8	zero	zero	JJ
work_ezwbhq2nrvadncyjpwjlluuona	162	9	eigenvalues	eigenvalue	NNS
work_ezwbhq2nrvadncyjpwjlluuona	162	10	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	162	11	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	163	1	Frequently	frequently	RB
work_ezwbhq2nrvadncyjpwjlluuona	163	2	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	163	3	multi-	multi-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	163	4	dimensional	dimensional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	163	5	scaling	scaling	NN
work_ezwbhq2nrvadncyjpwjlluuona	163	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	163	7	however	however	RB
work_ezwbhq2nrvadncyjpwjlluuona	163	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	163	9	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	163	10	acceptably	acceptably	RB
work_ezwbhq2nrvadncyjpwjlluuona	163	11	large	large	JJ
work_ezwbhq2nrvadncyjpwjlluuona	163	12	proportion	proportion	NN
work_ezwbhq2nrvadncyjpwjlluuona	163	13	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	163	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	163	15	variance	variance	NN
work_ezwbhq2nrvadncyjpwjlluuona	163	16	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	163	17	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	163	18	accounted	account	VBN
work_ezwbhq2nrvadncyjpwjlluuona	163	19	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	163	20	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	163	21	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	163	22	very	very	RB
work_ezwbhq2nrvadncyjpwjlluuona	163	23	small	small	JJ
work_ezwbhq2nrvadncyjpwjlluuona	163	24	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	163	25	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	163	26	dimensions	dimension	NNS
work_ezwbhq2nrvadncyjpwjlluuona	163	27	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	163	28	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	163	29	thus	thus	RB
work_ezwbhq2nrvadncyjpwjlluuona	163	30	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	163	31	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	163	32	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	163	33	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	163	34	smaller	small	JJR
work_ezwbhq2nrvadncyjpwjlluuona	163	35	eigenvalues	eigenvalue	NNS
work_ezwbhq2nrvadncyjpwjlluuona	163	36	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	163	37	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	163	38	dis-	dis-	RB
work_ezwbhq2nrvadncyjpwjlluuona	163	39	carded	card	VBN
work_ezwbhq2nrvadncyjpwjlluuona	163	40	without	without	IN
work_ezwbhq2nrvadncyjpwjlluuona	163	41	losing	lose	VBG
work_ezwbhq2nrvadncyjpwjlluuona	163	42	much	much	JJ
work_ezwbhq2nrvadncyjpwjlluuona	163	43	information	information	NN
work_ezwbhq2nrvadncyjpwjlluuona	163	44	about	about	IN
work_ezwbhq2nrvadncyjpwjlluuona	163	45	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	163	46	structure	structure	NN
work_ezwbhq2nrvadncyjpwjlluuona	163	47	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	163	48	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	163	49	stimulus	stimulus	NN
work_ezwbhq2nrvadncyjpwjlluuona	163	50	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	163	51	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	164	1	Likewise	likewise	RB
work_ezwbhq2nrvadncyjpwjlluuona	164	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	164	3	recall	recall	NN
work_ezwbhq2nrvadncyjpwjlluuona	164	4	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	164	5	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	164	6	associative	associative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	164	7	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	164	8	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	164	9	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	164	10	carried	carry	VBN
work_ezwbhq2nrvadncyjpwjlluuona	164	11	out	out	RP
work_ezwbhq2nrvadncyjpwjlluuona	164	12	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	164	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	164	14	smaller	small	JJR
work_ezwbhq2nrvadncyjpwjlluuona	164	15	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	164	16	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	164	17	eigen-	eigen-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	164	18	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	164	19	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	164	20	cf	cf	NNP
work_ezwbhq2nrvadncyjpwjlluuona	164	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	165	1	equation	equation	NN
work_ezwbhq2nrvadncyjpwjlluuona	165	2	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	165	3	4	4	CD
work_ezwbhq2nrvadncyjpwjlluuona	165	4	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	165	5	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	165	6	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	166	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	166	2	criterion	criterion	NN
work_ezwbhq2nrvadncyjpwjlluuona	166	3	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	166	4	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	166	5	ac-	ac-	RB
work_ezwbhq2nrvadncyjpwjlluuona	166	6	ceptable	ceptable	JJ
work_ezwbhq2nrvadncyjpwjlluuona	166	7	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	166	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	166	9	dimensions	dimension	NNS
work_ezwbhq2nrvadncyjpwjlluuona	166	10	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	166	11	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	166	12	case	case	NN
work_ezwbhq2nrvadncyjpwjlluuona	166	13	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	166	14	how-	how-	NN
work_ezwbhq2nrvadncyjpwjlluuona	166	15	ever	ever	RB
work_ezwbhq2nrvadncyjpwjlluuona	166	16	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	166	17	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	166	18	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	166	19	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	166	20	maintains	maintain	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	166	21	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	166	22	acceptable	acceptable	JJ
work_ezwbhq2nrvadncyjpwjlluuona	166	23	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	166	24	but	but	CC
work_ezwbhq2nrvadncyjpwjlluuona	166	25	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	166	26	perfect	perfect	JJ
work_ezwbhq2nrvadncyjpwjlluuona	166	27	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	166	28	level	level	NN
work_ezwbhq2nrvadncyjpwjlluuona	166	29	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	166	30	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	166	31	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	166	32	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	167	1	Recognition	Recognition	NNP
work_ezwbhq2nrvadncyjpwjlluuona	167	2	Memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	167	3	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	167	4	Same-	Same-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	167	5	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	167	6	Other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	167	7	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	167	8	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	167	9	Faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	167	10	To	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	167	11	simulate	simulate	VB
work_ezwbhq2nrvadncyjpwjlluuona	167	12	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	167	13	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	167	14	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	167	15	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	167	16	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	167	17	need	need	VBP
work_ezwbhq2nrvadncyjpwjlluuona	167	18	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	167	19	model	model	VB
work_ezwbhq2nrvadncyjpwjlluuona	167	20	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	167	21	components	component	NNS
work_ezwbhq2nrvadncyjpwjlluuona	167	22	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	167	23	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	167	24	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	167	25	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	167	26	long	long	JJ
work_ezwbhq2nrvadncyjpwjlluuona	167	27	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	167	28	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	167	29	experience	experience	NN
work_ezwbhq2nrvadncyjpwjlluuona	167	30	component	component	NN
work_ezwbhq2nrvadncyjpwjlluuona	167	31	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	167	32	i.e.	i.e.	FW
work_ezwbhq2nrvadncyjpwjlluuona	168	1	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	168	2	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	168	3	history	history	NN
work_ezwbhq2nrvadncyjpwjlluuona	168	4	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	168	5	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	168	6	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	168	7	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	168	8	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	168	9	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	168	10	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	168	11	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	168	12	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	168	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	169	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	169	2	expect	expect	VBP
work_ezwbhq2nrvadncyjpwjlluuona	169	3	expe-	expe-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	169	4	rience	rience	NN
work_ezwbhq2nrvadncyjpwjlluuona	169	5	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	169	6	affect	affect	VB
work_ezwbhq2nrvadncyjpwjlluuona	169	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	169	8	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	169	9	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	169	10	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	169	11	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	169	12	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	169	13	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	169	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	169	15	ways	way	NNS
work_ezwbhq2nrvadncyjpwjlluuona	169	16	outlined	outline	VBN
work_ezwbhq2nrvadncyjpwjlluuona	169	17	above	above	RB
work_ezwbhq2nrvadncyjpwjlluuona	169	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	170	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	170	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	170	3	purpose	purpose	NN
work_ezwbhq2nrvadncyjpwjlluuona	170	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	170	5	com-	com-	NN
work_ezwbhq2nrvadncyjpwjlluuona	170	6	pleteness	pleteness	NN
work_ezwbhq2nrvadncyjpwjlluuona	170	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	170	8	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	170	9	report	report	VBP
work_ezwbhq2nrvadncyjpwjlluuona	170	10	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	170	11	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	170	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	171	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	171	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	171	3	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	171	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	171	5	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	171	6	tested	test	VBD
work_ezwbhq2nrvadncyjpwjlluuona	171	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	171	8	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	171	9	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	171	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	171	11	autoassociative	autoassociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	171	12	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	171	13	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	171	14	distinguish	distinguish	VB
work_ezwbhq2nrvadncyjpwjlluuona	171	15	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	171	16	OLD	OLD	NNP
work_ezwbhq2nrvadncyjpwjlluuona	171	17	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	171	18	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	171	19	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	171	20	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	171	21	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	171	22	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	171	23	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	171	24	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	172	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	172	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	172	3	second	second	JJ
work_ezwbhq2nrvadncyjpwjlluuona	172	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	172	5	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	172	6	added	add	VBD
work_ezwbhq2nrvadncyjpwjlluuona	172	7	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	172	8	short-	short-	NN
work_ezwbhq2nrvadncyjpwjlluuona	172	9	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	172	10	component	component	NN
work_ezwbhq2nrvadncyjpwjlluuona	172	11	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	172	12	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	172	13	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	172	14	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	172	15	which	which	WDT
work_ezwbhq2nrvadncyjpwjlluuona	172	16	consisted	consist	VBD
work_ezwbhq2nrvadncyjpwjlluuona	172	17	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	172	18	half	half	NN
work_ezwbhq2nrvadncyjpwjlluuona	172	19	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	172	20	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	172	21	half	half	JJ
work_ezwbhq2nrvadncyjpwjlluuona	172	22	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	172	23	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	172	24	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	173	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	173	2	then	then	RB
work_ezwbhq2nrvadncyjpwjlluuona	173	3	examined	examine	VBD
work_ezwbhq2nrvadncyjpwjlluuona	173	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	173	5	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	173	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	173	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	173	8	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	173	9	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	173	10	discriminate	discriminate	VB
work_ezwbhq2nrvadncyjpwjlluuona	173	11	OLD	OLD	NNP
work_ezwbhq2nrvadncyjpwjlluuona	173	12	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	173	13	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	173	14	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	173	15	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	173	16	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	173	17	additional	additional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	173	18	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	173	19	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	174	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	174	2	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	174	3	note	note	VB
work_ezwbhq2nrvadncyjpwjlluuona	174	4	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	174	5	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	174	6	do	do	VBP
work_ezwbhq2nrvadncyjpwjlluuona	174	7	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	174	8	believe	believe	VB
work_ezwbhq2nrvadncyjpwjlluuona	174	9	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	174	10	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	174	11	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	174	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	174	13	only	only	JJ
work_ezwbhq2nrvadncyjpwjlluuona	174	14	or	or	CC
work_ezwbhq2nrvadncyjpwjlluuona	174	15	even	even	RB
work_ezwbhq2nrvadncyjpwjlluuona	174	16	best	good	JJS
work_ezwbhq2nrvadncyjpwjlluuona	174	17	way	way	NN
work_ezwbhq2nrvadncyjpwjlluuona	174	18	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	174	19	simulate	simulate	VB
work_ezwbhq2nrvadncyjpwjlluuona	174	20	such	such	PDT
work_ezwbhq2nrvadncyjpwjlluuona	174	21	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	174	22	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	174	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	175	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	175	2	feel	feel	VBP
work_ezwbhq2nrvadncyjpwjlluuona	175	3	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	175	4	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	175	5	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	175	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	175	7	simplest	simple	JJS
work_ezwbhq2nrvadncyjpwjlluuona	175	8	way	way	NN
work_ezwbhq2nrvadncyjpwjlluuona	175	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	175	10	however	however	RB
work_ezwbhq2nrvadncyjpwjlluuona	175	11	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	175	12	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	175	13	so	so	RB
work_ezwbhq2nrvadncyjpwjlluuona	175	14	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	175	15	chose	choose	VBD
work_ezwbhq2nrvadncyjpwjlluuona	175	16	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	175	17	explore	explore	VB
work_ezwbhq2nrvadncyjpwjlluuona	175	18	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	175	19	method	method	NN
work_ezwbhq2nrvadncyjpwjlluuona	175	20	first	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	175	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	176	1	Method	Method	NNP
work_ezwbhq2nrvadncyjpwjlluuona	176	2	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	176	3	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	176	4	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	176	5	tested	test	VBN
work_ezwbhq2nrvadncyjpwjlluuona	176	6	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	176	7	accuracy	accuracy	NN
work_ezwbhq2nrvadncyjpwjlluuona	176	8	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	176	9	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	176	10	Yes	yes	UH
work_ezwbhq2nrvadncyjpwjlluuona	176	11	/	/	SYM
work_ezwbhq2nrvadncyjpwjlluuona	176	12	No	no	DT
work_ezwbhq2nrvadncyjpwjlluuona	176	13	procedure	procedure	NN
work_ezwbhq2nrvadncyjpwjlluuona	176	14	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	176	15	follows	follow	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	176	16	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	177	1	Learned	learn	VBN
work_ezwbhq2nrvadncyjpwjlluuona	177	2	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	177	3	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	177	4	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	177	5	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	177	6	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	177	7	OLD	OLD	NNP
work_ezwbhq2nrvadncyjpwjlluuona	177	8	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	177	9	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	177	10	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	177	11	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	177	12	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	177	13	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	177	14	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	177	15	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	177	16	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	177	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	178	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	178	2	qual-	qual-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	178	3	ity	ity	NN
work_ezwbhq2nrvadncyjpwjlluuona	178	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	178	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	178	6	reconstructions	reconstruction	NNS
work_ezwbhq2nrvadncyjpwjlluuona	178	7	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	178	8	measured	measure	VBN
work_ezwbhq2nrvadncyjpwjlluuona	178	9	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	178	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	178	11	co-	co-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	178	12	sine	sine	NN
work_ezwbhq2nrvadncyjpwjlluuona	178	13	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	178	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	178	15	original	original	JJ
work_ezwbhq2nrvadncyjpwjlluuona	178	16	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	178	17	reconstructed	reconstruct	VBD
work_ezwbhq2nrvadncyjpwjlluuona	178	18	im-	im-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	178	19	ages	age	NNS
work_ezwbhq2nrvadncyjpwjlluuona	178	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	179	1	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	179	2	Yes	yes	UH
work_ezwbhq2nrvadncyjpwjlluuona	179	3	/	/	SYM
work_ezwbhq2nrvadncyjpwjlluuona	179	4	No	no	DT
work_ezwbhq2nrvadncyjpwjlluuona	179	5	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	6	procedure	procedure	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	7	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	179	8	imple-	imple-	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	9	mented	mente	VBN
work_ezwbhq2nrvadncyjpwjlluuona	179	10	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	179	11	setting	set	VBG
work_ezwbhq2nrvadncyjpwjlluuona	179	12	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	179	13	criterion	criterion	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	14	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	15	value	value	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	16	β	β	NNP
work_ezwbhq2nrvadncyjpwjlluuona	179	17	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	179	18	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	179	19	assigning	assign	VBG
work_ezwbhq2nrvadncyjpwjlluuona	179	20	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	179	21	‘	'	``
work_ezwbhq2nrvadncyjpwjlluuona	179	22	Yes	yes	UH
work_ezwbhq2nrvadncyjpwjlluuona	179	23	’	'	''
work_ezwbhq2nrvadncyjpwjlluuona	179	24	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	179	25	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	179	26	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	179	27	which	which	WDT
work_ezwbhq2nrvadncyjpwjlluuona	179	28	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	179	29	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	30	between	between	IN
work_ezwbhq2nrvadncyjpwjlluuona	179	31	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	179	32	original	original	JJ
work_ezwbhq2nrvadncyjpwjlluuona	179	33	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	179	34	reconstructed	reconstructed	JJ
work_ezwbhq2nrvadncyjpwjlluuona	179	35	image	image	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	36	ex-	ex-	XX
work_ezwbhq2nrvadncyjpwjlluuona	179	37	ceeded	ceede	VBD
work_ezwbhq2nrvadncyjpwjlluuona	179	38	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	179	39	criterion	criterion	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	40	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	179	41	‘	'	``
work_ezwbhq2nrvadncyjpwjlluuona	179	42	No	no	UH
work_ezwbhq2nrvadncyjpwjlluuona	179	43	’	'	''
work_ezwbhq2nrvadncyjpwjlluuona	179	44	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	179	45	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	179	46	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	179	47	which	which	WDT
work_ezwbhq2nrvadncyjpwjlluuona	179	48	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	179	49	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	50	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	179	51	less	less	JJR
work_ezwbhq2nrvadncyjpwjlluuona	179	52	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	179	53	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	179	54	criterion	criterion	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	55	β	β	NN
work_ezwbhq2nrvadncyjpwjlluuona	179	56	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	180	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	180	2	most	most	RBS
work_ezwbhq2nrvadncyjpwjlluuona	180	3	direct	direct	JJ
work_ezwbhq2nrvadncyjpwjlluuona	180	4	choice	choice	NN
work_ezwbhq2nrvadncyjpwjlluuona	180	5	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	180	6	β	β	NNP
work_ezwbhq2nrvadncyjpwjlluuona	180	7	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	180	8	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	180	9	mean	mean	NN
work_ezwbhq2nrvadncyjpwjlluuona	180	10	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	180	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	180	12	cosine	cosine	NN
work_ezwbhq2nrvadncyjpwjlluuona	180	13	dis-	dis-	IN
work_ezwbhq2nrvadncyjpwjlluuona	180	14	tribution	tribution	NN
work_ezwbhq2nrvadncyjpwjlluuona	180	15	means	mean	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	180	16	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	180	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	180	18	reconstructed	reconstruct	VBN
work_ezwbhq2nrvadncyjpwjlluuona	180	19	OLD	old	JJ
work_ezwbhq2nrvadncyjpwjlluuona	180	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	180	21	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	180	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	180	23	reconstructed	reconstructed	JJ
work_ezwbhq2nrvadncyjpwjlluuona	180	24	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	180	25	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	180	26	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	181	1	Signal	Signal	NNP
work_ezwbhq2nrvadncyjpwjlluuona	181	2	detec-	detec-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	181	3	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	181	4	methodology	methodology	NN
work_ezwbhq2nrvadncyjpwjlluuona	181	5	maps	map	NNS
work_ezwbhq2nrvadncyjpwjlluuona	181	6	easily	easily	RB
work_ezwbhq2nrvadncyjpwjlluuona	181	7	onto	onto	IN
work_ezwbhq2nrvadncyjpwjlluuona	181	8	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	181	9	Yes	yes	UH
work_ezwbhq2nrvadncyjpwjlluuona	181	10	/	/	SYM
work_ezwbhq2nrvadncyjpwjlluuona	181	11	No	no	DT
work_ezwbhq2nrvadncyjpwjlluuona	181	12	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	181	13	since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	181	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	181	15	distribution	distribution	NN
work_ezwbhq2nrvadncyjpwjlluuona	181	16	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	181	17	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	181	18	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	181	19	OLD	old	JJ
work_ezwbhq2nrvadncyjpwjlluuona	181	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	181	21	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	181	22	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	181	23	thought	think	VBN
work_ezwbhq2nrvadncyjpwjlluuona	181	24	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	181	25	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	181	26	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	181	27	signal	signal	NN
work_ezwbhq2nrvadncyjpwjlluuona	181	28	distribution	distribution	NN
work_ezwbhq2nrvadncyjpwjlluuona	181	29	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	181	30	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	181	31	distribution	distribution	NN
work_ezwbhq2nrvadncyjpwjlluuona	181	32	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	181	33	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	181	34	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	181	35	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	181	36	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	181	37	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	181	38	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	181	39	noise	noise	NN
work_ezwbhq2nrvadncyjpwjlluuona	181	40	distribution	distribution	NN
work_ezwbhq2nrvadncyjpwjlluuona	181	41	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	182	1	OLD	old	JJ
work_ezwbhq2nrvadncyjpwjlluuona	182	2	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	182	3	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	182	4	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	182	5	greater	great	JJR
work_ezwbhq2nrvadncyjpwjlluuona	182	6	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	182	7	β	β	NNP
work_ezwbhq2nrvadncyjpwjlluuona	182	8	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	182	9	considered	consider	VBN
work_ezwbhq2nrvadncyjpwjlluuona	182	10	hits	hit	NNS
work_ezwbhq2nrvadncyjpwjlluuona	182	11	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	182	12	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	182	13	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	182	14	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	182	15	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	182	16	greater	great	JJR
work_ezwbhq2nrvadncyjpwjlluuona	182	17	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	182	18	β	β	NNP
work_ezwbhq2nrvadncyjpwjlluuona	182	19	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	182	20	considered	consider	VBN
work_ezwbhq2nrvadncyjpwjlluuona	182	21	false	false	JJ
work_ezwbhq2nrvadncyjpwjlluuona	182	22	alarms	alarm	NNS
work_ezwbhq2nrvadncyjpwjlluuona	182	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	183	1	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	183	2	d′	d′	CD
work_ezwbhq2nrvadncyjpwjlluuona	183	3	score	score	NN
work_ezwbhq2nrvadncyjpwjlluuona	183	4	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	183	5	then	then	RB
work_ezwbhq2nrvadncyjpwjlluuona	183	6	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	183	7	computed	compute	VBN
work_ezwbhq2nrvadncyjpwjlluuona	183	8	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	183	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	183	10	standard	standard	JJ
work_ezwbhq2nrvadncyjpwjlluuona	183	11	way	way	NN
work_ezwbhq2nrvadncyjpwjlluuona	183	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	184	1	Also	also	RB
work_ezwbhq2nrvadncyjpwjlluuona	184	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	184	3	since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	184	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	184	5	distribution	distribution	NN
work_ezwbhq2nrvadncyjpwjlluuona	184	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	184	7	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	184	8	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	184	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	184	10	signal	signal	NN
work_ezwbhq2nrvadncyjpwjlluuona	184	11	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	184	12	i.e.	i.e.	FW
work_ezwbhq2nrvadncyjpwjlluuona	185	1	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	185	2	OLD	OLD	NNP
work_ezwbhq2nrvadncyjpwjlluuona	185	3	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	185	4	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	185	5	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	185	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	185	7	distribution	distribution	NN
work_ezwbhq2nrvadncyjpwjlluuona	185	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	185	9	cosines	cosine	NNS
work_ezwbhq2nrvadncyjpwjlluuona	185	10	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	185	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	185	12	noise	noise	NN
work_ezwbhq2nrvadncyjpwjlluuona	185	13	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	185	14	i.e.	i.e.	FW
work_ezwbhq2nrvadncyjpwjlluuona	186	1	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	186	2	NEW	NEW	NNP
work_ezwbhq2nrvadncyjpwjlluuona	186	3	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	186	4	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	186	5	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	186	6	known	know	VBN
work_ezwbhq2nrvadncyjpwjlluuona	186	7	com-	com-	NN
work_ezwbhq2nrvadncyjpwjlluuona	186	8	pletely	pletely	RB
work_ezwbhq2nrvadncyjpwjlluuona	186	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	186	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	186	11	ROC	ROC	NNP
work_ezwbhq2nrvadncyjpwjlluuona	186	12	curve	curve	NN
work_ezwbhq2nrvadncyjpwjlluuona	186	13	may	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	186	14	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	186	15	plotted	plot	VBN
work_ezwbhq2nrvadncyjpwjlluuona	186	16	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	186	17	choosing	choose	VBG
work_ezwbhq2nrvadncyjpwjlluuona	186	18	β	β	NNP
work_ezwbhq2nrvadncyjpwjlluuona	186	19	values	value	NNS
work_ezwbhq2nrvadncyjpwjlluuona	186	20	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	186	21	calculating	calculate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	186	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	186	23	hit	hit	JJ
work_ezwbhq2nrvadncyjpwjlluuona	186	24	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	186	25	false	false	JJ
work_ezwbhq2nrvadncyjpwjlluuona	186	26	alarm	alarm	NN
work_ezwbhq2nrvadncyjpwjlluuona	186	27	rates	rate	NNS
work_ezwbhq2nrvadncyjpwjlluuona	186	28	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	186	29	would	would	MD
work_ezwbhq2nrvadncyjpwjlluuona	186	30	result	result	VB
work_ezwbhq2nrvadncyjpwjlluuona	186	31	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	186	32	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	186	33	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	186	34	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	186	35	crite-	crite-	NNS
work_ezwbhq2nrvadncyjpwjlluuona	186	36	ria	ria	NN
work_ezwbhq2nrvadncyjpwjlluuona	186	37	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	187	1	Results	result	NNS
work_ezwbhq2nrvadncyjpwjlluuona	187	2	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	187	3	Discussion	discussion	NN
work_ezwbhq2nrvadncyjpwjlluuona	187	4	To	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	187	5	test	test	VB
work_ezwbhq2nrvadncyjpwjlluuona	187	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	187	7	accuracy	accuracy	NN
work_ezwbhq2nrvadncyjpwjlluuona	187	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	187	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	187	10	models	model	NNS
work_ezwbhq2nrvadncyjpwjlluuona	187	11	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	187	12	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	187	13	used	use	VBD
work_ezwbhq2nrvadncyjpwjlluuona	187	14	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	187	15	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	187	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	187	17	OLD	old	JJ
work_ezwbhq2nrvadncyjpwjlluuona	187	18	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	187	19	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	187	20	100	100	CD
work_ezwbhq2nrvadncyjpwjlluuona	187	21	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	187	22	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	187	23	95	95	CD
work_ezwbhq2nrvadncyjpwjlluuona	187	24	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	187	25	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	187	26	five	five	CD
work_ezwbhq2nrvadncyjpwjlluuona	187	27	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	187	28	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	187	29	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	187	30	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	187	31	sample	sample	NN
work_ezwbhq2nrvadncyjpwjlluuona	187	32	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	187	33	120	120	CD
work_ezwbhq2nrvadncyjpwjlluuona	187	34	NEW	new	JJ
work_ezwbhq2nrvadncyjpwjlluuona	187	35	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	187	36	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	187	37	approx-	approx-	VBG
work_ezwbhq2nrvadncyjpwjlluuona	187	38	imately	imately	RB
work_ezwbhq2nrvadncyjpwjlluuona	187	39	half	half	PDT
work_ezwbhq2nrvadncyjpwjlluuona	187	40	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	187	41	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	187	42	half	half	NN
work_ezwbhq2nrvadncyjpwjlluuona	187	43	Caucasian	Caucasian	NNP
work_ezwbhq2nrvadncyjpwjlluuona	187	44	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	188	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	188	2	ac-	ac-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	188	3	curacy	curacy	NN
work_ezwbhq2nrvadncyjpwjlluuona	188	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	188	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	188	6	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	188	7	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	188	8	all	all	PDT
work_ezwbhq2nrvadncyjpwjlluuona	188	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	188	10	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	188	11	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	188	12	essentially	essentially	RB
work_ezwbhq2nrvadncyjpwjlluuona	188	13	perfect	perfect	JJ
work_ezwbhq2nrvadncyjpwjlluuona	188	14	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	189	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	189	2	degraded	degrade	VBD
work_ezwbhq2nrvadncyjpwjlluuona	189	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	189	4	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	189	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	189	6	therefore	therefore	RB
work_ezwbhq2nrvadncyjpwjlluuona	189	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	189	8	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	189	9	using	use	VBG
work_ezwbhq2nrvadncyjpwjlluuona	189	10	smaller	small	JJR
work_ezwbhq2nrvadncyjpwjlluuona	189	11	numbers	number	NNS
work_ezwbhq2nrvadncyjpwjlluuona	189	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	189	13	eigenvec-	eigenvec-	NN
work_ezwbhq2nrvadncyjpwjlluuona	189	14	tors	tor	NNS
work_ezwbhq2nrvadncyjpwjlluuona	189	15	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	190	1	Figure	figure	NN
work_ezwbhq2nrvadncyjpwjlluuona	190	2	3	3	CD
work_ezwbhq2nrvadncyjpwjlluuona	190	3	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	190	4	a	a	NN
work_ezwbhq2nrvadncyjpwjlluuona	190	5	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	190	6	b	b	LS
work_ezwbhq2nrvadncyjpwjlluuona	190	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	190	8	displays	display	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	190	9	ROC	ROC	NNP
work_ezwbhq2nrvadncyjpwjlluuona	190	10	curves	curve	NNS
work_ezwbhq2nrvadncyjpwjlluuona	190	11	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	190	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	190	13	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	190	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	190	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	190	16	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	190	17	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	190	18	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	190	19	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	190	20	models	model	NNS
work_ezwbhq2nrvadncyjpwjlluuona	190	21	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	190	22	respectively	respectively	RB
work_ezwbhq2nrvadncyjpwjlluuona	190	23	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	190	24	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	190	25	three	three	CD
work_ezwbhq2nrvadncyjpwjlluuona	190	26	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	190	27	numbers	number	NNS
work_ezwbhq2nrvadncyjpwjlluuona	190	28	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	190	29	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	190	30	contributing	contribute	VBG
work_ezwbhq2nrvadncyjpwjlluuona	190	31	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	190	32	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	190	33	recon-	recon-	NN
work_ezwbhq2nrvadncyjpwjlluuona	190	34	struction	struction	NN
work_ezwbhq2nrvadncyjpwjlluuona	190	35	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	191	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	191	2	both	both	DT
work_ezwbhq2nrvadncyjpwjlluuona	191	3	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	191	4	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	191	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	191	6	10	10	CD
work_ezwbhq2nrvadncyjpwjlluuona	191	7	eigen-	eigen-	NN
work_ezwbhq2nrvadncyjpwjlluuona	191	8	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	191	9	yielded	yield	VBD
work_ezwbhq2nrvadncyjpwjlluuona	191	10	excellent	excellent	JJ
work_ezwbhq2nrvadncyjpwjlluuona	191	11	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	191	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	192	1	Dividing	divide	VBG
work_ezwbhq2nrvadncyjpwjlluuona	192	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	192	3	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	192	4	into	into	IN
work_ezwbhq2nrvadncyjpwjlluuona	192	5	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	192	6	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	192	7	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	192	8	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	192	9	groups	group	NNS
work_ezwbhq2nrvadncyjpwjlluuona	192	10	did	do	VBD
work_ezwbhq2nrvadncyjpwjlluuona	192	11	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	192	12	show	show	VB
work_ezwbhq2nrvadncyjpwjlluuona	192	13	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	192	14	cross	cross	JJ
work_ezwbhq2nrvadncyjpwjlluuona	192	15	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	192	16	race	race	JJ
work_ezwbhq2nrvadncyjpwjlluuona	192	17	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	192	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	193	1	That	that	RB
work_ezwbhq2nrvadncyjpwjlluuona	193	2	is	is	RB
work_ezwbhq2nrvadncyjpwjlluuona	193	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	193	4	major-	major-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	193	5	ity	ity	NN
work_ezwbhq2nrvadncyjpwjlluuona	193	6	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	193	7	did	do	VBD
work_ezwbhq2nrvadncyjpwjlluuona	193	8	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	193	9	yield	yield	VB
work_ezwbhq2nrvadncyjpwjlluuona	193	10	larger	large	JJR
work_ezwbhq2nrvadncyjpwjlluuona	193	11	values	value	NNS
work_ezwbhq2nrvadncyjpwjlluuona	193	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	193	13	d′.	d′.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	194	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	194	2	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	194	3	likely	likely	JJ
work_ezwbhq2nrvadncyjpwjlluuona	194	4	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	194	5	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	194	6	due	due	IN
work_ezwbhq2nrvadncyjpwjlluuona	194	7	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	194	8	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	194	9	fact	fact	NN
work_ezwbhq2nrvadncyjpwjlluuona	194	10	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	194	11	only	only	RB
work_ezwbhq2nrvadncyjpwjlluuona	194	12	five	five	CD
work_ezwbhq2nrvadncyjpwjlluuona	194	13	minority-	minority-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	194	14	SIMULATING	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	194	15	THE	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	194	16	OTHER	OTHER	NNP
work_ezwbhq2nrvadncyjpwjlluuona	194	17	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	194	18	RACE	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	194	19	EFFECT	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	194	20	171	171	CD
work_ezwbhq2nrvadncyjpwjlluuona	194	21	Figure	figure	NN
work_ezwbhq2nrvadncyjpwjlluuona	194	22	3	3	CD
work_ezwbhq2nrvadncyjpwjlluuona	194	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	195	1	ROC	ROC	NNP
work_ezwbhq2nrvadncyjpwjlluuona	195	2	curves	curve	NNS
work_ezwbhq2nrvadncyjpwjlluuona	195	3	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	195	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	195	5	performance	performance	NN
work_ezwbhq2nrvadncyjpwjlluuona	195	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	195	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	195	8	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	195	9	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	195	10	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	195	11	Caucasian	Caucasian	NNP
work_ezwbhq2nrvadncyjpwjlluuona	195	12	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	195	13	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	195	14	b	b	LS
work_ezwbhq2nrvadncyjpwjlluuona	195	15	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	195	16	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	195	17	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	195	18	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	195	19	95	95	CD
work_ezwbhq2nrvadncyjpwjlluuona	195	20	%	%	NN
work_ezwbhq2nrvadncyjpwjlluuona	195	21	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	195	22	models	model	NNS
work_ezwbhq2nrvadncyjpwjlluuona	195	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	196	1	Perfor-	Perfor-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	196	2	mance	mance	NN
work_ezwbhq2nrvadncyjpwjlluuona	196	3	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	196	4	plotted	plot	VBN
work_ezwbhq2nrvadncyjpwjlluuona	196	5	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	196	6	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	196	7	numbers	number	NNS
work_ezwbhq2nrvadncyjpwjlluuona	196	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	196	9	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	196	10	contributing	contribute	VBG
work_ezwbhq2nrvadncyjpwjlluuona	196	11	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	196	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	196	13	reconstructions	reconstruction	NNS
work_ezwbhq2nrvadncyjpwjlluuona	196	14	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	197	1	172	172	CD
work_ezwbhq2nrvadncyjpwjlluuona	197	2	O’TOOLE	O’TOOLE	NNP
work_ezwbhq2nrvadncyjpwjlluuona	197	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	197	4	DEFFENBACHER	DEFFENBACHER	NNP
work_ezwbhq2nrvadncyjpwjlluuona	197	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	197	6	ABDI	ABDI	NNP
work_ezwbhq2nrvadncyjpwjlluuona	197	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	197	8	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	197	9	BARTLETT	BARTLETT	NNP
work_ezwbhq2nrvadncyjpwjlluuona	197	10	Figure	Figure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	197	11	4	4	CD
work_ezwbhq2nrvadncyjpwjlluuona	197	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	198	1	ROC	ROC	NNP
work_ezwbhq2nrvadncyjpwjlluuona	198	2	curves	curve	NNS
work_ezwbhq2nrvadncyjpwjlluuona	198	3	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	198	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	198	5	long	long	JJ
work_ezwbhq2nrvadncyjpwjlluuona	198	6	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	198	7	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	198	8	experience	experience	NN
work_ezwbhq2nrvadncyjpwjlluuona	198	9	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	198	10	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	198	11	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	198	12	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	198	13	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	198	14	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	198	15	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	198	16	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	198	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	199	1	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	199	2	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	199	3	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	199	4	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	199	5	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	199	6	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	199	7	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	199	8	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	199	9	that	that	DT
work_ezwbhq2nrvadncyjpwjlluuona	199	10	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	199	11	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	199	12	probably	probably	RB
work_ezwbhq2nrvadncyjpwjlluuona	199	13	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	199	14	largest	large	JJS
work_ezwbhq2nrvadncyjpwjlluuona	199	15	category	category	NN
work_ezwbhq2nrvadncyjpwjlluuona	199	16	difference	difference	NN
work_ezwbhq2nrvadncyjpwjlluuona	199	17	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	199	18	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	199	19	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	199	20	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	199	21	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	199	22	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	199	23	therefore	therefore	RB
work_ezwbhq2nrvadncyjpwjlluuona	199	24	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	199	25	likely	likely	JJ
work_ezwbhq2nrvadncyjpwjlluuona	199	26	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	199	27	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	199	28	represented	represent	VBN
work_ezwbhq2nrvadncyjpwjlluuona	199	29	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	199	30	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	199	31	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	199	32	few	few	JJ
work_ezwbhq2nrvadncyjpwjlluuona	199	33	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	199	34	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	200	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	200	2	then	then	RB
work_ezwbhq2nrvadncyjpwjlluuona	200	3	simulated	simulate	VBD
work_ezwbhq2nrvadncyjpwjlluuona	200	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	200	5	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	200	6	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	200	7	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	8	component	component	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	9	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	200	10	recalling	recall	VBG
work_ezwbhq2nrvadncyjpwjlluuona	200	11	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	200	12	combining	combine	VBG
work_ezwbhq2nrvadncyjpwjlluuona	200	13	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	200	14	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	200	15	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	200	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	200	17	long	long	JJ
work_ezwbhq2nrvadncyjpwjlluuona	200	18	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	200	19	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	20	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	21	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	22	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	200	23	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	200	24	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	200	25	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	200	26	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	27	half	half	JJ
work_ezwbhq2nrvadncyjpwjlluuona	200	28	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	200	29	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	200	30	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	200	31	n	n	NNP
work_ezwbhq2nrvadncyjpwjlluuona	200	32	=	=	SYM
work_ezwbhq2nrvadncyjpwjlluuona	200	33	40	40	CD
work_ezwbhq2nrvadncyjpwjlluuona	200	34	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	200	35	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	200	36	half	half	JJ
work_ezwbhq2nrvadncyjpwjlluuona	200	37	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	200	38	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	200	39	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	200	40	n	n	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	41	=	=	SYM
work_ezwbhq2nrvadncyjpwjlluuona	200	42	40	40	CD
work_ezwbhq2nrvadncyjpwjlluuona	200	43	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	200	44	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	45	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	200	46	weighting	weight	VBG
work_ezwbhq2nrvadncyjpwjlluuona	200	47	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	200	48	long	long	JJ
work_ezwbhq2nrvadncyjpwjlluuona	200	49	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	200	50	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	51	component	component	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	52	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	200	53	0.75	0.75	CD
work_ezwbhq2nrvadncyjpwjlluuona	200	54	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	200	55	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	200	56	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	200	57	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	200	58	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	59	component	component	NN
work_ezwbhq2nrvadncyjpwjlluuona	200	60	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	200	61	0.25	0.25	CD
work_ezwbhq2nrvadncyjpwjlluuona	200	62	5	5	CD
work_ezwbhq2nrvadncyjpwjlluuona	200	63	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	201	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	201	2	report	report	VBP
work_ezwbhq2nrvadncyjpwjlluuona	201	3	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	201	4	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	201	5	only	only	RB
work_ezwbhq2nrvadncyjpwjlluuona	201	6	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	201	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	201	8	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	201	9	major-	major-	NN
work_ezwbhq2nrvadncyjpwjlluuona	201	10	ity	ity	NN
work_ezwbhq2nrvadncyjpwjlluuona	201	11	matrix6	matrix6	NN
work_ezwbhq2nrvadncyjpwjlluuona	201	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	202	1	Here	here	RB
work_ezwbhq2nrvadncyjpwjlluuona	202	2	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	202	3	see	see	VBP
work_ezwbhq2nrvadncyjpwjlluuona	202	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	202	5	classic	classic	JJ
work_ezwbhq2nrvadncyjpwjlluuona	202	6	cross	cross	JJ
work_ezwbhq2nrvadncyjpwjlluuona	202	7	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	202	8	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	202	9	ef-	ef-	XX
work_ezwbhq2nrvadncyjpwjlluuona	202	10	fect	fect	RB
work_ezwbhq2nrvadncyjpwjlluuona	202	11	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	202	12	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	202	13	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	202	14	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	202	15	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	202	16	being	be	VBG
work_ezwbhq2nrvadncyjpwjlluuona	202	17	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	202	18	difficult	difficult	JJ
work_ezwbhq2nrvadncyjpwjlluuona	202	19	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	202	20	recognize	recognize	VB
work_ezwbhq2nrvadncyjpwjlluuona	202	21	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	202	22	i.e.	i.e.	FW
work_ezwbhq2nrvadncyjpwjlluuona	203	1	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	203	2	separate	separate	VB
work_ezwbhq2nrvadncyjpwjlluuona	203	3	OLD	OLD	NNP
work_ezwbhq2nrvadncyjpwjlluuona	203	4	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	203	5	NEW	NEW	NNP
work_ezwbhq2nrvadncyjpwjlluuona	203	6	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	203	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	203	8	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	203	9	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	203	10	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	203	11	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	203	12	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	203	13	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	203	14	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	203	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	203	16	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	203	17	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	203	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	204	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	204	2	ROC	ROC	NNP
work_ezwbhq2nrvadncyjpwjlluuona	204	3	curves	curve	NNS
work_ezwbhq2nrvadncyjpwjlluuona	204	4	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	204	5	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	204	6	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	204	7	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	204	8	dis-	dis-	RB
work_ezwbhq2nrvadncyjpwjlluuona	204	9	played	play	VBN
work_ezwbhq2nrvadncyjpwjlluuona	204	10	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	204	11	Figure	Figure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	204	12	4	4	CD
work_ezwbhq2nrvadncyjpwjlluuona	204	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	205	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	205	2	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	205	3	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	205	4	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	205	5	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	206	1	Recalling	recalling	NN
work_ezwbhq2nrvadncyjpwjlluuona	206	2	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	206	3	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	206	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	206	5	autoassociative	autoassociative	NNP
work_ezwbhq2nrvadncyjpwjlluuona	206	6	matrix	matrix	NNP
work_ezwbhq2nrvadncyjpwjlluuona	206	7	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	206	8	carried	carry	VBN
work_ezwbhq2nrvadncyjpwjlluuona	206	9	out	out	RP
work_ezwbhq2nrvadncyjpwjlluuona	206	10	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	206	11	summing	sum	VBG
work_ezwbhq2nrvadncyjpwjlluuona	206	12	together	together	RB
work_ezwbhq2nrvadncyjpwjlluuona	206	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	206	14	weighted	weighted	JJ
work_ezwbhq2nrvadncyjpwjlluuona	206	15	combination	combination	NN
work_ezwbhq2nrvadncyjpwjlluuona	206	16	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	206	17	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	206	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	207	1	That	that	RB
work_ezwbhq2nrvadncyjpwjlluuona	207	2	is	is	RB
work_ezwbhq2nrvadncyjpwjlluuona	207	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	207	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	207	5	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	207	6	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	207	7	‘	'	``
work_ezwbhq2nrvadncyjpwjlluuona	207	8	put	put	VB
work_ezwbhq2nrvadncyjpwjlluuona	207	9	together	together	RB
work_ezwbhq2nrvadncyjpwjlluuona	207	10	’	'	''
work_ezwbhq2nrvadncyjpwjlluuona	207	11	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	207	12	adding	add	VBG
work_ezwbhq2nrvadncyjpwjlluuona	207	13	up	up	RP
work_ezwbhq2nrvadncyjpwjlluuona	207	14	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	207	15	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	207	16	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	207	17	differentially	differentially	RB
work_ezwbhq2nrvadncyjpwjlluuona	207	18	weighted	weight	VBN
work_ezwbhq2nrvadncyjpwjlluuona	207	19	combinations	combination	NNS
work_ezwbhq2nrvadncyjpwjlluuona	207	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	208	1	As	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	208	2	such	such	JJ
work_ezwbhq2nrvadncyjpwjlluuona	208	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	208	4	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	208	5	most	most	JJS
work_ezwbhq2nrvadncyjpwjlluuona	208	6	psycho-	psycho-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	208	7	logical	logical	JJ
work_ezwbhq2nrvadncyjpwjlluuona	208	8	definitions	definition	NNS
work_ezwbhq2nrvadncyjpwjlluuona	208	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	208	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	208	11	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	208	12	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	208	13	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	208	14	thought	think	VBN
work_ezwbhq2nrvadncyjpwjlluuona	208	15	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	208	16	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	208	17	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	208	18	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	208	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	208	20	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	208	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	209	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	209	2	interpretation	interpretation	NN
work_ezwbhq2nrvadncyjpwjlluuona	209	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	209	4	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	209	5	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	209	6	associative	associative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	209	7	matrices	matrix	NNS
work_ezwbhq2nrvadncyjpwjlluuona	209	8	has	have	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	209	9	been	be	VBN
work_ezwbhq2nrvadncyjpwjlluuona	209	10	pointed	point	VBN
work_ezwbhq2nrvadncyjpwjlluuona	209	11	out	out	RP
work_ezwbhq2nrvadncyjpwjlluuona	209	12	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	209	13	Anderson	Anderson	NNP
work_ezwbhq2nrvadncyjpwjlluuona	209	14	et	et	FW
work_ezwbhq2nrvadncyjpwjlluuona	209	15	al	al	NNP
work_ezwbhq2nrvadncyjpwjlluuona	209	16	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	210	1	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	210	2	1977	1977	CD
work_ezwbhq2nrvadncyjpwjlluuona	210	3	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	210	4	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	211	1	Also	also	RB
work_ezwbhq2nrvadncyjpwjlluuona	211	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	211	3	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	211	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	211	5	context	context	NN
work_ezwbhq2nrvadncyjpwjlluuona	211	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	211	7	low	low	JJ
work_ezwbhq2nrvadncyjpwjlluuona	211	8	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	211	9	dimensional	dimensional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	211	10	representation	representation	NN
work_ezwbhq2nrvadncyjpwjlluuona	211	11	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	211	12	images	image	NNS
work_ezwbhq2nrvadncyjpwjlluuona	211	13	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	211	14	Sirovitch	Sirovitch	NNP
work_ezwbhq2nrvadncyjpwjlluuona	211	15	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	211	16	Kirby	Kirby	NNP
work_ezwbhq2nrvadncyjpwjlluuona	211	17	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	211	18	1987	1987	CD
work_ezwbhq2nrvadncyjpwjlluuona	211	19	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	211	20	suggest	suggest	VBP
work_ezwbhq2nrvadncyjpwjlluuona	211	21	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	211	22	5	5	CD
work_ezwbhq2nrvadncyjpwjlluuona	211	23	These	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	211	24	numbers	number	NNS
work_ezwbhq2nrvadncyjpwjlluuona	211	25	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	211	26	arbitrary	arbitrary	JJ
work_ezwbhq2nrvadncyjpwjlluuona	211	27	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	211	28	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	211	29	simply	simply	RB
work_ezwbhq2nrvadncyjpwjlluuona	211	30	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	211	31	at-	at-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	211	32	tempt	tempt	NN
work_ezwbhq2nrvadncyjpwjlluuona	211	33	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	211	34	give	give	VB
work_ezwbhq2nrvadncyjpwjlluuona	211	35	more	more	JJR
work_ezwbhq2nrvadncyjpwjlluuona	211	36	weight	weight	NN
work_ezwbhq2nrvadncyjpwjlluuona	211	37	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	211	38	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	211	39	long	long	JJ
work_ezwbhq2nrvadncyjpwjlluuona	211	40	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	211	41	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	211	42	experience	experience	NN
work_ezwbhq2nrvadncyjpwjlluuona	211	43	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	211	44	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	211	45	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	211	46	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	211	47	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	211	48	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	211	49	task	task	NN
work_ezwbhq2nrvadncyjpwjlluuona	211	50	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	212	1	6	6	CD
work_ezwbhq2nrvadncyjpwjlluuona	212	2	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	212	3	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	212	4	because	because	IN
work_ezwbhq2nrvadncyjpwjlluuona	212	5	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	212	6	do	do	VBP
work_ezwbhq2nrvadncyjpwjlluuona	212	7	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	212	8	yet	yet	RB
work_ezwbhq2nrvadncyjpwjlluuona	212	9	have	have	VB
work_ezwbhq2nrvadncyjpwjlluuona	212	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	212	11	sufficient	sufficient	JJ
work_ezwbhq2nrvadncyjpwjlluuona	212	12	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	212	13	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	212	14	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	212	15	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	212	16	available	available	JJ
work_ezwbhq2nrvadncyjpwjlluuona	212	17	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	212	18	complete	complete	VB
work_ezwbhq2nrvadncyjpwjlluuona	212	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	212	20	analysis	analysis	NN
work_ezwbhq2nrvadncyjpwjlluuona	212	21	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	212	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	212	23	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	212	24	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	212	25	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	213	1	SIMULATING	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	213	2	THE	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	213	3	OTHER	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	213	4	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	213	5	RACE	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	213	6	EFFECT	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	213	7	173	173	CD
work_ezwbhq2nrvadncyjpwjlluuona	213	8	Figure	figure	NN
work_ezwbhq2nrvadncyjpwjlluuona	213	9	5	5	CD
work_ezwbhq2nrvadncyjpwjlluuona	213	10	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	214	1	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	214	2	a	a	LS
work_ezwbhq2nrvadncyjpwjlluuona	214	3	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	214	4	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	214	5	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	214	6	four	four	CD
work_ezwbhq2nrvadncyjpwjlluuona	214	7	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	214	8	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	214	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	214	10	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	214	11	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	214	12	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	214	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	215	1	174	174	CD
work_ezwbhq2nrvadncyjpwjlluuona	215	2	O’TOOLE	O’TOOLE	NNP
work_ezwbhq2nrvadncyjpwjlluuona	215	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	215	4	DEFFENBACHER	DEFFENBACHER	NNP
work_ezwbhq2nrvadncyjpwjlluuona	215	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	215	6	ABDI	ABDI	NNP
work_ezwbhq2nrvadncyjpwjlluuona	215	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	215	8	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	215	9	BARTLETT	BARTLETT	NNP
work_ezwbhq2nrvadncyjpwjlluuona	215	10	Figure	Figure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	215	11	5	5	CD
work_ezwbhq2nrvadncyjpwjlluuona	215	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	216	1	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	216	2	b	b	NN
work_ezwbhq2nrvadncyjpwjlluuona	216	3	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	216	4	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	216	5	b	b	LS
work_ezwbhq2nrvadncyjpwjlluuona	216	6	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	216	7	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	216	8	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	216	9	four	four	CD
work_ezwbhq2nrvadncyjpwjlluuona	216	10	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	216	11	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	216	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	216	13	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	216	14	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	216	15	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	216	16	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	217	1	SIMULATING	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	217	2	THE	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	217	3	OTHER	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	217	4	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	217	5	RACE	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	217	6	EFFECT	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	217	7	175	175	CD
work_ezwbhq2nrvadncyjpwjlluuona	217	8	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	217	9	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	217	10	based	base	VBN
work_ezwbhq2nrvadncyjpwjlluuona	217	11	description	description	NN
work_ezwbhq2nrvadncyjpwjlluuona	217	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	218	1	Applied	apply	VBN
work_ezwbhq2nrvadncyjpwjlluuona	218	2	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	218	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	218	4	current	current	JJ
work_ezwbhq2nrvadncyjpwjlluuona	218	5	work	work	NN
work_ezwbhq2nrvadncyjpwjlluuona	218	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	218	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	218	8	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	218	9	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	218	10	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	218	11	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	218	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	218	13	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	218	14	images	image	NNS
work_ezwbhq2nrvadncyjpwjlluuona	218	15	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	218	16	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	218	17	different	different	JJ
work_ezwbhq2nrvadncyjpwjlluuona	218	18	sort	sort	NN
work_ezwbhq2nrvadncyjpwjlluuona	218	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	218	20	feature	feature	NN
work_ezwbhq2nrvadncyjpwjlluuona	218	21	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	218	22	has	have	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	218	23	generally	generally	RB
work_ezwbhq2nrvadncyjpwjlluuona	218	24	been	be	VBN
work_ezwbhq2nrvadncyjpwjlluuona	218	25	used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	218	26	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	218	27	describ-	describ-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	218	28	ing	ing	NN
work_ezwbhq2nrvadncyjpwjlluuona	218	29	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	218	30	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	219	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	219	2	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	219	3	thing	thing	NN
work_ezwbhq2nrvadncyjpwjlluuona	219	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	219	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	219	6	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	219	7	repre-	repre-	NN
work_ezwbhq2nrvadncyjpwjlluuona	219	8	sent	send	VBD
work_ezwbhq2nrvadncyjpwjlluuona	219	9	global	global	JJ
work_ezwbhq2nrvadncyjpwjlluuona	219	10	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	219	11	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	219	12	local	local	JJ
work_ezwbhq2nrvadncyjpwjlluuona	219	13	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	219	14	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	219	15	since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	219	16	they	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	219	17	span	span	VBP
work_ezwbhq2nrvadncyjpwjlluuona	219	18	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	219	19	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	219	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	220	1	Secondly	secondly	RB
work_ezwbhq2nrvadncyjpwjlluuona	220	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	220	3	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	220	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	220	5	exception	exception	NN
work_ezwbhq2nrvadncyjpwjlluuona	220	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	220	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	220	8	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	220	9	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	220	10	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	220	11	each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	220	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	220	13	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	220	14	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	220	15	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	220	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	220	17	eigen-	eigen-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	220	18	vectors	vector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	220	19	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	220	20	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	220	21	readily	readily	RB
work_ezwbhq2nrvadncyjpwjlluuona	220	22	interpretable	interpretable	JJ
work_ezwbhq2nrvadncyjpwjlluuona	220	23	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	220	24	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	220	25	traditional	traditional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	220	26	feature	feature	NN
work_ezwbhq2nrvadncyjpwjlluuona	220	27	sense	sense	NN
work_ezwbhq2nrvadncyjpwjlluuona	220	28	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	221	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	221	2	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	221	3	four	four	CD
work_ezwbhq2nrvadncyjpwjlluuona	221	4	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	221	5	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	221	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	221	7	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	221	8	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	221	9	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	221	10	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	221	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	221	12	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	221	13	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	221	14	simulation	simulation	NN
work_ezwbhq2nrvadncyjpwjlluuona	221	15	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	221	16	displayed	display	VBN
work_ezwbhq2nrvadncyjpwjlluuona	221	17	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	221	18	Figure	Figure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	221	19	5	5	CD
work_ezwbhq2nrvadncyjpwjlluuona	221	20	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	221	21	a	a	NN
work_ezwbhq2nrvadncyjpwjlluuona	221	22	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	221	23	b	b	NN
work_ezwbhq2nrvadncyjpwjlluuona	221	24	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	221	25	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	222	1	It	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	222	2	should	should	MD
work_ezwbhq2nrvadncyjpwjlluuona	222	3	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	222	4	noted	note	VBN
work_ezwbhq2nrvadncyjpwjlluuona	222	5	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	222	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	222	7	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	222	8	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	222	9	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	222	10	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	222	11	like	like	JJ
work_ezwbhq2nrvadncyjpwjlluuona	222	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	223	1	Furthermore	furthermore	RB
work_ezwbhq2nrvadncyjpwjlluuona	223	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	223	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	223	4	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	223	5	resemble	resemble	VBP
work_ezwbhq2nrvadncyjpwjlluuona	223	6	somewhat	somewhat	RB
work_ezwbhq2nrvadncyjpwjlluuona	223	7	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	223	8	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	223	9	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	223	10	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	223	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	223	12	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	223	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	224	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	224	2	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	224	3	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	224	4	contains	contain	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	224	5	characteristics	characteristic	NNS
work_ezwbhq2nrvadncyjpwjlluuona	224	6	typical	typical	JJ
work_ezwbhq2nrvadncyjpwjlluuona	224	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	224	8	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	224	9	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	224	10	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	224	11	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	224	12	e.g.	e.g.	RB
work_ezwbhq2nrvadncyjpwjlluuona	225	1	note	note	VB
work_ezwbhq2nrvadncyjpwjlluuona	225	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	225	3	roundness	roundness	NN
work_ezwbhq2nrvadncyjpwjlluuona	225	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	225	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	225	6	eyes	eye	NNS
work_ezwbhq2nrvadncyjpwjlluuona	225	7	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	225	8	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	225	9	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	225	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	225	11	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	225	12	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	225	13	eigenvectors	eigenvector	NNS
work_ezwbhq2nrvadncyjpwjlluuona	225	14	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	225	15	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	225	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	225	17	squareness	squareness	NN
work_ezwbhq2nrvadncyjpwjlluuona	225	18	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	225	19	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	225	20	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	225	21	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	225	22	distinctiveness	distinctiveness	NN
work_ezwbhq2nrvadncyjpwjlluuona	225	23	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	225	24	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	225	25	nostrils	nostril	NNS
work_ezwbhq2nrvadncyjpwjlluuona	225	26	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	225	27	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	225	28	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	225	29	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	225	30	eigenvec-	eigenvec-	NN
work_ezwbhq2nrvadncyjpwjlluuona	225	31	tors	tor	NNS
work_ezwbhq2nrvadncyjpwjlluuona	225	32	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	225	33	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	226	1	Finally	finally	RB
work_ezwbhq2nrvadncyjpwjlluuona	226	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	226	3	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	226	4	completeness	completeness	NN
work_ezwbhq2nrvadncyjpwjlluuona	226	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	226	6	Figure	figure	NN
work_ezwbhq2nrvadncyjpwjlluuona	226	7	6	6	CD
work_ezwbhq2nrvadncyjpwjlluuona	226	8	shows	show	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	226	9	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	226	10	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	226	11	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	226	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	226	13	each	each	DT
work_ezwbhq2nrvadncyjpwjlluuona	226	14	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	226	15	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	226	16	made	make	VBN
work_ezwbhq2nrvadncyjpwjlluuona	226	17	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	226	18	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	226	19	pixel	pixel	NN
work_ezwbhq2nrvadncyjpwjlluuona	226	20	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	226	21	based	base	VBN
work_ezwbhq2nrvadncyjpwjlluuona	226	22	code	code	NN
work_ezwbhq2nrvadncyjpwjlluuona	226	23	without	without	IN
work_ezwbhq2nrvadncyjpwjlluuona	226	24	spatial	spatial	JJ
work_ezwbhq2nrvadncyjpwjlluuona	226	25	differen-	differen-	NN
work_ezwbhq2nrvadncyjpwjlluuona	226	26	tiation	tiation	NN
work_ezwbhq2nrvadncyjpwjlluuona	226	27	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	227	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	227	2	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	227	3	differences	difference	NNS
work_ezwbhq2nrvadncyjpwjlluuona	227	4	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	227	5	even	even	RB
work_ezwbhq2nrvadncyjpwjlluuona	227	6	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	227	7	strik-	strik-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	227	8	ing	ing	NN
work_ezwbhq2nrvadncyjpwjlluuona	227	9	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	227	10	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	227	11	cases	case	NNS
work_ezwbhq2nrvadncyjpwjlluuona	227	12	since	since	IN
work_ezwbhq2nrvadncyjpwjlluuona	227	13	shading	shade	VBG
work_ezwbhq2nrvadncyjpwjlluuona	227	14	information	information	NN
work_ezwbhq2nrvadncyjpwjlluuona	227	15	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	227	16	pre-	pre-	RB
work_ezwbhq2nrvadncyjpwjlluuona	227	17	served	serve	VBN
work_ezwbhq2nrvadncyjpwjlluuona	227	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	228	1	General	General	NNP
work_ezwbhq2nrvadncyjpwjlluuona	228	2	Summary	Summary	NNP
work_ezwbhq2nrvadncyjpwjlluuona	228	3	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	228	4	Discussion	Discussion	NNP
work_ezwbhq2nrvadncyjpwjlluuona	228	5	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	228	6	purpose	purpose	NN
work_ezwbhq2nrvadncyjpwjlluuona	228	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	228	8	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	228	9	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	228	10	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	228	11	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	228	12	model	model	VB
work_ezwbhq2nrvadncyjpwjlluuona	228	13	some	some	DT
work_ezwbhq2nrvadncyjpwjlluuona	228	14	common	common	JJ
work_ezwbhq2nrvadncyjpwjlluuona	228	15	effects	effect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	228	16	associated	associate	VBN
work_ezwbhq2nrvadncyjpwjlluuona	228	17	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	228	18	processing	process	VBG
work_ezwbhq2nrvadncyjpwjlluuona	228	19	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	228	20	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	228	21	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	228	22	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	228	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	229	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	229	2	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	229	3	tried	try	VBN
work_ezwbhq2nrvadncyjpwjlluuona	229	4	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	229	5	show	show	VB
work_ezwbhq2nrvadncyjpwjlluuona	229	6	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	229	7	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	229	8	effects	effect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	229	9	can	can	MD
work_ezwbhq2nrvadncyjpwjlluuona	229	10	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	229	11	modeled	model	VBN
work_ezwbhq2nrvadncyjpwjlluuona	229	12	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	229	13	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	229	14	part	part	NN
work_ezwbhq2nrvadncyjpwjlluuona	229	15	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	229	16	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	229	17	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	229	18	process	process	NN
work_ezwbhq2nrvadncyjpwjlluuona	229	19	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	229	20	fine	fine	JJ
work_ezwbhq2nrvadncyjpwjlluuona	229	21	tuning	tuning	NN
work_ezwbhq2nrvadncyjpwjlluuona	229	22	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	229	23	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	229	24	information	information	NN
work_ezwbhq2nrvadncyjpwjlluuona	229	25	most	most	RBS
work_ezwbhq2nrvadncyjpwjlluuona	229	26	useful	useful	JJ
work_ezwbhq2nrvadncyjpwjlluuona	229	27	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	229	28	distin-	distin-	NN
work_ezwbhq2nrvadncyjpwjlluuona	229	29	guishing	guishe	VBG
work_ezwbhq2nrvadncyjpwjlluuona	229	30	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	229	31	within	within	IN
work_ezwbhq2nrvadncyjpwjlluuona	229	32	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	229	33	homogenous	homogenous	JJ
work_ezwbhq2nrvadncyjpwjlluuona	229	34	set	set	NN
work_ezwbhq2nrvadncyjpwjlluuona	229	35	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	229	36	i.e.	i.e.	FW
work_ezwbhq2nrvadncyjpwjlluuona	230	1	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	230	2	sin-	sin-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	230	3	gle	gle	JJ
work_ezwbhq2nrvadncyjpwjlluuona	230	4	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	230	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	230	6	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	230	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	230	8	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	231	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	231	2	tuning	tuning	NN
work_ezwbhq2nrvadncyjpwjlluuona	231	3	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	231	4	suboptimal	suboptimal	JJ
work_ezwbhq2nrvadncyjpwjlluuona	231	5	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	231	6	processing	process	VBG
work_ezwbhq2nrvadncyjpwjlluuona	231	7	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	231	8	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	231	9	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	231	10	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	231	11	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	231	12	however	however	RB
work_ezwbhq2nrvadncyjpwjlluuona	231	13	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	231	14	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	231	15	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	231	16	sys-	sys-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	231	17	tem	tem	NN
work_ezwbhq2nrvadncyjpwjlluuona	231	18	shows	show	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	231	19	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	231	20	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	231	21	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	231	22	shortcomings	shortcoming	NNS
work_ezwbhq2nrvadncyjpwjlluuona	231	23	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	231	24	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	231	25	mi-	mi-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	231	26	nority	nority	NN
work_ezwbhq2nrvadncyjpwjlluuona	231	27	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	231	28	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	231	29	compared	compare	VBN
work_ezwbhq2nrvadncyjpwjlluuona	231	30	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	231	31	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	231	32	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	231	33	faces	face	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	231	34	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	232	1	Our	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	232	2	simulations	simulation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	232	3	produced	produce	VBD
work_ezwbhq2nrvadncyjpwjlluuona	232	4	three	three	CD
work_ezwbhq2nrvadncyjpwjlluuona	232	5	results	result	NNS
work_ezwbhq2nrvadncyjpwjlluuona	232	6	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	233	1	First	first	RB
work_ezwbhq2nrvadncyjpwjlluuona	233	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	233	3	when	when	WRB
work_ezwbhq2nrvadncyjpwjlluuona	233	4	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	233	5	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	233	6	history	history	NN
work_ezwbhq2nrvadncyjpwjlluuona	233	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	233	8	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	233	9	network	network	NN
work_ezwbhq2nrvadncyjpwjlluuona	233	10	was	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	233	11	strongly	strongly	RB
work_ezwbhq2nrvadncyjpwjlluuona	233	12	biased	bias	VBN
work_ezwbhq2nrvadncyjpwjlluuona	233	13	to-	to-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	233	14	ward	ward	NNP
work_ezwbhq2nrvadncyjpwjlluuona	233	15	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	233	16	single	single	JJ
work_ezwbhq2nrvadncyjpwjlluuona	233	17	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	233	18	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	233	19	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	233	20	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	233	21	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	233	22	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	233	23	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	233	24	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	233	25	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	233	26	represent	represent	VB
work_ezwbhq2nrvadncyjpwjlluuona	233	27	novel	novel	JJ
work_ezwbhq2nrvadncyjpwjlluuona	233	28	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	233	29	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	233	30	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	233	31	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	233	32	exceeded	exceed	VBD
work_ezwbhq2nrvadncyjpwjlluuona	233	33	its	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	233	34	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	233	35	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	233	36	represent	represent	VB
work_ezwbhq2nrvadncyjpwjlluuona	233	37	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	233	38	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	233	39	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	233	40	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	233	41	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	234	1	Sec-	sec-	VB
work_ezwbhq2nrvadncyjpwjlluuona	234	2	ondly	ondly	RB
work_ezwbhq2nrvadncyjpwjlluuona	234	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	234	4	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	234	5	autoassociative	autoassociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	234	6	network	network	NN
work_ezwbhq2nrvadncyjpwjlluuona	234	7	trained	train	VBN
work_ezwbhq2nrvadncyjpwjlluuona	234	8	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	234	9	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	234	10	ma-	ma-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	234	11	jority	jority	NN
work_ezwbhq2nrvadncyjpwjlluuona	234	12	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	234	13	produced	produce	VBD
work_ezwbhq2nrvadncyjpwjlluuona	234	14	codings	coding	NNS
work_ezwbhq2nrvadncyjpwjlluuona	234	15	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	234	16	were	be	VBD
work_ezwbhq2nrvadncyjpwjlluuona	234	17	more	more	JJR
work_ezwbhq2nrvadncyjpwjlluuona	234	18	sim-	sim-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	234	19	ilar	ilar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	234	20	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	234	21	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	234	22	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	234	23	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	234	24	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	234	25	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	234	26	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	234	27	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	234	28	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	234	29	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	234	30	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	234	31	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	234	32	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	234	33	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	234	34	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	234	35	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	234	36	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	235	1	This	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	235	2	simulates	simulate	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	235	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	235	4	well	well	RB
work_ezwbhq2nrvadncyjpwjlluuona	235	5	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	235	6	known	know	VBN
work_ezwbhq2nrvadncyjpwjlluuona	235	7	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	235	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	235	9	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	235	10	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	235	11	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	235	12	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	235	13	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	235	14	appearing	appear	VBG
work_ezwbhq2nrvadncyjpwjlluuona	235	15	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	235	16	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	235	17	one	one	CD
work_ezwbhq2nrvadncyjpwjlluuona	235	18	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	235	19	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	236	1	Finally	finally	RB
work_ezwbhq2nrvadncyjpwjlluuona	236	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	236	3	by	by	IN
work_ezwbhq2nrvadncyjpwjlluuona	236	4	com-	com-	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	5	bining	bin	VBG
work_ezwbhq2nrvadncyjpwjlluuona	236	6	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	236	7	long	long	JJ
work_ezwbhq2nrvadncyjpwjlluuona	236	8	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	236	9	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	10	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	11	history	history	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	12	experience	experience	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	13	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	14	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	236	15	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	236	16	short	short	JJ
work_ezwbhq2nrvadncyjpwjlluuona	236	17	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	236	18	term	term	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	19	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	20	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	21	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	236	22	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	236	23	simulated	simulate	VBD
work_ezwbhq2nrvadncyjpwjlluuona	236	24	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	236	25	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	236	26	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	236	27	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	28	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	29	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	236	30	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	31	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	236	32	being	be	VBG
work_ezwbhq2nrvadncyjpwjlluuona	236	33	bet-	bet-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	236	34	ter	ter	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	35	recognized	recognize	VBN
work_ezwbhq2nrvadncyjpwjlluuona	236	36	than	than	IN
work_ezwbhq2nrvadncyjpwjlluuona	236	37	minority	minority	NN
work_ezwbhq2nrvadncyjpwjlluuona	236	38	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	236	39	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	237	1	While	while	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	2	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	237	3	system	system	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	4	produced	produce	VBD
work_ezwbhq2nrvadncyjpwjlluuona	237	5	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	237	6	number	number	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	8	effects	effect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	237	9	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	237	10	are	be	VBP
work_ezwbhq2nrvadncyjpwjlluuona	237	11	qualitatively	qualitatively	RB
work_ezwbhq2nrvadncyjpwjlluuona	237	12	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	237	13	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	14	those	those	DT
work_ezwbhq2nrvadncyjpwjlluuona	237	15	seen	see	VBN
work_ezwbhq2nrvadncyjpwjlluuona	237	16	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	17	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	237	18	psychological	psychological	JJ
work_ezwbhq2nrvadncyjpwjlluuona	237	19	literature	literature	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	20	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	237	21	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	237	22	caution	caution	VBP
work_ezwbhq2nrvadncyjpwjlluuona	237	23	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	24	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	237	25	ap-	ap-	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	26	proach	proach	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	27	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	237	28	perhaps	perhaps	RB
work_ezwbhq2nrvadncyjpwjlluuona	237	29	best	good	JJS
work_ezwbhq2nrvadncyjpwjlluuona	237	30	thought	think	VBN
work_ezwbhq2nrvadncyjpwjlluuona	237	31	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	32	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	237	33	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	237	34	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	35	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	237	36	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	37	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	38	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	39	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	40	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	237	41	but	but	CC
work_ezwbhq2nrvadncyjpwjlluuona	237	42	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	43	an	an	DT
work_ezwbhq2nrvadncyjpwjlluuona	237	44	exploratory	exploratory	JJ
work_ezwbhq2nrvadncyjpwjlluuona	237	45	tool	tool	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	46	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	47	quantifying	quantifying	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	48	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	237	49	processing	process	VBG
work_ezwbhq2nrvadncyjpwjlluuona	237	50	subtle	subtle	JJ
work_ezwbhq2nrvadncyjpwjlluuona	237	51	perceptual	perceptual	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	52	infor-	infor-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	237	53	mation	mation	NN
work_ezwbhq2nrvadncyjpwjlluuona	237	54	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	55	complex	complex	JJ
work_ezwbhq2nrvadncyjpwjlluuona	237	56	images	image	NNS
work_ezwbhq2nrvadncyjpwjlluuona	237	57	such	such	JJ
work_ezwbhq2nrvadncyjpwjlluuona	237	58	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	237	59	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	237	60	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	238	1	It	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	238	2	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	238	3	also	also	RB
work_ezwbhq2nrvadncyjpwjlluuona	238	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	238	5	not	not	RB
work_ezwbhq2nrvadncyjpwjlluuona	238	6	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	238	7	only	only	JJ
work_ezwbhq2nrvadncyjpwjlluuona	238	8	approach	approach	NN
work_ezwbhq2nrvadncyjpwjlluuona	238	9	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	238	10	simulating	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	238	11	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	238	12	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	238	13	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	238	14	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	238	15	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	238	16	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	239	1	Used	use	VBN
work_ezwbhq2nrvadncyjpwjlluuona	239	2	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	239	3	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	239	4	context	context	NN
work_ezwbhq2nrvadncyjpwjlluuona	239	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	239	6	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	239	7	provides	provide	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	239	8	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	239	9	method	method	NN
work_ezwbhq2nrvadncyjpwjlluuona	239	10	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	239	11	examining	examine	VBG
work_ezwbhq2nrvadncyjpwjlluuona	239	12	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	239	13	kinds	kind	NNS
work_ezwbhq2nrvadncyjpwjlluuona	239	14	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	239	15	codings	coding	NNS
work_ezwbhq2nrvadncyjpwjlluuona	239	16	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	239	17	might	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	239	18	account	account	VB
work_ezwbhq2nrvadncyjpwjlluuona	239	19	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	239	20	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	239	21	effects	effect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	239	22	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	239	23	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	239	24	similar	similar	JJ
work_ezwbhq2nrvadncyjpwjlluuona	239	25	fashion	fashion	NN
work_ezwbhq2nrvadncyjpwjlluuona	239	26	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	240	1	Fur-	Fur-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	240	2	thermore	thermore	NN
work_ezwbhq2nrvadncyjpwjlluuona	240	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	240	4	its	-PRON-	PRP$
work_ezwbhq2nrvadncyjpwjlluuona	240	5	application	application	NN
work_ezwbhq2nrvadncyjpwjlluuona	240	6	might	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	240	7	give	give	VB
work_ezwbhq2nrvadncyjpwjlluuona	240	8	insight	insight	NN
work_ezwbhq2nrvadncyjpwjlluuona	240	9	into	into	IN
work_ezwbhq2nrvadncyjpwjlluuona	240	10	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	240	11	constraints	constraint	NNS
work_ezwbhq2nrvadncyjpwjlluuona	240	12	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	240	13	extensive	extensive	JJ
work_ezwbhq2nrvadncyjpwjlluuona	240	14	experience	experience	NN
work_ezwbhq2nrvadncyjpwjlluuona	240	15	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	240	16	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	240	17	given	give	VBN
work_ezwbhq2nrvadncyjpwjlluuona	240	18	stimulus	stimulus	NN
work_ezwbhq2nrvadncyjpwjlluuona	240	19	category	category	NN
work_ezwbhq2nrvadncyjpwjlluuona	240	20	place	place	NN
work_ezwbhq2nrvadncyjpwjlluuona	240	21	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	240	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	240	23	processing	processing	NN
work_ezwbhq2nrvadncyjpwjlluuona	240	24	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	240	25	stim-	stim-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	240	26	uli	uli	NNP
work_ezwbhq2nrvadncyjpwjlluuona	240	27	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	240	28	another	another	DT
work_ezwbhq2nrvadncyjpwjlluuona	240	29	category	category	NN
work_ezwbhq2nrvadncyjpwjlluuona	240	30	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	241	1	We	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	241	2	think	think	VBP
work_ezwbhq2nrvadncyjpwjlluuona	241	3	that	that	IN
work_ezwbhq2nrvadncyjpwjlluuona	241	4	this	this	DT
work_ezwbhq2nrvadncyjpwjlluuona	241	5	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	241	6	especially	especially	RB
work_ezwbhq2nrvadncyjpwjlluuona	241	7	important	important	JJ
work_ezwbhq2nrvadncyjpwjlluuona	241	8	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	241	9	cases	case	NNS
work_ezwbhq2nrvadncyjpwjlluuona	241	10	where	where	WRB
work_ezwbhq2nrvadncyjpwjlluuona	241	11	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	241	12	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	241	13	difficult	difficult	JJ
work_ezwbhq2nrvadncyjpwjlluuona	241	14	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	241	15	quantify	quantify	VB
work_ezwbhq2nrvadncyjpwjlluuona	241	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	241	17	subtle	subtle	JJ
work_ezwbhq2nrvadncyjpwjlluuona	241	18	visual	visual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	241	19	information	information	NN
work_ezwbhq2nrvadncyjpwjlluuona	241	20	that	that	WDT
work_ezwbhq2nrvadncyjpwjlluuona	241	21	sepa-	sepa-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	241	22	rates	rate	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	241	23	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	241	24	categories	category	NNS
work_ezwbhq2nrvadncyjpwjlluuona	241	25	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	242	1	Finally	finally	RB
work_ezwbhq2nrvadncyjpwjlluuona	242	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	242	3	we	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	242	4	think	think	VBP
work_ezwbhq2nrvadncyjpwjlluuona	242	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	242	6	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	7	also	also	RB
work_ezwbhq2nrvadncyjpwjlluuona	242	8	has	have	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	242	9	potential	potential	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	10	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	242	11	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	242	12	tool	tool	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	13	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	242	14	simulating	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	242	15	some	some	DT
work_ezwbhq2nrvadncyjpwjlluuona	242	16	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	242	17	well	well	RB
work_ezwbhq2nrvadncyjpwjlluuona	242	18	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	242	19	known	know	VBN
work_ezwbhq2nrvadncyjpwjlluuona	242	20	ef-	ef-	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	21	fects	fect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	242	22	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	242	23	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	24	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	25	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	242	26	such	such	JJ
work_ezwbhq2nrvadncyjpwjlluuona	242	27	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	242	28	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	242	29	relationship	relationship	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	30	be-	be-	XX
work_ezwbhq2nrvadncyjpwjlluuona	242	31	tween	tween	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	32	typicality	typicality	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	33	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	242	34	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	35	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	242	36	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	243	1	Further-	Further-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	243	2	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	243	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	243	4	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	243	5	might	may	MD
work_ezwbhq2nrvadncyjpwjlluuona	243	6	be	be	VB
work_ezwbhq2nrvadncyjpwjlluuona	243	7	useful	useful	JJ
work_ezwbhq2nrvadncyjpwjlluuona	243	8	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	243	9	giving	give	VBG
work_ezwbhq2nrvadncyjpwjlluuona	243	10	insight	insight	NN
work_ezwbhq2nrvadncyjpwjlluuona	243	11	into	into	IN
work_ezwbhq2nrvadncyjpwjlluuona	243	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	243	13	perceptual	perceptual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	243	14	components	component	NNS
work_ezwbhq2nrvadncyjpwjlluuona	243	15	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	243	16	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	243	17	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	243	18	difficul-	difficul-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	243	19	ties	tie	NNS
work_ezwbhq2nrvadncyjpwjlluuona	243	20	encountered	encounter	VBN
work_ezwbhq2nrvadncyjpwjlluuona	243	21	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	243	22	inverted	invert	VBN
work_ezwbhq2nrvadncyjpwjlluuona	243	23	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	243	24	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	243	25	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	243	26	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	243	27	presented	present	VBN
work_ezwbhq2nrvadncyjpwjlluuona	243	28	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	243	29	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	243	30	photographic	photographic	JJ
work_ezwbhq2nrvadncyjpwjlluuona	243	31	negative	negative	NN
work_ezwbhq2nrvadncyjpwjlluuona	243	32	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	244	1	For	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	244	2	these	these	DT
work_ezwbhq2nrvadncyjpwjlluuona	244	3	effects	effect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	244	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	244	5	it	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	244	6	is	be	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	244	7	instructive	instructive	JJ
work_ezwbhq2nrvadncyjpwjlluuona	244	8	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	244	9	pursue	pursue	VB
work_ezwbhq2nrvadncyjpwjlluuona	244	10	some	some	DT
work_ezwbhq2nrvadncyjpwjlluuona	244	11	simple	simple	JJ
work_ezwbhq2nrvadncyjpwjlluuona	244	12	per-	per-	RB
work_ezwbhq2nrvadncyjpwjlluuona	244	13	ceptual	ceptual	JJ
work_ezwbhq2nrvadncyjpwjlluuona	244	14	explanations	explanation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	244	15	before	before	IN
work_ezwbhq2nrvadncyjpwjlluuona	244	16	looking	look	VBG
work_ezwbhq2nrvadncyjpwjlluuona	244	17	at	at	IN
work_ezwbhq2nrvadncyjpwjlluuona	244	18	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	244	19	more	more	RBR
work_ezwbhq2nrvadncyjpwjlluuona	244	20	complicated	complicated	JJ
work_ezwbhq2nrvadncyjpwjlluuona	244	21	explanations	explanation	NNS
work_ezwbhq2nrvadncyjpwjlluuona	244	22	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	245	1	References	References	NNPS
work_ezwbhq2nrvadncyjpwjlluuona	245	2	Abdi	Abdi	NNP
work_ezwbhq2nrvadncyjpwjlluuona	245	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	245	4	H.	H.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	245	5	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	245	6	1988	1988	CD
work_ezwbhq2nrvadncyjpwjlluuona	245	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	245	8	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	246	1	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	246	2	generalized	generalize	VBN
work_ezwbhq2nrvadncyjpwjlluuona	246	3	approach	approach	NN
work_ezwbhq2nrvadncyjpwjlluuona	246	4	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	246	5	connectionist	connectionist	JJ
work_ezwbhq2nrvadncyjpwjlluuona	246	6	auto	auto	NN
work_ezwbhq2nrvadncyjpwjlluuona	246	7	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	246	8	associative	associative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	246	9	memories	memory	NNS
work_ezwbhq2nrvadncyjpwjlluuona	246	10	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	246	11	interpre-	interpre-	NN
work_ezwbhq2nrvadncyjpwjlluuona	246	12	tation	tation	NN
work_ezwbhq2nrvadncyjpwjlluuona	246	13	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	246	14	implications	implication	NNS
work_ezwbhq2nrvadncyjpwjlluuona	246	15	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	246	16	illustration	illustration	NN
work_ezwbhq2nrvadncyjpwjlluuona	246	17	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	246	18	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	246	19	pro-	pro-	NN
work_ezwbhq2nrvadncyjpwjlluuona	246	20	cessing	cessing	NN
work_ezwbhq2nrvadncyjpwjlluuona	246	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	247	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	247	2	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	247	3	Demongeot	Demongeot	NNP
work_ezwbhq2nrvadncyjpwjlluuona	247	4	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	247	5	Ed	Ed	NNP
work_ezwbhq2nrvadncyjpwjlluuona	247	6	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	247	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	248	1	Artificial	artificial	JJ
work_ezwbhq2nrvadncyjpwjlluuona	248	2	In-	In-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	248	3	telligence	telligence	NN
work_ezwbhq2nrvadncyjpwjlluuona	248	4	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	248	5	Cognitive	Cognitive	NNP
work_ezwbhq2nrvadncyjpwjlluuona	248	6	Sciences	Sciences	NNPS
work_ezwbhq2nrvadncyjpwjlluuona	248	7	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	249	1	Manchester	manchester	NN
work_ezwbhq2nrvadncyjpwjlluuona	249	2	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	249	3	Manchester	Manchester	NNP
work_ezwbhq2nrvadncyjpwjlluuona	249	4	University	University	NNP
work_ezwbhq2nrvadncyjpwjlluuona	249	5	Press	Press	NNP
work_ezwbhq2nrvadncyjpwjlluuona	249	6	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	250	1	Anderson	Anderson	NNP
work_ezwbhq2nrvadncyjpwjlluuona	250	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	250	3	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	251	1	A.	A.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	251	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	251	3	Silverstein	Silverstein	NNP
work_ezwbhq2nrvadncyjpwjlluuona	251	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	251	5	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	251	6	W.	W.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	251	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	251	8	Ritz	Ritz	NNP
work_ezwbhq2nrvadncyjpwjlluuona	251	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	251	10	S.	S.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	251	11	A.	A.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	252	1	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	252	2	Jones	Jones	NNP
work_ezwbhq2nrvadncyjpwjlluuona	252	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	252	4	R.	R.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	252	5	S.	S.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	252	6	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	252	7	1977	1977	CD
work_ezwbhq2nrvadncyjpwjlluuona	252	8	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	252	9	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	253	1	Distinctive	distinctive	JJ
work_ezwbhq2nrvadncyjpwjlluuona	253	2	features	feature	NNS
work_ezwbhq2nrvadncyjpwjlluuona	253	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	253	4	cate-	cate-	NN
work_ezwbhq2nrvadncyjpwjlluuona	253	5	gorical	gorical	JJ
work_ezwbhq2nrvadncyjpwjlluuona	253	6	perception	perception	NN
work_ezwbhq2nrvadncyjpwjlluuona	253	7	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	253	8	probability	probability	NN
work_ezwbhq2nrvadncyjpwjlluuona	253	9	learning	learning	NN
work_ezwbhq2nrvadncyjpwjlluuona	253	10	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	253	11	some	some	DT
work_ezwbhq2nrvadncyjpwjlluuona	253	12	176	176	CD
work_ezwbhq2nrvadncyjpwjlluuona	253	13	O’TOOLE	O’TOOLE	NNP
work_ezwbhq2nrvadncyjpwjlluuona	253	14	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	253	15	DEFFENBACHER	DEFFENBACHER	NNP
work_ezwbhq2nrvadncyjpwjlluuona	253	16	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	253	17	ABDI	ABDI	NNP
work_ezwbhq2nrvadncyjpwjlluuona	253	18	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	253	19	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	253	20	BARTLETT	BARTLETT	NNP
work_ezwbhq2nrvadncyjpwjlluuona	253	21	Figure	Figure	NNP
work_ezwbhq2nrvadncyjpwjlluuona	253	22	6	6	CD
work_ezwbhq2nrvadncyjpwjlluuona	253	23	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	254	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	254	2	first	first	JJ
work_ezwbhq2nrvadncyjpwjlluuona	254	3	eigenvector	eigenvector	NN
work_ezwbhq2nrvadncyjpwjlluuona	254	4	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	254	5	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	254	6	Caucasian	caucasian	JJ
work_ezwbhq2nrvadncyjpwjlluuona	254	7	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	254	8	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	254	9	majority	majority	NN
work_ezwbhq2nrvadncyjpwjlluuona	254	10	matrix	matrix	NN
work_ezwbhq2nrvadncyjpwjlluuona	254	11	made	make	VBN
work_ezwbhq2nrvadncyjpwjlluuona	254	12	from	from	IN
work_ezwbhq2nrvadncyjpwjlluuona	254	13	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	254	14	pixel	pixel	NN
work_ezwbhq2nrvadncyjpwjlluuona	254	15	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	254	16	based	base	VBN
work_ezwbhq2nrvadncyjpwjlluuona	254	17	code	code	NN
work_ezwbhq2nrvadncyjpwjlluuona	254	18	without	without	IN
work_ezwbhq2nrvadncyjpwjlluuona	254	19	spatial	spatial	JJ
work_ezwbhq2nrvadncyjpwjlluuona	254	20	differentiation	differentiation	NN
work_ezwbhq2nrvadncyjpwjlluuona	254	21	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	255	1	applications	application	NNS
work_ezwbhq2nrvadncyjpwjlluuona	255	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	255	3	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	255	4	neural	neural	JJ
work_ezwbhq2nrvadncyjpwjlluuona	255	5	model	model	NN
work_ezwbhq2nrvadncyjpwjlluuona	255	6	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	256	1	Psychological	Psychological	NNP
work_ezwbhq2nrvadncyjpwjlluuona	256	2	Re-	Re-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	256	3	view	view	NN
work_ezwbhq2nrvadncyjpwjlluuona	256	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	256	5	84	84	CD
work_ezwbhq2nrvadncyjpwjlluuona	256	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	256	7	413–451	413–451	CD
work_ezwbhq2nrvadncyjpwjlluuona	256	8	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	257	1	Bothwell	Bothwell	NNP
work_ezwbhq2nrvadncyjpwjlluuona	257	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	257	3	R.	R.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	257	4	K.	K.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	257	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	257	6	Brigham	Brigham	NNP
work_ezwbhq2nrvadncyjpwjlluuona	257	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	257	8	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	257	9	C.	C.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	257	10	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	257	11	Malpass	Malpass	NNP
work_ezwbhq2nrvadncyjpwjlluuona	257	12	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	257	13	R.	R.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	257	14	S.	S.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	257	15	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	257	16	1989	1989	CD
work_ezwbhq2nrvadncyjpwjlluuona	257	17	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	257	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	258	1	Cross	cross	JJ
work_ezwbhq2nrvadncyjpwjlluuona	258	2	-	-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	258	3	racial	racial	JJ
work_ezwbhq2nrvadncyjpwjlluuona	258	4	identification	identification	NN
work_ezwbhq2nrvadncyjpwjlluuona	258	5	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	259	1	Personality	Personality	NNP
work_ezwbhq2nrvadncyjpwjlluuona	259	2	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	259	3	Social	Social	NNP
work_ezwbhq2nrvadncyjpwjlluuona	259	4	Psychology	Psychology	NNP
work_ezwbhq2nrvadncyjpwjlluuona	259	5	Bulletin	Bulletin	NNP
work_ezwbhq2nrvadncyjpwjlluuona	259	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	259	7	15	15	CD
work_ezwbhq2nrvadncyjpwjlluuona	259	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	259	9	19–25	19–25	CD
work_ezwbhq2nrvadncyjpwjlluuona	259	10	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	260	1	Brigham	Brigham	NNP
work_ezwbhq2nrvadncyjpwjlluuona	260	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	260	3	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	260	4	C.	C.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	260	5	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	260	6	1986	1986	CD
work_ezwbhq2nrvadncyjpwjlluuona	260	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	260	8	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	260	9	influence	influence	NN
work_ezwbhq2nrvadncyjpwjlluuona	260	10	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	260	11	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	260	12	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	260	13	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	260	14	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	260	15	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	261	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	261	2	H.	H.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	261	3	D.	D.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	261	4	Ellis	Ellis	NNP
work_ezwbhq2nrvadncyjpwjlluuona	261	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	261	6	M.	M.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	261	7	A.	A.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	261	8	Jeeves	Jeeves	NNP
work_ezwbhq2nrvadncyjpwjlluuona	261	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	261	10	F.	F.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	261	11	Newcombe	Newcombe	NNP
work_ezwbhq2nrvadncyjpwjlluuona	261	12	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	261	13	A.	a.	NN
work_ezwbhq2nrvadncyjpwjlluuona	262	1	Young	Young	NNP
work_ezwbhq2nrvadncyjpwjlluuona	262	2	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	262	3	Eds	Eds	NNP
work_ezwbhq2nrvadncyjpwjlluuona	262	4	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	262	5	Aspects	aspect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	262	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	262	7	Face	Face	NNP
work_ezwbhq2nrvadncyjpwjlluuona	262	8	Processing	Processing	NNP
work_ezwbhq2nrvadncyjpwjlluuona	262	9	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	263	1	Dordrecht	Dordrecht	NNP
work_ezwbhq2nrvadncyjpwjlluuona	263	2	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	263	3	Martinus	Martinus	NNP
work_ezwbhq2nrvadncyjpwjlluuona	263	4	Nijhoff	Nijhoff	NNP
work_ezwbhq2nrvadncyjpwjlluuona	263	5	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	264	1	Brigham	Brigham	NNP
work_ezwbhq2nrvadncyjpwjlluuona	264	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	264	3	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	264	4	C.	C.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	264	5	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	264	6	Barkowitz	Barkowitz	NNP
work_ezwbhq2nrvadncyjpwjlluuona	264	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	264	8	P.	P.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	264	9	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	264	10	1978	1978	CD
work_ezwbhq2nrvadncyjpwjlluuona	264	11	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	264	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	265	1	Do	do	VBP
work_ezwbhq2nrvadncyjpwjlluuona	265	2	“	"	``
work_ezwbhq2nrvadncyjpwjlluuona	265	3	they	-PRON-	PRP
work_ezwbhq2nrvadncyjpwjlluuona	265	4	all	all	DT
work_ezwbhq2nrvadncyjpwjlluuona	265	5	look	look	VBP
work_ezwbhq2nrvadncyjpwjlluuona	265	6	alike	alike	RB
work_ezwbhq2nrvadncyjpwjlluuona	265	7	?	?	.
work_ezwbhq2nrvadncyjpwjlluuona	265	8	”	"	''
work_ezwbhq2nrvadncyjpwjlluuona	265	9	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	265	10	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	265	11	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	265	12	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	265	13	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	265	14	sex	sex	NN
work_ezwbhq2nrvadncyjpwjlluuona	265	15	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	265	16	ex-	ex-	DT
work_ezwbhq2nrvadncyjpwjlluuona	265	17	perience	perience	NN
work_ezwbhq2nrvadncyjpwjlluuona	265	18	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	265	19	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	265	20	attitudes	attitude	NNS
work_ezwbhq2nrvadncyjpwjlluuona	265	21	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	265	22	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	265	23	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	265	24	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	265	25	recognize	recognize	VB
work_ezwbhq2nrvadncyjpwjlluuona	265	26	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	265	27	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	266	1	Journal	Journal	NNP
work_ezwbhq2nrvadncyjpwjlluuona	266	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	266	3	Applied	Applied	NNP
work_ezwbhq2nrvadncyjpwjlluuona	266	4	Social	Social	NNP
work_ezwbhq2nrvadncyjpwjlluuona	266	5	Psychology	Psychology	NNP
work_ezwbhq2nrvadncyjpwjlluuona	266	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	266	7	8	8	CD
work_ezwbhq2nrvadncyjpwjlluuona	266	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	266	9	306–318	306–318	CD
work_ezwbhq2nrvadncyjpwjlluuona	266	10	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	267	1	Brigham	Brigham	NNP
work_ezwbhq2nrvadncyjpwjlluuona	267	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	267	3	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	268	1	C.	C.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	268	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	268	3	Malpass	Malpass	NNP
work_ezwbhq2nrvadncyjpwjlluuona	268	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	268	5	A.	a.	NN
work_ezwbhq2nrvadncyjpwjlluuona	268	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	268	7	Snyder	Snyder	NNP
work_ezwbhq2nrvadncyjpwjlluuona	268	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	268	9	L.	L.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	268	10	S.	S.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	268	11	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	268	12	Spaulding	Spaulding	NNP
work_ezwbhq2nrvadncyjpwjlluuona	268	13	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	268	14	K.	K.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	268	15	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	268	16	1982	1982	CD
work_ezwbhq2nrvadncyjpwjlluuona	268	17	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	268	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	269	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	269	2	accuracy	accuracy	NN
work_ezwbhq2nrvadncyjpwjlluuona	269	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	269	4	eyewitness	eyewitness	NN
work_ezwbhq2nrvadncyjpwjlluuona	269	5	identification	identification	NN
work_ezwbhq2nrvadncyjpwjlluuona	269	6	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	269	7	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	269	8	field	field	NN
work_ezwbhq2nrvadncyjpwjlluuona	269	9	setting	set	VBG
work_ezwbhq2nrvadncyjpwjlluuona	269	10	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	270	1	Journal	Journal	NNP
work_ezwbhq2nrvadncyjpwjlluuona	270	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	270	3	Person-	Person-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	270	4	ality	ality	NN
work_ezwbhq2nrvadncyjpwjlluuona	270	5	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	270	6	Social	Social	NNP
work_ezwbhq2nrvadncyjpwjlluuona	270	7	Psychology	Psychology	NNP
work_ezwbhq2nrvadncyjpwjlluuona	270	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	270	9	42	42	CD
work_ezwbhq2nrvadncyjpwjlluuona	270	10	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	270	11	673–681	673–681	CD
work_ezwbhq2nrvadncyjpwjlluuona	270	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	271	1	Chance	chance	NN
work_ezwbhq2nrvadncyjpwjlluuona	271	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	271	3	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	271	4	E.	E.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	271	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	271	6	Turner	Turner	NNP
work_ezwbhq2nrvadncyjpwjlluuona	271	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	271	8	A.	A.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	271	9	L.	L.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	271	10	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	271	11	Goldstein	Goldstein	NNP
work_ezwbhq2nrvadncyjpwjlluuona	271	12	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	271	13	A.	a.	NN
work_ezwbhq2nrvadncyjpwjlluuona	271	14	G.	G.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	271	15	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	271	16	1982	1982	CD
work_ezwbhq2nrvadncyjpwjlluuona	271	17	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	271	18	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	272	1	Development	development	NN
work_ezwbhq2nrvadncyjpwjlluuona	272	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	272	3	differential	differential	JJ
work_ezwbhq2nrvadncyjpwjlluuona	272	4	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	272	5	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	272	6	own	own	JJ
work_ezwbhq2nrvadncyjpwjlluuona	272	7	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	272	8	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	272	9	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	272	10	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	272	11	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	272	12	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	272	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	273	1	Journal	Journal	NNP
work_ezwbhq2nrvadncyjpwjlluuona	273	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	273	3	Psychology	Psychology	NNP
work_ezwbhq2nrvadncyjpwjlluuona	273	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	273	5	112	112	CD
work_ezwbhq2nrvadncyjpwjlluuona	273	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	273	7	29–37	29–37	CD
work_ezwbhq2nrvadncyjpwjlluuona	273	8	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	274	1	Cross	Cross	NNP
work_ezwbhq2nrvadncyjpwjlluuona	274	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	274	3	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	274	4	G.	G.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	274	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	274	6	Cross	Cross	NNP
work_ezwbhq2nrvadncyjpwjlluuona	274	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	274	8	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	275	1	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	275	2	Daly	Daly	NNP
work_ezwbhq2nrvadncyjpwjlluuona	275	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	275	4	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	276	1	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	276	2	1971	1971	CD
work_ezwbhq2nrvadncyjpwjlluuona	276	3	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	276	4	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	277	1	Sex	sex	NN
work_ezwbhq2nrvadncyjpwjlluuona	277	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	277	3	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	277	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	277	5	age	age	NN
work_ezwbhq2nrvadncyjpwjlluuona	277	6	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	277	7	beauty	beauty	NN
work_ezwbhq2nrvadncyjpwjlluuona	277	8	as	as	IN
work_ezwbhq2nrvadncyjpwjlluuona	277	9	factors	factor	NNS
work_ezwbhq2nrvadncyjpwjlluuona	277	10	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	277	11	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	277	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	277	13	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	277	14	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	278	1	Perception	Perception	NNP
work_ezwbhq2nrvadncyjpwjlluuona	278	2	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	278	3	Psychophysics	Psychophysics	NNP
work_ezwbhq2nrvadncyjpwjlluuona	278	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	278	5	10	10	CD
work_ezwbhq2nrvadncyjpwjlluuona	278	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	278	7	393–396	393–396	CD
work_ezwbhq2nrvadncyjpwjlluuona	278	8	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	279	1	Deffenbacher	Deffenbacher	NNP
work_ezwbhq2nrvadncyjpwjlluuona	279	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	279	3	K.	K.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	279	4	A.	a.	NN
work_ezwbhq2nrvadncyjpwjlluuona	280	1	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	280	2	Loftus	Loftus	NNP
work_ezwbhq2nrvadncyjpwjlluuona	280	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	280	4	E.	E.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	280	5	F.	F.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	280	6	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	280	7	1982	1982	CD
work_ezwbhq2nrvadncyjpwjlluuona	280	8	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	280	9	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	281	1	Do	do	VBP
work_ezwbhq2nrvadncyjpwjlluuona	281	2	jurors	juror	NNS
work_ezwbhq2nrvadncyjpwjlluuona	281	3	share	share	VB
work_ezwbhq2nrvadncyjpwjlluuona	281	4	a	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	281	5	common	common	JJ
work_ezwbhq2nrvadncyjpwjlluuona	281	6	understanding	understanding	NN
work_ezwbhq2nrvadncyjpwjlluuona	281	7	concerning	concern	VBG
work_ezwbhq2nrvadncyjpwjlluuona	281	8	eyewitness	eyewitness	NN
work_ezwbhq2nrvadncyjpwjlluuona	281	9	behavior	behavior	NN
work_ezwbhq2nrvadncyjpwjlluuona	281	10	?	?	.
work_ezwbhq2nrvadncyjpwjlluuona	282	1	Law	Law	NNP
work_ezwbhq2nrvadncyjpwjlluuona	282	2	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	282	3	Human	Human	NNP
work_ezwbhq2nrvadncyjpwjlluuona	282	4	Behavior	Behavior	NNP
work_ezwbhq2nrvadncyjpwjlluuona	282	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	282	6	6	6	CD
work_ezwbhq2nrvadncyjpwjlluuona	282	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	282	8	15–30	15–30	CD
work_ezwbhq2nrvadncyjpwjlluuona	282	9	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	283	1	Feingold	Feingold	NNP
work_ezwbhq2nrvadncyjpwjlluuona	283	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	283	3	C.	C.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	283	4	A.	A.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	284	1	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	284	2	1914	1914	CD
work_ezwbhq2nrvadncyjpwjlluuona	284	3	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	284	4	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	285	1	The	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	285	2	influence	influence	NN
work_ezwbhq2nrvadncyjpwjlluuona	285	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	285	4	envi-	envi-	FW
work_ezwbhq2nrvadncyjpwjlluuona	285	5	ronment	ronment	NN
work_ezwbhq2nrvadncyjpwjlluuona	285	6	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	285	7	identification	identification	NN
work_ezwbhq2nrvadncyjpwjlluuona	285	8	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	285	9	persons	person	NNS
work_ezwbhq2nrvadncyjpwjlluuona	285	10	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	285	11	things	thing	NNS
work_ezwbhq2nrvadncyjpwjlluuona	285	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	286	1	Journal	Journal	NNP
work_ezwbhq2nrvadncyjpwjlluuona	286	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	286	3	Criminal	Criminal	NNP
work_ezwbhq2nrvadncyjpwjlluuona	286	4	Law	Law	NNP
work_ezwbhq2nrvadncyjpwjlluuona	286	5	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	286	6	Police	Police	NNP
work_ezwbhq2nrvadncyjpwjlluuona	286	7	Science	Science	NNP
work_ezwbhq2nrvadncyjpwjlluuona	286	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	286	9	5	5	CD
work_ezwbhq2nrvadncyjpwjlluuona	286	10	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	286	11	39–51	39–51	CD
work_ezwbhq2nrvadncyjpwjlluuona	286	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	287	1	Feinman	Feinman	NNP
work_ezwbhq2nrvadncyjpwjlluuona	287	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	287	3	S.	S.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	287	4	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	287	5	Entwistle	Entwistle	NNP
work_ezwbhq2nrvadncyjpwjlluuona	287	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	287	7	D.	D.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	287	8	R.	R.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	287	9	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	287	10	1976	1976	CD
work_ezwbhq2nrvadncyjpwjlluuona	287	11	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	287	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	288	1	Chil-	Chil-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	288	2	dren	dren	NNP
work_ezwbhq2nrvadncyjpwjlluuona	288	3	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	288	4	ability	ability	NN
work_ezwbhq2nrvadncyjpwjlluuona	288	5	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	288	6	recognize	recognize	VB
work_ezwbhq2nrvadncyjpwjlluuona	288	7	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	288	8	children	child	NNS
work_ezwbhq2nrvadncyjpwjlluuona	288	9	’s	’s	POS
work_ezwbhq2nrvadncyjpwjlluuona	288	10	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	288	11	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	289	1	Child	Child	NNP
work_ezwbhq2nrvadncyjpwjlluuona	289	2	Development	Development	NNP
work_ezwbhq2nrvadncyjpwjlluuona	289	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	289	4	47	47	CD
work_ezwbhq2nrvadncyjpwjlluuona	289	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	289	6	506–510	506–510	CD
work_ezwbhq2nrvadncyjpwjlluuona	289	7	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	290	1	SIMULATING	simulate	VBG
work_ezwbhq2nrvadncyjpwjlluuona	290	2	THE	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	290	3	OTHER	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	290	4	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	290	5	RACE	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	290	6	EFFECT	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	290	7	177	177	CD
work_ezwbhq2nrvadncyjpwjlluuona	290	8	Goldstein	Goldstein	NNP
work_ezwbhq2nrvadncyjpwjlluuona	290	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	290	10	A.	A.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	290	11	G.	G.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	290	12	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	290	13	Chance	Chance	NNP
work_ezwbhq2nrvadncyjpwjlluuona	290	14	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	290	15	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	290	16	E.	E.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	290	17	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	290	18	1985	1985	CD
work_ezwbhq2nrvadncyjpwjlluuona	290	19	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	290	20	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	291	1	Effects	effect	NNS
work_ezwbhq2nrvadncyjpwjlluuona	291	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	291	3	training	training	NN
work_ezwbhq2nrvadncyjpwjlluuona	291	4	on	on	IN
work_ezwbhq2nrvadncyjpwjlluuona	291	5	Japanese	japanese	JJ
work_ezwbhq2nrvadncyjpwjlluuona	291	6	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	291	7	recognition	recognition	NN
work_ezwbhq2nrvadncyjpwjlluuona	291	8	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	291	9	Reduc-	Reduc-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	291	10	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	291	11	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	291	12	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	291	13	other	other	JJ
work_ezwbhq2nrvadncyjpwjlluuona	291	14	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	291	15	race	race	NN
work_ezwbhq2nrvadncyjpwjlluuona	291	16	effect	effect	NN
work_ezwbhq2nrvadncyjpwjlluuona	291	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	292	1	Bulletin	bulletin	NN
work_ezwbhq2nrvadncyjpwjlluuona	292	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	292	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	292	4	Psy-	Psy-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	292	5	chonomic	chonomic	JJ
work_ezwbhq2nrvadncyjpwjlluuona	292	6	Society	Society	NNP
work_ezwbhq2nrvadncyjpwjlluuona	292	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	292	8	23	23	CD
work_ezwbhq2nrvadncyjpwjlluuona	292	9	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	292	10	211–214	211–214	CD
work_ezwbhq2nrvadncyjpwjlluuona	292	11	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	293	1	Hopfield	Hopfield	NNP
work_ezwbhq2nrvadncyjpwjlluuona	293	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	293	3	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	293	4	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	294	1	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	294	2	1984	1984	CD
work_ezwbhq2nrvadncyjpwjlluuona	294	3	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	294	4	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	295	1	Neurons	neuron	NNS
work_ezwbhq2nrvadncyjpwjlluuona	295	2	with	with	IN
work_ezwbhq2nrvadncyjpwjlluuona	295	3	graded	grade	VBN
work_ezwbhq2nrvadncyjpwjlluuona	295	4	re-	re-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	295	5	sponses	sponse	NNS
work_ezwbhq2nrvadncyjpwjlluuona	295	6	have	have	VBP
work_ezwbhq2nrvadncyjpwjlluuona	295	7	collective	collective	JJ
work_ezwbhq2nrvadncyjpwjlluuona	295	8	computational	computational	JJ
work_ezwbhq2nrvadncyjpwjlluuona	295	9	properties	property	NNS
work_ezwbhq2nrvadncyjpwjlluuona	295	10	like	like	IN
work_ezwbhq2nrvadncyjpwjlluuona	295	11	those	those	DT
work_ezwbhq2nrvadncyjpwjlluuona	295	12	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	295	13	two	two	CD
work_ezwbhq2nrvadncyjpwjlluuona	295	14	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	295	15	state	state	NN
work_ezwbhq2nrvadncyjpwjlluuona	295	16	neurons	neuron	NNS
work_ezwbhq2nrvadncyjpwjlluuona	295	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	296	1	Proceedings	proceeding	NNS
work_ezwbhq2nrvadncyjpwjlluuona	296	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	296	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	296	4	National	National	NNP
work_ezwbhq2nrvadncyjpwjlluuona	296	5	Academy	Academy	NNP
work_ezwbhq2nrvadncyjpwjlluuona	296	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	296	7	Sciences	Sciences	NNPS
work_ezwbhq2nrvadncyjpwjlluuona	296	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	296	9	81	81	CD
work_ezwbhq2nrvadncyjpwjlluuona	296	10	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	296	11	3088–3092	3088–3092	CD
work_ezwbhq2nrvadncyjpwjlluuona	296	12	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	297	1	Kohonen	Kohonen	NNP
work_ezwbhq2nrvadncyjpwjlluuona	297	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	297	3	T.	T.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	297	4	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	297	5	1984	1984	CD
work_ezwbhq2nrvadncyjpwjlluuona	297	6	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	297	7	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	298	1	Self	Self	NNP
work_ezwbhq2nrvadncyjpwjlluuona	298	2	Organization	Organization	NNP
work_ezwbhq2nrvadncyjpwjlluuona	298	3	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	298	4	As-	As-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	298	5	sociative	sociative	JJ
work_ezwbhq2nrvadncyjpwjlluuona	298	6	Memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	298	7	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	299	1	Berlin	Berlin	NNP
work_ezwbhq2nrvadncyjpwjlluuona	299	2	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	299	3	Springer	Springer	NNP
work_ezwbhq2nrvadncyjpwjlluuona	299	4	Verlag	Verlag	NNP
work_ezwbhq2nrvadncyjpwjlluuona	299	5	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	300	1	Malpass	Malpass	NNP
work_ezwbhq2nrvadncyjpwjlluuona	300	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	300	3	R.	R.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	300	4	S.	S.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	300	5	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	300	6	1981	1981	CD
work_ezwbhq2nrvadncyjpwjlluuona	300	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	300	8	Training	training	NN
work_ezwbhq2nrvadncyjpwjlluuona	300	9	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	300	10	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	300	11	recog-	recog-	XX
work_ezwbhq2nrvadncyjpwjlluuona	300	12	nition	nition	NN
work_ezwbhq2nrvadncyjpwjlluuona	300	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	301	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	301	2	G.	G.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	301	3	Davies	Davies	NNP
work_ezwbhq2nrvadncyjpwjlluuona	301	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	301	5	H.	H.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	301	6	Ellis	Ellis	NNP
work_ezwbhq2nrvadncyjpwjlluuona	301	7	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	301	8	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	301	9	Shepherd	Shepherd	NNP
work_ezwbhq2nrvadncyjpwjlluuona	301	10	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	301	11	Eds	Eds	NNP
work_ezwbhq2nrvadncyjpwjlluuona	301	12	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	301	13	Perceiving	Perceiving	NNP
work_ezwbhq2nrvadncyjpwjlluuona	301	14	and	and	CC
work_ezwbhq2nrvadncyjpwjlluuona	301	15	Remembering	Remembering	NNP
work_ezwbhq2nrvadncyjpwjlluuona	301	16	Faces	Faces	NNPS
work_ezwbhq2nrvadncyjpwjlluuona	301	17	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	302	1	Lon-	Lon-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	302	2	don	don	NN
work_ezwbhq2nrvadncyjpwjlluuona	302	3	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	302	4	Academic	Academic	NNP
work_ezwbhq2nrvadncyjpwjlluuona	302	5	Press	Press	NNP
work_ezwbhq2nrvadncyjpwjlluuona	302	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	302	7	pp.271	pp.271	NNP
work_ezwbhq2nrvadncyjpwjlluuona	302	8	-	-	HYPH
work_ezwbhq2nrvadncyjpwjlluuona	302	9	285	285	NNP
work_ezwbhq2nrvadncyjpwjlluuona	302	10	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	303	1	O’Toole	O’Toole	NNP
work_ezwbhq2nrvadncyjpwjlluuona	303	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	303	3	A.	a.	NN
work_ezwbhq2nrvadncyjpwjlluuona	303	4	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	304	1	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	304	2	Abdi	Abdi	NNP
work_ezwbhq2nrvadncyjpwjlluuona	304	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	304	4	H.	H.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	304	5	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	304	6	1989	1989	CD
work_ezwbhq2nrvadncyjpwjlluuona	304	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	304	8	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	305	1	Connection-	Connection-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	305	2	ists	ist	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	305	3	approaches	approach	VBZ
work_ezwbhq2nrvadncyjpwjlluuona	305	4	to	to	TO
work_ezwbhq2nrvadncyjpwjlluuona	305	5	visually	visually	RB
work_ezwbhq2nrvadncyjpwjlluuona	305	6	based	base	VBN
work_ezwbhq2nrvadncyjpwjlluuona	305	7	feature	feature	NN
work_ezwbhq2nrvadncyjpwjlluuona	305	8	extrac-	extrac-	XX
work_ezwbhq2nrvadncyjpwjlluuona	305	9	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	305	10	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	306	1	In	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	306	2	G.	G.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	306	3	Tiberghien	Tiberghien	NNP
work_ezwbhq2nrvadncyjpwjlluuona	306	4	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	306	5	Ed	Ed	NNP
work_ezwbhq2nrvadncyjpwjlluuona	306	6	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	306	7	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	307	1	Advances	advance	NNS
work_ezwbhq2nrvadncyjpwjlluuona	307	2	in	in	IN
work_ezwbhq2nrvadncyjpwjlluuona	307	3	Cognitive	Cognitive	NNP
work_ezwbhq2nrvadncyjpwjlluuona	307	4	Psychology	Psychology	NNP
work_ezwbhq2nrvadncyjpwjlluuona	307	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	307	6	Vol	Vol	NNP
work_ezwbhq2nrvadncyjpwjlluuona	307	7	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	308	1	2	2	LS
work_ezwbhq2nrvadncyjpwjlluuona	308	2	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	309	1	London	London	NNP
work_ezwbhq2nrvadncyjpwjlluuona	309	2	:	:	:
work_ezwbhq2nrvadncyjpwjlluuona	309	3	Wiley	Wiley	NNP
work_ezwbhq2nrvadncyjpwjlluuona	309	4	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	310	1	O’Toole	O’Toole	NNP
work_ezwbhq2nrvadncyjpwjlluuona	310	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	310	3	A.	A.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	310	4	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	310	5	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	310	6	Millward	Millward	NNP
work_ezwbhq2nrvadncyjpwjlluuona	310	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	310	8	R.	R.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	310	9	B.	B.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	311	1	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	311	2	Anderson	Anderson	NNP
work_ezwbhq2nrvadncyjpwjlluuona	311	3	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	311	4	J.	J.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	312	1	A.	A.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	313	1	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	313	2	1988	1988	CD
work_ezwbhq2nrvadncyjpwjlluuona	313	3	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	313	4	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	314	1	A	a	DT
work_ezwbhq2nrvadncyjpwjlluuona	314	2	physical	physical	JJ
work_ezwbhq2nrvadncyjpwjlluuona	314	3	system	system	NN
work_ezwbhq2nrvadncyjpwjlluuona	314	4	approach	approach	NN
work_ezwbhq2nrvadncyjpwjlluuona	314	5	to	to	IN
work_ezwbhq2nrvadncyjpwjlluuona	314	6	recogni-	recogni-	JJ
work_ezwbhq2nrvadncyjpwjlluuona	314	7	tion	tion	NN
work_ezwbhq2nrvadncyjpwjlluuona	314	8	memory	memory	NN
work_ezwbhq2nrvadncyjpwjlluuona	314	9	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	314	10	spatially	spatially	RB
work_ezwbhq2nrvadncyjpwjlluuona	314	11	transformed	transform	VBN
work_ezwbhq2nrvadncyjpwjlluuona	314	12	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	314	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	315	1	Neu-	Neu-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	315	2	ral	ral	NNP
work_ezwbhq2nrvadncyjpwjlluuona	315	3	Networks	Networks	NNP
work_ezwbhq2nrvadncyjpwjlluuona	315	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	315	5	1	1	CD
work_ezwbhq2nrvadncyjpwjlluuona	315	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	315	7	179–199	179–199	CD
work_ezwbhq2nrvadncyjpwjlluuona	315	8	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	316	1	Shapiro	Shapiro	NNP
work_ezwbhq2nrvadncyjpwjlluuona	316	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	316	3	P.	P.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	316	4	N.	N.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	316	5	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	316	6	Penod	Penod	NNP
work_ezwbhq2nrvadncyjpwjlluuona	316	7	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	316	8	S.	S.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	316	9	D.	D.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	316	10	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	316	11	1986	1986	CD
work_ezwbhq2nrvadncyjpwjlluuona	316	12	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	316	13	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	317	1	Meta-	meta-	NN
work_ezwbhq2nrvadncyjpwjlluuona	317	2	analysis	analysis	NN
work_ezwbhq2nrvadncyjpwjlluuona	317	3	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	317	4	face	face	NN
work_ezwbhq2nrvadncyjpwjlluuona	317	5	identification	identification	NN
work_ezwbhq2nrvadncyjpwjlluuona	317	6	studies	study	NNS
work_ezwbhq2nrvadncyjpwjlluuona	317	7	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	318	1	Psychologi-	Psychologi-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	318	2	cal	cal	NNP
work_ezwbhq2nrvadncyjpwjlluuona	318	3	Bulletin	Bulletin	NNP
work_ezwbhq2nrvadncyjpwjlluuona	318	4	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	318	5	100	100	CD
work_ezwbhq2nrvadncyjpwjlluuona	318	6	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	318	7	139–156	139–156	CD
work_ezwbhq2nrvadncyjpwjlluuona	318	8	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	319	1	Sirovitch	Sirovitch	NNP
work_ezwbhq2nrvadncyjpwjlluuona	319	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	319	3	L.	L.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	319	4	&	&	CC
work_ezwbhq2nrvadncyjpwjlluuona	319	5	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	320	1	Kirby	Kirby	NNP
work_ezwbhq2nrvadncyjpwjlluuona	320	2	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	320	3	M.	M.	NNP
work_ezwbhq2nrvadncyjpwjlluuona	320	4	(	(	-LRB-
work_ezwbhq2nrvadncyjpwjlluuona	320	5	1987	1987	CD
work_ezwbhq2nrvadncyjpwjlluuona	320	6	)	)	-RRB-
work_ezwbhq2nrvadncyjpwjlluuona	320	7	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	321	1	Low-	Low-	NNP
work_ezwbhq2nrvadncyjpwjlluuona	321	2	dimensional	dimensional	JJ
work_ezwbhq2nrvadncyjpwjlluuona	321	3	procedure	procedure	NN
work_ezwbhq2nrvadncyjpwjlluuona	321	4	for	for	IN
work_ezwbhq2nrvadncyjpwjlluuona	321	5	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	321	6	characterization	characterization	NN
work_ezwbhq2nrvadncyjpwjlluuona	321	7	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	321	8	human	human	JJ
work_ezwbhq2nrvadncyjpwjlluuona	321	9	faces	face	NNS
work_ezwbhq2nrvadncyjpwjlluuona	321	10	.	.	.
work_ezwbhq2nrvadncyjpwjlluuona	322	1	Journal	Journal	NNP
work_ezwbhq2nrvadncyjpwjlluuona	322	2	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	322	3	the	the	DT
work_ezwbhq2nrvadncyjpwjlluuona	322	4	Optical	Optical	NNP
work_ezwbhq2nrvadncyjpwjlluuona	322	5	Society	Society	NNP
work_ezwbhq2nrvadncyjpwjlluuona	322	6	of	of	IN
work_ezwbhq2nrvadncyjpwjlluuona	322	7	America	America	NNP
work_ezwbhq2nrvadncyjpwjlluuona	322	8	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	322	9	3	3	CD
work_ezwbhq2nrvadncyjpwjlluuona	322	10	,	,	,
work_ezwbhq2nrvadncyjpwjlluuona	322	11	519–524	519–524	CD
work_ezwbhq2nrvadncyjpwjlluuona	322	12	.	.	.