id	sid	tid	token	lemma	pos
ajema-5904	1	1	49	49	NUM
ajema-5904	1	2	-	-	SYM
ajema-5904	1	3	54	54	NUM
ajema-5904	1	4	abnormal	abnormal	ADJ
ajema-5904	1	5	files	file	NOUN
ajema-5904	1	6	classification	classification	NOUN
ajema-5904	1	7	using	use	VERB
ajema-5904	1	8	convolutional	convolutional	ADJ
ajema-5904	1	9	49	49	NUM
ajema-5904	1	10	journal	journal	NOUN
ajema-5904	1	11	of	of	ADP
ajema-5904	1	12	engineering	engineering	NOUN
ajema-5904	1	13	,	,	PUNCT
ajema-5904	1	14	mechanics	mechanic	NOUN
ajema-5904	1	15	and	and	CCONJ
ajema-5904	1	16	architecture	architecture	NOUN
ajema-5904	1	17	www	www	NOUN
ajema-5904	1	18	.	.	PUNCT
ajema-5904	2	1	grnjournal.us	grnjournal.us	PROPN
ajema-5904	2	2	american	american	PROPN
ajema-5904	2	3	journal	journal	PROPN
ajema-5904	2	4	of	of	ADP
ajema-5904	2	5	engineering	engineering	NOUN
ajema-5904	2	6	,	,	PUNCT
ajema-5904	2	7	mechanics	mechanic	NOUN
ajema-5904	2	8	and	and	CCONJ
ajema-5904	2	9	architecture	architecture	NOUN
ajema-5904	2	10	volume	volume	NOUN
ajema-5904	2	11	2	2	NUM
ajema-5904	2	12	,	,	PUNCT
ajema-5904	2	13	issue	issue	NOUN
ajema-5904	2	14	10	10	NUM
ajema-5904	2	15	,	,	PUNCT
ajema-5904	2	16	2024	2024	NUM
ajema-5904	2	17	issn	issn	X
ajema-5904	2	18	(	(	PUNCT
ajema-5904	2	19	e	e	NOUN
ajema-5904	2	20	):	):	PUNCT
ajema-5904	2	21	2993	2993	NUM
ajema-5904	2	22	-	-	SYM
ajema-5904	2	23	2637	2637	NUM
ajema-5904	2	24	abnormal	abnormal	ADJ
ajema-5904	2	25	files	file	NOUN
ajema-5904	2	26	classification	classification	NOUN
ajema-5904	2	27	using	use	VERB
ajema-5904	2	28	convolutional	convolutional	ADJ
ajema-5904	2	29	neural	neural	ADJ
ajema-5904	2	30	network	network	NOUN
ajema-5904	2	31	cnn	cnn	PROPN
ajema-5904	2	32	mohammad	mohammad	PROPN
ajema-5904	2	33	musaddak	musaddak	PROPN
ajema-5904	2	34	al	al	PROPN
ajema-5904	2	35	-	-	PUNCT
ajema-5904	2	36	farabi	farabi	PROPN
ajema-5904	2	37	university	university	PROPN
ajema-5904	2	38	,	,	PUNCT
ajema-5904	2	39	iraq	iraq	PROPN
ajema-5904	2	40	,	,	PUNCT
ajema-5904	2	41	baghdad	baghdad	PROPN
ajema-5904	2	42	abstract	abstract	NOUN
ajema-5904	2	43	:	:	PUNCT
ajema-5904	2	44	as	as	ADP
ajema-5904	2	45	of	of	ADP
ajema-5904	2	46	late	late	ADJ
ajema-5904	2	47	,	,	PUNCT
ajema-5904	2	48	turning	turn	VERB
ajema-5904	2	49	windows	window	VERB
ajema-5904	2	50	files	file	NOUN
ajema-5904	2	51	into	into	ADP
ajema-5904	2	52	pictures	picture	NOUN
ajema-5904	2	53	and	and	CCONJ
ajema-5904	2	54	analyzing	analyze	VERB
ajema-5904	2	55	them	they	PRON
ajema-5904	2	56	with	with	ADP
ajema-5904	2	57	deep	deep	ADJ
ajema-5904	2	58	learning	learning	NOUN
ajema-5904	2	59	and	and	CCONJ
ajema-5904	2	60	machine	machine	NOUN
ajema-5904	2	61	learning	learning	NOUN
ajema-5904	2	62	have	have	AUX
ajema-5904	2	63	been	be	AUX
ajema-5904	2	64	regarded	regard	VERB
ajema-5904	2	65	as	as	ADP
ajema-5904	2	66	state	state	NOUN
ajema-5904	2	67	-	-	PUNCT
ajema-5904	2	68	of	of	ADP
ajema-5904	2	69	-	-	PUNCT
ajema-5904	2	70	the	the	DET
ajema-5904	2	71	-	-	PUNCT
ajema-5904	2	72	art	art	NOUN
ajema-5904	2	73	methods	method	NOUN
ajema-5904	2	74	for	for	ADP
ajema-5904	2	75	identifying	identify	VERB
ajema-5904	2	76	and	and	CCONJ
ajema-5904	2	77	classifying	classify	VERB
ajema-5904	2	78	malware	malware	NOUN
ajema-5904	2	79	.	.	PUNCT
ajema-5904	3	1	this	this	PRON
ajema-5904	3	2	is	be	AUX
ajema-5904	3	3	mostly	mostly	ADV
ajema-5904	3	4	because	because	SCONJ
ajema-5904	3	5	deep	deep	ADJ
ajema-5904	3	6	learning	learning	NOUN
ajema-5904	3	7	model	model	NOUN
ajema-5904	3	8	performance	performance	NOUN
ajema-5904	3	9	in	in	ADP
ajema-5904	3	10	image	image	NOUN
ajema-5904	3	11	classification	classification	NOUN
ajema-5904	3	12	has	have	AUX
ajema-5904	3	13	recently	recently	ADV
ajema-5904	3	14	experienced	experience	VERB
ajema-5904	3	15	a	a	DET
ajema-5904	3	16	boom	boom	NOUN
ajema-5904	3	17	in	in	ADP
ajema-5904	3	18	success	success	NOUN
ajema-5904	3	19	,	,	PUNCT
ajema-5904	3	20	and	and	CCONJ
ajema-5904	3	21	image	image	NOUN
ajema-5904	3	22	-	-	PUNCT
ajema-5904	3	23	based	base	VERB
ajema-5904	3	24	malware	malware	NOUN
ajema-5904	3	25	detection	detection	NOUN
ajema-5904	3	26	and	and	CCONJ
ajema-5904	3	27	classification	classification	NOUN
ajema-5904	3	28	is	be	AUX
ajema-5904	3	29	platform	platform	NOUN
ajema-5904	3	30	independent	independent	ADJ
ajema-5904	3	31	.	.	PUNCT
ajema-5904	4	1	convolutional	convolutional	ADJ
ajema-5904	4	2	neural	neural	ADJ
ajema-5904	4	3	network	network	NOUN
ajema-5904	4	4	(	(	PUNCT
ajema-5904	4	5	cnn	cnn	PROPN
ajema-5904	4	6	)	)	PUNCT
ajema-5904	4	7	deep	deep	ADJ
ajema-5904	4	8	learning	learning	NOUN
ajema-5904	4	9	techniques	technique	NOUN
ajema-5904	4	10	are	be	AUX
ajema-5904	4	11	successfully	successfully	ADV
ajema-5904	4	12	used	use	VERB
ajema-5904	4	13	for	for	ADP
ajema-5904	4	14	image	image	NOUN
ajema-5904	4	15	-	-	PUNCT
ajema-5904	4	16	based	base	VERB
ajema-5904	4	17	windows	windows	PROPN
ajema-5904	4	18	malware	malware	NOUN
ajema-5904	4	19	classification	classification	NOUN
ajema-5904	4	20	,	,	PUNCT
ajema-5904	4	21	according	accord	VERB
ajema-5904	4	22	to	to	ADP
ajema-5904	4	23	a	a	DET
ajema-5904	4	24	review	review	NOUN
ajema-5904	4	25	of	of	ADP
ajema-5904	4	26	the	the	DET
ajema-5904	4	27	literature	literature	NOUN
ajema-5904	4	28	.	.	PUNCT
ajema-5904	5	1	nevertheless	nevertheless	ADV
ajema-5904	5	2	,	,	PUNCT
ajema-5904	5	3	just	just	ADV
ajema-5904	5	4	a	a	DET
ajema-5904	5	5	small	small	ADJ
ajema-5904	5	6	percentage	percentage	NOUN
ajema-5904	5	7	of	of	ADP
ajema-5904	5	8	the	the	DET
ajema-5904	5	9	entire	entire	ADJ
ajema-5904	5	10	picture	picture	NOUN
ajema-5904	5	11	representation	representation	NOUN
ajema-5904	5	12	had	have	VERB
ajema-5904	5	13	the	the	DET
ajema-5904	5	14	infection	infection	NOUN
ajema-5904	5	15	.	.	PUNCT
ajema-5904	6	1	finding	find	VERB
ajema-5904	6	2	and	and	CCONJ
ajema-5904	6	3	identifying	identify	VERB
ajema-5904	6	4	these	these	PRON
ajema-5904	6	5	impacted	impact	VERB
ajema-5904	6	6	little	little	ADJ
ajema-5904	6	7	areas	area	NOUN
ajema-5904	6	8	is	be	AUX
ajema-5904	6	9	crucial	crucial	ADJ
ajema-5904	6	10	to	to	ADP
ajema-5904	6	11	achieving	achieve	VERB
ajema-5904	6	12	a	a	DET
ajema-5904	6	13	high	high	ADJ
ajema-5904	6	14	degree	degree	NOUN
ajema-5904	6	15	of	of	ADP
ajema-5904	6	16	malware	malware	NOUN
ajema-5904	6	17	classification	classification	NOUN
ajema-5904	6	18	accuracy	accuracy	NOUN
ajema-5904	6	19	.	.	PUNCT
ajema-5904	7	1	this	this	DET
ajema-5904	7	2	study	study	NOUN
ajema-5904	7	3	locates	locate	VERB
ajema-5904	7	4	and	and	CCONJ
ajema-5904	7	5	identifies	identify	VERB
ajema-5904	7	6	the	the	DET
ajema-5904	7	7	little	little	ADJ
ajema-5904	7	8	contaminated	contaminate	VERB
ajema-5904	7	9	spots	spot	NOUN
ajema-5904	7	10	in	in	ADP
ajema-5904	7	11	the	the	DET
ajema-5904	7	12	overall	overall	ADJ
ajema-5904	7	13	picture	picture	NOUN
ajema-5904	7	14	by	by	ADP
ajema-5904	7	15	integrating	integrate	VERB
ajema-5904	7	16	a	a	DET
ajema-5904	7	17	data	data	NOUN
ajema-5904	7	18	augmentation	augmentation	NOUN
ajema-5904	7	19	technique	technique	NOUN
ajema-5904	7	20	with	with	ADP
ajema-5904	7	21	a	a	DET
ajema-5904	7	22	cnn	cnn	PROPN
ajema-5904	7	23	.	.	PUNCT
ajema-5904	8	1	on	on	ADP
ajema-5904	8	2	a	a	DET
ajema-5904	8	3	dataset	dataset	NOUN
ajema-5904	8	4	of	of	ADP
ajema-5904	8	5	malware	malware	NOUN
ajema-5904	8	6	images	image	NOUN
ajema-5904	8	7	,	,	PUNCT
ajema-5904	8	8	a	a	DET
ajema-5904	8	9	thorough	thorough	ADJ
ajema-5904	8	10	examination	examination	NOUN
ajema-5904	8	11	and	and	CCONJ
ajema-5904	8	12	analysis	analysis	NOUN
ajema-5904	8	13	of	of	ADP
ajema-5904	8	14	the	the	DET
ajema-5904	8	15	suggested	suggest	VERB
ajema-5904	8	16	technique	technique	NOUN
ajema-5904	8	17	were	be	AUX
ajema-5904	8	18	conducted	conduct	VERB
ajema-5904	8	19	.	.	PUNCT
ajema-5904	9	1	the	the	DET
ajema-5904	9	2	effectiveness	effectiveness	NOUN
ajema-5904	9	3	of	of	ADP
ajema-5904	9	4	the	the	DET
ajema-5904	9	5	suggested	suggest	VERB
ajema-5904	9	6	data	data	NOUN
ajema-5904	9	7	augmentation	augmentation	NOUN
ajema-5904	9	8	-	-	PUNCT
ajema-5904	9	9	based	base	VERB
ajema-5904	9	10	cnn	cnn	PROPN
ajema-5904	9	11	method	method	NOUN
ajema-5904	9	12	was	be	AUX
ajema-5904	9	13	evaluated	evaluate	VERB
ajema-5904	9	14	against	against	ADP
ajema-5904	9	15	many	many	ADJ
ajema-5904	9	16	non	non	ADJ
ajema-5904	9	17	-	-	ADJ
ajema-5904	9	18	attention	attention	ADJ
ajema-5904	9	19	-	-	PUNCT
ajema-5904	9	20	cnn	cnn	NOUN
ajema-5904	9	21	-	-	PUNCT
ajema-5904	9	22	based	base	VERB
ajema-5904	9	23	methods	method	NOUN
ajema-5904	9	24	using	use	VERB
ajema-5904	9	25	different	different	ADJ
ajema-5904	9	26	data	datum	NOUN
ajema-5904	9	27	splits	split	VERB
ajema-5904	9	28	from	from	ADP
ajema-5904	9	29	the	the	DET
ajema-5904	9	30	benchmark	benchmark	ADJ
ajema-5904	9	31	malware	malware	NOUN
ajema-5904	9	32	picture	picture	NOUN
ajema-5904	9	33	training	training	NOUN
ajema-5904	9	34	and	and	CCONJ
ajema-5904	9	35	testing	testing	NOUN
ajema-5904	9	36	dataset	dataset	NOUN
ajema-5904	9	37	.	.	PUNCT
ajema-5904	10	1	the	the	DET
ajema-5904	10	2	proposed	propose	VERB
ajema-5904	10	3	cnn	cnn	PROPN
ajema-5904	10	4	approach	approach	NOUN
ajema-5904	10	5	ensured	ensure	VERB
ajema-5904	10	6	computational	computational	ADJ
ajema-5904	10	7	efficiency	efficiency	NOUN
ajema-5904	10	8	and	and	CCONJ
ajema-5904	10	9	outperformed	outperform	VERB
ajema-5904	10	10	non	non	ADJ
ajema-5904	10	11	-	-	ADJ
ajema-5904	10	12	attention	attention	NOUN
ajema-5904	10	13	-	-	PUNCT
ajema-5904	10	14	based	base	VERB
ajema-5904	10	15	cnn	cnn	PROPN
ajema-5904	10	16	methods	method	NOUN
ajema-5904	10	17	in	in	ADP
ajema-5904	10	18	all	all	DET
ajema-5904	10	19	the	the	DET
ajema-5904	10	20	data	data	NOUN
ajema-5904	10	21	-	-	PUNCT
ajema-5904	10	22	splits	split	VERB
ajema-5904	10	23	.	.	PUNCT
ajema-5904	11	1	most	most	ADV
ajema-5904	11	2	notably	notably	ADV
ajema-5904	11	3	,	,	PUNCT
ajema-5904	11	4	the	the	DET
ajema-5904	11	5	majority	majority	NOUN
ajema-5904	11	6	of	of	ADP
ajema-5904	11	7	the	the	DET
ajema-5904	11	8	techniques	technique	NOUN
ajema-5904	11	9	demonstrate	demonstrate	VERB
ajema-5904	11	10	consistent	consistent	ADJ
ajema-5904	11	11	performance	performance	NOUN
ajema-5904	11	12	across	across	ADP
ajema-5904	11	13	all	all	DET
ajema-5904	11	14	training	training	NOUN
ajema-5904	11	15	and	and	CCONJ
ajema-5904	11	16	testing	testing	NOUN
ajema-5904	11	17	data	datum	NOUN
ajema-5904	11	18	splits	split	VERB
ajema-5904	11	19	,	,	PUNCT
ajema-5904	11	20	illuminating	illuminate	VERB
ajema-5904	11	21	cnn	cnn	PROPN
ajema-5904	11	22	's	's	PART
ajema-5904	11	23	multi	multi	ADJ
ajema-5904	11	24	-	-	ADJ
ajema-5904	11	25	headed	headed	ADJ
ajema-5904	11	26	attention	attention	NOUN
ajema-5904	11	27	.	.	PUNCT
ajema-5904	12	1	model	model	PROPN
ajema-5904	12	2	’s	’s	PART
ajema-5904	12	3	generalizability	generalizability	NOUN
ajema-5904	12	4	to	to	PART
ajema-5904	12	5	perform	perform	VERB
ajema-5904	12	6	on	on	ADP
ajema-5904	12	7	the	the	DET
ajema-5904	12	8	diverse	diverse	ADJ
ajema-5904	12	9	datasets	dataset	NOUN
ajema-5904	12	10	.	.	PUNCT
ajema-5904	13	1	introduction	introduction	NOUN
ajema-5904	13	2	the	the	DET
ajema-5904	13	3	majority	majority	NOUN
ajema-5904	13	4	of	of	ADP
ajema-5904	13	5	anti	anti	ADJ
ajema-5904	13	6	-	-	ADJ
ajema-5904	13	7	virus	virus	ADJ
ajema-5904	13	8	products	product	NOUN
ajema-5904	13	9	on	on	ADP
ajema-5904	13	10	the	the	DET
ajema-5904	13	11	market	market	NOUN
ajema-5904	13	12	today	today	NOUN
ajema-5904	13	13	are	be	AUX
ajema-5904	13	14	signature	signature	NOUN
ajema-5904	13	15	-	-	PUNCT
ajema-5904	13	16	based	base	VERB
ajema-5904	13	17	,	,	PUNCT
ajema-5904	13	18	having	having	AUX
ajema-5904	13	19	gained	gain	VERB
ajema-5904	13	20	popularity	popularity	NOUN
ajema-5904	13	21	in	in	ADP
ajema-5904	13	22	the	the	DET
ajema-5904	13	23	early	early	ADJ
ajema-5904	13	24	days	day	NOUN
ajema-5904	13	25	of	of	ADP
ajema-5904	13	26	malware	malware	NOUN
ajema-5904	13	27	detection	detection	NOUN
ajema-5904	13	28	.	.	PUNCT
ajema-5904	14	1	raw	raw	ADJ
ajema-5904	14	2	binaries	binary	NOUN
ajema-5904	14	3	were	be	AUX
ajema-5904	14	4	examined	examine	VERB
ajema-5904	14	5	and	and	CCONJ
ajema-5904	14	6	a	a	DET
ajema-5904	14	7	signature	signature	NOUN
ajema-5904	14	8	based	base	VERB
ajema-5904	14	9	on	on	ADP
ajema-5904	14	10	n	n	PRON
ajema-5904	14	11	-	-	PUNCT
ajema-5904	14	12	grams	gram	NOUN
ajema-5904	14	13	was	be	AUX
ajema-5904	14	14	created	create	VERB
ajema-5904	14	15	for	for	ADP
ajema-5904	14	16	signature	signature	NOUN
ajema-5904	14	17	-	-	PUNCT
ajema-5904	14	18	based	base	VERB
ajema-5904	14	19	malware	malware	NOUN
ajema-5904	14	20	detection	detection	NOUN
ajema-5904	15	1	[	[	X
ajema-5904	15	2	1	1	NUM
ajema-5904	15	3	]	]	PUNCT
ajema-5904	15	4	.	.	PUNCT
ajema-5904	16	1	this	this	DET
ajema-5904	16	2	kind	kind	NOUN
ajema-5904	16	3	of	of	ADP
ajema-5904	16	4	strategy	strategy	NOUN
ajema-5904	16	5	is	be	AUX
ajema-5904	16	6	straightforward	straightforward	ADJ
ajema-5904	16	7	and	and	CCONJ
ajema-5904	16	8	precise	precise	ADJ
ajema-5904	16	9	.	.	PUNCT
ajema-5904	17	1	nevertheless	nevertheless	ADV
ajema-5904	17	2	,	,	PUNCT
ajema-5904	17	3	in	in	ADP
ajema-5904	17	4	order	order	NOUN
ajema-5904	17	5	to	to	PART
ajema-5904	17	6	handle	handle	VERB
ajema-5904	17	7	newly	newly	ADV
ajema-5904	17	8	discovered	discover	VERB
ajema-5904	17	9	malware	malware	NOUN
ajema-5904	17	10	or	or	CCONJ
ajema-5904	17	11	variations	variation	NOUN
ajema-5904	17	12	of	of	ADP
ajema-5904	17	13	already	already	ADV
ajema-5904	17	14	-	-	PUNCT
ajema-5904	17	15	existing	exist	VERB
ajema-5904	17	16	malware	malware	NOUN
ajema-5904	17	17	,	,	PUNCT
ajema-5904	17	18	the	the	DET
ajema-5904	17	19	signature	signature	NOUN
ajema-5904	17	20	database	database	NOUN
ajema-5904	17	21	must	must	AUX
ajema-5904	17	22	be	be	AUX
ajema-5904	17	23	updated	update	VERB
ajema-5904	17	24	on	on	ADP
ajema-5904	17	25	a	a	DET
ajema-5904	17	26	regular	regular	ADJ
ajema-5904	17	27	basis	basis	NOUN
ajema-5904	17	28	.	.	PUNCT
ajema-5904	18	1	the	the	DET
ajema-5904	18	2	typical	typical	ADJ
ajema-5904	18	3	methodologies	methodology	NOUN
ajema-5904	18	4	for	for	ADP
ajema-5904	18	5	malware	malware	NOUN
ajema-5904	18	6	analysis	analysis	NOUN
ajema-5904	18	7	that	that	PRON
ajema-5904	18	8	were	be	AUX
ajema-5904	18	9	most	most	ADV
ajema-5904	18	10	often	often	ADV
ajema-5904	18	11	used	use	VERB
ajema-5904	18	12	were	be	AUX
ajema-5904	18	13	heuristicbased	heuristicbase	VERB
ajema-5904	18	14	techniques	technique	NOUN
ajema-5904	18	15	,	,	PUNCT
ajema-5904	18	16	hybrid	hybrid	ADJ
ajema-5904	18	17	analysis	analysis	NOUN
ajema-5904	18	18	,	,	PUNCT
ajema-5904	18	19	static	static	ADJ
ajema-5904	18	20	analysis	analysis	NOUN
ajema-5904	18	21	,	,	PUNCT
ajema-5904	18	22	and	and	CCONJ
ajema-5904	18	23	dynamic	dynamic	ADJ
ajema-5904	18	24	analysis	analysis	NOUN
ajema-5904	18	25	[	[	X
ajema-5904	18	26	2	2	NUM
ajema-5904	18	27	]	]	PUNCT
ajema-5904	18	28	.	.	PUNCT
ajema-5904	19	1	.all	.all	PUNCT
ajema-5904	19	2	of	of	ADP
ajema-5904	19	3	these	these	DET
ajema-5904	19	4	techniques	technique	NOUN
ajema-5904	19	5	,	,	PUNCT
ajema-5904	19	6	taken	take	VERB
ajema-5904	19	7	together	together	ADV
ajema-5904	19	8	,	,	PUNCT
ajema-5904	19	9	entail	entail	VERB
ajema-5904	19	10	scanning	scan	VERB
ajema-5904	19	11	systems	system	NOUN
ajema-5904	19	12	to	to	PART
ajema-5904	19	13	find	find	VERB
ajema-5904	19	14	harmful	harmful	ADJ
ajema-5904	19	15	or	or	CCONJ
ajema-5904	19	16	unauthorized	unauthorized	ADJ
ajema-5904	19	17	activity	activity	NOUN
ajema-5904	19	18	.	.	PUNCT
ajema-5904	20	1	the	the	DET
ajema-5904	20	2	suspicious	suspicious	ADJ
ajema-5904	20	3	files	file	NOUN
ajema-5904	20	4	are	be	AUX
ajema-5904	20	5	then	then	ADV
ajema-5904	20	6	examined	examine	VERB
ajema-5904	20	7	by	by	ADP
ajema-5904	20	8	the	the	DET
ajema-5904	20	9	security	security	NOUN
ajema-5904	20	10	administrator	administrator	NOUN
ajema-5904	20	11	,	,	PUNCT
ajema-5904	20	12	and	and	CCONJ
ajema-5904	20	13	either	either	CCONJ
ajema-5904	20	14	a	a	DET
ajema-5904	20	15	quarantine	quarantine	NOUN
ajema-5904	20	16	is	be	AUX
ajema-5904	20	17	placed	place	VERB
ajema-5904	20	18	on	on	ADP
ajema-5904	20	19	them	they	PRON
ajema-5904	20	20	or	or	CCONJ
ajema-5904	20	21	an	an	DET
ajema-5904	20	22	update	update	NOUN
ajema-5904	20	23	is	be	AUX
ajema-5904	20	24	applied	apply	VERB
ajema-5904	20	25	to	to	ADP
ajema-5904	20	26	the	the	DET
ajema-5904	20	27	susceptible	susceptible	ADJ
ajema-5904	20	28	system	system	NOUN
ajema-5904	20	29	.	.	PUNCT
ajema-5904	21	1	every	every	DET
ajema-5904	21	2	strategy	strategy	NOUN
ajema-5904	21	3	frequently	frequently	ADV
ajema-5904	21	4	reacts	react	VERB
ajema-5904	21	5	slowly	slowly	ADV
ajema-5904	21	6	to	to	ADP
ajema-5904	21	7	fresh	fresh	ADJ
ajema-5904	21	8	dangers	danger	NOUN
ajema-5904	21	9	and	and	CCONJ
ajema-5904	21	10	attacks	attack	NOUN
ajema-5904	21	11	.	.	PUNCT
ajema-5904	22	1	by	by	ADP
ajema-5904	22	2	examining	examine	VERB
ajema-5904	22	3	the	the	DET
ajema-5904	22	4	computer	computer	NOUN
ajema-5904	22	5	codes	code	NOUN
ajema-5904	22	6	and	and	CCONJ
ajema-5904	22	7	creating	create	VERB
ajema-5904	22	8	control	control	NOUN
ajema-5904	22	9	flow	flow	NOUN
ajema-5904	22	10	graphs	graph	NOUN
ajema-5904	22	11	without	without	ADP
ajema-5904	22	12	running	run	VERB
ajema-5904	22	13	the	the	DET
ajema-5904	22	14	program	program	NOUN
ajema-5904	22	15	,	,	PUNCT
ajema-5904	22	16	a	a	DET
ajema-5904	22	17	method	method	NOUN
ajema-5904	22	18	known	know	VERB
ajema-5904	22	19	as	as	ADP
ajema-5904	22	20	static	static	ADJ
ajema-5904	22	21	analysis	analysis	NOUN
ajema-5904	22	22	is	be	AUX
ajema-5904	22	23	used	use	VERB
ajema-5904	22	24	.	.	PUNCT
ajema-5904	23	1	nevertheless	nevertheless	ADV
ajema-5904	23	2	,	,	PUNCT
ajema-5904	23	3	code	code	NOUN
ajema-5904	23	4	obfuscation	obfuscation	NOUN
ajema-5904	23	5	is	be	AUX
ajema-5904	23	6	outside	outside	ADP
ajema-5904	23	7	the	the	DET
ajema-5904	23	8	scope	scope	NOUN
ajema-5904	23	9	of	of	ADP
ajema-5904	23	10	static	static	ADJ
ajema-5904	23	11	analysis	analysis	NOUN
ajema-5904	23	12	.	.	PUNCT
ajema-5904	24	1	with	with	ADP
ajema-5904	24	2	dynamic	dynamic	ADJ
ajema-5904	24	3	analysis	analysis	NOUN
ajema-5904	24	4	,	,	PUNCT
ajema-5904	24	5	the	the	DET
ajema-5904	24	6	malware	malware	NOUN
ajema-5904	24	7	is	be	AUX
ajema-5904	24	8	run	run	VERB
ajema-5904	24	9	in	in	ADP
ajema-5904	24	10	a	a	DET
ajema-5904	24	11	sandbox	sandbox	NOUN
ajema-5904	24	12	,	,	PUNCT
ajema-5904	24	13	simulator	simulator	NOUN
ajema-5904	24	14	,	,	PUNCT
ajema-5904	24	15	or	or	CCONJ
ajema-5904	24	16	virtual	virtual	ADJ
ajema-5904	24	17	environment	environment	NOUN
ajema-5904	24	18	,	,	PUNCT
ajema-5904	24	19	and	and	CCONJ
ajema-5904	24	20	its	its	PRON
ajema-5904	24	21	execution	execution	NOUN
ajema-5904	24	22	track	track	NOUN
ajema-5904	24	23	is	be	AUX
ajema-5904	24	24	examined	examine	VERB
ajema-5904	24	25	(	(	PUNCT
ajema-5904	24	26	behavior	behavior	NOUN
ajema-5904	24	27	analysis	analysis	NOUN
ajema-5904	24	28	)	)	PUNCT
ajema-5904	24	29	.	.	PUNCT
ajema-5904	25	1	while	while	SCONJ
ajema-5904	25	2	it	it	PRON
ajema-5904	25	3	is	be	AUX
ajema-5904	25	4	a	a	DET
ajema-5904	25	5	promising	promising	ADJ
ajema-5904	25	6	method	method	NOUN
ajema-5904	25	7	for	for	ADP
ajema-5904	25	8	handling	handle	VERB
ajema-5904	25	9	code	code	NOUN
ajema-5904	25	10	obfuscation	obfuscation	NOUN
ajema-5904	25	11	,	,	PUNCT
ajema-5904	25	12	it	it	PRON
ajema-5904	25	13	is	be	AUX
ajema-5904	25	14	difficult	difficult	ADJ
ajema-5904	25	15	,	,	PUNCT
ajema-5904	25	16	timeconsuming	timeconsuming	ADJ
ajema-5904	25	17	,	,	PUNCT
ajema-5904	25	18	and	and	CCONJ
ajema-5904	25	19	computationally	computationally	ADV
ajema-5904	25	20	costly	costly	ADJ
ajema-5904	25	21	.	.	PUNCT
ajema-5904	26	1	the	the	DET
ajema-5904	26	2	malwares	malware	NOUN
ajema-5904	26	3	were	be	AUX
ajema-5904	26	4	first	first	ADV
ajema-5904	26	5	converted	convert	VERB
ajema-5904	26	6	into	into	ADP
ajema-5904	26	7	grayscale	grayscale	NOUN
ajema-5904	26	8	50	50	NUM
ajema-5904	26	9	journal	journal	NOUN
ajema-5904	26	10	of	of	ADP
ajema-5904	26	11	engineering	engineering	NOUN
ajema-5904	26	12	,	,	PUNCT
ajema-5904	26	13	mechanics	mechanic	NOUN
ajema-5904	26	14	and	and	CCONJ
ajema-5904	26	15	architecture	architecture	NOUN
ajema-5904	26	16	www	www	NOUN
ajema-5904	26	17	.	.	PUNCT
ajema-5904	27	1	grnjournal.us	grnjournal.us	X
ajema-5904	27	2	images	image	NOUN
ajema-5904	27	3	,	,	PUNCT
ajema-5904	27	4	and	and	CCONJ
ajema-5904	27	5	[	[	X
ajema-5904	27	6	4	4	X
ajema-5904	27	7	]	]	PUNCT
ajema-5904	27	8	used	use	VERB
ajema-5904	27	9	the	the	DET
ajema-5904	27	10	gist	gist	NOUN
ajema-5904	27	11	feature	feature	NOUN
ajema-5904	27	12	in	in	ADP
ajema-5904	27	13	conjunction	conjunction	NOUN
ajema-5904	27	14	with	with	ADP
ajema-5904	27	15	the	the	DET
ajema-5904	27	16	k	k	NOUN
ajema-5904	27	17	-	-	PUNCT
ajema-5904	27	18	nearest	near	ADJ
ajema-5904	27	19	neighbor	neighbor	NOUN
ajema-5904	27	20	classifier	classifier	NOUN
ajema-5904	27	21	to	to	PART
ajema-5904	27	22	classify	classify	VERB
ajema-5904	27	23	the	the	DET
ajema-5904	27	24	trojan	trojan	NOUN
ajema-5904	27	25	.	.	PUNCT
ajema-5904	28	1	the	the	DET
ajema-5904	28	2	malimg	malimg	NOUN
ajema-5904	28	3	dataset	dataset	VERB
ajema-5904	28	4	and	and	CCONJ
ajema-5904	28	5	the	the	DET
ajema-5904	28	6	author	author	NOUN
ajema-5904	28	7	's	's	PART
ajema-5904	28	8	thorough	thorough	ADJ
ajema-5904	28	9	examination	examination	NOUN
ajema-5904	28	10	and	and	CCONJ
ajema-5904	28	11	analysis	analysis	NOUN
ajema-5904	28	12	were	be	AUX
ajema-5904	28	13	made	make	VERB
ajema-5904	28	14	freely	freely	ADV
ajema-5904	28	15	available	available	ADJ
ajema-5904	28	16	for	for	ADP
ajema-5904	28	17	study	study	NOUN
ajema-5904	28	18	.	.	PUNCT
ajema-5904	29	1	owing	owe	VERB
ajema-5904	29	2	to	to	ADP
ajema-5904	29	3	the	the	DET
ajema-5904	29	4	current	current	ADJ
ajema-5904	29	5	explosion	explosion	NOUN
ajema-5904	29	6	in	in	ADP
ajema-5904	29	7	deep	deep	ADJ
ajema-5904	29	8	learning	learning	NOUN
ajema-5904	29	9	[	[	X
ajema-5904	29	10	7	7	NUM
ajema-5904	29	11	,	,	PUNCT
ajema-5904	29	12	8	8	NUM
ajema-5904	29	13	]	]	PUNCT
ajema-5904	29	14	,	,	PUNCT
ajema-5904	29	15	primarily	primarily	ADV
ajema-5904	29	16	to	to	ADP
ajema-5904	29	17	cnn	cnn	PROPN
ajema-5904	29	18	's	's	PART
ajema-5904	29	19	success	success	NOUN
ajema-5904	29	20	,	,	PUNCT
ajema-5904	29	21	numerous	numerous	ADJ
ajema-5904	29	22	techniques	technique	NOUN
ajema-5904	29	23	have	have	AUX
ajema-5904	29	24	been	be	AUX
ajema-5904	29	25	put	put	VERB
ajema-5904	29	26	out	out	ADP
ajema-5904	29	27	and	and	CCONJ
ajema-5904	29	28	assessed	assess	VERB
ajema-5904	29	29	by	by	ADP
ajema-5904	29	30	researchers	researcher	NOUN
ajema-5904	29	31	for	for	ADP
ajema-5904	29	32	the	the	DET
ajema-5904	29	33	malimg	malimg	PROPN
ajema-5904	29	34	dataset	dataset	NOUN
ajema-5904	29	35	's	's	PART
ajema-5904	29	36	use	use	NOUN
ajema-5904	29	37	in	in	ADP
ajema-5904	29	38	malware	malware	NOUN
ajema-5904	29	39	classification	classification	NOUN
ajema-5904	29	40	.	.	PUNCT
ajema-5904	30	1	furthermore	furthermore	ADV
ajema-5904	30	2	,	,	PUNCT
ajema-5904	30	3	for	for	ADP
ajema-5904	30	4	a	a	DET
ajema-5904	30	5	variety	variety	NOUN
ajema-5904	30	6	of	of	ADP
ajema-5904	30	7	applications	application	NOUN
ajema-5904	30	8	,	,	PUNCT
ajema-5904	30	9	deep	deep	ADJ
ajema-5904	30	10	learning	learning	NOUN
ajema-5904	30	11	-	-	PUNCT
ajema-5904	30	12	based	base	VERB
ajema-5904	30	13	approaches	approach	NOUN
ajema-5904	30	14	outperformed	outperform	VERB
ajema-5904	30	15	the	the	DET
ajema-5904	30	16	other	other	ADJ
ajema-5904	30	17	techniques	technique	NOUN
ajema-5904	30	18	[	[	X
ajema-5904	30	19	9	9	NUM
ajema-5904	30	20	]	]	PUNCT
ajema-5904	30	21	,	,	PUNCT
ajema-5904	30	22	[	[	X
ajema-5904	30	23	10	10	NUM
ajema-5904	30	24	]	]	PUNCT
ajema-5904	30	25	,	,	PUNCT
ajema-5904	30	26	[	[	X
ajema-5904	30	27	11	11	NUM
ajema-5904	30	28	]	]	PUNCT
ajema-5904	30	29	,	,	PUNCT
ajema-5904	30	30	[	[	X
ajema-5904	30	31	12	12	NUM
ajema-5904	30	32	]	]	PUNCT
ajema-5904	30	33	.	.	PUNCT
ajema-5904	31	1	the	the	DET
ajema-5904	31	2	comprehensive	comprehensive	ADJ
ajema-5904	31	3	review	review	NOUN
ajema-5904	31	4	of	of	ADP
ajema-5904	31	5	the	the	DET
ajema-5904	31	6	literature	literature	NOUN
ajema-5904	31	7	found	find	VERB
ajema-5904	31	8	in	in	ADP
ajema-5904	31	9	section	section	NOUN
ajema-5904	31	10	2	2	NUM
ajema-5904	31	11	.	.	X
ajema-5904	31	12	for	for	ADP
ajema-5904	31	13	malware	malware	NOUN
ajema-5904	31	14	detection	detection	NOUN
ajema-5904	31	15	and	and	CCONJ
ajema-5904	31	16	classification	classification	NOUN
ajema-5904	31	17	,	,	PUNCT
ajema-5904	31	18	the	the	DET
ajema-5904	31	19	majority	majority	NOUN
ajema-5904	31	20	of	of	ADP
ajema-5904	31	21	currently	currently	ADV
ajema-5904	31	22	used	use	VERB
ajema-5904	31	23	techniques	technique	NOUN
ajema-5904	31	24	rely	rely	VERB
ajema-5904	31	25	on	on	ADP
ajema-5904	31	26	cnn	cnn	PROPN
ajema-5904	31	27	or	or	CCONJ
ajema-5904	31	28	feature	feature	NOUN
ajema-5904	31	29	extraction	extraction	NOUN
ajema-5904	31	30	with	with	ADP
ajema-5904	31	31	machine	machine	NOUN
ajema-5904	31	32	learning	learn	VERB
ajema-5904	31	33	classifiers	classifier	NOUN
ajema-5904	31	34	.	.	PUNCT
ajema-5904	32	1	these	these	DET
ajema-5904	32	2	days	day	NOUN
ajema-5904	32	3	,	,	PUNCT
ajema-5904	32	4	malware	malware	NOUN
ajema-5904	32	5	detection	detection	NOUN
ajema-5904	32	6	and	and	CCONJ
ajema-5904	32	7	classification	classification	NOUN
ajema-5904	32	8	using	use	VERB
ajema-5904	32	9	deep	deep	ADJ
ajema-5904	32	10	learning	learning	NOUN
ajema-5904	32	11	-	-	PUNCT
ajema-5904	32	12	based	base	VERB
ajema-5904	32	13	techniques	technique	NOUN
ajema-5904	32	14	that	that	PRON
ajema-5904	32	15	turn	turn	VERB
ajema-5904	32	16	malware	malware	NOUN
ajema-5904	32	17	into	into	ADP
ajema-5904	32	18	images	image	NOUN
ajema-5904	32	19	work	work	VERB
ajema-5904	32	20	well	well	ADV
ajema-5904	32	21	.	.	PUNCT
ajema-5904	33	1	this	this	PRON
ajema-5904	33	2	is	be	AUX
ajema-5904	33	3	primarily	primarily	ADV
ajema-5904	33	4	because	because	SCONJ
ajema-5904	33	5	these	these	DET
ajema-5904	33	6	techniques	technique	NOUN
ajema-5904	33	7	fully	fully	ADV
ajema-5904	33	8	steer	steer	VERB
ajema-5904	33	9	clear	clear	ADJ
ajema-5904	33	10	of	of	ADP
ajema-5904	33	11	manual	manual	ADJ
ajema-5904	33	12	feature	feature	NOUN
ajema-5904	33	13	engineering	engineering	NOUN
ajema-5904	33	14	techniques	technique	NOUN
ajema-5904	33	15	.	.	PUNCT
ajema-5904	34	1	because	because	SCONJ
ajema-5904	34	2	manual	manual	ADJ
ajema-5904	34	3	feature	feature	NOUN
ajema-5904	34	4	engineering	engineering	NOUN
ajema-5904	34	5	techniques	technique	NOUN
ajema-5904	34	6	necessitate	necessitate	ADJ
ajema-5904	34	7	domain	domain	NOUN
ajema-5904	34	8	expertise	expertise	NOUN
ajema-5904	34	9	and	and	CCONJ
ajema-5904	34	10	perform	perform	VERB
ajema-5904	34	11	poorly	poorly	ADV
ajema-5904	34	12	when	when	SCONJ
ajema-5904	34	13	applied	apply	VERB
ajema-5904	34	14	to	to	ADP
ajema-5904	34	15	complicated	complicated	ADJ
ajema-5904	34	16	and	and	CCONJ
ajema-5904	34	17	unknown	unknown	ADJ
ajema-5904	34	18	data	datum	NOUN
ajema-5904	34	19	.	.	PUNCT
ajema-5904	35	1	nevertheless	nevertheless	ADV
ajema-5904	35	2	,	,	PUNCT
ajema-5904	35	3	these	these	DET
ajema-5904	35	4	techniques	technique	NOUN
ajema-5904	35	5	fail	fail	VERB
ajema-5904	35	6	to	to	PART
ajema-5904	35	7	recognize	recognize	VERB
ajema-5904	35	8	even	even	ADV
ajema-5904	35	9	a	a	DET
ajema-5904	35	10	small	small	ADJ
ajema-5904	35	11	portion	portion	NOUN
ajema-5904	35	12	of	of	ADP
ajema-5904	35	13	the	the	DET
ajema-5904	35	14	virus	virus	NOUN
ajema-5904	35	15	image	image	NOUN
ajema-5904	35	16	's	's	PART
ajema-5904	35	17	affected	affect	VERB
ajema-5904	35	18	areas	area	NOUN
ajema-5904	35	19	.	.	PUNCT
ajema-5904	36	1	the	the	DET
ajema-5904	36	2	majority	majority	NOUN
ajema-5904	36	3	of	of	ADP
ajema-5904	36	4	the	the	DET
ajema-5904	36	5	time	time	NOUN
ajema-5904	36	6	,	,	PUNCT
ajema-5904	36	7	virus	virus	NOUN
ajema-5904	36	8	writers	writer	NOUN
ajema-5904	36	9	merely	merely	ADV
ajema-5904	36	10	slightly	slightly	ADV
ajema-5904	36	11	alter	alter	VERB
ajema-5904	36	12	the	the	DET
ajema-5904	36	13	original	original	ADJ
ajema-5904	36	14	files	file	NOUN
ajema-5904	36	15	.	.	PUNCT
ajema-5904	37	1	by	by	ADP
ajema-5904	37	2	applying	apply	VERB
ajema-5904	37	3	filters	filter	NOUN
ajema-5904	37	4	to	to	ADP
ajema-5904	37	5	a	a	DET
ajema-5904	37	6	picture	picture	NOUN
ajema-5904	37	7	,	,	PUNCT
ajema-5904	37	8	cnn	cnn	PROPN
ajema-5904	37	9	can	can	AUX
ajema-5904	37	10	improve	improve	VERB
ajema-5904	37	11	its	its	PRON
ajema-5904	37	12	feature	feature	NOUN
ajema-5904	37	13	representation	representation	NOUN
ajema-5904	37	14	;	;	PUNCT
ajema-5904	37	15	but	but	CCONJ
ajema-5904	37	16	,	,	PUNCT
ajema-5904	37	17	its	its	PRON
ajema-5904	37	18	ability	ability	NOUN
ajema-5904	37	19	to	to	PART
ajema-5904	37	20	recognize	recognize	VERB
ajema-5904	37	21	afflicted	afflict	VERB
ajema-5904	37	22	areas	area	NOUN
ajema-5904	37	23	is	be	AUX
ajema-5904	37	24	severely	severely	ADV
ajema-5904	37	25	restricted	restrict	VERB
ajema-5904	37	26	.	.	PUNCT
ajema-5904	38	1	security	security	NOUN
ajema-5904	38	2	experts	expert	NOUN
ajema-5904	38	3	increased	increase	VERB
ajema-5904	38	4	the	the	DET
ajema-5904	38	5	breadth	breadth	NOUN
ajema-5904	38	6	and	and	CCONJ
ajema-5904	38	7	depth	depth	NOUN
ajema-5904	38	8	of	of	ADP
ajema-5904	38	9	the	the	DET
ajema-5904	38	10	cnn	cnn	PROPN
ajema-5904	38	11	network	network	NOUN
ajema-5904	38	12	in	in	ADP
ajema-5904	38	13	order	order	NOUN
ajema-5904	38	14	to	to	PART
ajema-5904	38	15	enhance	enhance	VERB
ajema-5904	38	16	cnn	cnn	PROPN
ajema-5904	38	17	's	's	PART
ajema-5904	38	18	effectiveness	effectiveness	NOUN
ajema-5904	38	19	on	on	ADP
ajema-5904	38	20	malware	malware	NOUN
ajema-5904	38	21	image	image	NOUN
ajema-5904	38	22	-	-	PUNCT
ajema-5904	38	23	based	base	VERB
ajema-5904	38	24	tasks	task	NOUN
ajema-5904	38	25	[	[	X
ajema-5904	38	26	13	13	NUM
ajema-5904	38	27	,	,	PUNCT
ajema-5904	38	28	14	14	NUM
ajema-5904	38	29	]	]	PUNCT
ajema-5904	38	30	.	.	PUNCT
ajema-5904	39	1	literature	literature	NOUN
ajema-5904	39	2	survey	survey	VERB
ajema-5904	39	3	an	an	DET
ajema-5904	39	4	technique	technique	NOUN
ajema-5904	39	5	to	to	ADP
ajema-5904	39	6	malware	malware	NOUN
ajema-5904	39	7	detection	detection	NOUN
ajema-5904	39	8	known	know	VERB
ajema-5904	39	9	as	as	ADP
ajema-5904	39	10	custom	custom	ADV
ajema-5904	39	11	-	-	PUNCT
ajema-5904	39	12	trained	train	VERB
ajema-5904	39	13	transfer	transfer	NOUN
ajema-5904	39	14	learning	learning	NOUN
ajema-5904	39	15	was	be	AUX
ajema-5904	39	16	introduced	introduce	VERB
ajema-5904	39	17	by	by	ADP
ajema-5904	39	18	marastoni	marastoni	PROPN
ajema-5904	39	19	et	et	PROPN
ajema-5904	39	20	al	al	PROPN
ajema-5904	39	21	.	.	PUNCT
ajema-5904	40	1	[	[	X
ajema-5904	40	2	34	34	NUM
ajema-5904	40	3	]	]	PUNCT
ajema-5904	40	4	.	.	PUNCT
ajema-5904	41	1	they	they	PRON
ajema-5904	41	2	trained	train	VERB
ajema-5904	41	3	the	the	DET
ajema-5904	41	4	cnn	cnn	PROPN
ajema-5904	41	5	model	model	NOUN
ajema-5904	41	6	as	as	ADP
ajema-5904	41	7	a	a	DET
ajema-5904	41	8	transfer	transfer	NOUN
ajema-5904	41	9	-	-	PUNCT
ajema-5904	41	10	learned	learn	VERB
ajema-5904	41	11	model	model	NOUN
ajema-5904	41	12	using	use	VERB
ajema-5904	41	13	a	a	DET
ajema-5904	41	14	bespoke	bespoke	NOUN
ajema-5904	41	15	obf	obf	NOUN
ajema-5904	41	16	dataset	dataset	VERB
ajema-5904	41	17	that	that	PRON
ajema-5904	41	18	was	be	AUX
ajema-5904	41	19	produced	produce	VERB
ajema-5904	41	20	utilizing	utilize	VERB
ajema-5904	41	21	obfuscation	obfuscation	NOUN
ajema-5904	41	22	techniques	technique	NOUN
ajema-5904	41	23	.	.	PUNCT
ajema-5904	42	1	the	the	DET
ajema-5904	42	2	malimg	malimg	PROPN
ajema-5904	42	3	and	and	CCONJ
ajema-5904	42	4	msm2015	msm2015	VERB
ajema-5904	42	5	datasets	dataset	NOUN
ajema-5904	42	6	were	be	AUX
ajema-5904	42	7	utilized	utilize	VERB
ajema-5904	42	8	to	to	PART
ajema-5904	42	9	train	train	VERB
ajema-5904	42	10	an	an	DET
ajema-5904	42	11	additional	additional	ADJ
ajema-5904	42	12	cnn	cnn	NOUN
ajema-5904	42	13	and	and	CCONJ
ajema-5904	42	14	lstm	lstm	PROPN
ajema-5904	42	15	model	model	NOUN
ajema-5904	42	16	,	,	PUNCT
ajema-5904	42	17	which	which	PRON
ajema-5904	42	18	was	be	AUX
ajema-5904	42	19	then	then	ADV
ajema-5904	42	20	used	use	VERB
ajema-5904	42	21	to	to	PART
ajema-5904	42	22	predict	predict	VERB
ajema-5904	42	23	each	each	DET
ajema-5904	42	24	other	other	ADJ
ajema-5904	42	25	's	's	PART
ajema-5904	42	26	malware	malware	NOUN
ajema-5904	42	27	dataset	dataset	VERB
ajema-5904	42	28	and	and	CCONJ
ajema-5904	42	29	confirm	confirm	VERB
ajema-5904	42	30	prediction	prediction	NOUN
ajema-5904	42	31	accuracy	accuracy	NOUN
ajema-5904	42	32	.	.	PUNCT
ajema-5904	43	1	they	they	PRON
ajema-5904	43	2	attained	attain	VERB
ajema-5904	43	3	lstm	lstm	ADJ
ajema-5904	43	4	-	-	PUNCT
ajema-5904	43	5	level	level	NOUN
ajema-5904	43	6	accuracy	accuracy	NOUN
ajema-5904	43	7	and	and	CCONJ
ajema-5904	43	8	suggested	suggest	VERB
ajema-5904	43	9	using	use	VERB
ajema-5904	43	10	bicubic	bicubic	ADJ
ajema-5904	43	11	interpolation	interpolation	NOUN
ajema-5904	43	12	to	to	PART
ajema-5904	43	13	obtain	obtain	VERB
ajema-5904	43	14	consistent	consistent	ADJ
ajema-5904	43	15	image	image	NOUN
ajema-5904	43	16	size	size	NOUN
ajema-5904	43	17	.	.	PUNCT
ajema-5904	44	1	but	but	CCONJ
ajema-5904	44	2	without	without	ADP
ajema-5904	44	3	data	data	NOUN
ajema-5904	44	4	balance	balance	NOUN
ajema-5904	44	5	,	,	PUNCT
ajema-5904	44	6	transfer	transfer	NOUN
ajema-5904	44	7	learning	learning	NOUN
ajema-5904	44	8	showed	show	VERB
ajema-5904	44	9	mediocre	mediocre	ADJ
ajema-5904	44	10	performance	performance	NOUN
ajema-5904	44	11	;	;	PUNCT
ajema-5904	44	12	moreover	moreover	ADV
ajema-5904	44	13	,	,	PUNCT
ajema-5904	44	14	the	the	DET
ajema-5904	44	15	models	model	NOUN
ajema-5904	44	16	required	require	VERB
ajema-5904	44	17	fixed	fix	VERB
ajema-5904	44	18	-	-	PUNCT
ajema-5904	44	19	size	size	NOUN
ajema-5904	44	20	pictures	picture	NOUN
ajema-5904	44	21	,	,	PUNCT
ajema-5904	44	22	which	which	PRON
ajema-5904	44	23	prolonged	prolong	VERB
ajema-5904	44	24	the	the	DET
ajema-5904	44	25	time	time	NOUN
ajema-5904	44	26	necessary	necessary	ADJ
ajema-5904	44	27	for	for	ADP
ajema-5904	44	28	data	datum	NOUN
ajema-5904	44	29	pre	pre	ADJ
ajema-5904	44	30	-	-	NOUN
ajema-5904	44	31	processing	processing	ADJ
ajema-5904	44	32	;	;	PUNCT
ajema-5904	44	33	and	and	CCONJ
ajema-5904	44	34	the	the	DET
ajema-5904	44	35	obfuscation	obfuscation	NOUN
ajema-5904	44	36	techniques	technique	NOUN
ajema-5904	44	37	needed	need	VERB
ajema-5904	44	38	to	to	PART
ajema-5904	44	39	be	be	AUX
ajema-5904	44	40	improved	improve	VERB
ajema-5904	44	41	.	.	PUNCT
ajema-5904	45	1	cnn	cnn	PROPN
ajema-5904	45	2	and	and	CCONJ
ajema-5904	45	3	mlp	mlp	PROPN
ajema-5904	45	4	models	model	NOUN
ajema-5904	45	5	should	should	AUX
ajema-5904	45	6	be	be	AUX
ajema-5904	45	7	used	use	VERB
ajema-5904	45	8	for	for	ADP
ajema-5904	45	9	malware	malware	NOUN
ajema-5904	45	10	sample	sample	NOUN
ajema-5904	45	11	analysis	analysis	NOUN
ajema-5904	45	12	and	and	CCONJ
ajema-5904	45	13	detection	detection	NOUN
ajema-5904	45	14	,	,	PUNCT
ajema-5904	45	15	according	accord	VERB
ajema-5904	45	16	to	to	ADP
ajema-5904	45	17	kim	kim	PROPN
ajema-5904	45	18	et	et	PROPN
ajema-5904	45	19	al	al	PROPN
ajema-5904	45	20	.	.	PUNCT
ajema-5904	46	1	[	[	X
ajema-5904	46	2	36	36	NUM
ajema-5904	46	3	]	]	PUNCT
ajema-5904	46	4	.	.	PUNCT
ajema-5904	47	1	they	they	PRON
ajema-5904	47	2	used	use	VERB
ajema-5904	47	3	their	their	PRON
ajema-5904	47	4	method	method	NOUN
ajema-5904	47	5	to	to	ADP
ajema-5904	47	6	the	the	DET
ajema-5904	47	7	publicly	publicly	ADV
ajema-5904	47	8	accessible	accessible	ADJ
ajema-5904	47	9	microsoft	microsoft	PROPN
ajema-5904	47	10	malware	malware	PROPN
ajema-5904	47	11	dataset	dataset	VERB
ajema-5904	47	12	.	.	PUNCT
ajema-5904	48	1	some	some	PRON
ajema-5904	48	2	of	of	ADP
ajema-5904	48	3	the	the	DET
ajema-5904	48	4	data	datum	NOUN
ajema-5904	48	5	in	in	ADP
ajema-5904	48	6	the	the	DET
ajema-5904	48	7	dataset	dataset	NOUN
ajema-5904	48	8	were	be	AUX
ajema-5904	48	9	eliminated	eliminate	VERB
ajema-5904	48	10	and	and	CCONJ
ajema-5904	48	11	converted	convert	VERB
ajema-5904	48	12	into	into	ADP
ajema-5904	48	13	an	an	DET
ajema-5904	48	14	abstract	abstract	ADJ
ajema-5904	48	15	visual	visual	ADJ
ajema-5904	48	16	graph	graph	NOUN
ajema-5904	48	17	in	in	ADP
ajema-5904	48	18	order	order	NOUN
ajema-5904	48	19	to	to	PART
ajema-5904	48	20	determine	determine	VERB
ajema-5904	48	21	which	which	DET
ajema-5904	48	22	category	category	NOUN
ajema-5904	48	23	the	the	DET
ajema-5904	48	24	virus	virus	NOUN
ajema-5904	48	25	is	be	AUX
ajema-5904	48	26	located	locate	VERB
ajema-5904	48	27	in	in	ADP
ajema-5904	48	28	.	.	PUNCT
ajema-5904	49	1	through	through	ADP
ajema-5904	49	2	comparison	comparison	NOUN
ajema-5904	49	3	investigation	investigation	NOUN
ajema-5904	49	4	,	,	PUNCT
ajema-5904	49	5	the	the	DET
ajema-5904	49	6	degree	degree	NOUN
ajema-5904	49	7	of	of	ADP
ajema-5904	49	8	similarity	similarity	NOUN
ajema-5904	49	9	between	between	ADP
ajema-5904	49	10	this	this	DET
ajema-5904	49	11	malware	malware	NOUN
ajema-5904	49	12	was	be	AUX
ajema-5904	49	13	discovered	discover	VERB
ajema-5904	49	14	.	.	PUNCT
ajema-5904	50	1	it	it	PRON
ajema-5904	50	2	has	have	AUX
ajema-5904	50	3	been	be	AUX
ajema-5904	50	4	discovered	discover	VERB
ajema-5904	50	5	that	that	SCONJ
ajema-5904	50	6	artificial	artificial	ADJ
ajema-5904	50	7	intelligence	intelligence	NOUN
ajema-5904	50	8	deep	deep	ADJ
ajema-5904	50	9	learning	learning	NOUN
ajema-5904	50	10	is	be	AUX
ajema-5904	50	11	a	a	DET
ajema-5904	50	12	useful	useful	ADJ
ajema-5904	50	13	technique	technique	NOUN
ajema-5904	50	14	for	for	ADP
ajema-5904	50	15	precisely	precisely	ADV
ajema-5904	50	16	and	and	CCONJ
ajema-5904	50	17	swiftly	swiftly	ADV
ajema-5904	50	18	identifying	identify	VERB
ajema-5904	50	19	malware	malware	NOUN
ajema-5904	50	20	using	use	VERB
ajema-5904	50	21	images	image	NOUN
ajema-5904	50	22	.	.	PUNCT
ajema-5904	51	1	this	this	DET
ajema-5904	51	2	technique	technique	NOUN
ajema-5904	51	3	outperformed	outperform	VERB
ajema-5904	51	4	the	the	DET
ajema-5904	51	5	conventional	conventional	ADJ
ajema-5904	51	6	signature	signature	NOUN
ajema-5904	51	7	-	-	PUNCT
ajema-5904	51	8	based	base	VERB
ajema-5904	51	9	strategy	strategy	NOUN
ajema-5904	51	10	in	in	ADP
ajema-5904	51	11	identifying	identify	VERB
ajema-5904	51	12	recently	recently	ADV
ajema-5904	51	13	discovered	discover	VERB
ajema-5904	51	14	malware	malware	NOUN
ajema-5904	51	15	.	.	PUNCT
ajema-5904	52	1	it	it	PRON
ajema-5904	52	2	was	be	AUX
ajema-5904	52	3	discovered	discover	VERB
ajema-5904	52	4	that	that	SCONJ
ajema-5904	52	5	the	the	DET
ajema-5904	52	6	mlp	mlp	PROPN
ajema-5904	52	7	classifier	classifier	NOUN
ajema-5904	52	8	needs	need	VERB
ajema-5904	52	9	to	to	PART
ajema-5904	52	10	be	be	AUX
ajema-5904	52	11	upgraded	upgrade	VERB
ajema-5904	52	12	since	since	SCONJ
ajema-5904	52	13	it	it	PRON
ajema-5904	52	14	was	be	AUX
ajema-5904	52	15	unreliable	unreliable	ADJ
ajema-5904	52	16	.	.	PUNCT
ajema-5904	53	1	in	in	ADP
ajema-5904	53	2	order	order	NOUN
ajema-5904	53	3	to	to	PART
ajema-5904	53	4	fully	fully	ADV
ajema-5904	53	5	utilize	utilize	VERB
ajema-5904	53	6	deep	deep	ADJ
ajema-5904	53	7	learning	learning	NOUN
ajema-5904	53	8	models	model	NOUN
ajema-5904	53	9	,	,	PUNCT
ajema-5904	53	10	sharma	sharma	PROPN
ajema-5904	53	11	et	et	PROPN
ajema-5904	53	12	al	al	PROPN
ajema-5904	53	13	.	.	PUNCT
ajema-5904	54	1	[	[	X
ajema-5904	54	2	40	40	NUM
ajema-5904	54	3	]	]	PUNCT
ajema-5904	54	4	used	use	VERB
ajema-5904	54	5	a	a	DET
ajema-5904	54	6	classification	classification	NOUN
ajema-5904	54	7	model	model	NOUN
ajema-5904	54	8	for	for	ADP
ajema-5904	54	9	malware	malware	NOUN
ajema-5904	54	10	that	that	PRON
ajema-5904	54	11	combined	combine	VERB
ajema-5904	54	12	them	they	PRON
ajema-5904	54	13	with	with	ADP
ajema-5904	54	14	cnn	cnn	PROPN
ajema-5904	54	15	and	and	CCONJ
ajema-5904	54	16	other	other	ADJ
ajema-5904	54	17	machine	machine	NOUN
ajema-5904	54	18	learning	learn	VERB
ajema-5904	54	19	classifiers	classifier	NOUN
ajema-5904	54	20	.	.	PUNCT
ajema-5904	55	1	by	by	ADP
ajema-5904	55	2	using	use	VERB
ajema-5904	55	3	mathematical	mathematical	ADJ
ajema-5904	55	4	functions	function	NOUN
ajema-5904	55	5	,	,	PUNCT
ajema-5904	55	6	these	these	DET
ajema-5904	55	7	models	model	NOUN
ajema-5904	55	8	made	make	VERB
ajema-5904	55	9	it	it	PRON
ajema-5904	55	10	possible	possible	ADJ
ajema-5904	55	11	to	to	PART
ajema-5904	55	12	identify	identify	VERB
ajema-5904	55	13	malware	malware	NOUN
ajema-5904	55	14	that	that	PRON
ajema-5904	55	15	was	be	AUX
ajema-5904	55	16	just	just	ADV
ajema-5904	55	17	released	release	VERB
ajema-5904	55	18	.	.	PUNCT
ajema-5904	56	1	in	in	ADP
ajema-5904	56	2	terms	term	NOUN
ajema-5904	56	3	of	of	ADP
ajema-5904	56	4	approach	approach	NOUN
ajema-5904	56	5	,	,	PUNCT
ajema-5904	56	6	cnn	cnn	PROPN
ajema-5904	56	7	alone	alone	ADV
ajema-5904	56	8	did	do	VERB
ajema-5904	56	9	the	the	DET
ajema-5904	56	10	best	good	ADJ
ajema-5904	56	11	on	on	ADP
ajema-5904	56	12	the	the	DET
ajema-5904	56	13	malimg	malimg	NOUN
ajema-5904	56	14	dataset	dataset	NOUN
ajema-5904	56	15	.	.	PUNCT
ajema-5904	57	1	architecture	architecture	NOUN
ajema-5904	57	2	for	for	ADP
ajema-5904	57	3	the	the	DET
ajema-5904	57	4	cnnsvm	cnnsvm	NOUN
ajema-5904	57	5	model	model	NOUN
ajema-5904	57	6	required	require	VERB
ajema-5904	57	7	to	to	PART
ajema-5904	57	8	be	be	AUX
ajema-5904	57	9	upgraded	upgrade	VERB
ajema-5904	57	10	.	.	PUNCT
ajema-5904	58	1	for	for	ADP
ajema-5904	58	2	multi	multi	ADJ
ajema-5904	58	3	-	-	ADJ
ajema-5904	58	4	class	class	ADJ
ajema-5904	58	5	situations	situation	NOUN
ajema-5904	58	6	,	,	PUNCT
ajema-5904	58	7	this	this	PRON
ajema-5904	58	8	would	would	AUX
ajema-5904	58	9	entail	entail	VERB
ajema-5904	58	10	utilizing	utilize	VERB
ajema-5904	58	11	many	many	ADJ
ajema-5904	58	12	svms	svms	NOUN
ajema-5904	58	13	,	,	PUNCT
ajema-5904	58	14	increasing	increase	VERB
ajema-5904	58	15	the	the	DET
ajema-5904	58	16	size	size	NOUN
ajema-5904	58	17	of	of	ADP
ajema-5904	58	18	the	the	DET
ajema-5904	58	19	model	model	NOUN
ajema-5904	58	20	and	and	CCONJ
ajema-5904	58	21	hence	hence	ADV
ajema-5904	58	22	the	the	DET
ajema-5904	58	23	calculation	calculation	NOUN
ajema-5904	58	24	time	time	NOUN
ajema-5904	58	25	.	.	PUNCT
ajema-5904	59	1	in	in	ADP
ajema-5904	59	2	[	[	PUNCT
ajema-5904	59	3	27	27	NUM
ajema-5904	59	4	p3	p3	NOUN
ajema-5904	59	5	]	]	PUNCT
ajema-5904	59	6	,	,	PUNCT
ajema-5904	59	7	a	a	DET
ajema-5904	59	8	cnn	cnn	PROPN
ajema-5904	59	9	is	be	AUX
ajema-5904	59	10	trained	train	VERB
ajema-5904	59	11	with	with	ADP
ajema-5904	59	12	a	a	DET
ajema-5904	59	13	multi	multi	ADJ
ajema-5904	59	14	-	-	ADJ
ajema-5904	59	15	headed	headed	ADJ
ajema-5904	59	16	attention	attention	NOUN
ajema-5904	59	17	-	-	PUNCT
ajema-5904	59	18	based	base	VERB
ajema-5904	59	19	approach	approach	NOUN
ajema-5904	59	20	to	to	PART
ajema-5904	59	21	identify	identify	VERB
ajema-5904	59	22	and	and	CCONJ
ajema-5904	59	23	categorize	categorize	VERB
ajema-5904	59	24	the	the	DET
ajema-5904	59	25	minuscule	minuscule	ADJ
ajema-5904	59	26	afflicted	afflict	VERB
ajema-5904	59	27	regions	region	NOUN
ajema-5904	59	28	within	within	ADP
ajema-5904	59	29	the	the	DET
ajema-5904	59	30	whole	whole	ADJ
ajema-5904	59	31	picture	picture	NOUN
ajema-5904	59	32	.	.	PUNCT
ajema-5904	60	1	on	on	ADP
ajema-5904	60	2	many	many	ADJ
ajema-5904	60	3	data	datum	NOUN
ajema-5904	60	4	splits	split	NOUN
ajema-5904	60	5	of	of	ADP
ajema-5904	60	6	the	the	DET
ajema-5904	60	7	testing	testing	NOUN
ajema-5904	60	8	and	and	CCONJ
ajema-5904	60	9	training	training	NOUN
ajema-5904	60	10	malware	malware	NOUN
ajema-5904	60	11	image	image	NOUN
ajema-5904	60	12	benchmark	benchmark	NOUN
ajema-5904	60	13	dataset	dataset	NOUN
ajema-5904	60	14	,	,	PUNCT
ajema-5904	60	15	the	the	DET
ajema-5904	60	16	anticipated	anticipate	VERB
ajema-5904	60	17	multiheaded	multiheade	VERB
ajema-5904	60	18	attentionbased	attentionbase	VERB
ajema-5904	60	19	cnn	cnn	PROPN
ajema-5904	60	20	approach	approach	NOUN
ajema-5904	60	21	performed	perform	VERB
ajema-5904	60	22	comparably	comparably	ADV
ajema-5904	60	23	to	to	ADP
ajema-5904	60	24	other	other	ADJ
ajema-5904	60	25	non	non	ADJ
ajema-5904	60	26	-	-	ADJ
ajema-5904	60	27	attention	attention	ADJ
ajema-5904	60	28	-	-	PUNCT
ajema-5904	60	29	cnn	cnn	NOUN
ajema-5904	60	30	-	-	PUNCT
ajema-5904	60	31	based	base	VERB
ajema-5904	60	32	strategies	strategy	NOUN
ajema-5904	60	33	.	.	PUNCT
ajema-5904	61	1	using	use	VERB
ajema-5904	61	2	the	the	DET
ajema-5904	61	3	fusion	fusion	NOUN
ajema-5904	61	4	feature	feature	NOUN
ajema-5904	61	5	set	set	NOUN
ajema-5904	61	6	paradigm	paradigm	NOUN
ajema-5904	61	7	,	,	PUNCT
ajema-5904	61	8	chaganti	chaganti	PROPN
ajema-5904	61	9	et	et	PROPN
ajema-5904	61	10	al	al	PROPN
ajema-5904	61	11	.	.	PUNCT
ajema-5904	62	1	[	[	X
ajema-5904	62	2	28	28	NUM
ajema-5904	62	3	p3	p3	PROPN
ajema-5904	62	4	]	]	PUNCT
ajema-5904	62	5	presented	present	VERB
ajema-5904	62	6	a	a	DET
ajema-5904	62	7	dl	dl	PROPN
ajema-5904	62	8	-	-	PUNCT
ajema-5904	62	9	based	base	VERB
ajema-5904	62	10	cnn	cnn	PROPN
ajema-5904	62	11	technique	technique	NOUN
ajema-5904	62	12	to	to	PART
ajema-5904	62	13	identify	identify	VERB
ajema-5904	62	14	malware	malware	NOUN
ajema-5904	62	15	on	on	ADP
ajema-5904	62	16	portable	portable	ADJ
ajema-5904	62	17	executable	executable	ADJ
ajema-5904	62	18	(	(	PUNCT
ajema-5904	62	19	pe	pe	INTJ
ajema-5904	62	20	)	)	PUNCT
ajema-5904	62	21	dual	dual	ADJ
ajema-5904	62	22	files	file	NOUN
ajema-5904	62	23	.	.	PUNCT
ajema-5904	63	1	we	we	PRON
ajema-5904	63	2	provided	provide	VERB
ajema-5904	63	3	a	a	DET
ajema-5904	63	4	comprehensive	comprehensive	ADJ
ajema-5904	63	5	evaluation	evaluation	NOUN
ajema-5904	63	6	of	of	ADP
ajema-5904	63	7	the	the	DET
ajema-5904	63	8	performance	performance	NOUN
ajema-5904	63	9	of	of	ADP
ajema-5904	63	10	several	several	ADJ
ajema-5904	63	11	dl	dl	NOUN
ajema-5904	63	12	technique	technique	NOUN
ajema-5904	63	13	structures	structure	NOUN
ajema-5904	63	14	and	and	CCONJ
ajema-5904	63	15	machine	machine	NOUN
ajema-5904	63	16	51	51	NUM
ajema-5904	63	17	journal	journal	NOUN
ajema-5904	63	18	of	of	ADP
ajema-5904	63	19	engineering	engineering	NOUN
ajema-5904	63	20	,	,	PUNCT
ajema-5904	63	21	mechanics	mechanic	NOUN
ajema-5904	63	22	and	and	CCONJ
ajema-5904	63	23	architecture	architecture	NOUN
ajema-5904	63	24	www	www	NOUN
ajema-5904	63	25	.	.	PUNCT
ajema-5904	64	1	grnjournal.us	grnjournal.us	X
ajema-5904	64	2	learning	learn	VERB
ajema-5904	64	3	classifiers	classifier	NOUN
ajema-5904	64	4	,	,	PUNCT
ajema-5904	64	5	such	such	ADJ
ajema-5904	64	6	as	as	ADP
ajema-5904	64	7	support	support	NOUN
ajema-5904	64	8	vector	vector	NOUN
ajema-5904	64	9	machines	machine	NOUN
ajema-5904	64	10	(	(	PUNCT
ajema-5904	64	11	svm	svm	PROPN
ajema-5904	64	12	)	)	PUNCT
ajema-5904	64	13	,	,	PUNCT
ajema-5904	64	14	using	use	VERB
ajema-5904	64	15	multi	multi	ADJ
ajema-5904	64	16	-	-	ADJ
ajema-5904	64	17	aspect	aspect	ADJ
ajema-5904	64	18	feature	feature	NOUN
ajema-5904	64	19	sets	set	NOUN
ajema-5904	64	20	that	that	PRON
ajema-5904	64	21	included	include	VERB
ajema-5904	64	22	static	static	ADJ
ajema-5904	64	23	,	,	PUNCT
ajema-5904	64	24	dynamic	dynamic	ADJ
ajema-5904	64	25	,	,	PUNCT
ajema-5904	64	26	and	and	CCONJ
ajema-5904	64	27	picture	picture	NOUN
ajema-5904	64	28	characteristics	characteristic	NOUN
ajema-5904	64	29	.	.	PUNCT
ajema-5904	65	1	the	the	DET
ajema-5904	65	2	created	create	VERB
ajema-5904	65	3	cnn	cnn	PROPN
ajema-5904	65	4	technique	technique	NOUN
ajema-5904	65	5	.	.	PUNCT
ajema-5904	66	1	opcode	opcode	PROPN
ajema-5904	66	2	sequences	sequence	NOUN
ajema-5904	66	3	were	be	AUX
ajema-5904	66	4	employed	employ	VERB
ajema-5904	66	5	in	in	ADP
ajema-5904	66	6	earlier	early	ADJ
ajema-5904	66	7	research	research	NOUN
ajema-5904	66	8	on	on	ADP
ajema-5904	66	9	word2vec	word2vec	ADV
ajema-5904	66	10	-	-	PUNCT
ajema-5904	66	11	based	base	VERB
ajema-5904	66	12	long	long	ADJ
ajema-5904	66	13	short	short	ADJ
ajema-5904	66	14	-	-	PUNCT
ajema-5904	66	15	term	term	NOUN
ajema-5904	66	16	memory	memory	NOUN
ajema-5904	66	17	(	(	PUNCT
ajema-5904	66	18	ltsm	ltsm	NOUN
ajema-5904	66	19	)	)	PUNCT
ajema-5904	67	1	[	[	X
ajema-5904	67	2	20	20	NUM
ajema-5904	67	3	p4	p4	ADJ
ajema-5904	67	4	]	]	X
ajema-5904	67	5	,	,	PUNCT
ajema-5904	67	6	and	and	CCONJ
ajema-5904	67	7	an	an	DET
ajema-5904	67	8	lstm	lstm	NOUN
ajema-5904	67	9	was	be	AUX
ajema-5904	67	10	utilized	utilize	VERB
ajema-5904	67	11	to	to	PART
ajema-5904	67	12	take	take	VERB
ajema-5904	67	13	into	into	ADP
ajema-5904	67	14	account	account	NOUN
ajema-5904	67	15	the	the	DET
ajema-5904	67	16	opcodes	opcode	NOUN
ajema-5904	67	17	'	'	PART
ajema-5904	67	18	sequential	sequential	ADJ
ajema-5904	67	19	order	order	NOUN
ajema-5904	67	20	in	in	ADP
ajema-5904	67	21	the	the	DET
ajema-5904	67	22	file	file	NOUN
ajema-5904	67	23	.	.	PUNCT
ajema-5904	68	1	they	they	PRON
ajema-5904	68	2	passed	pass	VERB
ajema-5904	68	3	the	the	DET
ajema-5904	68	4	obtained	obtain	VERB
ajema-5904	68	5	representation	representation	NOUN
ajema-5904	68	6	to	to	ADP
ajema-5904	68	7	the	the	DET
ajema-5904	68	8	lstm	lstm	NOUN
ajema-5904	68	9	after	after	ADP
ajema-5904	68	10	encoding	encode	VERB
ajema-5904	68	11	opcodes	opcode	NOUN
ajema-5904	68	12	using	use	VERB
ajema-5904	68	13	the	the	DET
ajema-5904	68	14	word2vec	word2vec	X
ajema-5904	68	15	embedding	embed	VERB
ajema-5904	68	16	technique	technique	NOUN
ajema-5904	68	17	.	.	PUNCT
ajema-5904	69	1	in	in	ADP
ajema-5904	69	2	order	order	NOUN
ajema-5904	69	3	to	to	PART
ajema-5904	69	4	classify	classify	VERB
ajema-5904	69	5	malware	malware	NOUN
ajema-5904	69	6	,	,	PUNCT
ajema-5904	69	7	qiao	qiao	PROPN
ajema-5904	69	8	et	et	PROPN
ajema-5904	69	9	al	al	PROPN
ajema-5904	69	10	.	.	PROPN
ajema-5904	69	11	suggested	suggest	VERB
ajema-5904	69	12	transforming	transform	VERB
ajema-5904	69	13	grayscale	grayscale	NOUN
ajema-5904	69	14	pictures	picture	NOUN
ajema-5904	69	15	using	use	VERB
ajema-5904	69	16	word	word	NOUN
ajema-5904	69	17	vector	vector	NOUN
ajema-5904	69	18	similarity	similarity	NOUN
ajema-5904	69	19	determined	determine	VERB
ajema-5904	69	20	from	from	ADP
ajema-5904	69	21	opcodes	opcode	NOUN
ajema-5904	69	22	and	and	CCONJ
ajema-5904	69	23	sending	send	VERB
ajema-5904	69	24	the	the	DET
ajema-5904	69	25	results	result	NOUN
ajema-5904	69	26	to	to	ADP
ajema-5904	69	27	a	a	DET
ajema-5904	69	28	cnn	cnn	NOUN
ajema-5904	70	1	[	[	X
ajema-5904	70	2	21	21	NUM
ajema-5904	70	3	p4	p4	ADJ
ajema-5904	70	4	]	]	X
ajema-5904	70	5	.	.	PUNCT
ajema-5904	71	1	in	in	ADP
ajema-5904	71	2	[	[	X
ajema-5904	71	3	18	18	NUM
ajema-5904	71	4	p5	p5	NOUN
ajema-5904	71	5	]	]	PUNCT
ajema-5904	71	6	,	,	PUNCT
ajema-5904	71	7	an	an	DET
ajema-5904	71	8	advanced	advanced	ADJ
ajema-5904	71	9	learning	learning	NOUN
ajema-5904	71	10	machine	machine	NOUN
ajema-5904	71	11	(	(	PUNCT
ajema-5904	71	12	elm	elm	PROPN
ajema-5904	71	13	)	)	PUNCT
ajema-5904	71	14	and	and	CCONJ
ajema-5904	71	15	a	a	DET
ajema-5904	71	16	cnn	cnn	NOUN
ajema-5904	71	17	were	be	AUX
ajema-5904	71	18	utilized	utilize	VERB
ajema-5904	71	19	to	to	PART
ajema-5904	71	20	classify	classify	VERB
ajema-5904	71	21	malware	malware	NOUN
ajema-5904	71	22	.	.	PUNCT
ajema-5904	72	1	comparing	compare	VERB
ajema-5904	72	2	the	the	DET
ajema-5904	72	3	two	two	NUM
ajema-5904	72	4	algorithms	algorithm	NOUN
ajema-5904	72	5	'	'	PART
ajema-5904	72	6	performance	performance	NOUN
ajema-5904	72	7	in	in	ADP
ajema-5904	72	8	identifying	identify	VERB
ajema-5904	72	9	malware	malware	NOUN
ajema-5904	72	10	photos	photo	NOUN
ajema-5904	72	11	is	be	AUX
ajema-5904	72	12	the	the	DET
ajema-5904	72	13	primary	primary	ADJ
ajema-5904	72	14	objective	objective	NOUN
ajema-5904	72	15	of	of	ADP
ajema-5904	72	16	this	this	DET
ajema-5904	72	17	paper	paper	NOUN
ajema-5904	72	18	.	.	PUNCT
ajema-5904	73	1	the	the	DET
ajema-5904	73	2	results	result	NOUN
ajema-5904	73	3	show	show	VERB
ajema-5904	73	4	that	that	SCONJ
ajema-5904	73	5	elms	elm	NOUN
ajema-5904	73	6	are	be	AUX
ajema-5904	73	7	better	well	ADV
ajema-5904	73	8	appropriate	appropriate	ADJ
ajema-5904	73	9	for	for	ADP
ajema-5904	73	10	the	the	DET
ajema-5904	73	11	task	task	NOUN
ajema-5904	73	12	at	at	ADP
ajema-5904	73	13	hand	hand	NOUN
ajema-5904	73	14	,	,	PUNCT
ajema-5904	73	15	being	be	AUX
ajema-5904	73	16	quicker	quick	ADJ
ajema-5904	73	17	and	and	CCONJ
ajema-5904	73	18	more	more	ADV
ajema-5904	73	19	accurate	accurate	ADJ
ajema-5904	73	20	than	than	ADP
ajema-5904	73	21	cnns	cnn	NOUN
ajema-5904	73	22	.	.	PUNCT
ajema-5904	74	1	this	this	DET
ajema-5904	74	2	study	study	NOUN
ajema-5904	74	3	trains	train	VERB
ajema-5904	74	4	their	their	PRON
ajema-5904	74	5	models	model	NOUN
ajema-5904	74	6	using	use	VERB
ajema-5904	74	7	photos	photo	NOUN
ajema-5904	74	8	from	from	ADP
ajema-5904	74	9	the	the	DET
ajema-5904	74	10	dataset	dataset	NOUN
ajema-5904	74	11	of	of	ADP
ajema-5904	74	12	nataraj	nataraj	PROPN
ajema-5904	74	13	et	et	PROPN
ajema-5904	74	14	al	al	PROPN
ajema-5904	74	15	.	.	PUNCT
ajema-5904	75	1	[	[	X
ajema-5904	75	2	30	30	NUM
ajema-5904	75	3	p5	p5	NOUN
ajema-5904	75	4	]	]	PUNCT
ajema-5904	75	5	.	.	PUNCT
ajema-5904	76	1	the	the	DET
ajema-5904	76	2	deep	deep	ADJ
ajema-5904	76	3	residual	residual	ADJ
ajema-5904	76	4	network	network	NOUN
ajema-5904	76	5	architecture	architecture	NOUN
ajema-5904	76	6	with	with	ADP
ajema-5904	76	7	50	50	NUM
ajema-5904	76	8	layers	layer	NOUN
ajema-5904	76	9	from	from	ADP
ajema-5904	76	10	[	[	PUNCT
ajema-5904	76	11	16	16	NUM
ajema-5904	76	12	p5	p5	NOUN
ajema-5904	76	13	]	]	PUNCT
ajema-5904	76	14	is	be	AUX
ajema-5904	76	15	used	use	VERB
ajema-5904	76	16	by	by	ADP
ajema-5904	76	17	the	the	DET
ajema-5904	76	18	authors	author	NOUN
ajema-5904	76	19	in	in	ADP
ajema-5904	76	20	[	[	X
ajema-5904	76	21	40	40	NUM
ajema-5904	76	22	p5	p5	NOUN
ajema-5904	76	23	]	]	PUNCT
ajema-5904	76	24	to	to	PART
ajema-5904	76	25	categorize	categorize	VERB
ajema-5904	76	26	malware	malware	NOUN
ajema-5904	76	27	from	from	ADP
ajema-5904	76	28	the	the	DET
ajema-5904	76	29	malimg	malimg	NOUN
ajema-5904	76	30	dataset	dataset	VERB
ajema-5904	76	31	.	.	PUNCT
ajema-5904	77	1	in	in	ADP
ajema-5904	77	2	order	order	NOUN
ajema-5904	77	3	to	to	PART
ajema-5904	77	4	categorize	categorize	VERB
ajema-5904	77	5	the	the	DET
ajema-5904	77	6	malware	malware	NOUN
ajema-5904	77	7	samples	sample	NOUN
ajema-5904	77	8	into	into	ADP
ajema-5904	77	9	their	their	PRON
ajema-5904	77	10	respective	respective	ADJ
ajema-5904	77	11	families	family	NOUN
ajema-5904	77	12	,	,	PUNCT
ajema-5904	77	13	the	the	DET
ajema-5904	77	14	network	network	NOUN
ajema-5904	77	15	is	be	AUX
ajema-5904	77	16	first	first	ADV
ajema-5904	77	17	trained	train	VERB
ajema-5904	77	18	on	on	ADP
ajema-5904	77	19	a	a	DET
ajema-5904	77	20	common	common	ADJ
ajema-5904	77	21	object	object	NOUN
ajema-5904	77	22	identification	identification	NOUN
ajema-5904	77	23	task	task	NOUN
ajema-5904	77	24	.	.	PUNCT
ajema-5904	78	1	after	after	ADP
ajema-5904	78	2	that	that	PRON
ajema-5904	78	3	,	,	PUNCT
ajema-5904	78	4	the	the	DET
ajema-5904	78	5	last	last	ADJ
ajema-5904	78	6	layer	layer	NOUN
ajema-5904	78	7	is	be	AUX
ajema-5904	78	8	removed	remove	VERB
ajema-5904	78	9	from	from	ADP
ajema-5904	78	10	the	the	DET
ajema-5904	78	11	model	model	NOUN
ajema-5904	78	12	and	and	CCONJ
ajema-5904	78	13	a	a	DET
ajema-5904	78	14	new	new	ADJ
ajema-5904	78	15	dense	dense	ADJ
ajema-5904	78	16	model	model	NOUN
ajema-5904	78	17	with	with	ADP
ajema-5904	78	18	25	25	NUM
ajema-5904	78	19	output	output	NOUN
ajema-5904	78	20	nodes	node	NOUN
ajema-5904	78	21	is	be	AUX
ajema-5904	78	22	created	create	VERB
ajema-5904	78	23	.	.	PUNCT
ajema-5904	79	1	slight	slight	ADJ
ajema-5904	79	2	modifications	modification	NOUN
ajema-5904	79	3	to	to	ADP
ajema-5904	79	4	the	the	DET
ajema-5904	79	5	previously	previously	ADV
ajema-5904	79	6	described	describe	VERB
ajema-5904	79	7	method	method	NOUN
ajema-5904	79	8	have	have	AUX
ajema-5904	79	9	been	be	AUX
ajema-5904	79	10	suggested	suggest	VERB
ajema-5904	79	11	by	by	ADP
ajema-5904	79	12	other	other	ADJ
ajema-5904	79	13	recent	recent	ADJ
ajema-5904	79	14	publications	publication	NOUN
ajema-5904	79	15	,	,	PUNCT
ajema-5904	79	16	such	such	ADJ
ajema-5904	79	17	[	[	PUNCT
ajema-5904	79	18	9	9	NUM
ajema-5904	79	19	p5	p5	NOUN
ajema-5904	79	20	]	]	PUNCT
ajema-5904	79	21	and	and	CCONJ
ajema-5904	79	22	[	[	X
ajema-5904	79	23	5	5	NUM
ajema-5904	79	24	p5	p5	NOUN
ajema-5904	79	25	]	]	PUNCT
ajema-5904	79	26	,	,	PUNCT
ajema-5904	79	27	who	who	PRON
ajema-5904	79	28	always	always	ADV
ajema-5904	79	29	pre	pre	VERB
ajema-5904	79	30	-	-	VERB
ajema-5904	79	31	train	train	VERB
ajema-5904	79	32	their	their	PRON
ajema-5904	79	33	model	model	NOUN
ajema-5904	79	34	on	on	ADP
ajema-5904	79	35	typical	typical	ADJ
ajema-5904	79	36	picture	picture	NOUN
ajema-5904	79	37	classification	classification	NOUN
ajema-5904	79	38	tasks	task	NOUN
ajema-5904	79	39	.	.	PUNCT
ajema-5904	80	1	convolutional	convolutional	ADJ
ajema-5904	80	2	neural	neural	ADJ
ajema-5904	80	3	network	network	NOUN
ajema-5904	80	4	cnns	cnn	NOUN
ajema-5904	80	5	are	be	AUX
ajema-5904	80	6	popularly	popularly	ADV
ajema-5904	80	7	employed	employ	VERB
ajema-5904	80	8	for	for	ADP
ajema-5904	80	9	image	image	NOUN
ajema-5904	80	10	identification	identification	NOUN
ajema-5904	80	11	and	and	CCONJ
ajema-5904	80	12	classification	classification	NOUN
ajema-5904	80	13	.	.	PUNCT
ajema-5904	81	1	they	they	PRON
ajema-5904	81	2	are	be	AUX
ajema-5904	81	3	feed	feed	NOUN
ajema-5904	81	4	-	-	PUNCT
ajema-5904	81	5	forward	forward	ADV
ajema-5904	81	6	neural	neural	ADJ
ajema-5904	81	7	network	network	NOUN
ajema-5904	81	8	models	model	NOUN
ajema-5904	81	9	that	that	PRON
ajema-5904	81	10	are	be	AUX
ajema-5904	81	11	inspired	inspire	VERB
ajema-5904	81	12	by	by	ADP
ajema-5904	81	13	the	the	DET
ajema-5904	81	14	human	human	ADJ
ajema-5904	81	15	visual	visual	ADJ
ajema-5904	81	16	cortex	cortex	NOUN
ajema-5904	81	17	.	.	PUNCT
ajema-5904	82	1	their	their	PRON
ajema-5904	82	2	translational	translational	ADJ
ajema-5904	82	3	invariance	invariance	NOUN
ajema-5904	82	4	has	have	AUX
ajema-5904	82	5	been	be	AUX
ajema-5904	82	6	credited	credit	VERB
ajema-5904	82	7	with	with	ADP
ajema-5904	82	8	their	their	PRON
ajema-5904	82	9	performance	performance	NOUN
ajema-5904	82	10	in	in	ADP
ajema-5904	82	11	image	image	NOUN
ajema-5904	82	12	-	-	PUNCT
ajema-5904	82	13	related	relate	VERB
ajema-5904	82	14	learning	learning	NOUN
ajema-5904	82	15	tasks	task	NOUN
ajema-5904	82	16	,	,	PUNCT
ajema-5904	82	17	enabling	enable	VERB
ajema-5904	82	18	them	they	PRON
ajema-5904	82	19	to	to	PART
ajema-5904	82	20	be	be	AUX
ajema-5904	82	21	applied	apply	VERB
ajema-5904	82	22	to	to	ADP
ajema-5904	82	23	issues	issue	NOUN
ajema-5904	82	24	like	like	ADP
ajema-5904	82	25	handwritten	handwritten	ADJ
ajema-5904	82	26	character	character	NOUN
ajema-5904	82	27	recognition	recognition	NOUN
ajema-5904	82	28	and	and	CCONJ
ajema-5904	82	29	face	face	NOUN
ajema-5904	82	30	recognition	recognition	NOUN
ajema-5904	82	31	.	.	PUNCT
ajema-5904	83	1	cnns	cnns	PROPN
ajema-5904	83	2	perform	perform	VERB
ajema-5904	83	3	a	a	DET
ajema-5904	83	4	great	great	ADJ
ajema-5904	83	5	job	job	NOUN
ajema-5904	83	6	of	of	ADP
ajema-5904	83	7	extracting	extract	VERB
ajema-5904	83	8	spatial	spatial	ADJ
ajema-5904	83	9	information	information	NOUN
ajema-5904	83	10	from	from	ADP
ajema-5904	83	11	the	the	DET
ajema-5904	83	12	data	datum	NOUN
ajema-5904	83	13	in	in	ADP
ajema-5904	83	14	general	general	ADJ
ajema-5904	83	15	.	.	PUNCT
ajema-5904	84	1	cnns	cnns	PROPN
ajema-5904	84	2	are	be	AUX
ajema-5904	84	3	fundamentally	fundamentally	ADV
ajema-5904	84	4	made	make	VERB
ajema-5904	84	5	up	up	ADP
ajema-5904	84	6	of	of	ADP
ajema-5904	84	7	a	a	DET
ajema-5904	84	8	dense	dense	ADJ
ajema-5904	84	9	layer	layer	NOUN
ajema-5904	84	10	coupled	couple	VERB
ajema-5904	84	11	to	to	ADP
ajema-5904	84	12	at	at	ADV
ajema-5904	84	13	least	least	ADV
ajema-5904	84	14	one	one	NUM
ajema-5904	84	15	convolutional	convolutional	ADJ
ajema-5904	84	16	layer	layer	NOUN
ajema-5904	84	17	.	.	PUNCT
ajema-5904	85	1	during	during	ADP
ajema-5904	85	2	the	the	DET
ajema-5904	85	3	training	training	NOUN
ajema-5904	85	4	phase	phase	NOUN
ajema-5904	85	5	,	,	PUNCT
ajema-5904	85	6	the	the	DET
ajema-5904	85	7	convolutional	convolutional	ADJ
ajema-5904	85	8	layer	layer	NOUN
ajema-5904	85	9	applies	apply	VERB
ajema-5904	85	10	a	a	DET
ajema-5904	85	11	convolution	convolution	NOUN
ajema-5904	85	12	to	to	ADP
ajema-5904	85	13	the	the	DET
ajema-5904	85	14	input	input	NOUN
ajema-5904	85	15	to	to	PART
ajema-5904	85	16	separate	separate	VERB
ajema-5904	85	17	the	the	DET
ajema-5904	85	18	features	feature	NOUN
ajema-5904	85	19	that	that	SCONJ
ajema-5904	85	20	the	the	DET
ajema-5904	85	21	network	network	NOUN
ajema-5904	85	22	considers	consider	VERB
ajema-5904	85	23	significant	significant	ADJ
ajema-5904	85	24	.	.	PUNCT
ajema-5904	86	1	the	the	DET
ajema-5904	86	2	input	input	NOUN
ajema-5904	86	3	photos	photo	NOUN
ajema-5904	86	4	are	be	AUX
ajema-5904	86	5	reduced	reduce	VERB
ajema-5904	86	6	to	to	ADP
ajema-5904	86	7	a	a	DET
ajema-5904	86	8	smaller	small	ADJ
ajema-5904	86	9	feature	feature	NOUN
ajema-5904	86	10	set	set	VERB
ajema-5904	86	11	by	by	ADP
ajema-5904	86	12	this	this	DET
ajema-5904	86	13	procedure	procedure	NOUN
ajema-5904	86	14	,	,	PUNCT
ajema-5904	86	15	which	which	PRON
ajema-5904	86	16	significantly	significantly	ADV
ajema-5904	86	17	lowers	lower	VERB
ajema-5904	86	18	the	the	DET
ajema-5904	86	19	number	number	NOUN
ajema-5904	86	20	of	of	ADP
ajema-5904	86	21	weights	weight	NOUN
ajema-5904	86	22	required	require	VERB
ajema-5904	86	23	[	[	PUNCT
ajema-5904	86	24	23	23	NUM
ajema-5904	86	25	p6	p6	PROPN
ajema-5904	86	26	]	]	X
ajema-5904	86	27	.	.	PUNCT
ajema-5904	87	1	the	the	DET
ajema-5904	87	2	max	max	PROPN
ajema-5904	87	3	pooling	pool	VERB
ajema-5904	87	4	layer	layer	NOUN
ajema-5904	87	5	,	,	PUNCT
ajema-5904	87	6	which	which	PRON
ajema-5904	87	7	is	be	AUX
ajema-5904	87	8	often	often	ADV
ajema-5904	87	9	coupled	couple	VERB
ajema-5904	87	10	to	to	ADP
ajema-5904	87	11	the	the	DET
ajema-5904	87	12	convolutional	convolutional	ADJ
ajema-5904	87	13	network	network	NOUN
ajema-5904	87	14	,	,	PUNCT
ajema-5904	87	15	helps	help	VERB
ajema-5904	87	16	further	far	ADV
ajema-5904	87	17	minimize	minimize	VERB
ajema-5904	87	18	the	the	DET
ajema-5904	87	19	size	size	NOUN
ajema-5904	87	20	of	of	ADP
ajema-5904	87	21	the	the	DET
ajema-5904	87	22	features	feature	NOUN
ajema-5904	87	23	identified	identify	VERB
ajema-5904	87	24	by	by	ADP
ajema-5904	87	25	aggregating	aggregate	VERB
ajema-5904	87	26	the	the	DET
ajema-5904	87	27	values	value	NOUN
ajema-5904	87	28	of	of	ADP
ajema-5904	87	29	several	several	ADJ
ajema-5904	87	30	neurons	neuron	NOUN
ajema-5904	87	31	into	into	ADP
ajema-5904	87	32	a	a	DET
ajema-5904	87	33	single	single	ADJ
ajema-5904	87	34	value	value	NOUN
ajema-5904	87	35	,	,	PUNCT
ajema-5904	87	36	which	which	PRON
ajema-5904	87	37	is	be	AUX
ajema-5904	87	38	typically	typically	ADV
ajema-5904	87	39	either	either	CCONJ
ajema-5904	87	40	the	the	DET
ajema-5904	87	41	average	average	ADJ
ajema-5904	87	42	value	value	NOUN
ajema-5904	87	43	or	or	CCONJ
ajema-5904	87	44	the	the	DET
ajema-5904	87	45	maximum	maximum	ADJ
ajema-5904	87	46	value	value	NOUN
ajema-5904	87	47	(	(	PUNCT
ajema-5904	87	48	hence	hence	ADV
ajema-5904	87	49	max	max	NOUN
ajema-5904	87	50	-	-	PUNCT
ajema-5904	87	51	pooling	pooling	NOUN
ajema-5904	87	52	)	)	PUNCT
ajema-5904	87	53	.	.	PUNCT
ajema-5904	88	1	at	at	ADV
ajema-5904	88	2	least	least	ADV
ajema-5904	88	3	one	one	NUM
ajema-5904	88	4	completely	completely	ADV
ajema-5904	88	5	linked	link	VERB
ajema-5904	88	6	(	(	PUNCT
ajema-5904	88	7	or	or	CCONJ
ajema-5904	88	8	dense	dense	ADJ
ajema-5904	88	9	)	)	PUNCT
ajema-5904	88	10	layer	layer	NOUN
ajema-5904	88	11	completes	complete	VERB
ajema-5904	88	12	the	the	DET
ajema-5904	88	13	cnn	cnn	PROPN
ajema-5904	88	14	architecture	architecture	NOUN
ajema-5904	88	15	after	after	ADP
ajema-5904	88	16	a	a	DET
ajema-5904	88	17	combination	combination	NOUN
ajema-5904	88	18	of	of	ADP
ajema-5904	88	19	the	the	DET
ajema-5904	88	20	aforementioned	aforementioned	ADJ
ajema-5904	88	21	layers	layer	NOUN
ajema-5904	88	22	.	.	PUNCT
ajema-5904	89	1	this	this	DET
ajema-5904	89	2	layer	layer	NOUN
ajema-5904	89	3	operates	operate	VERB
ajema-5904	89	4	as	as	ADP
ajema-5904	89	5	a	a	DET
ajema-5904	89	6	multi	multi	ADJ
ajema-5904	89	7	-	-	ADJ
ajema-5904	89	8	layer	layer	ADJ
ajema-5904	89	9	perceptron	perceptron	NOUN
ajema-5904	89	10	and	and	CCONJ
ajema-5904	89	11	is	be	AUX
ajema-5904	89	12	in	in	ADP
ajema-5904	89	13	charge	charge	NOUN
ajema-5904	89	14	of	of	ADP
ajema-5904	89	15	the	the	DET
ajema-5904	89	16	actual	actual	ADJ
ajema-5904	89	17	classification	classification	NOUN
ajema-5904	89	18	process	process	NOUN
ajema-5904	89	19	.	.	PUNCT
ajema-5904	90	1	it	it	PRON
ajema-5904	90	2	receives	receive	VERB
ajema-5904	90	3	as	as	ADP
ajema-5904	90	4	input	input	NOUN
ajema-5904	90	5	the	the	DET
ajema-5904	90	6	characteristics	characteristic	NOUN
ajema-5904	90	7	that	that	PRON
ajema-5904	90	8	were	be	AUX
ajema-5904	90	9	taken	take	VERB
ajema-5904	90	10	out	out	ADP
ajema-5904	90	11	of	of	ADP
ajema-5904	90	12	the	the	DET
ajema-5904	90	13	earlier	early	ADJ
ajema-5904	90	14	layers	layer	NOUN
ajema-5904	90	15	[	[	X
ajema-5904	90	16	24	24	NUM
ajema-5904	90	17	p6	p6	PROPN
ajema-5904	90	18	]	]	PUNCT
ajema-5904	90	19	.	.	PUNCT
ajema-5904	91	1	regularization	regularization	NOUN
ajema-5904	91	2	methods	method	NOUN
ajema-5904	91	3	are	be	AUX
ajema-5904	91	4	widely	widely	ADV
ajema-5904	91	5	used	use	VERB
ajema-5904	91	6	and	and	CCONJ
ajema-5904	91	7	may	may	AUX
ajema-5904	91	8	be	be	AUX
ajema-5904	91	9	implemented	implement	VERB
ajema-5904	91	10	anywhere	anywhere	ADV
ajema-5904	91	11	in	in	ADP
ajema-5904	91	12	the	the	DET
ajema-5904	91	13	network	network	NOUN
ajema-5904	91	14	to	to	PART
ajema-5904	91	15	prevent	prevent	VERB
ajema-5904	91	16	overfitting	overfitting	NOUN
ajema-5904	91	17	.	.	PUNCT
ajema-5904	92	1	better	well	ADJ
ajema-5904	92	2	generalization	generalization	NOUN
ajema-5904	92	3	is	be	AUX
ajema-5904	92	4	made	make	VERB
ajema-5904	92	5	possible	possible	ADJ
ajema-5904	92	6	as	as	ADP
ajema-5904	92	7	a	a	DET
ajema-5904	92	8	result	result	NOUN
ajema-5904	92	9	of	of	ADP
ajema-5904	92	10	the	the	DET
ajema-5904	92	11	network	network	NOUN
ajema-5904	92	12	being	be	AUX
ajema-5904	92	13	less	less	ADV
ajema-5904	92	14	likely	likely	ADJ
ajema-5904	92	15	to	to	PART
ajema-5904	92	16	memorize	memorize	VERB
ajema-5904	92	17	the	the	DET
ajema-5904	92	18	training	training	NOUN
ajema-5904	92	19	set	set	NOUN
ajema-5904	92	20	[	[	PUNCT
ajema-5904	92	21	22	22	NUM
ajema-5904	92	22	p6	p6	PROPN
ajema-5904	92	23	]	]	PUNCT
ajema-5904	92	24	.	.	PUNCT
ajema-5904	93	1	cnn	cnn	PROPN
ajema-5904	93	2	consist	consist	VERB
ajema-5904	93	3	of	of	ADP
ajema-5904	93	4	:	:	PUNCT
ajema-5904	93	5	convolutional	convolutional	ADJ
ajema-5904	93	6	layer	layer	NOUN
ajema-5904	93	7	:	:	PUNCT
ajema-5904	93	8	first	first	ADV
ajema-5904	93	9	,	,	PUNCT
ajema-5904	93	10	the	the	DET
ajema-5904	93	11	cnn	cnn	PROPN
ajema-5904	93	12	input	input	NOUN
ajema-5904	93	13	format	format	NOUN
ajema-5904	93	14	is	be	AUX
ajema-5904	93	15	covered	cover	VERB
ajema-5904	93	16	.	.	PUNCT
ajema-5904	94	1	while	while	SCONJ
ajema-5904	94	2	cnn	cnn	PROPN
ajema-5904	94	3	employs	employ	VERB
ajema-5904	94	4	a	a	DET
ajema-5904	94	5	multichanneled	multichannele	VERB
ajema-5904	94	6	image	image	NOUN
ajema-5904	94	7	as	as	ADP
ajema-5904	94	8	input	input	NOUN
ajema-5904	94	9	,	,	PUNCT
ajema-5904	94	10	typical	typical	ADJ
ajema-5904	94	11	neural	neural	ADJ
ajema-5904	94	12	networks	network	NOUN
ajema-5904	94	13	use	use	VERB
ajema-5904	94	14	vector	vector	NOUN
ajema-5904	94	15	formats	format	NOUN
ajema-5904	94	16	.	.	PUNCT
ajema-5904	95	1	for	for	ADP
ajema-5904	95	2	instance	instance	NOUN
ajema-5904	95	3	,	,	PUNCT
ajema-5904	95	4	the	the	DET
ajema-5904	95	5	rgb	rgb	PROPN
ajema-5904	95	6	image	image	NOUN
ajema-5904	95	7	format	format	NOUN
ajema-5904	95	8	has	have	VERB
ajema-5904	95	9	three	three	NUM
ajema-5904	95	10	channels	channel	NOUN
ajema-5904	95	11	,	,	PUNCT
ajema-5904	95	12	but	but	CCONJ
ajema-5904	95	13	the	the	DET
ajema-5904	95	14	grayscale	grayscale	NOUN
ajema-5904	95	15	picture	picture	NOUN
ajema-5904	95	16	format	format	NOUN
ajema-5904	95	17	just	just	ADV
ajema-5904	95	18	has	have	VERB
ajema-5904	95	19	one	one	NUM
ajema-5904	95	20	.	.	PUNCT
ajema-5904	96	1	to	to	PART
ajema-5904	96	2	further	far	ADV
ajema-5904	96	3	understand	understand	VERB
ajema-5904	96	4	the	the	DET
ajema-5904	96	5	convolutional	convolutional	ADJ
ajema-5904	96	6	process	process	NOUN
ajema-5904	96	7	,	,	PUNCT
ajema-5904	96	8	consider	consider	VERB
ajema-5904	96	9	a	a	DET
ajema-5904	96	10	4	4	NUM
ajema-5904	96	11	x	x	SYM
ajema-5904	96	12	4	4	NUM
ajema-5904	96	13	grayscale	grayscale	NOUN
ajema-5904	96	14	image	image	NOUN
ajema-5904	96	15	with	with	ADP
ajema-5904	96	16	a	a	DET
ajema-5904	96	17	2	2	NUM
ajema-5904	96	18	x	x	SYM
ajema-5904	96	19	2	2	NUM
ajema-5904	96	20	random	random	ADJ
ajema-5904	96	21	weight	weight	NOUN
ajema-5904	96	22	-	-	PUNCT
ajema-5904	96	23	initialized	initialize	VERB
ajema-5904	96	24	kernel	kernel	NOUN
ajema-5904	96	25	.	.	PUNCT
ajema-5904	97	1	the	the	DET
ajema-5904	97	2	kernel	kernel	PROPN
ajema-5904	97	3	first	first	PROPN
ajema-5904	97	4	scans	scan	VERB
ajema-5904	97	5	the	the	DET
ajema-5904	97	6	whole	whole	ADJ
ajema-5904	97	7	picture	picture	NOUN
ajema-5904	97	8	in	in	ADP
ajema-5904	97	9	both	both	CCONJ
ajema-5904	97	10	horizontal	horizontal	ADJ
ajema-5904	97	11	and	and	CCONJ
ajema-5904	97	12	vertical	vertical	ADJ
ajema-5904	97	13	directions	direction	NOUN
ajema-5904	97	14	.	.	PUNCT
ajema-5904	98	1	furthermore	furthermore	ADV
ajema-5904	98	2	,	,	PUNCT
ajema-5904	98	3	the	the	DET
ajema-5904	98	4	kernel	kernel	NOUN
ajema-5904	98	5	's	's	PART
ajema-5904	98	6	dot	dot	NOUN
ajema-5904	98	7	product	product	NOUN
ajema-5904	98	8	and	and	CCONJ
ajema-5904	98	9	the	the	DET
ajema-5904	98	10	input	input	NOUN
ajema-5904	98	11	image	image	NOUN
ajema-5904	98	12	are	be	AUX
ajema-5904	98	13	computed	compute	VERB
ajema-5904	98	14	,	,	PUNCT
ajema-5904	98	15	which	which	PRON
ajema-5904	98	16	involves	involve	VERB
ajema-5904	98	17	multiplying	multiply	VERB
ajema-5904	98	18	and	and	CCONJ
ajema-5904	98	19	adding	add	VERB
ajema-5904	98	20	each	each	PRON
ajema-5904	98	21	of	of	ADP
ajema-5904	98	22	their	their	PRON
ajema-5904	98	23	distinct	distinct	ADJ
ajema-5904	98	24	values	value	NOUN
ajema-5904	98	25	to	to	PART
ajema-5904	98	26	produce	produce	VERB
ajema-5904	98	27	a	a	DET
ajema-5904	98	28	single	single	ADJ
ajema-5904	98	29	scalar	scalar	ADJ
ajema-5904	98	30	result	result	NOUN
ajema-5904	98	31	.	.	PUNCT
ajema-5904	99	1	the	the	DET
ajema-5904	99	2	procedure	procedure	NOUN
ajema-5904	99	3	is	be	AUX
ajema-5904	99	4	then	then	ADV
ajema-5904	99	5	carried	carry	VERB
ajema-5904	99	6	out	out	ADP
ajema-5904	99	7	again	again	ADV
ajema-5904	99	8	until	until	SCONJ
ajema-5904	99	9	sliding	slide	VERB
ajema-5904	99	10	is	be	AUX
ajema-5904	99	11	no	no	ADV
ajema-5904	99	12	longer	long	ADV
ajema-5904	99	13	feasible	feasible	ADJ
ajema-5904	99	14	.	.	PUNCT
ajema-5904	100	1	the	the	DET
ajema-5904	100	2	calculated	calculate	VERB
ajema-5904	100	3	dot	dot	NOUN
ajema-5904	100	4	product	product	NOUN
ajema-5904	100	5	values	value	NOUN
ajema-5904	100	6	reflect	reflect	VERB
ajema-5904	100	7	the	the	DET
ajema-5904	100	8	output	output	NOUN
ajema-5904	100	9	feature	feature	NOUN
ajema-5904	100	10	map	map	NOUN
ajema-5904	100	11	.	.	PUNCT
ajema-5904	101	1	in	in	ADP
ajema-5904	101	2	figure	figure	NOUN
ajema-5904	101	3	1	1	NUM
ajema-5904	101	4	,	,	PUNCT
ajema-5904	101	5	the	the	DET
ajema-5904	101	6	fundamental	fundamental	ADJ
ajema-5904	101	7	calculations	calculation	NOUN
ajema-5904	101	8	carried	carry	VERB
ajema-5904	101	9	out	out	ADP
ajema-5904	101	10	at	at	ADP
ajema-5904	101	11	every	every	DET
ajema-5904	101	12	step	step	NOUN
ajema-5904	101	13	are	be	AUX
ajema-5904	101	14	displayed	display	VERB
ajema-5904	101	15	clearly	clearly	ADV
ajema-5904	101	16	.	.	PUNCT
ajema-5904	102	1	in	in	ADP
ajema-5904	102	2	this	this	DET
ajema-5904	102	3	illustration	illustration	NOUN
ajema-5904	102	4	,	,	PUNCT
ajema-5904	102	5	the	the	DET
ajema-5904	102	6	2	2	NUM
ajema-5904	102	7	2	2	NUM
ajema-5904	102	8	kernel	kernel	NOUN
ajema-5904	102	9	is	be	AUX
ajema-5904	102	10	represented	represent	VERB
ajema-5904	102	11	by	by	ADP
ajema-5904	102	12	the	the	DET
ajema-5904	102	13	52	52	NUM
ajema-5904	102	14	journal	journal	NOUN
ajema-5904	102	15	of	of	ADP
ajema-5904	102	16	engineering	engineering	NOUN
ajema-5904	102	17	,	,	PUNCT
ajema-5904	102	18	mechanics	mechanic	NOUN
ajema-5904	102	19	and	and	CCONJ
ajema-5904	102	20	architecture	architecture	NOUN
ajema-5904	102	21	www	www	NOUN
ajema-5904	102	22	.	.	PUNCT
ajema-5904	103	1	grnjournal.us	grnjournal.us	X
ajema-5904	103	2	light	light	VERB
ajema-5904	103	3	green	green	ADJ
ajema-5904	103	4	hue	hue	NOUN
ajema-5904	103	5	,	,	PUNCT
ajema-5904	103	6	while	while	SCONJ
ajema-5904	103	7	a	a	DET
ajema-5904	103	8	piece	piece	NOUN
ajema-5904	103	9	of	of	ADP
ajema-5904	103	10	the	the	DET
ajema-5904	103	11	same	same	ADJ
ajema-5904	103	12	-	-	PUNCT
ajema-5904	103	13	sized	sized	ADJ
ajema-5904	103	14	input	input	NOUN
ajema-5904	103	15	image	image	NOUN
ajema-5904	103	16	is	be	AUX
ajema-5904	103	17	represented	represent	VERB
ajema-5904	103	18	by	by	ADP
ajema-5904	103	19	the	the	DET
ajema-5904	103	20	light	light	ADJ
ajema-5904	103	21	blue	blue	ADJ
ajema-5904	103	22	hue	hue	NOUN
ajema-5904	103	23	.	.	PUNCT
ajema-5904	104	1	both	both	PRON
ajema-5904	104	2	are	be	AUX
ajema-5904	104	3	multiplied	multiply	VERB
ajema-5904	104	4	,	,	PUNCT
ajema-5904	104	5	and	and	CCONJ
ajema-5904	104	6	after	after	ADP
ajema-5904	104	7	adding	add	VERB
ajema-5904	104	8	together	together	ADV
ajema-5904	104	9	the	the	DET
ajema-5904	104	10	final	final	ADJ
ajema-5904	104	11	product	product	NOUN
ajema-5904	104	12	values	value	NOUN
ajema-5904	104	13	,	,	PUNCT
ajema-5904	104	14	the	the	DET
ajema-5904	104	15	result	result	NOUN
ajema-5904	104	16	(	(	PUNCT
ajema-5904	104	17	shown	show	VERB
ajema-5904	104	18	in	in	ADP
ajema-5904	104	19	light	light	ADJ
ajema-5904	104	20	orange	orange	NOUN
ajema-5904	104	21	)	)	PUNCT
ajema-5904	104	22	represents	represent	VERB
ajema-5904	104	23	an	an	DET
ajema-5904	104	24	input	input	NOUN
ajema-5904	104	25	value	value	NOUN
ajema-5904	104	26	into	into	ADP
ajema-5904	104	27	the	the	DET
ajema-5904	104	28	output	output	NOUN
ajema-5904	104	29	feature	feature	NOUN
ajema-5904	104	30	map	map	NOUN
ajema-5904	104	31	.	.	PUNCT
ajema-5904	105	1	in	in	ADP
ajema-5904	105	2	the	the	DET
ajema-5904	105	3	example	example	NOUN
ajema-5904	105	4	above	above	ADV
ajema-5904	105	5	,	,	PUNCT
ajema-5904	105	6	a	a	DET
ajema-5904	105	7	stride	stride	NOUN
ajema-5904	105	8	of	of	ADP
ajema-5904	105	9	one	one	NUM
ajema-5904	105	10	is	be	AUX
ajema-5904	105	11	applied	apply	VERB
ajema-5904	105	12	to	to	ADP
ajema-5904	105	13	the	the	DET
ajema-5904	105	14	kernel	kernel	NOUN
ajema-5904	105	15	(	(	PUNCT
ajema-5904	105	16	represented	represent	VERB
ajema-5904	105	17	by	by	ADP
ajema-5904	105	18	the	the	DET
ajema-5904	105	19	chosen	choose	VERB
ajema-5904	105	20	step	step	NOUN
ajema-5904	105	21	-	-	PUNCT
ajema-5904	105	22	size	size	NOUN
ajema-5904	105	23	over	over	ADP
ajema-5904	105	24	all	all	DET
ajema-5904	105	25	vertical	vertical	ADJ
ajema-5904	105	26	and	and	CCONJ
ajema-5904	105	27	horizontal	horizontal	ADJ
ajema-5904	105	28	places	place	NOUN
ajema-5904	105	29	)	)	PUNCT
ajema-5904	105	30	instead	instead	ADV
ajema-5904	105	31	of	of	ADP
ajema-5904	105	32	padding	pad	VERB
ajema-5904	105	33	the	the	DET
ajema-5904	105	34	input	input	NOUN
ajema-5904	105	35	picture	picture	NOUN
ajema-5904	105	36	.	.	PUNCT
ajema-5904	106	1	using	use	VERB
ajema-5904	106	2	a	a	DET
ajema-5904	106	3	different	different	ADJ
ajema-5904	106	4	stride	stride	NOUN
ajema-5904	106	5	value	value	NOUN
ajema-5904	106	6	is	be	AUX
ajema-5904	106	7	also	also	ADV
ajema-5904	106	8	possible	possible	ADJ
ajema-5904	106	9	.	.	PUNCT
ajema-5904	107	1	raising	raise	VERB
ajema-5904	107	2	the	the	DET
ajema-5904	107	3	stride	stride	ADJ
ajema-5904	107	4	value	value	NOUN
ajema-5904	107	5	also	also	ADV
ajema-5904	107	6	results	result	VERB
ajema-5904	107	7	in	in	ADP
ajema-5904	107	8	a	a	DET
ajema-5904	107	9	feature	feature	NOUN
ajema-5904	107	10	map	map	NOUN
ajema-5904	107	11	that	that	PRON
ajema-5904	107	12	is	be	AUX
ajema-5904	107	13	smaller	small	ADJ
ajema-5904	107	14	in	in	ADP
ajema-5904	107	15	size	size	NOUN
ajema-5904	107	16	.	.	PUNCT
ajema-5904	108	1	figure	figure	VERB
ajema-5904	108	2	1	1	NUM
ajema-5904	108	3	convolutional	convolutional	ADJ
ajema-5904	108	4	operation	operation	NOUN
ajema-5904	108	5	pooling	pool	VERB
ajema-5904	108	6	layer	layer	NOUN
ajema-5904	108	7	:	:	PUNCT
ajema-5904	108	8	subsampling	subsample	VERB
ajema-5904	108	9	is	be	AUX
ajema-5904	108	10	the	the	DET
ajema-5904	108	11	main	main	ADJ
ajema-5904	108	12	function	function	NOUN
ajema-5904	108	13	of	of	ADP
ajema-5904	108	14	the	the	DET
ajema-5904	108	15	pooling	pooling	NOUN
ajema-5904	108	16	layer	layer	NOUN
ajema-5904	108	17	for	for	ADP
ajema-5904	108	18	the	the	DET
ajema-5904	108	19	feature	feature	NOUN
ajema-5904	108	20	maps	map	NOUN
ajema-5904	108	21	.	.	PUNCT
ajema-5904	109	1	convolutional	convolutional	ADJ
ajema-5904	109	2	methods	method	NOUN
ajema-5904	109	3	are	be	AUX
ajema-5904	109	4	used	use	VERB
ajema-5904	109	5	in	in	ADP
ajema-5904	109	6	the	the	DET
ajema-5904	109	7	production	production	NOUN
ajema-5904	109	8	of	of	ADP
ajema-5904	109	9	these	these	DET
ajema-5904	109	10	maps	map	NOUN
ajema-5904	109	11	.	.	PUNCT
ajema-5904	110	1	put	put	VERB
ajema-5904	110	2	otherwise	otherwise	ADV
ajema-5904	110	3	,	,	PUNCT
ajema-5904	110	4	this	this	DET
ajema-5904	110	5	method	method	NOUN
ajema-5904	110	6	reduces	reduce	VERB
ajema-5904	110	7	the	the	DET
ajema-5904	110	8	size	size	NOUN
ajema-5904	110	9	of	of	ADP
ajema-5904	110	10	bigger	big	ADJ
ajema-5904	110	11	feature	feature	NOUN
ajema-5904	110	12	maps	map	NOUN
ajema-5904	110	13	into	into	ADP
ajema-5904	110	14	smaller	small	ADJ
ajema-5904	110	15	ones	one	NOUN
ajema-5904	110	16	.	.	PUNCT
ajema-5904	111	1	it	it	PRON
ajema-5904	111	2	also	also	ADV
ajema-5904	111	3	keeps	keep	VERB
ajema-5904	111	4	most	most	ADJ
ajema-5904	111	5	of	of	ADP
ajema-5904	111	6	the	the	DET
ajema-5904	111	7	dominant	dominant	ADJ
ajema-5904	111	8	information	information	NOUN
ajema-5904	111	9	(	(	PUNCT
ajema-5904	111	10	or	or	CCONJ
ajema-5904	111	11	attributes	attribute	VERB
ajema-5904	111	12	)	)	PUNCT
ajema-5904	111	13	during	during	ADP
ajema-5904	111	14	the	the	DET
ajema-5904	111	15	pooling	pooling	NOUN
ajema-5904	111	16	process	process	NOUN
ajema-5904	111	17	.	.	PUNCT
ajema-5904	112	1	before	before	ADP
ajema-5904	112	2	the	the	DET
ajema-5904	112	3	pooling	pooling	NOUN
ajema-5904	112	4	procedure	procedure	NOUN
ajema-5904	112	5	,	,	PUNCT
ajema-5904	112	6	the	the	DET
ajema-5904	112	7	size	size	NOUN
ajema-5904	112	8	of	of	ADP
ajema-5904	112	9	the	the	DET
ajema-5904	112	10	kernel	kernel	NOUN
ajema-5904	112	11	and	and	CCONJ
ajema-5904	112	12	the	the	DET
ajema-5904	112	13	stride	stride	NOUN
ajema-5904	112	14	are	be	AUX
ajema-5904	112	15	allocated	allocate	VERB
ajema-5904	112	16	.	.	PUNCT
ajema-5904	113	1	different	different	ADJ
ajema-5904	113	2	pooling	pool	VERB
ajema-5904	113	3	layers	layer	NOUN
ajema-5904	113	4	can	can	AUX
ajema-5904	113	5	employ	employ	VERB
ajema-5904	113	6	a	a	DET
ajema-5904	113	7	variety	variety	NOUN
ajema-5904	113	8	of	of	ADP
ajema-5904	113	9	pooling	pool	VERB
ajema-5904	113	10	algorithm	algorithm	NOUN
ajema-5904	113	11	types	type	NOUN
ajema-5904	113	12	.	.	PUNCT
ajema-5904	114	1	global	global	ADJ
ajema-5904	114	2	max	max	PROPN
ajema-5904	114	3	pooling	pooling	PROPN
ajema-5904	114	4	,	,	PUNCT
ajema-5904	114	5	min	min	NOUN
ajema-5904	114	6	pooling	pooling	NOUN
ajema-5904	114	7	,	,	PUNCT
ajema-5904	114	8	max	max	PROPN
ajema-5904	114	9	pooling	pooling	NOUN
ajema-5904	114	10	,	,	PUNCT
ajema-5904	114	11	average	average	ADJ
ajema-5904	114	12	pooling	pooling	NOUN
ajema-5904	114	13	,	,	PUNCT
ajema-5904	114	14	and	and	CCONJ
ajema-5904	114	15	global	global	ADJ
ajema-5904	114	16	average	average	ADJ
ajema-5904	114	17	pooling	pooling	NOUN
ajema-5904	114	18	(	(	PUNCT
ajema-5904	114	19	gap	gap	NOUN
ajema-5904	114	20	)	)	PUNCT
ajema-5904	114	21	are	be	AUX
ajema-5904	114	22	a	a	DET
ajema-5904	114	23	few	few	ADJ
ajema-5904	114	24	of	of	ADP
ajema-5904	114	25	these	these	DET
ajema-5904	114	26	options	option	NOUN
ajema-5904	114	27	.	.	PUNCT
ajema-5904	115	1	the	the	DET
ajema-5904	115	2	maximum	maximum	ADJ
ajema-5904	115	3	,	,	PUNCT
ajema-5904	115	4	minimum	minimum	ADJ
ajema-5904	115	5	,	,	PUNCT
ajema-5904	115	6	and	and	CCONJ
ajema-5904	115	7	gap	gap	NOUN
ajema-5904	115	8	techniques	technique	NOUN
ajema-5904	115	9	are	be	AUX
ajema-5904	115	10	the	the	DET
ajema-5904	115	11	most	most	ADV
ajema-5904	115	12	popular	popular	ADJ
ajema-5904	115	13	and	and	CCONJ
ajema-5904	115	14	often	often	ADV
ajema-5904	115	15	applied	apply	VERB
ajema-5904	115	16	pooling	pool	VERB
ajema-5904	115	17	algorithms	algorithm	NOUN
ajema-5904	115	18	.	.	PUNCT
ajema-5904	116	1	these	these	DET
ajema-5904	116	2	three	three	NUM
ajema-5904	116	3	pooling	pool	VERB
ajema-5904	116	4	strategies	strategy	NOUN
ajema-5904	116	5	are	be	AUX
ajema-5904	116	6	shown	show	VERB
ajema-5904	116	7	in	in	ADP
ajema-5904	116	8	figure	figure	NOUN
ajema-5904	116	9	2	2	NUM
ajema-5904	116	10	.	.	PUNCT
ajema-5904	116	11	figure	figure	VERB
ajema-5904	116	12	2	2	NUM
ajema-5904	116	13	pooling	pool	VERB
ajema-5904	116	14	layer	layer	NOUN
ajema-5904	116	15	53	53	NUM
ajema-5904	116	16	journal	journal	NOUN
ajema-5904	116	17	of	of	ADP
ajema-5904	116	18	engineering	engineering	NOUN
ajema-5904	116	19	,	,	PUNCT
ajema-5904	116	20	mechanics	mechanic	NOUN
ajema-5904	116	21	and	and	CCONJ
ajema-5904	116	22	architecture	architecture	NOUN
ajema-5904	116	23	www	www	NOUN
ajema-5904	116	24	.	.	PUNCT
ajema-5904	117	1	grnjournal.us	grnjournal.us	X
ajema-5904	117	2	fully	fully	ADV
ajema-5904	117	3	connected	connected	ADJ
ajema-5904	117	4	layer	layer	NOUN
ajema-5904	117	5	:	:	PUNCT
ajema-5904	117	6	in	in	ADP
ajema-5904	117	7	most	most	ADJ
ajema-5904	117	8	cnn	cnn	PROPN
ajema-5904	117	9	architectures	architecture	NOUN
ajema-5904	117	10	,	,	PUNCT
ajema-5904	117	11	this	this	DET
ajema-5904	117	12	layer	layer	NOUN
ajema-5904	117	13	is	be	AUX
ajema-5904	117	14	found	find	VERB
ajema-5904	117	15	at	at	ADP
ajema-5904	117	16	the	the	DET
ajema-5904	117	17	end	end	NOUN
ajema-5904	117	18	.	.	PUNCT
ajema-5904	118	1	the	the	DET
ajema-5904	118	2	fully	fully	ADV
ajema-5904	118	3	connected	connect	VERB
ajema-5904	118	4	(	(	PUNCT
ajema-5904	118	5	fc	fc	NOUN
ajema-5904	118	6	)	)	PUNCT
ajema-5904	118	7	technique	technique	NOUN
ajema-5904	118	8	connects	connect	VERB
ajema-5904	118	9	every	every	DET
ajema-5904	118	10	cell	cell	NOUN
ajema-5904	118	11	in	in	ADP
ajema-5904	118	12	this	this	DET
ajema-5904	118	13	layer	layer	NOUN
ajema-5904	118	14	to	to	ADP
ajema-5904	118	15	every	every	DET
ajema-5904	118	16	cell	cell	NOUN
ajema-5904	118	17	in	in	ADP
ajema-5904	118	18	the	the	DET
ajema-5904	118	19	layer	layer	NOUN
ajema-5904	118	20	above	above	ADV
ajema-5904	118	21	.	.	PUNCT
ajema-5904	119	1	it	it	PRON
ajema-5904	119	2	functions	function	VERB
ajema-5904	119	3	as	as	ADP
ajema-5904	119	4	a	a	DET
ajema-5904	119	5	classifier	classifier	NOUN
ajema-5904	119	6	for	for	ADP
ajema-5904	119	7	cnn	cnn	PROPN
ajema-5904	119	8	.	.	PUNCT
ajema-5904	120	1	because	because	SCONJ
ajema-5904	120	2	it	it	PRON
ajema-5904	120	3	is	be	AUX
ajema-5904	120	4	a	a	DET
ajema-5904	120	5	feed	feed	NOUN
ajema-5904	120	6	-	-	PUNCT
ajema-5904	120	7	forward	forward	NOUN
ajema-5904	120	8	ann	ann	PROPN
ajema-5904	120	9	,	,	PUNCT
ajema-5904	120	10	it	it	PRON
ajema-5904	120	11	uses	use	VERB
ajema-5904	120	12	the	the	DET
ajema-5904	120	13	same	same	ADJ
ajema-5904	120	14	fundamental	fundamental	ADJ
ajema-5904	120	15	method	method	NOUN
ajema-5904	120	16	as	as	ADP
ajema-5904	120	17	a	a	DET
ajema-5904	120	18	traditional	traditional	ADJ
ajema-5904	120	19	multiple	multiple	ADJ
ajema-5904	120	20	-	-	PUNCT
ajema-5904	120	21	layer	layer	NOUN
ajema-5904	120	22	perceptron	perceptron	PROPN
ajema-5904	120	23	neural	neural	ADJ
ajema-5904	120	24	network	network	NOUN
ajema-5904	120	25	.	.	PUNCT
ajema-5904	121	1	the	the	DET
ajema-5904	121	2	fc	fc	PROPN
ajema-5904	121	3	layer	layer	NOUN
ajema-5904	121	4	receives	receive	VERB
ajema-5904	121	5	its	its	PRON
ajema-5904	121	6	input	input	NOUN
ajema-5904	121	7	from	from	ADP
ajema-5904	121	8	the	the	DET
ajema-5904	121	9	preceding	precede	VERB
ajema-5904	121	10	pooling	pooling	NOUN
ajema-5904	121	11	or	or	CCONJ
ajema-5904	121	12	convolutional	convolutional	ADJ
ajema-5904	121	13	layer	layer	NOUN
ajema-5904	121	14	.	.	PUNCT
ajema-5904	122	1	to	to	PART
ajema-5904	122	2	construct	construct	VERB
ajema-5904	122	3	a	a	DET
ajema-5904	122	4	vector	vector	NOUN
ajema-5904	122	5	representing	represent	VERB
ajema-5904	122	6	the	the	DET
ajema-5904	122	7	shape	shape	NOUN
ajema-5904	122	8	of	of	ADP
ajema-5904	122	9	this	this	DET
ajema-5904	122	10	input	input	NOUN
ajema-5904	122	11	,	,	PUNCT
ajema-5904	122	12	the	the	DET
ajema-5904	122	13	feature	feature	NOUN
ajema-5904	122	14	maps	map	NOUN
ajema-5904	122	15	are	be	AUX
ajema-5904	122	16	flattened	flatten	VERB
ajema-5904	122	17	.	.	PUNCT
ajema-5904	123	1	the	the	DET
ajema-5904	123	2	final	final	ADJ
ajema-5904	123	3	cnn	cnn	PROPN
ajema-5904	123	4	output	output	NOUN
ajema-5904	123	5	is	be	AUX
ajema-5904	123	6	represented	represent	VERB
ajema-5904	123	7	by	by	ADP
ajema-5904	123	8	the	the	DET
ajema-5904	123	9	fc	fc	PROPN
ajema-5904	123	10	layer	layer	NOUN
ajema-5904	123	11	's	's	PART
ajema-5904	123	12	output	output	NOUN
ajema-5904	123	13	.	.	PUNCT
ajema-5904	124	1	as	as	SCONJ
ajema-5904	124	2	shown	show	VERB
ajema-5904	124	3	in	in	ADP
ajema-5904	124	4	figure	figure	NOUN
ajema-5904	124	5	3	3	NUM
ajema-5904	124	6	figure	figure	NOUN
ajema-5904	124	7	3	3	NUM
ajema-5904	124	8	fully	fully	ADV
ajema-5904	124	9	connected	connect	VERB
ajema-5904	124	10	layer	layer	NOUN
ajema-5904	124	11	evaluation	evaluation	NOUN
ajema-5904	124	12	metrics	metric	NOUN
ajema-5904	124	13	determining	determine	VERB
ajema-5904	124	14	the	the	DET
ajema-5904	124	15	effectiveness	effectiveness	NOUN
ajema-5904	124	16	of	of	ADP
ajema-5904	124	17	trained	train	VERB
ajema-5904	124	18	classifiers	classifier	NOUN
ajema-5904	124	19	or	or	CCONJ
ajema-5904	124	20	learning	learn	VERB
ajema-5904	124	21	algorithms	algorithm	NOUN
ajema-5904	124	22	on	on	ADP
ajema-5904	124	23	various	various	ADJ
ajema-5904	124	24	data	datum	NOUN
ajema-5904	124	25	sets	set	NOUN
ajema-5904	124	26	is	be	AUX
ajema-5904	124	27	the	the	DET
ajema-5904	124	28	aim	aim	NOUN
ajema-5904	124	29	of	of	ADP
ajema-5904	124	30	assessment	assessment	NOUN
ajema-5904	124	31	in	in	ADP
ajema-5904	124	32	deep	deep	ADJ
ajema-5904	124	33	learning	learning	NOUN
ajema-5904	124	34	.	.	PUNCT
ajema-5904	125	1	most	most	ADJ
ajema-5904	125	2	existing	exist	VERB
ajema-5904	125	3	metrics	metric	NOUN
ajema-5904	125	4	concentrate	concentrate	VERB
ajema-5904	125	5	on	on	ADP
ajema-5904	125	6	the	the	DET
ajema-5904	125	7	accuracy	accuracy	NOUN
ajema-5904	125	8	of	of	ADP
ajema-5904	125	9	a	a	DET
ajema-5904	125	10	classifier	classifier	NOUN
ajema-5904	125	11	's	's	PART
ajema-5904	125	12	class	class	NOUN
ajema-5904	125	13	identification	identification	NOUN
ajema-5904	125	14	.	.	PUNCT
ajema-5904	126	1	metrics	metric	NOUN
ajema-5904	126	2	for	for	ADP
ajema-5904	126	3	evaluating	evaluate	VERB
ajema-5904	126	4	classification	classification	NOUN
ajema-5904	126	5	performance	performance	NOUN
ajema-5904	126	6	are	be	AUX
ajema-5904	126	7	essential	essential	ADJ
ajema-5904	126	8	for	for	ADP
ajema-5904	126	9	guiding	guide	VERB
ajema-5904	126	10	the	the	DET
ajema-5904	126	11	construction	construction	NOUN
ajema-5904	126	12	of	of	ADP
ajema-5904	126	13	classifiers	classifier	NOUN
ajema-5904	126	14	.	.	PUNCT
ajema-5904	127	1	there	there	PRON
ajema-5904	127	2	are	be	VERB
ajema-5904	127	3	significant	significant	ADJ
ajema-5904	127	4	disadvantages	disadvantage	NOUN
ajema-5904	127	5	to	to	ADP
ajema-5904	127	6	even	even	ADV
ajema-5904	127	7	the	the	DET
ajema-5904	127	8	most	most	ADV
ajema-5904	127	9	widely	widely	ADV
ajema-5904	127	10	used	use	VERB
ajema-5904	127	11	techniques	technique	NOUN
ajema-5904	127	12	,	,	PUNCT
ajema-5904	127	13	including	include	VERB
ajema-5904	127	14	figuring	figure	VERB
ajema-5904	127	15	the	the	DET
ajema-5904	127	16	accuracy	accuracy	NOUN
ajema-5904	127	17	or	or	CCONJ
ajema-5904	127	18	error	error	NOUN
ajema-5904	127	19	rate	rate	NOUN
ajema-5904	127	20	on	on	ADP
ajema-5904	127	21	a	a	DET
ajema-5904	127	22	test	test	NOUN
ajema-5904	127	23	set	set	NOUN
ajema-5904	127	24	.	.	PUNCT
ajema-5904	128	1	consequently	consequently	ADV
ajema-5904	128	2	,	,	PUNCT
ajema-5904	128	3	optimizing	optimize	VERB
ajema-5904	128	4	criteria	criterion	NOUN
ajema-5904	128	5	is	be	AUX
ajema-5904	128	6	correlated	correlate	VERB
ajema-5904	128	7	with	with	ADP
ajema-5904	128	8	modifications	modification	NOUN
ajema-5904	128	9	to	to	ADP
ajema-5904	128	10	classification	classification	NOUN
ajema-5904	128	11	algorithms	algorithm	NOUN
ajema-5904	128	12	to	to	ADP
ajema-5904	128	13	some	some	DET
ajema-5904	128	14	extent	extent	NOUN
ajema-5904	128	15	.	.	PUNCT
ajema-5904	129	1	a	a	DET
ajema-5904	129	2	lot	lot	NOUN
ajema-5904	129	3	of	of	ADP
ajema-5904	129	4	work	work	NOUN
ajema-5904	129	5	has	have	AUX
ajema-5904	129	6	gone	go	VERB
ajema-5904	129	7	into	into	ADP
ajema-5904	129	8	creating	create	VERB
ajema-5904	129	9	more	more	ADV
ajema-5904	129	10	intricate	intricate	ADJ
ajema-5904	129	11	algorithms	algorithm	NOUN
ajema-5904	129	12	to	to	PART
ajema-5904	129	13	deal	deal	VERB
ajema-5904	129	14	with	with	ADP
ajema-5904	129	15	the	the	DET
ajema-5904	129	16	categorization	categorization	NOUN
ajema-5904	129	17	issue	issue	NOUN
ajema-5904	129	18	.	.	PUNCT
ajema-5904	130	1	at	at	ADV
ajema-5904	130	2	least	least	ADJ
ajema-5904	130	3	as	as	ADV
ajema-5904	130	4	important	important	ADJ
ajema-5904	130	5	as	as	SCONJ
ajema-5904	130	6	the	the	DET
ajema-5904	130	7	algorithm	algorithm	NOUN
ajema-5904	130	8	is	be	AUX
ajema-5904	130	9	the	the	DET
ajema-5904	130	10	evaluation	evaluation	NOUN
ajema-5904	130	11	metrics	metric	NOUN
ajema-5904	130	12	phase	phase	NOUN
ajema-5904	130	13	,	,	PUNCT
ajema-5904	130	14	which	which	PRON
ajema-5904	130	15	comes	come	VERB
ajema-5904	130	16	first	first	ADV
ajema-5904	130	17	in	in	ADP
ajema-5904	130	18	the	the	DET
ajema-5904	130	19	learning	learning	NOUN
ajema-5904	130	20	process	process	NOUN
ajema-5904	130	21	.	.	PUNCT
ajema-5904	131	1	classifier	classifier	NOUN
ajema-5904	131	2	performance	performance	NOUN
ajema-5904	131	3	may	may	AUX
ajema-5904	131	4	be	be	AUX
ajema-5904	131	5	measured	measure	VERB
ajema-5904	131	6	using	use	VERB
ajema-5904	131	7	two	two	NUM
ajema-5904	131	8	different	different	ADJ
ajema-5904	131	9	approaches	approach	NOUN
ajema-5904	131	10	:	:	PUNCT
ajema-5904	131	11	graphical	graphical	ADJ
ajema-5904	131	12	and	and	CCONJ
ajema-5904	131	13	numerical	numerical	ADJ
ajema-5904	131	14	.	.	PUNCT
ajema-5904	132	1	a	a	DET
ajema-5904	132	2	classifier	classifier	NOUN
ajema-5904	132	3	's	's	PART
ajema-5904	132	4	performance	performance	NOUN
ajema-5904	132	5	is	be	AUX
ajema-5904	132	6	expressed	express	VERB
ajema-5904	132	7	as	as	ADP
ajema-5904	132	8	a	a	DET
ajema-5904	132	9	single	single	ADJ
ajema-5904	132	10	number	number	NOUN
ajema-5904	132	11	in	in	ADP
ajema-5904	132	12	numerical	numerical	ADJ
ajema-5904	132	13	assessments	assessment	NOUN
ajema-5904	132	14	,	,	PUNCT
ajema-5904	132	15	but	but	CCONJ
ajema-5904	132	16	performance	performance	NOUN
ajema-5904	132	17	is	be	AUX
ajema-5904	132	18	shown	show	VERB
ajema-5904	132	19	on	on	ADP
ajema-5904	132	20	a	a	DET
ajema-5904	132	21	plot	plot	NOUN
ajema-5904	132	22	with	with	ADP
ajema-5904	132	23	just	just	ADV
ajema-5904	132	24	two	two	NUM
ajema-5904	132	25	or	or	CCONJ
ajema-5904	132	26	three	three	NUM
ajema-5904	132	27	dimensions	dimension	NOUN
ajema-5904	132	28	in	in	ADP
ajema-5904	132	29	graphical	graphical	ADJ
ajema-5904	132	30	approaches	approach	NOUN
ajema-5904	132	31	,	,	PUNCT
ajema-5904	132	32	which	which	PRON
ajema-5904	132	33	makes	make	VERB
ajema-5904	132	34	it	it	PRON
ajema-5904	132	35	easier	easy	ADJ
ajema-5904	132	36	for	for	SCONJ
ajema-5904	132	37	humans	human	NOUN
ajema-5904	132	38	to	to	PART
ajema-5904	132	39	verify	verify	VERB
ajema-5904	132	40	.	.	PUNCT
ajema-5904	133	1	while	while	SCONJ
ajema-5904	133	2	cost	cost	NOUN
ajema-5904	133	3	curves	curve	NOUN
ajema-5904	133	4	are	be	AUX
ajema-5904	133	5	examples	example	NOUN
ajema-5904	133	6	of	of	ADP
ajema-5904	133	7	graphical	graphical	ADJ
ajema-5904	133	8	methodologies	methodology	NOUN
ajema-5904	133	9	,	,	PUNCT
ajema-5904	133	10	accuracy	accuracy	NOUN
ajema-5904	133	11	,	,	PUNCT
ajema-5904	133	12	precision	precision	NOUN
ajema-5904	133	13	,	,	PUNCT
ajema-5904	133	14	recall	recall	NOUN
ajema-5904	133	15	,	,	PUNCT
ajema-5904	133	16	and	and	CCONJ
ajema-5904	133	17	f1	f1	NOUN
ajema-5904	133	18	-	-	PUNCT
ajema-5904	133	19	score	score	NOUN
ajema-5904	133	20	are	be	AUX
ajema-5904	133	21	examples	example	NOUN
ajema-5904	133	22	of	of	ADP
ajema-5904	133	23	numerical	numerical	ADJ
ajema-5904	133	24	performance	performance	NOUN
ajema-5904	133	25	evaluations	evaluation	NOUN
ajema-5904	133	26	.	.	PUNCT
ajema-5904	134	1	the	the	DET
ajema-5904	134	2	percentage	percentage	NOUN
ajema-5904	134	3	of	of	ADP
ajema-5904	134	4	correctly	correctly	ADV
ajema-5904	134	5	classified	classified	ADJ
ajema-5904	134	6	records	record	NOUN
ajema-5904	134	7	to	to	ADP
ajema-5904	134	8	all	all	DET
ajema-5904	134	9	records	record	NOUN
ajema-5904	134	10	is	be	AUX
ajema-5904	134	11	referred	refer	VERB
ajema-5904	134	12	to	to	ADP
ajema-5904	134	13	as	as	ADP
ajema-5904	134	14	"	"	PUNCT
ajema-5904	134	15	accuracy	accuracy	NOUN
ajema-5904	134	16	"	"	PUNCT
ajema-5904	134	17	[	[	X
ajema-5904	134	18	84	84	NUM
ajema-5904	134	19	]	]	PUNCT
ajema-5904	134	20	.	.	PUNCT
ajema-5904	135	1	𝑎𝑐𝑐𝑢𝑟𝑎𝑐𝑦	𝑎𝑐𝑐𝑢𝑟𝑎𝑐𝑦	ADJ
ajema-5904	135	2	=	=	PUNCT
ajema-5904	135	3	𝑇𝑃	𝑇𝑃	PROPN
ajema-5904	135	4	+	+	CCONJ
ajema-5904	135	5	𝑇𝑁	𝑇𝑁	PROPN
ajema-5904	135	6	𝑇𝑃	𝑇𝑃	PROPN
ajema-5904	135	7	+	+	CCONJ
ajema-5904	135	8	𝑇𝑁	𝑇𝑁	PROPN
ajema-5904	135	9	+	+	NUM
ajema-5904	135	10	𝐹𝑃	𝐹𝑃	NOUN
ajema-5904	136	1	+	+	CCONJ
ajema-5904	136	2	𝐹𝑁	𝐹𝑁	PROPN
ajema-5904	136	3	precision	precision	NOUN
ajema-5904	136	4	quantifies	quantify	VERB
ajema-5904	136	5	the	the	DET
ajema-5904	136	6	proportion	proportion	NOUN
ajema-5904	136	7	of	of	ADP
ajema-5904	136	8	attack	attack	NOUN
ajema-5904	136	9	class	class	NOUN
ajema-5904	136	10	predictions	prediction	NOUN
ajema-5904	136	11	that	that	PRON
ajema-5904	136	12	fall	fall	VERB
ajema-5904	136	13	into	into	ADP
ajema-5904	136	14	the	the	DET
ajema-5904	136	15	anomaly	anomaly	NOUN
ajema-5904	136	16	class	class	NOUN
ajema-5904	136	17	[	[	X
ajema-5904	136	18	84	84	NUM
ajema-5904	136	19	]	]	PUNCT
ajema-5904	136	20	.	.	PUNCT
ajema-5904	137	1	𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛	𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛	PROPN
ajema-5904	137	2	=	=	SYM
ajema-5904	137	3	𝑇𝑃	𝑇𝑃	PROPN
ajema-5904	137	4	𝑇𝑃	𝑇𝑃	PROPN
ajema-5904	137	5	+	+	CCONJ
ajema-5904	137	6	𝐹𝑃	𝐹𝑃	PROPN
ajema-5904	137	7	recall	recall	NOUN
ajema-5904	137	8	is	be	AUX
ajema-5904	137	9	defined	define	VERB
ajema-5904	137	10	as	as	ADP
ajema-5904	137	11	the	the	DET
ajema-5904	137	12	percentage	percentage	NOUN
ajema-5904	137	13	of	of	ADP
ajema-5904	137	14	anomalous	anomalous	ADJ
ajema-5904	137	15	records	record	NOUN
ajema-5904	137	16	that	that	PRON
ajema-5904	137	17	are	be	AUX
ajema-5904	137	18	successfully	successfully	ADV
ajema-5904	137	19	identified	identify	VERB
ajema-5904	137	20	divided	divide	VERB
ajema-5904	137	21	by	by	ADP
ajema-5904	137	22	the	the	DET
ajema-5904	137	23	total	total	ADJ
ajema-5904	137	24	number	number	NOUN
ajema-5904	137	25	of	of	ADP
ajema-5904	137	26	anomalous	anomalous	ADJ
ajema-5904	137	27	records	record	NOUN
ajema-5904	137	28	[	[	X
ajema-5904	137	29	84	84	NUM
ajema-5904	137	30	]	]	PUNCT
ajema-5904	137	31	.	.	PUNCT
ajema-5904	138	1	54	54	NUM
ajema-5904	138	2	journal	journal	NOUN
ajema-5904	138	3	of	of	ADP
ajema-5904	138	4	engineering	engineering	NOUN
ajema-5904	138	5	,	,	PUNCT
ajema-5904	138	6	mechanics	mechanic	NOUN
ajema-5904	138	7	and	and	CCONJ
ajema-5904	138	8	architecture	architecture	NOUN
ajema-5904	138	9	www	www	NOUN
ajema-5904	138	10	.	.	PUNCT
ajema-5904	139	1	grnjournal.us	grnjournal.us	X
ajema-5904	139	2	𝑅𝑒𝑐𝑎𝑙𝑙	𝑅𝑒𝑐𝑎𝑙𝑙	PROPN
ajema-5904	139	3	=	=	SYM
ajema-5904	139	4	𝑇𝑃	𝑇𝑃	PROPN
ajema-5904	139	5	𝑇𝑃	𝑇𝑃	PROPN
ajema-5904	139	6	+	+	CCONJ
ajema-5904	139	7	𝐹𝑁	𝐹𝑁	ADV
ajema-5904	139	8	the	the	DET
ajema-5904	139	9	harmonic	harmonic	ADJ
ajema-5904	139	10	mean	mean	NOUN
ajema-5904	139	11	of	of	ADP
ajema-5904	139	12	accuracy	accuracy	NOUN
ajema-5904	139	13	and	and	CCONJ
ajema-5904	139	14	recall	recall	NOUN
ajema-5904	139	15	,	,	PUNCT
ajema-5904	139	16	or	or	CCONJ
ajema-5904	139	17	f1	f1	NOUN
ajema-5904	139	18	score	score	NOUN
ajema-5904	139	19	,	,	PUNCT
ajema-5904	139	20	is	be	AUX
ajema-5904	139	21	what	what	PRON
ajema-5904	139	22	's	be	AUX
ajema-5904	139	23	utilized	utilize	VERB
ajema-5904	139	24	to	to	PART
ajema-5904	139	25	determine	determine	VERB
ajema-5904	139	26	derived	derive	VERB
ajema-5904	139	27	efficacy	efficacy	NOUN
ajema-5904	139	28	.	.	PUNCT
ajema-5904	140	1	a	a	DET
ajema-5904	140	2	traffic	traffic	NOUN
ajema-5904	140	3	categorization	categorization	NOUN
ajema-5904	140	4	model	model	NOUN
ajema-5904	140	5	has	have	VERB
ajema-5904	140	6	to	to	PART
ajema-5904	140	7	have	have	VERB
ajema-5904	140	8	high	high	ADJ
ajema-5904	140	9	recall	recall	NOUN
ajema-5904	140	10	,	,	PUNCT
ajema-5904	140	11	precision	precision	NOUN
ajema-5904	140	12	,	,	PUNCT
ajema-5904	140	13	and	and	CCONJ
ajema-5904	140	14	accuracy	accuracy	NOUN
ajema-5904	140	15	in	in	ADP
ajema-5904	140	16	order	order	NOUN
ajema-5904	140	17	to	to	PART
ajema-5904	140	18	be	be	AUX
ajema-5904	140	19	deemed	deem	VERB
ajema-5904	140	20	excellent	excellent	ADJ
ajema-5904	140	21	[	[	X
ajema-5904	140	22	85	85	NUM
ajema-5904	140	23	]	]	PUNCT
ajema-5904	140	24	.	.	PUNCT
ajema-5904	141	1	𝐹1	𝐹1	PROPN
ajema-5904	141	2	−	−	PROPN
ajema-5904	141	3	𝑆𝑐𝑜𝑟𝑒	𝑆𝑐𝑜𝑟𝑒	PROPN
ajema-5904	141	4	=	=	PUNCT
ajema-5904	141	5	2𝑋	2𝑋	PROPN
ajema-5904	141	6	𝑅𝑒𝑐𝑎𝑙𝑙	𝑅𝑒𝑐𝑎𝑙𝑙	PROPN
ajema-5904	141	7	𝑋	𝑋	PROPN
ajema-5904	141	8	𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛	𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛	PROPN
ajema-5904	141	9	𝑅𝑒𝑐𝑎𝑙𝑙	𝑅𝑒𝑐𝑎𝑙𝑙	PROPN
ajema-5904	141	10	+	+	PROPN
ajema-5904	141	11	𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛	𝑃𝑟𝑒𝑐𝑖𝑠𝑖𝑜𝑛	PROPN
ajema-5904	141	12	results	result	VERB
ajema-5904	141	13	unlike	unlike	ADP
ajema-5904	141	14	previous	previous	ADJ
ajema-5904	141	15	research	research	NOUN
ajema-5904	141	16	,	,	PUNCT
ajema-5904	141	17	we	we	PRON
ajema-5904	141	18	proposed	propose	VERB
ajema-5904	141	19	a	a	DET
ajema-5904	141	20	2d	2d	NUM
ajema-5904	141	21	-	-	PUNCT
ajema-5904	141	22	cnn	cnn	NOUN
ajema-5904	141	23	that	that	PRON
ajema-5904	141	24	makes	make	VERB
ajema-5904	141	25	the	the	DET
ajema-5904	141	26	most	most	ADJ
ajema-5904	141	27	use	use	NOUN
ajema-5904	141	28	of	of	ADP
ajema-5904	141	29	the	the	DET
ajema-5904	141	30	dataset	dataset	NOUN
ajema-5904	141	31	by	by	ADP
ajema-5904	141	32	employing	employ	VERB
ajema-5904	141	33	a	a	DET
ajema-5904	141	34	data	data	NOUN
ajema-5904	141	35	augmentation	augmentation	NOUN
ajema-5904	141	36	approach	approach	NOUN
ajema-5904	141	37	.	.	PUNCT
ajema-5904	142	1	prior	prior	ADJ
ajema-5904	142	2	studies	study	NOUN
ajema-5904	142	3	trained	train	VERB
ajema-5904	142	4	the	the	DET
ajema-5904	142	5	gist	gist	NOUN
ajema-5904	142	6	+	+	CCONJ
ajema-5904	142	7	svm	svm	ADJ
ajema-5904	142	8	classifier	classifier	NOUN
ajema-5904	142	9	by	by	ADP
ajema-5904	142	10	selecting	select	VERB
ajema-5904	142	11	many	many	ADJ
ajema-5904	142	12	characteristics	characteristic	NOUN
ajema-5904	142	13	linked	link	VERB
ajema-5904	142	14	to	to	ADP
ajema-5904	142	15	malware	malware	NOUN
ajema-5904	142	16	files	file	NOUN
ajema-5904	142	17	.	.	PUNCT
ajema-5904	143	1	secondly	secondly	ADV
ajema-5904	143	2	,	,	PUNCT
ajema-5904	143	3	training	train	VERB
ajema-5904	143	4	the	the	DET
ajema-5904	143	5	2d	2d	NOUN
ajema-5904	143	6	-	-	PUNCT
ajema-5904	143	7	cnn	cnn	PROPN
ajema-5904	143	8	to	to	PART
ajema-5904	143	9	enable	enable	VERB
ajema-5904	143	10	autonomous	autonomous	ADJ
ajema-5904	143	11	feature	feature	NOUN
ajema-5904	143	12	extraction	extraction	NOUN
ajema-5904	143	13	from	from	ADP
ajema-5904	143	14	pictures	picture	NOUN
ajema-5904	143	15	was	be	AUX
ajema-5904	143	16	proposed	propose	VERB
ajema-5904	143	17	.	.	PUNCT
ajema-5904	144	1	our	our	PRON
ajema-5904	144	2	solution	solution	NOUN
ajema-5904	144	3	outperforms	outperform	VERB
ajema-5904	144	4	these	these	DET
ajema-5904	144	5	two	two	NUM
ajema-5904	144	6	methods	method	NOUN
ajema-5904	144	7	in	in	ADP
ajema-5904	144	8	terms	term	NOUN
ajema-5904	144	9	of	of	ADP
ajema-5904	144	10	performance	performance	NOUN
ajema-5904	144	11	(	(	PUNCT
ajema-5904	144	12	less	less	ADJ
ajema-5904	144	13	than	than	ADP
ajema-5904	144	14	that	that	PRON
ajema-5904	144	15	of	of	ADP
ajema-5904	144	16	earlier	early	ADJ
ajema-5904	144	17	work	work	NOUN
ajema-5904	144	18	)	)	PUNCT
ajema-5904	144	19	,	,	PUNCT
ajema-5904	144	20	using	use	VERB
ajema-5904	144	21	data	datum	NOUN
ajema-5904	144	22	augmentation	augmentation	NOUN
ajema-5904	144	23	techniques	technique	NOUN
ajema-5904	144	24	with	with	ADP
ajema-5904	144	25	10	10	NUM
ajema-5904	144	26	epochs	epoch	NOUN
ajema-5904	144	27	.	.	PUNCT
ajema-5904	145	1	we	we	PRON
ajema-5904	145	2	were	be	AUX
ajema-5904	145	3	able	able	ADJ
ajema-5904	145	4	to	to	PART
ajema-5904	145	5	effectively	effectively	ADV
ajema-5904	145	6	use	use	VERB
ajema-5904	145	7	the	the	DET
ajema-5904	145	8	dataset	dataset	NOUN
ajema-5904	145	9	and	and	CCONJ
ajema-5904	145	10	prevent	prevent	VERB
ajema-5904	145	11	overfitting	overfitting	NOUN
ajema-5904	145	12	in	in	ADP
ajema-5904	145	13	this	this	DET
ajema-5904	145	14	way	way	NOUN
ajema-5904	145	15	as	as	SCONJ
ajema-5904	145	16	shown	show	VERB
ajema-5904	145	17	in	in	ADP
ajema-5904	145	18	table	table	NOUN
ajema-5904	145	19	1	1	NUM
ajema-5904	145	20	.	.	PUNCT
ajema-5904	145	21	table	table	NOUN
ajema-5904	145	22	1	1	NUM
ajema-5904	145	23	method	method	NOUN
ajema-5904	145	24	accuracy	accuracy	NOUN
ajema-5904	145	25	of	of	ADP
ajema-5904	145	26	method	method	NOUN
ajema-5904	145	27	model	model	NOUN
ajema-5904	145	28	in	in	ADP
ajema-5904	145	29	[	[	X
ajema-5904	145	30	3	3	NUM
ajema-5904	145	31	]	]	PUNCT
ajema-5904	145	32	96.46	96.46	NUM
ajema-5904	145	33	%	%	NOUN
ajema-5904	145	34	gist	gist	NOUN
ajema-5904	145	35	+	+	CCONJ
ajema-5904	145	36	svm	svm	ADJ
ajema-5904	145	37	[	[	X
ajema-5904	145	38	6	6	NUM
ajema-5904	145	39	]	]	PUNCT
ajema-5904	145	40	92.27	92.27	NUM
ajema-5904	145	41	%	%	NOUN
ajema-5904	145	42	2d	2d	NOUN
ajema-5904	145	43	-	-	PUNCT
ajema-5904	145	44	cnn	cnn	PROPN
ajema-5904	146	1	[	[	X
ajema-5904	146	2	6	6	NUM
ajema-5904	146	3	]	]	PUNCT
ajema-5904	146	4	93.56	93.56	NUM
ajema-5904	146	5	%	%	NOUN
ajema-5904	146	6	our	our	PRON
ajema-5904	146	7	model	model	NOUN
ajema-5904	146	8	99.12	99.12	NUM
ajema-5904	146	9	%	%	NOUN