id	sid	tid	token	lemma	pos
prudhomme	1	1	Taking	take	VBG
prudhomme	1	2	a	a	DT
prudhomme	1	3	Leap	Leap	NNP
prudhomme	1	4	Forward	Forward	NNP
prudhomme	1	5	:	:	:
prudhomme	1	6	Machine	Machine	NNP
prudhomme	1	7	Learning	Learning	NNP
prudhomme	1	8	for	for	IN
prudhomme	1	9	New	New	NNP
prudhomme	1	10	Limits	Limits	NNP
prudhomme	1	11	Introduction	Introduction	NNP
prudhomme	1	12	Today	Today	NNP
prudhomme	1	13	,	,	,
prudhomme	1	14	machines	machine	NNS
prudhomme	1	15	can	can	MD
prudhomme	1	16	analyze	analyze	VB
prudhomme	1	17	vast	vast	JJ
prudhomme	1	18	amounts	amount	NNS
prudhomme	1	19	of	of	IN
prudhomme	1	20	data	datum	NNS
prudhomme	1	21	and	and	CC
prudhomme	1	22	increasingly	increasingly	RB
prudhomme	1	23	produce	produce	VBP
prudhomme	1	24	accurate	accurate	JJ
prudhomme	1	25	results	result	NNS
prudhomme	1	26	through	through	IN
prudhomme	1	27	the	the	DT
prudhomme	1	28	repetition	repetition	NN
prudhomme	1	29	of	of	IN
prudhomme	1	30	mathematical	mathematical	JJ
prudhomme	1	31	or	or	CC
prudhomme	1	32	computational	computational	JJ
prudhomme	1	33	procedures	procedure	NNS
prudhomme	1	34	.	.	.
prudhomme	2	1	With	with	IN
prudhomme	2	2	the	the	DT
prudhomme	2	3	increasing	increase	VBG
prudhomme	2	4	computing	computing	NN
prudhomme	2	5	capabilities	capability	NNS
prudhomme	2	6	available	available	JJ
prudhomme	2	7	to	to	IN
prudhomme	2	8	us	-PRON-	PRP
prudhomme	2	9	today	today	NN
prudhomme	2	10	,	,	,
prudhomme	2	11	artificial	artificial	JJ
prudhomme	2	12	intelligence	intelligence	NN
prudhomme	2	13	(	(	-LRB-
prudhomme	2	14	AI	AI	NNP
prudhomme	2	15	)	)	-RRB-
prudhomme	2	16	and	and	CC
prudhomme	2	17	machine	machine	NN
prudhomme	2	18	applications	application	NNS
prudhomme	2	19	have	have	VBP
prudhomme	2	20	made	make	VBN
prudhomme	2	21	a	a	DT
prudhomme	2	22	leap	leap	NN
prudhomme	2	23	forward	forward	RB
prudhomme	2	24	.	.	.
prudhomme	3	1	These	these	DT
prudhomme	3	2	rapid	rapid	JJ
prudhomme	3	3	technological	technological	JJ
prudhomme	3	4	changes	change	NNS
prudhomme	3	5	are	be	VBP
prudhomme	3	6	inevitably	inevitably	RB
prudhomme	3	7	influencing	influence	VBG
prudhomme	3	8	our	-PRON-	PRP$
prudhomme	3	9	interpretation	interpretation	NN
prudhomme	3	10	of	of	IN
prudhomme	3	11	what	what	WP
prudhomme	3	12	AI	AI	NNP
prudhomme	3	13	can	can	MD
prudhomme	3	14	do	do	VB
prudhomme	3	15	and	and	CC
prudhomme	3	16	how	how	WRB
prudhomme	3	17	it	-PRON-	PRP
prudhomme	3	18	can	can	MD
prudhomme	3	19	affect	affect	VB
prudhomme	3	20	people	people	NNS
prudhomme	3	21	’s	’s	POS
prudhomme	3	22	lives	life	NNS
prudhomme	3	23	.	.	.
prudhomme	4	1	Machine	machine	NN
prudhomme	4	2	learning	learning	NN
prudhomme	4	3	models	model	NNS
prudhomme	4	4	that	that	WDT
prudhomme	4	5	are	be	VBP
prudhomme	4	6	developed	develop	VBN
prudhomme	4	7	on	on	IN
prudhomme	4	8	the	the	DT
prudhomme	4	9	basis	basis	NN
prudhomme	4	10	of	of	IN
prudhomme	4	11	statistical	statistical	JJ
prudhomme	4	12	patterns	pattern	NNS
prudhomme	4	13	from	from	IN
prudhomme	4	14	observed	observed	JJ
prudhomme	4	15	data	datum	NNS
prudhomme	4	16	provide	provide	VBP
prudhomme	4	17	new	new	JJ
prudhomme	4	18	opportunities	opportunity	NNS
prudhomme	4	19	to	to	TO
prudhomme	4	20	augment	augment	VB
prudhomme	4	21	our	-PRON-	PRP$
prudhomme	4	22	knowledge	knowledge	NN
prudhomme	4	23	of	of	IN
prudhomme	4	24	text	text	NN
prudhomme	4	25	,	,	,
prudhomme	4	26	photographs	photograph	NNS
prudhomme	4	27	,	,	,
prudhomme	4	28	and	and	CC
prudhomme	4	29	other	other	JJ
prudhomme	4	30	types	type	NNS
prudhomme	4	31	of	of	IN
prudhomme	4	32	data	datum	NNS
prudhomme	4	33	in	in	IN
prudhomme	4	34	support	support	NN
prudhomme	4	35	of	of	IN
prudhomme	4	36	research	research	NN
prudhomme	4	37	and	and	CC
prudhomme	4	38	education	education	NN
prudhomme	4	39	.	.	.
prudhomme	5	1	However	however	RB
prudhomme	5	2	,	,	,
prudhomme	5	3	“	"	``
prudhomme	5	4	the	the	DT
prudhomme	5	5	viability	viability	NN
prudhomme	5	6	of	of	IN
prudhomme	5	7	machine	machine	NN
prudhomme	5	8	learning	learning	NN
prudhomme	5	9	and	and	CC
prudhomme	5	10	artificial	artificial	JJ
prudhomme	5	11	intelligence	intelligence	NN
prudhomme	5	12	is	be	VBZ
prudhomme	5	13	predicated	predicate	VBN
prudhomme	5	14	on	on	IN
prudhomme	5	15	the	the	DT
prudhomme	5	16	representativeness	representativeness	NN
prudhomme	5	17	and	and	CC
prudhomme	5	18	quality	quality	NN
prudhomme	5	19	of	of	IN
prudhomme	5	20	the	the	DT
prudhomme	5	21	data	datum	NNS
prudhomme	5	22	that	that	WDT
prudhomme	5	23	they	-PRON-	PRP
prudhomme	5	24	are	be	VBP
prudhomme	5	25	trained	train	VBN
prudhomme	5	26	on	on	IN
prudhomme	5	27	,	,	,
prudhomme	5	28	”	"	''
prudhomme	5	29	as	as	IN
prudhomme	5	30	Thomas	Thomas	NNP
prudhomme	5	31	Padilla	Padilla	NNP
prudhomme	5	32	,	,	,
prudhomme	5	33	Interim	Interim	NNP
prudhomme	5	34	Head	Head	NNP
prudhomme	5	35	,	,	,
prudhomme	5	36	Knowledge	Knowledge	NNP
prudhomme	5	37	Production	Production	NNP
prudhomme	5	38	at	at	IN
prudhomme	5	39	the	the	DT
prudhomme	5	40	University	University	NNP
prudhomme	5	41	of	of	IN
prudhomme	5	42	Nevada	Nevada	NNP
prudhomme	5	43	Las	Las	NNP
prudhomme	5	44	Vegas	Vegas	NNP
prudhomme	5	45	,	,	,
prudhomme	5	46	asserts	assert	VBZ
prudhomme	5	47	(	(	-LRB-
prudhomme	5	48	2019	2019	CD
prudhomme	5	49	,	,	,
prudhomme	5	50	14	14	CD
prudhomme	5	51	)	)	-RRB-
prudhomme	5	52	.	.	.
prudhomme	6	1	With	with	IN
prudhomme	6	2	that	that	DT
prudhomme	6	3	in	in	IN
prudhomme	6	4	mind	mind	NN
prudhomme	6	5	,	,	,
prudhomme	6	6	these	these	DT
prudhomme	6	7	technologies	technology	NNS
prudhomme	6	8	and	and	CC
prudhomme	6	9	methodologies	methodology	NNS
prudhomme	6	10	could	could	MD
prudhomme	6	11	help	help	VB
prudhomme	6	12	augment	augment	VB
prudhomme	6	13	the	the	DT
prudhomme	6	14	capacity	capacity	NN
prudhomme	6	15	of	of	IN
prudhomme	6	16	archives	archive	NNS
prudhomme	6	17	and	and	CC
prudhomme	6	18	libraries	library	NNS
prudhomme	6	19	to	to	TO
prudhomme	6	20	leverage	leverage	VB
prudhomme	6	21	their	-PRON-	PRP$
prudhomme	6	22	creation	creation	NN
prudhomme	6	23	-	-	HYPH
prudhomme	6	24	value	value	NN
prudhomme	6	25	and	and	CC
prudhomme	6	26	minimize	minimize	VB
prudhomme	6	27	their	-PRON-	PRP$
prudhomme	6	28	institutional	institutional	JJ
prudhomme	6	29	memory	memory	NN
prudhomme	6	30	loss	loss	NN
prudhomme	6	31	while	while	IN
prudhomme	6	32	enhancing	enhance	VBG
prudhomme	6	33	the	the	DT
prudhomme	6	34	interdisciplinary	interdisciplinary	JJ
prudhomme	6	35	approach	approach	NN
prudhomme	6	36	to	to	IN
prudhomme	6	37	research	research	NN
prudhomme	6	38	and	and	CC
prudhomme	6	39	scholarship	scholarship	NN
prudhomme	6	40	.	.	.
prudhomme	7	1	In	in	IN
prudhomme	7	2	this	this	DT
prudhomme	7	3	essay	essay	NN
prudhomme	7	4	,	,	,
prudhomme	7	5	I	-PRON-	PRP
prudhomme	7	6	begin	begin	VBP
prudhomme	7	7	by	by	IN
prudhomme	7	8	placing	place	VBG
prudhomme	7	9	artificial	artificial	JJ
prudhomme	7	10	intelligence	intelligence	NN
prudhomme	7	11	and	and	CC
prudhomme	7	12	machine	machine	NN
prudhomme	7	13	learning	learning	NN
prudhomme	7	14	in	in	IN
prudhomme	7	15	context	context	NN
prudhomme	7	16	,	,	,
prudhomme	7	17	then	then	RB
prudhomme	7	18	proceed	proceed	VB
prudhomme	7	19	by	by	IN
prudhomme	7	20	discussing	discuss	VBG
prudhomme	7	21	why	why	WRB
prudhomme	7	22	AI	AI	NNP
prudhomme	7	23	matters	matter	VBZ
prudhomme	7	24	for	for	IN
prudhomme	7	25	archives	archive	NNS
prudhomme	7	26	and	and	CC
prudhomme	7	27	libraries	library	NNS
prudhomme	7	28	,	,	,
prudhomme	7	29	and	and	CC
prudhomme	7	30	describing	describe	VBG
prudhomme	7	31	the	the	DT
prudhomme	7	32	techniques	technique	NNS
prudhomme	7	33	used	use	VBN
prudhomme	7	34	in	in	IN
prudhomme	7	35	a	a	DT
prudhomme	7	36	pilot	pilot	JJ
prudhomme	7	37	automation	automation	NN
prudhomme	7	38	project	project	NN
prudhomme	7	39	from	from	IN
prudhomme	7	40	the	the	DT
prudhomme	7	41	perspective	perspective	NN
prudhomme	7	42	of	of	IN
prudhomme	7	43	digital	digital	JJ
prudhomme	7	44	curation	curation	NN
prudhomme	7	45	at	at	IN
prudhomme	7	46	Oklahoma	Oklahoma	NNP
prudhomme	7	47	State	State	NNP
prudhomme	7	48	University	University	NNP
prudhomme	7	49	Archives	Archives	NNPS
prudhomme	7	50	.	.	.
prudhomme	8	1	Lastly	lastly	RB
prudhomme	8	2	,	,	,
prudhomme	8	3	I	-PRON-	PRP
prudhomme	8	4	end	end	VBP
prudhomme	8	5	by	by	IN
prudhomme	8	6	challenging	challenge	VBG
prudhomme	8	7	other	other	JJ
prudhomme	8	8	areas	area	NNS
prudhomme	8	9	in	in	IN
prudhomme	8	10	the	the	DT
prudhomme	8	11	library	library	NN
prudhomme	8	12	and	and	CC
prudhomme	8	13	adjacent	adjacent	JJ
prudhomme	8	14	fields	field	NNS
prudhomme	8	15	to	to	TO
prudhomme	8	16	join	join	VB
prudhomme	8	17	in	in	IN
prudhomme	8	18	the	the	DT
prudhomme	8	19	dialogue	dialogue	NN
prudhomme	8	20	,	,	,
prudhomme	8	21	to	to	TO
prudhomme	8	22	develop	develop	VB
prudhomme	8	23	a	a	DT
prudhomme	8	24	machine	machine	NN
prudhomme	8	25	learning	learn	VBG
prudhomme	8	26	solution	solution	NN
prudhomme	8	27	more	more	RBR
prudhomme	8	28	broadly	broadly	RB
prudhomme	8	29	,	,	,
prudhomme	8	30	and	and	CC
prudhomme	8	31	to	to	TO
prudhomme	8	32	explore	explore	VB
prudhomme	8	33	opportunities	opportunity	NNS
prudhomme	8	34	that	that	WDT
prudhomme	8	35	we	-PRON-	PRP
prudhomme	8	36	can	can	MD
prudhomme	8	37	reap	reap	VB
prudhomme	8	38	by	by	IN
prudhomme	8	39	reaching	reach	VBG
prudhomme	8	40	out	out	RP
prudhomme	8	41	to	to	IN
prudhomme	8	42	others	other	NNS
prudhomme	8	43	who	who	WP
prudhomme	8	44	share	share	VBP
prudhomme	8	45	a	a	DT
prudhomme	8	46	similar	similar	JJ
prudhomme	8	47	interest	interest	NN
prudhomme	8	48	in	in	IN
prudhomme	8	49	connecting	connect	VBG
prudhomme	8	50	people	people	NNS
prudhomme	8	51	to	to	TO
prudhomme	8	52	build	build	VB
prudhomme	8	53	knowledge	knowledge	NN
prudhomme	8	54	.	.	.
prudhomme	9	1	Artificial	Artificial	NNP
prudhomme	9	2	Intelligence	Intelligence	NNP
prudhomme	9	3	and	and	CC
prudhomme	9	4	Machine	Machine	NNP
prudhomme	9	5	Learning	Learning	NNP
prudhomme	9	6	.	.	.
prudhomme	10	1	Why	why	WRB
prudhomme	10	2	do	do	VBP
prudhomme	10	3	they	-PRON-	PRP
prudhomme	10	4	Matter	Matter	NNP
prudhomme	10	5	?	?	.
prudhomme	11	1	Artificial	artificial	JJ
prudhomme	11	2	intelligence	intelligence	NN
prudhomme	11	3	has	have	VBZ
prudhomme	11	4	seen	see	VBN
prudhomme	11	5	a	a	DT
prudhomme	11	6	resurging	resurge	VBG
prudhomme	11	7	interest	interest	NN
prudhomme	11	8	in	in	IN
prudhomme	11	9	the	the	DT
prudhomme	11	10	recent	recent	JJ
prudhomme	11	11	past	past	NN
prudhomme	11	12	-	-	:
prudhomme	11	13	in	in	IN
prudhomme	11	14	the	the	DT
prudhomme	11	15	news	news	NN
prudhomme	11	16	,	,	,
prudhomme	11	17	in	in	IN
prudhomme	11	18	the	the	DT
prudhomme	11	19	literature	literature	NN
prudhomme	11	20	,	,	,
prudhomme	11	21	in	in	IN
prudhomme	11	22	academic	academic	JJ
prudhomme	11	23	libraries	library	NNS
prudhomme	11	24	and	and	CC
prudhomme	11	25	archives	archive	NNS
prudhomme	11	26	,	,	,
prudhomme	11	27	and	and	CC
prudhomme	11	28	in	in	IN
prudhomme	11	29	other	other	JJ
prudhomme	11	30	fields	field	NNS
prudhomme	11	31	,	,	,
prudhomme	11	32	such	such	JJ
prudhomme	11	33	as	as	IN
prudhomme	11	34	medical	medical	JJ
prudhomme	11	35	imaging	imaging	NN
prudhomme	11	36	,	,	,
prudhomme	11	37	inspection	inspection	NN
prudhomme	11	38	of	of	IN
prudhomme	11	39	steel	steel	NN
prudhomme	11	40	corrosion	corrosion	NN
prudhomme	11	41	,	,	,
prudhomme	11	42	and	and	CC
prudhomme	11	43	more	more	JJR
prudhomme	11	44	.	.	.
prudhomme	12	1	John	John	NNP
prudhomme	12	2	McCarthy	McCarthy	NNP
prudhomme	12	3	,	,	,
prudhomme	12	4	American	american	JJ
prudhomme	12	5	computer	computer	NN
prudhomme	12	6	scientist	scientist	NN
prudhomme	12	7	,	,	,
prudhomme	12	8	defined	define	VBD
prudhomme	12	9	artificial	artificial	JJ
prudhomme	12	10	intelligence	intelligence	NN
prudhomme	12	11	as	as	IN
prudhomme	12	12	“	"	``
prudhomme	12	13	the	the	DT
prudhomme	12	14	science	science	NN
prudhomme	12	15	and	and	CC
prudhomme	12	16	engineering	engineering	NN
prudhomme	12	17	of	of	IN
prudhomme	12	18	making	make	VBG
prudhomme	12	19	intelligent	intelligent	JJ
prudhomme	12	20	machines	machine	NNS
prudhomme	12	21	,	,	,
prudhomme	12	22	especially	especially	RB
prudhomme	12	23	intelligent	intelligent	JJ
prudhomme	12	24	computer	computer	NN
prudhomme	12	25	programs	program	NNS
prudhomme	12	26	.	.	.
prudhomme	13	1	It	-PRON-	PRP
prudhomme	13	2	is	be	VBZ
prudhomme	13	3	related	relate	VBN
prudhomme	13	4	to	to	IN
prudhomme	13	5	the	the	DT
prudhomme	13	6	similar	similar	JJ
prudhomme	13	7	task	task	NN
prudhomme	13	8	of	of	IN
prudhomme	13	9	using	use	VBG
prudhomme	13	10	computers	computer	NNS
prudhomme	13	11	to	to	TO
prudhomme	13	12	understand	understand	VB
prudhomme	13	13	human	human	JJ
prudhomme	13	14	intelligence	intelligence	NN
prudhomme	13	15	,	,	,
prudhomme	13	16	but	but	CC
prudhomme	13	17	AI	AI	NNP
prudhomme	13	18	does	do	VBZ
prudhomme	13	19	not	not	RB
prudhomme	13	20	have	have	VB
prudhomme	13	21	to	to	TO
prudhomme	13	22	confine	confine	VB
prudhomme	13	23	itself	-PRON-	PRP
prudhomme	13	24	to	to	IN
prudhomme	13	25	methods	method	NNS
prudhomme	13	26	that	that	WDT
prudhomme	13	27	are	be	VBP
prudhomme	13	28	biologically	biologically	RB
prudhomme	13	29	observable	observable	JJ
prudhomme	13	30	”	"	''
prudhomme	13	31	(	(	-LRB-
prudhomme	13	32	2007	2007	CD
prudhomme	13	33	,	,	,
prudhomme	13	34	2	2	CD
prudhomme	13	35	)	)	-RRB-
prudhomme	13	36	.	.	.
prudhomme	14	1	This	this	DT
prudhomme	14	2	definition	definition	NN
prudhomme	14	3	has	have	VBZ
prudhomme	14	4	since	since	IN
prudhomme	14	5	been	be	VBN
prudhomme	14	6	extended	extend	VBN
prudhomme	14	7	to	to	TO
prudhomme	14	8	reflect	reflect	VB
prudhomme	14	9	a	a	DT
prudhomme	14	10	deeper	deep	JJR
prudhomme	14	11	understanding	understanding	NN
prudhomme	14	12	of	of	IN
prudhomme	14	13	AI	AI	NNP
prudhomme	14	14	today	today	NN
prudhomme	14	15	and	and	CC
prudhomme	14	16	what	what	WDT
prudhomme	14	17	systems	system	NNS
prudhomme	14	18	run	run	VBN
prudhomme	14	19	by	by	IN
prudhomme	14	20	computers	computer	NNS
prudhomme	14	21	are	be	VBP
prudhomme	14	22	now	now	RB
prudhomme	14	23	able	able	JJ
prudhomme	14	24	to	to	TO
prudhomme	14	25	do	do	VB
prudhomme	14	26	.	.	.
prudhomme	15	1	Dr.	Dr.	NNP
prudhomme	15	2	Carmel	Carmel	NNP
prudhomme	15	3	Kent	Kent	NNP
prudhomme	15	4	notes	note	VBZ
prudhomme	15	5	that	that	IN
prudhomme	15	6	“	"	``
prudhomme	15	7	AI	AI	NNP
prudhomme	15	8	feels	feel	VBZ
prudhomme	15	9	like	like	IN
prudhomme	15	10	a	a	DT
prudhomme	15	11	moving	move	VBG
prudhomme	15	12	target	target	NN
prudhomme	15	13	”	"	''
prudhomme	15	14	as	as	IN
prudhomme	15	15	we	-PRON-	PRP
prudhomme	15	16	still	still	RB
prudhomme	15	17	need	need	VBP
prudhomme	15	18	to	to	TO
prudhomme	15	19	learn	learn	VB
prudhomme	15	20	how	how	WRB
prudhomme	15	21	it	-PRON-	PRP
prudhomme	15	22	affects	affect	VBZ
prudhomme	15	23	our	-PRON-	PRP$
prudhomme	15	24	lives	life	NNS
prudhomme	15	25	(	(	-LRB-
prudhomme	15	26	2019	2019	CD
prudhomme	15	27	)	)	-RRB-
prudhomme	15	28	.	.	.
prudhomme	16	1	Within	within	IN
prudhomme	16	2	the	the	DT
prudhomme	16	3	last	last	JJ
prudhomme	16	4	decades	decade	NNS
prudhomme	16	5	,	,	,
prudhomme	16	6	the	the	DT
prudhomme	16	7	amazing	amazing	JJ
prudhomme	16	8	jump	jump	NN
prudhomme	16	9	in	in	IN
prudhomme	16	10	computing	computing	NN
prudhomme	16	11	capabilities	capability	NNS
prudhomme	16	12	has	have	VBZ
prudhomme	16	13	been	be	VBN
prudhomme	16	14	quite	quite	RB
prudhomme	16	15	transformative	transformative	JJ
prudhomme	16	16	in	in	IN
prudhomme	16	17	that	that	DT
prudhomme	16	18	machines	machine	NNS
prudhomme	16	19	are	be	VBP
prudhomme	16	20	increasingly	increasingly	RB
prudhomme	16	21	able	able	JJ
prudhomme	16	22	to	to	TO
prudhomme	16	23	ingest	ingest	VB
prudhomme	16	24	and	and	CC
prudhomme	16	25	analyze	analyze	VB
prudhomme	16	26	large	large	JJ
prudhomme	16	27	amounts	amount	NNS
prudhomme	16	28	of	of	IN
prudhomme	16	29	data	datum	NNS
prudhomme	16	30	and	and	CC
prudhomme	16	31	more	more	RBR
prudhomme	16	32	complex	complex	JJ
prudhomme	16	33	data	datum	NNS
prudhomme	16	34	to	to	TO
prudhomme	16	35	automatically	automatically	RB
prudhomme	16	36	produce	produce	VB
prudhomme	16	37	models	model	NNS
prudhomme	16	38	that	that	WDT
prudhomme	16	39	can	can	MD
prudhomme	16	40	deliver	deliver	VB
prudhomme	16	41	faster	fast	RBR
prudhomme	16	42	and	and	CC
prudhomme	16	43	more	more	RBR
prudhomme	16	44	accurate	accurate	JJ
prudhomme	16	45	results	result	NNS
prudhomme	16	46	.	.	.
prudhomme	17	1	[	[	-LRB-
prudhomme	17	2	footnoteRef:0	footnoteref:0	NN
prudhomme	17	3	]	]	-RRB-
prudhomme	17	4	Their	-PRON-	PRP$
prudhomme	17	5	“	"	``
prudhomme	17	6	power	power	NN
prudhomme	17	7	lies	lie	VBZ
prudhomme	17	8	in	in	IN
prudhomme	17	9	the	the	DT
prudhomme	17	10	fact	fact	NN
prudhomme	17	11	that	that	IN
prudhomme	17	12	machines	machine	NNS
prudhomme	17	13	can	can	MD
prudhomme	17	14	recognize	recognize	VB
prudhomme	17	15	patterns	pattern	NNS
prudhomme	17	16	efficiently	efficiently	RB
prudhomme	17	17	and	and	CC
prudhomme	17	18	routinely	routinely	RB
prudhomme	17	19	,	,	,
prudhomme	17	20	at	at	IN
prudhomme	17	21	a	a	DT
prudhomme	17	22	scale	scale	NN
prudhomme	17	23	and	and	CC
prudhomme	17	24	speed	speed	NN
prudhomme	17	25	that	that	WDT
prudhomme	17	26	humans	human	NNS
prudhomme	17	27	can	can	MD
prudhomme	17	28	not	not	RB
prudhomme	17	29	approach	approach	VB
prudhomme	17	30	,	,	,
prudhomme	17	31	”	"	''
prudhomme	17	32	writes	write	VBZ
prudhomme	17	33	Catherine	Catherine	NNP
prudhomme	17	34	Nicole	Nicole	NNP
prudhomme	17	35	Coleman	Coleman	NNP
prudhomme	17	36	,	,	,
prudhomme	17	37	digital	digital	JJ
prudhomme	17	38	research	research	NN
prudhomme	17	39	architect	architect	NN
prudhomme	17	40	for	for	IN
prudhomme	17	41	Stanford	Stanford	NNP
prudhomme	17	42	University	University	NNP
prudhomme	17	43	(	(	-LRB-
prudhomme	17	44	2017	2017	CD
prudhomme	17	45	)	)	-RRB-
prudhomme	17	46	.	.	.
prudhomme	18	1	[	[	-LRB-
prudhomme	18	2	0	0	CD
prudhomme	18	3	:	:	:
prudhomme	18	4	See	see	VB
prudhomme	18	5	SAS	SAS	NNP
prudhomme	18	6	n.d	n.d	NNP
prudhomme	18	7	.	.	NNP
prudhomme	18	8	and	and	CC
prudhomme	18	9	Brennan	Brennan	NNP
prudhomme	18	10	2019	2019	CD
prudhomme	18	11	.	.	.
prudhomme	18	12	]	]	-RRB-
prudhomme	19	1	A	a	DT
prudhomme	19	2	Paradigm	Paradigm	NNP
prudhomme	19	3	Shift	shift	NN
prudhomme	19	4	for	for	IN
prudhomme	19	5	Archives	Archives	NNPS
prudhomme	19	6	and	and	CC
prudhomme	19	7	Libraries	library	NNS
prudhomme	19	8	Within	within	IN
prudhomme	19	9	the	the	DT
prudhomme	19	10	context	context	NN
prudhomme	19	11	of	of	IN
prudhomme	19	12	university	university	NN
prudhomme	19	13	archives	archive	NNS
prudhomme	19	14	,	,	,
prudhomme	19	15	this	this	DT
prudhomme	19	16	paradigm	paradigm	NN
prudhomme	19	17	shift	shift	NN
prudhomme	19	18	has	have	VBZ
prudhomme	19	19	been	be	VBN
prudhomme	19	20	transforming	transform	VBG
prudhomme	19	21	the	the	DT
prudhomme	19	22	way	way	NN
prudhomme	19	23	we	-PRON-	PRP
prudhomme	19	24	interpret	interpret	VBP
prudhomme	19	25	archival	archival	NNP
prudhomme	19	26	data	datum	NNS
prudhomme	19	27	.	.	.
prudhomme	20	1	Artificial	artificial	JJ
prudhomme	20	2	intelligence	intelligence	NN
prudhomme	20	3	,	,	,
prudhomme	20	4	and	and	CC
prudhomme	20	5	specifically	specifically	RB
prudhomme	20	6	machine	machine	VB
prudhomme	20	7	learning	learn	VBG
prudhomme	20	8	as	as	IN
prudhomme	20	9	a	a	DT
prudhomme	20	10	subfield	subfield	NN
prudhomme	20	11	of	of	IN
prudhomme	20	12	AI	AI	NNP
prudhomme	20	13	,	,	,
prudhomme	20	14	has	have	VBZ
prudhomme	20	15	direct	direct	JJ
prudhomme	20	16	applications	application	NNS
prudhomme	20	17	through	through	IN
prudhomme	20	18	pattern	pattern	NN
prudhomme	20	19	recognition	recognition	NN
prudhomme	20	20	techniques	technique	NNS
prudhomme	20	21	that	that	WDT
prudhomme	20	22	predict	predict	VBP
prudhomme	20	23	the	the	DT
prudhomme	20	24	labeling	labeling	NN
prudhomme	20	25	values	value	NNS
prudhomme	20	26	for	for	IN
prudhomme	20	27	unlabeled	unlabeled	JJ
prudhomme	20	28	data	datum	NNS
prudhomme	20	29	.	.	.
prudhomme	21	1	As	as	IN
prudhomme	21	2	the	the	DT
prudhomme	21	3	software	software	NN
prudhomme	21	4	analytics	analytic	NNS
prudhomme	21	5	company	company	NN
prudhomme	21	6	SAS	SAS	NNP
prudhomme	21	7	argues	argue	VBZ
prudhomme	21	8	,	,	,
prudhomme	21	9	it	-PRON-	PRP
prudhomme	21	10	is	be	VBZ
prudhomme	21	11	“	"	``
prudhomme	21	12	the	the	DT
prudhomme	21	13	iterative	iterative	JJ
prudhomme	21	14	aspect	aspect	NN
prudhomme	21	15	of	of	IN
prudhomme	21	16	machine	machine	NN
prudhomme	21	17	learning	learn	VBG
prudhomme	21	18	[	[	-LRB-
prudhomme	21	19	that	that	IN
prudhomme	21	20	]	]	-RRB-
prudhomme	21	21	is	be	VBZ
prudhomme	21	22	important	important	JJ
prudhomme	21	23	because	because	IN
prudhomme	21	24	as	as	IN
prudhomme	21	25	models	model	NNS
prudhomme	21	26	are	be	VBP
prudhomme	21	27	exposed	expose	VBN
prudhomme	21	28	to	to	IN
prudhomme	21	29	new	new	JJ
prudhomme	21	30	data	datum	NNS
prudhomme	21	31	,	,	,
prudhomme	21	32	they	-PRON-	PRP
prudhomme	21	33	are	be	VBP
prudhomme	21	34	able	able	JJ
prudhomme	21	35	to	to	TO
prudhomme	21	36	independently	independently	RB
prudhomme	21	37	adapt	adapt	VB
prudhomme	21	38	.	.	.
prudhomme	22	1	They	-PRON-	PRP
prudhomme	22	2	learn	learn	VBP
prudhomme	22	3	from	from	IN
prudhomme	22	4	previous	previous	JJ
prudhomme	22	5	computations	computation	NNS
prudhomme	22	6	to	to	TO
prudhomme	22	7	produce	produce	VB
prudhomme	22	8	reliable	reliable	JJ
prudhomme	22	9	,	,	,
prudhomme	22	10	repeatable	repeatable	JJ
prudhomme	22	11	decisions	decision	NNS
prudhomme	22	12	and	and	CC
prudhomme	22	13	results	result	NNS
prudhomme	22	14	”	"	''
prudhomme	22	15	(	(	-LRB-
prudhomme	22	16	n.d	n.d	NNP
prudhomme	22	17	.	.	NNP
prudhomme	22	18	)	)	-RRB-
prudhomme	22	19	.	.	.
prudhomme	23	1	Case	case	NN
prudhomme	23	2	in	in	IN
prudhomme	23	3	point	point	NN
prudhomme	23	4	,	,	,
prudhomme	23	5	how	how	WRB
prudhomme	23	6	can	can	MD
prudhomme	23	7	we	-PRON-	PRP
prudhomme	23	8	use	use	VB
prudhomme	23	9	machine	machine	NN
prudhomme	23	10	learning	learn	VBG
prudhomme	23	11	to	to	TO
prudhomme	23	12	train	train	VB
prudhomme	23	13	machines	machine	NNS
prudhomme	23	14	and	and	CC
prudhomme	23	15	apply	apply	VB
prudhomme	23	16	facial	facial	JJ
prudhomme	23	17	and	and	CC
prudhomme	23	18	text	text	NN
prudhomme	23	19	recognition	recognition	NN
prudhomme	23	20	techniques	technique	NNS
prudhomme	23	21	to	to	TO
prudhomme	23	22	interpret	interpret	VB
prudhomme	23	23	the	the	DT
prudhomme	23	24	sheer	sheer	JJ
prudhomme	23	25	number	number	NN
prudhomme	23	26	of	of	IN
prudhomme	23	27	photographs	photograph	NNS
prudhomme	23	28	and	and	CC
prudhomme	23	29	texts	text	NNS
prudhomme	23	30	in	in	IN
prudhomme	23	31	either	either	DT
prudhomme	23	32	analog	analog	NN
prudhomme	23	33	or	or	CC
prudhomme	23	34	born	bear	VBN
prudhomme	23	35	-	-	HYPH
prudhomme	23	36	digital	digital	JJ
prudhomme	23	37	formats	format	NNS
prudhomme	23	38	held	hold	VBN
prudhomme	23	39	in	in	IN
prudhomme	23	40	archives	archive	NNS
prudhomme	23	41	and	and	CC
prudhomme	23	42	libraries	library	NNS
prudhomme	23	43	?	?	.
prudhomme	24	1	Combining	combine	VBG
prudhomme	24	2	automatic	automatic	JJ
prudhomme	24	3	processes	process	NNS
prudhomme	24	4	to	to	TO
prudhomme	24	5	assist	assist	VB
prudhomme	24	6	in	in	IN
prudhomme	24	7	supporting	support	VBG
prudhomme	24	8	inventory	inventory	NN
prudhomme	24	9	management	management	NN
prudhomme	24	10	with	with	IN
prudhomme	24	11	a	a	DT
prudhomme	24	12	focus	focus	NN
prudhomme	24	13	on	on	IN
prudhomme	24	14	descriptive	descriptive	JJ
prudhomme	24	15	metadata	metadata	NN
prudhomme	24	16	,	,	,
prudhomme	24	17	a	a	DT
prudhomme	24	18	machine	machine	NN
prudhomme	24	19	learning	learning	NN
prudhomme	24	20	solution	solution	NN
prudhomme	24	21	could	could	MD
prudhomme	24	22	help	help	VB
prudhomme	24	23	alleviate	alleviate	VB
prudhomme	24	24	time	time	NN
prudhomme	24	25	-	-	HYPH
prudhomme	24	26	consuming	consume	VBG
prudhomme	24	27	and	and	CC
prudhomme	24	28	relatively	relatively	RB
prudhomme	24	29	expensive	expensive	JJ
prudhomme	24	30	metadata	metadata	NN
prudhomme	24	31	tagging	tag	VBG
prudhomme	24	32	tasks	task	NNS
prudhomme	24	33	,	,	,
prudhomme	24	34	and	and	CC
prudhomme	24	35	thus	thus	RB
prudhomme	24	36	scale	scale	VBP
prudhomme	24	37	the	the	DT
prudhomme	24	38	process	process	NN
prudhomme	24	39	more	more	RBR
prudhomme	24	40	effectively	effectively	RB
prudhomme	24	41	using	use	VBG
prudhomme	24	42	relatively	relatively	RB
prudhomme	24	43	small	small	JJ
prudhomme	24	44	amounts	amount	NNS
prudhomme	24	45	of	of	IN
prudhomme	24	46	data	datum	NNS
prudhomme	24	47	.	.	.
prudhomme	25	1	However	however	RB
prudhomme	25	2	,	,	,
prudhomme	25	3	the	the	DT
prudhomme	25	4	traditional	traditional	JJ
prudhomme	25	5	approach	approach	NN
prudhomme	25	6	of	of	IN
prudhomme	25	7	machine	machine	NN
prudhomme	25	8	learning	learning	NN
prudhomme	25	9	would	would	MD
prudhomme	25	10	still	still	RB
prudhomme	25	11	require	require	VB
prudhomme	25	12	a	a	DT
prudhomme	25	13	significant	significant	JJ
prudhomme	25	14	time	time	NN
prudhomme	25	15	commitment	commitment	NN
prudhomme	25	16	by	by	IN
prudhomme	25	17	archivists	archivist	NNS
prudhomme	25	18	and	and	CC
prudhomme	25	19	curators	curator	NNS
prudhomme	25	20	to	to	TO
prudhomme	25	21	identify	identify	VB
prudhomme	25	22	essential	essential	JJ
prudhomme	25	23	features	feature	NNS
prudhomme	25	24	to	to	TO
prudhomme	25	25	make	make	VB
prudhomme	25	26	patterns	pattern	NNS
prudhomme	25	27	usable	usable	JJ
prudhomme	25	28	for	for	IN
prudhomme	25	29	data	datum	NNS
prudhomme	25	30	training	training	NN
prudhomme	25	31	.	.	.
prudhomme	26	1	By	by	IN
prudhomme	26	2	contrast	contrast	NN
prudhomme	26	3	,	,	,
prudhomme	26	4	deep	deep	JJ
prudhomme	26	5	learning	learning	NN
prudhomme	26	6	algorithms	algorithm	NNS
prudhomme	26	7	are	be	VBP
prudhomme	26	8	able	able	JJ
prudhomme	26	9	“	"	``
prudhomme	26	10	to	to	TO
prudhomme	26	11	learn	learn	VB
prudhomme	26	12	high	high	JJ
prudhomme	26	13	-	-	HYPH
prudhomme	26	14	level	level	NN
prudhomme	26	15	features	feature	NNS
prudhomme	26	16	from	from	IN
prudhomme	26	17	data	datum	NNS
prudhomme	26	18	in	in	IN
prudhomme	26	19	an	an	DT
prudhomme	26	20	incremental	incremental	JJ
prudhomme	26	21	manner	manner	NN
prudhomme	26	22	.	.	.
prudhomme	27	1	This	this	DT
prudhomme	27	2	eliminates	eliminate	VBZ
prudhomme	27	3	the	the	DT
prudhomme	27	4	need	need	NN
prudhomme	27	5	of	of	IN
prudhomme	27	6	domain	domain	NN
prudhomme	27	7	expertise	expertise	NN
prudhomme	27	8	and	and	CC
prudhomme	27	9	hard	hard	JJ
prudhomme	27	10	core	core	NN
prudhomme	27	11	feature	feature	NN
prudhomme	27	12	extraction	extraction	NN
prudhomme	27	13	”	"	''
prudhomme	27	14	(	(	-LRB-
prudhomme	27	15	Mahapatra	Mahapatra	NNP
prudhomme	27	16	2018	2018	CD
prudhomme	27	17	)	)	-RRB-
prudhomme	27	18	.	.	.
prudhomme	28	1	Deep	deep	JJ
prudhomme	28	2	learning	learning	NN
prudhomme	28	3	has	have	VBZ
prudhomme	28	4	regained	regain	VBN
prudhomme	28	5	popularity	popularity	NN
prudhomme	28	6	since	since	IN
prudhomme	28	7	the	the	DT
prudhomme	28	8	mid-2000s	mid-2000	NNS
prudhomme	28	9	due	due	IN
prudhomme	28	10	to	to	IN
prudhomme	28	11	“	"	``
prudhomme	28	12	fast	fast	JJ
prudhomme	28	13	development	development	NN
prudhomme	28	14	of	of	IN
prudhomme	28	15	high	high	JJ
prudhomme	28	16	-	-	HYPH
prudhomme	28	17	performance	performance	NN
prudhomme	28	18	parallel	parallel	JJ
prudhomme	28	19	computing	computing	NN
prudhomme	28	20	systems	system	NNS
prudhomme	28	21	,	,	,
prudhomme	28	22	such	such	JJ
prudhomme	28	23	as	as	IN
prudhomme	28	24	GPU	GPU	NNP
prudhomme	28	25	clusters	cluster	NNS
prudhomme	28	26	”	"	''
prudhomme	28	27	(	(	-LRB-
prudhomme	28	28	Zhao	Zhao	NNP
prudhomme	28	29	2019	2019	CD
prudhomme	28	30	,	,	,
prudhomme	28	31	3213	3213	CD
prudhomme	28	32	)	)	-RRB-
prudhomme	28	33	.	.	.
prudhomme	29	1	Deep	deep	JJ
prudhomme	29	2	learning	learn	VBG
prudhomme	29	3	neural	neural	JJ
prudhomme	29	4	networks	network	NNS
prudhomme	29	5	are	be	VBP
prudhomme	29	6	more	more	RBR
prudhomme	29	7	effective	effective	JJ
prudhomme	29	8	in	in	IN
prudhomme	29	9	feature	feature	NN
prudhomme	29	10	detection	detection	NN
prudhomme	29	11	as	as	IN
prudhomme	29	12	they	-PRON-	PRP
prudhomme	29	13	are	be	VBP
prudhomme	29	14	able	able	JJ
prudhomme	29	15	to	to	TO
prudhomme	29	16	solve	solve	VB
prudhomme	29	17	complex	complex	JJ
prudhomme	29	18	problems	problem	NNS
prudhomme	29	19	such	such	JJ
prudhomme	29	20	as	as	IN
prudhomme	29	21	image	image	NN
prudhomme	29	22	classification	classification	NN
prudhomme	29	23	with	with	IN
prudhomme	29	24	greater	great	JJR
prudhomme	29	25	accuracy	accuracy	NN
prudhomme	29	26	when	when	WRB
prudhomme	29	27	trained	train	VBN
prudhomme	29	28	with	with	IN
prudhomme	29	29	large	large	JJ
prudhomme	29	30	datasets	dataset	NNS
prudhomme	29	31	.	.	.
prudhomme	30	1	The	the	DT
prudhomme	30	2	challenge	challenge	NN
prudhomme	30	3	is	be	VBZ
prudhomme	30	4	whether	whether	IN
prudhomme	30	5	archives	archive	NNS
prudhomme	30	6	and	and	CC
prudhomme	30	7	libraries	library	NNS
prudhomme	30	8	can	can	MD
prudhomme	30	9	afford	afford	VB
prudhomme	30	10	to	to	TO
prudhomme	30	11	take	take	VB
prudhomme	30	12	advantage	advantage	NN
prudhomme	30	13	of	of	IN
prudhomme	30	14	greater	great	JJR
prudhomme	30	15	computing	computing	NN
prudhomme	30	16	capabilities	capability	NNS
prudhomme	30	17	to	to	TO
prudhomme	30	18	develop	develop	VB
prudhomme	30	19	sophisticated	sophisticated	JJ
prudhomme	30	20	techniques	technique	NNS
prudhomme	30	21	and	and	CC
prudhomme	30	22	make	make	VB
prudhomme	30	23	complex	complex	JJ
prudhomme	30	24	patterns	pattern	NNS
prudhomme	30	25	from	from	IN
prudhomme	30	26	thousands	thousand	NNS
prudhomme	30	27	of	of	IN
prudhomme	30	28	digital	digital	JJ
prudhomme	30	29	works	work	NNS
prudhomme	30	30	.	.	.
prudhomme	31	1	The	the	DT
prudhomme	31	2	sheer	sheer	JJ
prudhomme	31	3	size	size	NN
prudhomme	31	4	of	of	IN
prudhomme	31	5	library	library	NN
prudhomme	31	6	and	and	CC
prudhomme	31	7	archive	archive	JJ
prudhomme	31	8	datasets	dataset	NNS
prudhomme	31	9	,	,	,
prudhomme	31	10	such	such	JJ
prudhomme	31	11	as	as	IN
prudhomme	31	12	university	university	NN
prudhomme	31	13	photograph	photograph	NN
prudhomme	31	14	collections	collection	NNS
prudhomme	31	15	,	,	,
prudhomme	31	16	presents	present	VBZ
prudhomme	31	17	challenges	challenge	NNS
prudhomme	31	18	to	to	TO
prudhomme	31	19	properly	properly	RB
prudhomme	31	20	using	use	VBG
prudhomme	31	21	these	these	DT
prudhomme	31	22	new	new	JJ
prudhomme	31	23	,	,	,
prudhomme	31	24	sophisticated	sophisticated	JJ
prudhomme	31	25	techniques	technique	NNS
prudhomme	31	26	.	.	.
prudhomme	32	1	As	as	IN
prudhomme	32	2	Jason	Jason	NNP
prudhomme	32	3	Griffey	Griffey	NNP
prudhomme	32	4	writes	write	VBZ
prudhomme	32	5	,	,	,
prudhomme	32	6	“	"	``
prudhomme	32	7	AI	AI	NNP
prudhomme	32	8	is	be	VBZ
prudhomme	32	9	only	only	RB
prudhomme	32	10	as	as	RB
prudhomme	32	11	good	good	JJ
prudhomme	32	12	as	as	IN
prudhomme	32	13	its	-PRON-	PRP$
prudhomme	32	14	training	training	NN
prudhomme	32	15	data	datum	NNS
prudhomme	32	16	and	and	CC
prudhomme	32	17	the	the	DT
prudhomme	32	18	weighting	weighting	NN
prudhomme	32	19	that	that	WDT
prudhomme	32	20	is	be	VBZ
prudhomme	32	21	given	give	VBN
prudhomme	32	22	to	to	IN
prudhomme	32	23	the	the	DT
prudhomme	32	24	system	system	NN
prudhomme	32	25	as	as	IN
prudhomme	32	26	it	-PRON-	PRP
prudhomme	32	27	learns	learn	VBZ
prudhomme	32	28	to	to	TO
prudhomme	32	29	make	make	VB
prudhomme	32	30	decisions	decision	NNS
prudhomme	32	31	.	.	.
prudhomme	33	1	If	if	IN
prudhomme	33	2	that	that	DT
prudhomme	33	3	data	data	NN
prudhomme	33	4	is	be	VBZ
prudhomme	33	5	biased	biased	JJ
prudhomme	33	6	,	,	,
prudhomme	33	7	contains	contain	VBZ
prudhomme	33	8	bad	bad	JJ
prudhomme	33	9	examples	example	NNS
prudhomme	33	10	of	of	IN
prudhomme	33	11	decision	decision	NN
prudhomme	33	12	-	-	HYPH
prudhomme	33	13	making	making	NN
prudhomme	33	14	,	,	,
prudhomme	33	15	or	or	CC
prudhomme	33	16	is	be	VBZ
prudhomme	33	17	simply	simply	RB
prudhomme	33	18	collected	collect	VBN
prudhomme	33	19	in	in	IN
prudhomme	33	20	such	such	PDT
prudhomme	33	21	a	a	DT
prudhomme	33	22	way	way	NN
prudhomme	33	23	that	that	IN
prudhomme	33	24	it	-PRON-	PRP
prudhomme	33	25	is	be	VBZ
prudhomme	33	26	n’t	not	RB
prudhomme	33	27	representative	representative	JJ
prudhomme	33	28	of	of	IN
prudhomme	33	29	the	the	DT
prudhomme	33	30	entirety	entirety	NN
prudhomme	33	31	of	of	IN
prudhomme	33	32	the	the	DT
prudhomme	33	33	problem	problem	NN
prudhomme	33	34	set	set	VBN
prudhomme	33	35	[	[	-LRB-
prudhomme	33	36	,	,	,
prudhomme	33	37	…	…	NFP
prudhomme	33	38	]	]	-RRB-
prudhomme	33	39	that	that	DT
prudhomme	33	40	system	system	NN
prudhomme	33	41	is	be	VBZ
prudhomme	33	42	going	go	VBG
prudhomme	33	43	to	to	TO
prudhomme	33	44	produce	produce	VB
prudhomme	33	45	broken	break	VBN
prudhomme	33	46	,	,	,
prudhomme	33	47	biased	biased	JJ
prudhomme	33	48	,	,	,
prudhomme	33	49	and	and	CC
prudhomme	33	50	bad	bad	JJ
prudhomme	33	51	outputs	output	NNS
prudhomme	33	52	”	"	''
prudhomme	33	53	(	(	-LRB-
prudhomme	33	54	2019	2019	CD
prudhomme	33	55	,	,	,
prudhomme	33	56	8)	8)	CD
prudhomme	33	57	.	.	.
prudhomme	34	1	How	how	WRB
prudhomme	34	2	can	can	MD
prudhomme	34	3	cultural	cultural	JJ
prudhomme	34	4	heritage	heritage	NN
prudhomme	34	5	institutions	institution	NNS
prudhomme	34	6	ensure	ensure	VBP
prudhomme	34	7	that	that	IN
prudhomme	34	8	their	-PRON-	PRP$
prudhomme	34	9	machine	machine	NN
prudhomme	34	10	learning	learn	VBG
prudhomme	34	11	algorithms	algorithm	NNS
prudhomme	34	12	avoid	avoid	VBP
prudhomme	34	13	such	such	JJ
prudhomme	34	14	bad	bad	JJ
prudhomme	34	15	outputs	output	NNS
prudhomme	34	16	?	?	.
prudhomme	35	1	Comment	comment	NN
prudhomme	35	2	by	by	IN
prudhomme	35	3	Daniel	Daniel	NNP
prudhomme	35	4	Johnson	Johnson	NNP
prudhomme	35	5	:	:	:
prudhomme	35	6	Not	not	RB
prudhomme	35	7	sure	sure	JJ
prudhomme	35	8	where	where	WRB
prudhomme	35	9	the	the	DT
prudhomme	35	10	discrepancy	discrepancy	NN
prudhomme	35	11	comes	come	VBZ
prudhomme	35	12	in	in	RP
prudhomme	35	13	,	,	,
prudhomme	35	14	but	but	CC
prudhomme	35	15	the	the	DT
prudhomme	35	16	changes	change	NNS
prudhomme	35	17	I	-PRON-	PRP
prudhomme	35	18	've	have	VB
prudhomme	35	19	made	make	VBN
prudhomme	35	20	are	be	VBP
prudhomme	35	21	pulled	pull	VBN
prudhomme	35	22	directly	directly	RB
prudhomme	35	23	from	from	IN
prudhomme	35	24	Griffey	Griffey	NNP
prudhomme	35	25	's	's	POS
prudhomme	35	26	report	report	NN
prudhomme	35	27	here	here	RB
prudhomme	35	28	:	:	:
prudhomme	35	29	https://journals.ala.org/index.php/ltr/issue/viewIssue/709/471	https://journals.ala.org/index.php/ltr/issue/viewIssue/709/471	,
prudhomme	35	30	.	.	.
prudhomme	36	1	Implications	implication	NNS
prudhomme	36	2	to	to	IN
prudhomme	36	3	Machine	Machine	NNP
prudhomme	36	4	Learning	Learning	NNP
prudhomme	36	5	Machine	Machine	NNP
prudhomme	36	6	learning	learning	NN
prudhomme	36	7	has	have	VBZ
prudhomme	36	8	the	the	DT
prudhomme	36	9	potential	potential	NN
prudhomme	36	10	to	to	TO
prudhomme	36	11	enrich	enrich	VB
prudhomme	36	12	the	the	DT
prudhomme	36	13	value	value	NN
prudhomme	36	14	of	of	IN
prudhomme	36	15	digital	digital	JJ
prudhomme	36	16	collections	collection	NNS
prudhomme	36	17	by	by	IN
prudhomme	36	18	building	build	VBG
prudhomme	36	19	upon	upon	IN
prudhomme	36	20	experts	expert	NNS
prudhomme	36	21	’	’	POS
prudhomme	36	22	knowledge	knowledge	NN
prudhomme	36	23	.	.	.
prudhomme	37	1	It	-PRON-	PRP
prudhomme	37	2	can	can	MD
prudhomme	37	3	also	also	RB
prudhomme	37	4	help	help	VB
prudhomme	37	5	identify	identify	VB
prudhomme	37	6	resources	resource	NNS
prudhomme	37	7	that	that	IN
prudhomme	37	8	archivists	archivist	NNS
prudhomme	37	9	and	and	CC
prudhomme	37	10	curators	curator	NNS
prudhomme	37	11	may	may	MD
prudhomme	37	12	never	never	RB
prudhomme	37	13	have	have	VB
prudhomme	37	14	the	the	DT
prudhomme	37	15	time	time	NN
prudhomme	37	16	for	for	IN
prudhomme	37	17	,	,	,
prudhomme	37	18	and	and	CC
prudhomme	37	19	at	at	IN
prudhomme	37	20	the	the	DT
prudhomme	37	21	same	same	JJ
prudhomme	37	22	time	time	NN
prudhomme	37	23	correct	correct	JJ
prudhomme	37	24	assumptions	assumption	NNS
prudhomme	37	25	about	about	IN
prudhomme	37	26	heritage	heritage	NN
prudhomme	37	27	materials	material	NNS
prudhomme	37	28	.	.	.
prudhomme	38	1	It	-PRON-	PRP
prudhomme	38	2	can	can	MD
prudhomme	38	3	generate	generate	VB
prudhomme	38	4	the	the	DT
prudhomme	38	5	necessary	necessary	JJ
prudhomme	38	6	added	add	VBN
prudhomme	38	7	value	value	NN
prudhomme	38	8	to	to	TO
prudhomme	38	9	support	support	VB
prudhomme	38	10	the	the	DT
prudhomme	38	11	mission	mission	NN
prudhomme	38	12	of	of	IN
prudhomme	38	13	archives	archive	NNS
prudhomme	38	14	and	and	CC
prudhomme	38	15	libraries	library	NNS
prudhomme	38	16	in	in	IN
prudhomme	38	17	providing	provide	VBG
prudhomme	38	18	a	a	DT
prudhomme	38	19	public	public	JJ
prudhomme	38	20	good	good	NN
prudhomme	38	21	.	.	.
prudhomme	39	1	Annie	Annie	NNP
prudhomme	39	2	Schweikert	Schweikert	NNP
prudhomme	39	3	states	state	VBZ
prudhomme	39	4	that	that	IN
prudhomme	39	5	“	"	``
prudhomme	39	6	artificial	artificial	JJ
prudhomme	39	7	intelligence	intelligence	NN
prudhomme	39	8	and	and	CC
prudhomme	39	9	machine	machine	NN
prudhomme	39	10	learning	learning	NN
prudhomme	39	11	tools	tool	NNS
prudhomme	39	12	are	be	VBP
prudhomme	39	13	considered	consider	VBN
prudhomme	39	14	by	by	IN
prudhomme	39	15	many	many	JJ
prudhomme	39	16	to	to	TO
prudhomme	39	17	be	be	VB
prudhomme	39	18	the	the	DT
prudhomme	39	19	next	next	JJ
prudhomme	39	20	step	step	NN
prudhomme	39	21	in	in	IN
prudhomme	39	22	streamlining	streamline	VBG
prudhomme	39	23	workflows	workflow	NNS
prudhomme	39	24	and	and	CC
prudhomme	39	25	easing	ease	VBG
prudhomme	39	26	workloads	workload	NNS
prudhomme	39	27	”	"	''
prudhomme	39	28	(	(	-LRB-
prudhomme	39	29	2019	2019	CD
prudhomme	39	30	,	,	,
prudhomme	39	31	6	6	CD
prudhomme	39	32	)	)	-RRB-
prudhomme	39	33	.	.	.
prudhomme	40	1	For	for	IN
prudhomme	40	2	images	image	NNS
prudhomme	40	3	,	,	,
prudhomme	40	4	how	how	WRB
prudhomme	40	5	can	can	MD
prudhomme	40	6	archives	archive	NNS
prudhomme	40	7	build	build	VB
prudhomme	40	8	a	a	DT
prudhomme	40	9	data	data	NN
prudhomme	40	10	-	-	HYPH
prudhomme	40	11	labeling	labeling	NN
prudhomme	40	12	pipeline	pipeline	NN
prudhomme	40	13	into	into	IN
prudhomme	40	14	their	-PRON-	PRP$
prudhomme	40	15	digital	digital	JJ
prudhomme	40	16	curation	curation	NN
prudhomme	40	17	workflow	workflow	NN
prudhomme	40	18	that	that	WDT
prudhomme	40	19	enables	enable	VBZ
prudhomme	40	20	machine	machine	NN
prudhomme	40	21	learning	learning	NN
prudhomme	40	22	of	of	IN
prudhomme	40	23	collections	collection	NNS
prudhomme	40	24	?	?	.
prudhomme	41	1	With	with	IN
prudhomme	41	2	the	the	DT
prudhomme	41	3	objective	objective	NN
prudhomme	41	4	being	be	VBG
prudhomme	41	5	to	to	TO
prudhomme	41	6	augment	augment	JJ
prudhomme	41	7	knowledge	knowledge	NN
prudhomme	41	8	and	and	CC
prudhomme	41	9	create	create	VB
prudhomme	41	10	value	value	NN
prudhomme	41	11	,	,	,
prudhomme	41	12	how	how	WRB
prudhomme	41	13	can	can	MD
prudhomme	41	14	archives	archive	NNS
prudhomme	41	15	and	and	CC
prudhomme	41	16	libraries	library	NNS
prudhomme	41	17	“	"	``
prudhomme	41	18	bring	bring	VBP
prudhomme	41	19	the	the	DT
prudhomme	41	20	skills	skill	NNS
prudhomme	41	21	and	and	CC
prudhomme	41	22	knowledge	knowledge	NN
prudhomme	41	23	of	of	IN
prudhomme	41	24	library	library	NN
prudhomme	41	25	staff	staff	NN
prudhomme	41	26	,	,	,
prudhomme	41	27	scholars	scholar	NNS
prudhomme	41	28	,	,	,
prudhomme	41	29	and	and	CC
prudhomme	41	30	students	student	NNS
prudhomme	41	31	together	together	RB
prudhomme	41	32	to	to	TO
prudhomme	41	33	design	design	VB
prudhomme	41	34	an	an	DT
prudhomme	41	35	intelligent	intelligent	JJ
prudhomme	41	36	information	information	NN
prudhomme	41	37	system	system	NN
prudhomme	41	38	”	"	''
prudhomme	41	39	(	(	-LRB-
prudhomme	41	40	Coleman	Coleman	NNP
prudhomme	41	41	2017	2017	CD
prudhomme	41	42	)	)	-RRB-
prudhomme	41	43	?	?	.
prudhomme	42	1	Despite	despite	IN
prudhomme	42	2	the	the	DT
prudhomme	42	3	opportunities	opportunity	NNS
prudhomme	42	4	to	to	IN
prudhomme	42	5	augment	augment	JJ
prudhomme	42	6	knowledge	knowledge	NN
prudhomme	42	7	from	from	IN
prudhomme	42	8	facial	facial	JJ
prudhomme	42	9	recognition	recognition	NN
prudhomme	42	10	,	,	,
prudhomme	42	11	models	model	NNS
prudhomme	42	12	generated	generate	VBN
prudhomme	42	13	by	by	IN
prudhomme	42	14	machine	machine	NN
prudhomme	42	15	learning	learning	NN
prudhomme	42	16	algorithms	algorithm	NNS
prudhomme	42	17	should	should	MD
prudhomme	42	18	be	be	VB
prudhomme	42	19	scrutinized	scrutinize	VBN
prudhomme	42	20	so	so	RB
prudhomme	42	21	long	long	RB
prudhomme	42	22	it	-PRON-	PRP
prudhomme	42	23	is	be	VBZ
prudhomme	42	24	unclear	unclear	JJ
prudhomme	42	25	how	how	WRB
prudhomme	42	26	choices	choice	NNS
prudhomme	42	27	are	be	VBP
prudhomme	42	28	made	make	VBN
prudhomme	42	29	in	in	IN
prudhomme	42	30	feature	feature	NN
prudhomme	42	31	selection	selection	NN
prudhomme	42	32	.	.	.
prudhomme	43	1	Machine	machine	NN
prudhomme	43	2	learning	learning	NN
prudhomme	43	3	“	"	``
prudhomme	43	4	has	have	VBZ
prudhomme	43	5	the	the	DT
prudhomme	43	6	potential	potential	NN
prudhomme	43	7	to	to	TO
prudhomme	43	8	reveal	reveal	VB
prudhomme	43	9	things	thing	NNS
prudhomme	43	10	...	...	NFP
prudhomme	43	11	that	that	IN
prudhomme	43	12	we	-PRON-	PRP
prudhomme	43	13	did	do	VBD
prudhomme	43	14	not	not	RB
prudhomme	43	15	know	know	VB
prudhomme	43	16	and	and	CC
prudhomme	43	17	did	do	VBD
prudhomme	43	18	not	not	RB
prudhomme	43	19	want	want	VB
prudhomme	43	20	to	to	TO
prudhomme	43	21	know	know	VB
prudhomme	43	22	”	"	''
prudhomme	43	23	as	as	IN
prudhomme	43	24	Charlie	Charlie	NNP
prudhomme	43	25	Harper	Harper	NNP
prudhomme	43	26	asserts	assert	VBZ
prudhomme	43	27	(	(	-LRB-
prudhomme	43	28	2018	2018	CD
prudhomme	43	29	)	)	-RRB-
prudhomme	43	30	.	.	.
prudhomme	44	1	It	-PRON-	PRP
prudhomme	44	2	can	can	MD
prudhomme	44	3	also	also	RB
prudhomme	44	4	have	have	VB
prudhomme	44	5	direct	direct	JJ
prudhomme	44	6	ethical	ethical	JJ
prudhomme	44	7	implications	implication	NNS
prudhomme	44	8	,	,	,
prudhomme	44	9	leading	lead	VBG
prudhomme	44	10	to	to	IN
prudhomme	44	11	biased	biased	JJ
prudhomme	44	12	interpretations	interpretation	NNS
prudhomme	44	13	for	for	IN
prudhomme	44	14	nefarious	nefarious	JJ
prudhomme	44	15	motives	motive	NNS
prudhomme	44	16	.	.	.
prudhomme	45	1	Machine	Machine	NNP
prudhomme	45	2	Learning	Learning	NNP
prudhomme	45	3	and	and	CC
prudhomme	45	4	Deep	Deep	NNP
prudhomme	45	5	Learning	Learning	NNP
prudhomme	45	6	on	on	IN
prudhomme	45	7	the	the	DT
prudhomme	45	8	Grounds	Grounds	NNPS
prudhomme	45	9	of	of	IN
prudhomme	45	10	Generating	Generating	NNP
prudhomme	45	11	Value	Value	NNP
prudhomme	45	12	In	in	IN
prudhomme	45	13	the	the	DT
prudhomme	45	14	fall	fall	NN
prudhomme	45	15	2018	2018	CD
prudhomme	45	16	,	,	,
prudhomme	45	17	Oklahoma	Oklahoma	NNP
prudhomme	45	18	State	State	NNP
prudhomme	45	19	University	University	NNP
prudhomme	45	20	Archives	Archives	NNPS
prudhomme	45	21	began	begin	VBD
prudhomme	45	22	to	to	TO
prudhomme	45	23	look	look	VB
prudhomme	45	24	more	more	RBR
prudhomme	45	25	closely	closely	RB
prudhomme	45	26	at	at	IN
prudhomme	45	27	a	a	DT
prudhomme	45	28	machine	machine	NN
prudhomme	45	29	learning	learn	VBG
prudhomme	45	30	solution	solution	NN
prudhomme	45	31	to	to	TO
prudhomme	45	32	facilitate	facilitate	VB
prudhomme	45	33	metadata	metadata	NN
prudhomme	45	34	creation	creation	NN
prudhomme	45	35	in	in	IN
prudhomme	45	36	support	support	NN
prudhomme	45	37	of	of	IN
prudhomme	45	38	curation	curation	NN
prudhomme	45	39	,	,	,
prudhomme	45	40	preservation	preservation	NN
prudhomme	45	41	,	,	,
prudhomme	45	42	and	and	CC
prudhomme	45	43	discovery	discovery	NN
prudhomme	45	44	.	.	.
prudhomme	46	1	Conceptually	conceptually	RB
prudhomme	46	2	,	,	,
prudhomme	46	3	we	-PRON-	PRP
prudhomme	46	4	envisioned	envision	VBD
prudhomme	46	5	boosting	boost	VBG
prudhomme	46	6	the	the	DT
prudhomme	46	7	curation	curation	NN
prudhomme	46	8	of	of	IN
prudhomme	46	9	digital	digital	JJ
prudhomme	46	10	assets	asset	NNS
prudhomme	46	11	,	,	,
prudhomme	46	12	setting	set	VBG
prudhomme	46	13	up	up	RP
prudhomme	46	14	policies	policy	NNS
prudhomme	46	15	to	to	TO
prudhomme	46	16	prioritize	prioritize	VB
prudhomme	46	17	digital	digital	JJ
prudhomme	46	18	preservation	preservation	NN
prudhomme	46	19	and	and	CC
prudhomme	46	20	access	access	NN
prudhomme	46	21	for	for	IN
prudhomme	46	22	education	education	NN
prudhomme	46	23	and	and	CC
prudhomme	46	24	research	research	NN
prudhomme	46	25	,	,	,
prudhomme	46	26	and	and	CC
prudhomme	46	27	enhancing	enhance	VBG
prudhomme	46	28	the	the	DT
prudhomme	46	29	long	long	JJ
prudhomme	46	30	-	-	HYPH
prudhomme	46	31	term	term	NN
prudhomme	46	32	value	value	NN
prudhomme	46	33	of	of	IN
prudhomme	46	34	those	those	DT
prudhomme	46	35	data	datum	NNS
prudhomme	46	36	.	.	.
prudhomme	47	1	In	in	IN
prudhomme	47	2	this	this	DT
prudhomme	47	3	section	section	NN
prudhomme	47	4	,	,	,
prudhomme	47	5	I	-PRON-	PRP
prudhomme	47	6	describe	describe	VBP
prudhomme	47	7	the	the	DT
prudhomme	47	8	parameters	parameter	NNS
prudhomme	47	9	of	of	IN
prudhomme	47	10	automation	automation	NN
prudhomme	47	11	and	and	CC
prudhomme	47	12	machine	machine	NN
prudhomme	47	13	learning	learning	NN
prudhomme	47	14	used	use	VBN
prudhomme	47	15	to	to	TO
prudhomme	47	16	support	support	VB
prudhomme	47	17	inventory	inventory	NN
prudhomme	47	18	work	work	NN
prudhomme	47	19	and	and	CC
prudhomme	47	20	experiment	experiment	NN
prudhomme	47	21	with	with	IN
prudhomme	47	22	face	face	NN
prudhomme	47	23	recognition	recognition	NN
prudhomme	47	24	models	model	NNS
prudhomme	47	25	to	to	TO
prudhomme	47	26	add	add	VB
prudhomme	47	27	contextualization	contextualization	NN
prudhomme	47	28	to	to	IN
prudhomme	47	29	digital	digital	JJ
prudhomme	47	30	objects	object	NNS
prudhomme	47	31	.	.	.
prudhomme	48	1	From	from	IN
prudhomme	48	2	a	a	DT
prudhomme	48	3	digital	digital	JJ
prudhomme	48	4	curation	curation	NN
prudhomme	48	5	perspective	perspective	NN
prudhomme	48	6	,	,	,
prudhomme	48	7	the	the	DT
prudhomme	48	8	objective	objective	NN
prudhomme	48	9	is	be	VBZ
prudhomme	48	10	to	to	TO
prudhomme	48	11	explore	explore	VB
prudhomme	48	12	ways	way	NNS
prudhomme	48	13	to	to	TO
prudhomme	48	14	add	add	VB
prudhomme	48	15	value	value	NN
prudhomme	48	16	to	to	IN
prudhomme	48	17	digital	digital	JJ
prudhomme	48	18	objects	object	NNS
prudhomme	48	19	for	for	IN
prudhomme	48	20	which	which	WDT
prudhomme	48	21	little	little	JJ
prudhomme	48	22	information	information	NN
prudhomme	48	23	is	be	VBZ
prudhomme	48	24	known	know	VBN
prudhomme	48	25	,	,	,
prudhomme	48	26	if	if	IN
prudhomme	48	27	any	any	DT
prudhomme	48	28	,	,	,
prudhomme	48	29	in	in	IN
prudhomme	48	30	order	order	NN
prudhomme	48	31	to	to	TO
prudhomme	48	32	increase	increase	VB
prudhomme	48	33	the	the	DT
prudhomme	48	34	visibility	visibility	NN
prudhomme	48	35	of	of	IN
prudhomme	48	36	archival	archival	NN
prudhomme	48	37	collections	collection	NNS
prudhomme	48	38	.	.	.
prudhomme	49	1	What	what	WP
prudhomme	49	2	started	start	VBD
prudhomme	49	3	this	this	DT
prudhomme	49	4	Pilot	Pilot	NNP
prudhomme	49	5	Project	Project	NNP
prudhomme	49	6	?	?	.
prudhomme	50	1	Before	before	IN
prudhomme	50	2	proceeding	proceeding	NN
prudhomme	50	3	,	,	,
prudhomme	50	4	we	-PRON-	PRP
prudhomme	50	5	needed	need	VBD
prudhomme	50	6	to	to	TO
prudhomme	50	7	gain	gain	VB
prudhomme	50	8	a	a	DT
prudhomme	50	9	deeper	deep	JJR
prudhomme	50	10	understanding	understanding	NN
prudhomme	50	11	of	of	IN
prudhomme	50	12	the	the	DT
prudhomme	50	13	large	large	JJ
prudhomme	50	14	quantity	quantity	NN
prudhomme	50	15	of	of	IN
prudhomme	50	16	files	file	NNS
prudhomme	50	17	held	hold	VBN
prudhomme	50	18	in	in	IN
prudhomme	50	19	the	the	DT
prudhomme	50	20	archives	archive	NNS
prudhomme	50	21	–	–	:
prudhomme	50	22	both	both	DT
prudhomme	50	23	types	type	NNS
prudhomme	50	24	of	of	IN
prudhomme	50	25	data	datum	NNS
prudhomme	50	26	and	and	CC
prudhomme	50	27	metadata	metadata	NN
prudhomme	50	28	.	.	.
prudhomme	51	1	The	the	DT
prudhomme	51	2	challenge	challenge	NN
prudhomme	51	3	was	be	VBD
prudhomme	51	4	that	that	DT
prudhomme	51	5	with	with	IN
prudhomme	51	6	so	so	RB
prudhomme	51	7	many	many	JJ
prudhomme	51	8	files	file	NNS
prudhomme	51	9	,	,	,
prudhomme	51	10	so	so	RB
prudhomme	51	11	many	many	JJ
prudhomme	51	12	formats	format	NNS
prudhomme	51	13	,	,	,
prudhomme	51	14	files	file	NNS
prudhomme	51	15	become	become	VBP
prudhomme	51	16	duplicated	duplicated	JJ
prudhomme	51	17	and	and	CC
prudhomme	51	18	renamed	rename	VBN
prudhomme	51	19	,	,	,
prudhomme	51	20	doctored	doctor	VBD
prudhomme	51	21	,	,	,
prudhomme	51	22	and	and	CC
prudhomme	51	23	scattered	scatter	VBN
prudhomme	51	24	throughout	throughout	IN
prudhomme	51	25	directories	directory	NNS
prudhomme	51	26	to	to	TO
prudhomme	51	27	accommodate	accommodate	VB
prudhomme	51	28	different	different	JJ
prudhomme	51	29	types	type	NNS
prudhomme	51	30	of	of	IN
prudhomme	51	31	projects	project	NNS
prudhomme	51	32	over	over	IN
prudhomme	51	33	time	time	NN
prudhomme	51	34	,	,	,
prudhomme	51	35	making	make	VBG
prudhomme	51	36	it	-PRON-	PRP
prudhomme	51	37	hard	hard	JJ
prudhomme	51	38	to	to	TO
prudhomme	51	39	sift	sift	VB
prudhomme	51	40	due	due	IN
prudhomme	51	41	to	to	IN
prudhomme	51	42	sparse	sparse	JJ
prudhomme	51	43	metadata	metadata	NN
prudhomme	51	44	tags	tag	NNS
prudhomme	51	45	that	that	WDT
prudhomme	51	46	may	may	MD
prudhomme	51	47	have	have	VB
prudhomme	51	48	differed	differ	VBN
prudhomme	51	49	from	from	IN
prudhomme	51	50	one	one	CD
prudhomme	51	51	system	system	NN
prudhomme	51	52	to	to	IN
prudhomme	51	53	another	another	DT
prudhomme	51	54	.	.	.
prudhomme	52	1	In	in	IN
prudhomme	52	2	short	short	JJ
prudhomme	52	3	,	,	,
prudhomme	52	4	how	how	WRB
prudhomme	52	5	could	could	MD
prudhomme	52	6	we	-PRON-	PRP
prudhomme	52	7	justify	justify	VB
prudhomme	52	8	the	the	DT
prudhomme	52	9	value	value	NN
prudhomme	52	10	of	of	IN
prudhomme	52	11	these	these	DT
prudhomme	52	12	digital	digital	JJ
prudhomme	52	13	assets	asset	NNS
prudhomme	52	14	for	for	IN
prudhomme	52	15	curatorial	curatorial	JJ
prudhomme	52	16	purposes	purpose	NNS
prudhomme	52	17	?	?	.
prudhomme	53	1	How	how	WRB
prudhomme	53	2	much	much	JJ
prudhomme	53	3	could	could	MD
prudhomme	53	4	we	-PRON-	PRP
prudhomme	53	5	rely	rely	VB
prudhomme	53	6	on	on	IN
prudhomme	53	7	the	the	DT
prudhomme	53	8	established	established	JJ
prudhomme	53	9	institutional	institutional	JJ
prudhomme	53	10	memory	memory	NN
prudhomme	53	11	within	within	IN
prudhomme	53	12	the	the	DT
prudhomme	53	13	archives	archive	NNS
prudhomme	53	14	?	?	.
prudhomme	54	1	Lastly	lastly	RB
prudhomme	54	2	,	,	,
prudhomme	54	3	could	could	MD
prudhomme	54	4	machine	machine	VB
prudhomme	54	5	learning	learning	NN
prudhomme	54	6	or	or	CC
prudhomme	54	7	deep	deep	JJ
prudhomme	54	8	learning	learning	NN
prudhomme	54	9	applications	application	NNS
prudhomme	54	10	help	help	VBP
prudhomme	54	11	us	-PRON-	PRP
prudhomme	54	12	build	build	VB
prudhomme	54	13	a	a	DT
prudhomme	54	14	greater	great	JJR
prudhomme	54	15	capacity	capacity	NN
prudhomme	54	16	to	to	IN
prudhomme	54	17	augment	augment	JJ
prudhomme	54	18	knowledge	knowledge	NN
prudhomme	54	19	?	?	.
prudhomme	55	1	In	in	IN
prudhomme	55	2	order	order	NN
prudhomme	55	3	to	to	TO
prudhomme	55	4	optimize	optimize	VB
prudhomme	55	5	resources	resource	NNS
prudhomme	55	6	and	and	CC
prudhomme	55	7	systematically	systematically	RB
prudhomme	55	8	make	make	VB
prudhomme	55	9	sense	sense	NN
prudhomme	55	10	of	of	IN
prudhomme	55	11	data	datum	NNS
prudhomme	55	12	,	,	,
prudhomme	55	13	we	-PRON-	PRP
prudhomme	55	14	needed	need	VBD
prudhomme	55	15	to	to	TO
prudhomme	55	16	determine	determine	VB
prudhomme	55	17	that	that	DT
prudhomme	55	18	machine	machine	NN
prudhomme	55	19	learning	learning	NN
prudhomme	55	20	could	could	MD
prudhomme	55	21	generate	generate	VB
prudhomme	55	22	value	value	NN
prudhomme	55	23	,	,	,
prudhomme	55	24	which	which	WDT
prudhomme	55	25	in	in	IN
prudhomme	55	26	turn	turn	NN
prudhomme	55	27	could	could	MD
prudhomme	55	28	help	help	VB
prudhomme	55	29	us	-PRON-	PRP
prudhomme	55	30	more	more	RBR
prudhomme	55	31	tightly	tightly	RB
prudhomme	55	32	integrate	integrate	VB
prudhomme	55	33	our	-PRON-	PRP$
prudhomme	55	34	digital	digital	JJ
prudhomme	55	35	initiatives	initiative	NNS
prudhomme	55	36	with	with	IN
prudhomme	55	37	machine	machine	NN
prudhomme	55	38	learning	learning	NN
prudhomme	55	39	applications	application	NNS
prudhomme	55	40	.	.	.
prudhomme	56	1	Such	such	JJ
prudhomme	56	2	applications	application	NNS
prudhomme	56	3	would	would	MD
prudhomme	56	4	only	only	RB
prudhomme	56	5	be	be	VB
prudhomme	56	6	as	as	RB
prudhomme	56	7	effective	effective	JJ
prudhomme	56	8	as	as	IN
prudhomme	56	9	the	the	DT
prudhomme	56	10	data	datum	NNS
prudhomme	56	11	are	be	VBP
prudhomme	56	12	good	good	JJ
prudhomme	56	13	for	for	IN
prudhomme	56	14	training	training	NN
prudhomme	56	15	and	and	CC
prudhomme	56	16	the	the	DT
prudhomme	56	17	value	value	NN
prudhomme	56	18	we	-PRON-	PRP
prudhomme	56	19	could	could	MD
prudhomme	56	20	derive	derive	VB
prudhomme	56	21	from	from	IN
prudhomme	56	22	them	-PRON-	PRP
prudhomme	56	23	.	.	.
prudhomme	57	1	Methodology	Methodology	NNP
prudhomme	57	2	and	and	CC
prudhomme	57	3	Plan	Plan	NNP
prudhomme	57	4	of	of	IN
prudhomme	57	5	Action	Action	NNP
prudhomme	57	6	First	first	RB
prudhomme	57	7	,	,	,
prudhomme	57	8	we	-PRON-	PRP
prudhomme	57	9	recruited	recruit	VBD
prudhomme	57	10	two	two	CD
prudhomme	57	11	student	student	NN
prudhomme	57	12	interns	intern	NNS
prudhomme	57	13	to	to	TO
prudhomme	57	14	create	create	VB
prudhomme	57	15	a	a	DT
prudhomme	57	16	series	series	NN
prudhomme	57	17	of	of	IN
prudhomme	57	18	processes	process	NNS
prudhomme	57	19	that	that	WDT
prudhomme	57	20	would	would	MD
prudhomme	57	21	automatically	automatically	RB
prudhomme	57	22	populate	populate	VB
prudhomme	57	23	a	a	DT
prudhomme	57	24	comprehensive	comprehensive	JJ
prudhomme	57	25	inventory	inventory	NN
prudhomme	57	26	of	of	IN
prudhomme	57	27	all	all	DT
prudhomme	57	28	digital	digital	JJ
prudhomme	57	29	collections	collection	NNS
prudhomme	57	30	,	,	,
prudhomme	57	31	including	include	VBG
prudhomme	57	32	finding	find	VBG
prudhomme	57	33	duplicate	duplicate	JJ
prudhomme	57	34	files	file	NNS
prudhomme	57	35	by	by	IN
prudhomme	57	36	hashing	hash	VBG
prudhomme	57	37	.	.	.
prudhomme	58	1	We	-PRON-	PRP
prudhomme	58	2	generated	generate	VBD
prudhomme	58	3	the	the	DT
prudhomme	58	4	inventory	inventory	NN
prudhomme	58	5	by	by	IN
prudhomme	58	6	developing	develop	VBG
prudhomme	58	7	a	a	DT
prudhomme	58	8	process	process	NN
prudhomme	58	9	that	that	WDT
prudhomme	58	10	could	could	MD
prudhomme	58	11	be	be	VB
prudhomme	58	12	universally	universally	RB
prudhomme	58	13	adapted	adapt	VBN
prudhomme	58	14	to	to	IN
prudhomme	58	15	all	all	DT
prudhomme	58	16	library	library	NN
prudhomme	58	17	digital	digital	JJ
prudhomme	58	18	collections	collection	NNS
prudhomme	58	19	,	,	,
prudhomme	58	20	setting	set	VBG
prudhomme	58	21	up	up	RP
prudhomme	58	22	a	a	DT
prudhomme	58	23	universal	universal	JJ
prudhomme	58	24	list	list	NN
prudhomme	58	25	of	of	IN
prudhomme	58	26	works	work	NNS
prudhomme	58	27	and	and	CC
prudhomme	58	28	their	-PRON-	PRP$
prudhomme	58	29	associated	associate	VBN
prudhomme	58	30	metadata	metadata	NN
prudhomme	58	31	,	,	,
prudhomme	58	32	with	with	IN
prudhomme	58	33	a	a	DT
prudhomme	58	34	focus	focus	NN
prudhomme	58	35	on	on	IN
prudhomme	58	36	descriptive	descriptive	JJ
prudhomme	58	37	metadata	metadata	NN
prudhomme	58	38	,	,	,
prudhomme	58	39	which	which	WDT
prudhomme	58	40	in	in	IN
prudhomme	58	41	turn	turn	NN
prudhomme	58	42	could	could	MD
prudhomme	58	43	support	support	VB
prudhomme	58	44	digital	digital	JJ
prudhomme	58	45	curation	curation	NN
prudhomme	58	46	and	and	CC
prudhomme	58	47	discovery	discovery	NN
prudhomme	58	48	of	of	IN
prudhomme	58	49	archival	archival	NN
prudhomme	58	50	materials	material	NNS
prudhomme	58	51	—	—	:
prudhomme	58	52	digitized	digitize	VBN
prudhomme	58	53	analog	analog	NN
prudhomme	58	54	materials	material	NNS
prudhomme	58	55	and	and	CC
prudhomme	58	56	born	bear	VBN
prudhomme	58	57	-	-	HYPH
prudhomme	58	58	digital	digital	JJ
prudhomme	58	59	materials	material	NNS
prudhomme	58	60	.	.	.
prudhomme	59	1	We	-PRON-	PRP
prudhomme	59	2	developed	develop	VBD
prudhomme	59	3	a	a	DT
prudhomme	59	4	universal	universal	JJ
prudhomme	59	5	policy	policy	NN
prudhomme	59	6	for	for	IN
prudhomme	59	7	digital	digital	JJ
prudhomme	59	8	archival	archival	NN
prudhomme	59	9	collections	collection	NNS
prudhomme	59	10	,	,	,
prudhomme	59	11	which	which	WDT
prudhomme	59	12	would	would	MD
prudhomme	59	13	allow	allow	VB
prudhomme	59	14	us	-PRON-	PRP
prudhomme	59	15	to	to	TO
prudhomme	59	16	incorporate	incorporate	VB
prudhomme	59	17	all	all	DT
prudhomme	59	18	forms	form	NNS
prudhomme	59	19	of	of	IN
prudhomme	59	20	metadata	metadata	NN
prudhomme	59	21	into	into	IN
prudhomme	59	22	a	a	DT
prudhomme	59	23	single	single	JJ
prudhomme	59	24	format	format	NN
prudhomme	59	25	to	to	IN
prudhomme	59	26	remedy	remedy	NN
prudhomme	59	27	inconsistencies	inconsistency	NNS
prudhomme	59	28	in	in	IN
prudhomme	59	29	existing	exist	VBG
prudhomme	59	30	metadata	metadata	NN
prudhomme	59	31	.	.	.
prudhomme	60	1	This	this	DT
prudhomme	60	2	first	first	JJ
prudhomme	60	3	phase	phase	NN
prudhomme	60	4	was	be	VBD
prudhomme	60	5	critical	critical	JJ
prudhomme	60	6	in	in	IN
prudhomme	60	7	the	the	DT
prudhomme	60	8	sense	sense	NN
prudhomme	60	9	that	that	IN
prudhomme	60	10	it	-PRON-	PRP
prudhomme	60	11	would	would	MD
prudhomme	60	12	condition	condition	VB
prudhomme	60	13	the	the	DT
prudhomme	60	14	cleansing	cleansing	NN
prudhomme	60	15	and	and	CC
prudhomme	60	16	organizing	organizing	NN
prudhomme	60	17	of	of	IN
prudhomme	60	18	data	datum	NNS
prudhomme	60	19	.	.	.
prudhomme	61	1	We	-PRON-	PRP
prudhomme	61	2	could	could	MD
prudhomme	61	3	then	then	RB
prudhomme	61	4	proceed	proceed	VB
prudhomme	61	5	with	with	IN
prudhomme	61	6	the	the	DT
prudhomme	61	7	design	design	NN
prudhomme	61	8	of	of	IN
prudhomme	61	9	a	a	DT
prudhomme	61	10	face	face	NN
prudhomme	61	11	recognition	recognition	NN
prudhomme	61	12	database	database	NN
prudhomme	61	13	,	,	,
prudhomme	61	14	with	with	IN
prudhomme	61	15	the	the	DT
prudhomme	61	16	intent	intent	NN
prudhomme	61	17	to	to	TO
prudhomme	61	18	trace	trace	VB
prudhomme	61	19	individuals	individual	NNS
prudhomme	61	20	featured	feature	VBN
prudhomme	61	21	in	in	IN
prudhomme	61	22	the	the	DT
prudhomme	61	23	inventory	inventory	NN
prudhomme	61	24	works	work	NNS
prudhomme	61	25	of	of	IN
prudhomme	61	26	the	the	DT
prudhomme	61	27	archives	archive	NNS
prudhomme	61	28	to	to	IN
prudhomme	61	29	the	the	DT
prudhomme	61	30	extent	extent	NN
prudhomme	61	31	that	that	IN
prudhomme	61	32	our	-PRON-	PRP$
prudhomme	61	33	data	datum	NNS
prudhomme	61	34	were	be	VBD
prudhomme	61	35	accurate	accurate	JJ
prudhomme	61	36	.	.	.
prudhomme	62	1	We	-PRON-	PRP
prudhomme	62	2	utilized	utilize	VBD
prudhomme	62	3	the	the	DT
prudhomme	62	4	Oklahoma	Oklahoma	NNP
prudhomme	62	5	State	State	NNP
prudhomme	62	6	University	University	NNP
prudhomme	62	7	Yearbook	Yearbook	NNP
prudhomme	62	8	collections	collection	NNS
prudhomme	62	9	and	and	CC
prudhomme	62	10	other	other	JJ
prudhomme	62	11	digital	digital	JJ
prudhomme	62	12	collections	collection	NNS
prudhomme	62	13	as	as	IN
prudhomme	62	14	authoritative	authoritative	JJ
prudhomme	62	15	references	reference	NNS
prudhomme	62	16	for	for	IN
prudhomme	62	17	other	other	JJ
prudhomme	62	18	works	work	NNS
prudhomme	62	19	,	,	,
prudhomme	62	20	for	for	IN
prudhomme	62	21	the	the	DT
prudhomme	62	22	purpose	purpose	NN
prudhomme	62	23	of	of	IN
prudhomme	62	24	contextualization	contextualization	NN
prudhomme	62	25	to	to	TO
prudhomme	62	26	augment	augment	VB
prudhomme	62	27	our	-PRON-	PRP$
prudhomme	62	28	data	data	NN
prudhomme	62	29	capacity	capacity	NN
prudhomme	62	30	.	.	.
prudhomme	63	1	Second	second	JJ
prudhomme	63	2	,	,	,
prudhomme	63	3	we	-PRON-	PRP
prudhomme	63	4	implemented	implement	VBD
prudhomme	63	5	our	-PRON-	PRP$
prudhomme	63	6	plan	plan	NN
prudhomme	63	7	;	;	,
prudhomme	63	8	worked	work	VBD
prudhomme	63	9	closely	closely	RB
prudhomme	63	10	with	with	IN
prudhomme	63	11	the	the	DT
prudhomme	63	12	Library	Library	NNP
prudhomme	63	13	Systems	Systems	NNPS
prudhomme	63	14	’	'	''
prudhomme	63	15	team	team	NN
prudhomme	63	16	within	within	IN
prudhomme	63	17	a	a	DT
prudhomme	63	18	Windows	Windows	NNP
prudhomme	63	19	-	-	HYPH
prudhomme	63	20	based	base	VBN
prudhomme	63	21	environment	environment	NN
prudhomme	63	22	;	;	:
prudhomme	63	23	decided	decide	VBD
prudhomme	63	24	on	on	IN
prudhomme	63	25	Graphics	Graphics	NNP
prudhomme	63	26	Processing	Processing	NNP
prudhomme	63	27	Unit	Unit	NNP
prudhomme	63	28	(	(	-LRB-
prudhomme	63	29	GPU	GPU	NNP
prudhomme	63	30	)	)	-RRB-
prudhomme	63	31	performance	performance	NN
prudhomme	63	32	and	and	CC
prudhomme	63	33	cost	cost	NN
prudhomme	63	34	,	,	,
prudhomme	63	35	taking	take	VBG
prudhomme	63	36	into	into	IN
prudhomme	63	37	consideration	consideration	NN
prudhomme	63	38	that	that	IN
prudhomme	63	39	training	training	NN
prudhomme	63	40	neural	neural	JJ
prudhomme	63	41	networks	network	NNS
prudhomme	63	42	necessitates	necessitate	VBZ
prudhomme	63	43	computing	compute	VBG
prudhomme	63	44	power	power	NN
prudhomme	63	45	;	;	:
prudhomme	63	46	determined	determine	VBN
prudhomme	63	47	storage	storage	NN
prudhomme	63	48	needs	need	VBZ
prudhomme	63	49	;	;	:
prudhomme	63	50	and	and	CC
prudhomme	63	51	fulfilled	fulfil	VBD
prudhomme	63	52	other	other	JJ
prudhomme	63	53	logistical	logistical	JJ
prudhomme	63	54	requirements	requirement	NNS
prudhomme	63	55	to	to	TO
prudhomme	63	56	begin	begin	VB
prudhomme	63	57	the	the	DT
prudhomme	63	58	step	step	NN
prudhomme	63	59	-	-	HYPH
prudhomme	63	60	by	by	IN
prudhomme	63	61	-	-	HYPH
prudhomme	63	62	step	step	NN
prudhomme	63	63	process	process	NN
prudhomme	63	64	of	of	IN
prudhomme	63	65	establishing	establish	VBG
prudhomme	63	66	a	a	DT
prudhomme	63	67	pattern	pattern	NN
prudhomme	63	68	recognition	recognition	NN
prudhomme	63	69	database	database	NN
prudhomme	63	70	.	.	.
prudhomme	64	1	We	-PRON-	PRP
prudhomme	64	2	designed	design	VBD
prudhomme	64	3	the	the	DT
prudhomme	64	4	database	database	NN
prudhomme	64	5	on	on	IN
prudhomme	64	6	known	know	VBN
prudhomme	64	7	objects	object	NNS
prudhomme	64	8	before	before	IN
prudhomme	64	9	introducing	introduce	VBG
prudhomme	64	10	and	and	CC
prudhomme	64	11	comparing	compare	VBG
prudhomme	64	12	new	new	JJ
prudhomme	64	13	data	datum	NNS
prudhomme	64	14	to	to	TO
prudhomme	64	15	contextualize	contextualize	VB
prudhomme	64	16	each	each	DT
prudhomme	64	17	entry	entry	NN
prudhomme	64	18	.	.	.
prudhomme	65	1	With	with	IN
prudhomme	65	2	this	this	DT
prudhomme	65	3	framework	framework	NN
prudhomme	65	4	,	,	,
prudhomme	65	5	we	-PRON-	PRP
prudhomme	65	6	would	would	MD
prudhomme	65	7	be	be	VB
prudhomme	65	8	able	able	JJ
prudhomme	65	9	to	to	TO
prudhomme	65	10	add	add	VB
prudhomme	65	11	general	general	JJ
prudhomme	65	12	metadata	metadata	NN
prudhomme	65	13	tags	tag	NNS
prudhomme	65	14	to	to	IN
prudhomme	65	15	a	a	DT
prudhomme	65	16	uniform	uniform	JJ
prudhomme	65	17	storage	storage	NN
prudhomme	65	18	system	system	NN
prudhomme	65	19	using	use	VBG
prudhomme	65	20	deep	deep	JJ
prudhomme	65	21	learning	learning	NN
prudhomme	65	22	technology	technology	NN
prudhomme	65	23	.	.	.
prudhomme	66	1	Third	third	JJ
prudhomme	66	2	,	,	,
prudhomme	66	3	we	-PRON-	PRP
prudhomme	66	4	applied	apply	VBD
prudhomme	66	5	Tesseract	Tesseract	NNP
prudhomme	66	6	OCR	OCR	NNP
prudhomme	66	7	on	on	IN
prudhomme	66	8	a	a	DT
prudhomme	66	9	series	series	NN
prudhomme	66	10	of	of	IN
prudhomme	66	11	archival	archival	NNP
prudhomme	66	12	image	image	NN
prudhomme	66	13	-	-	HYPH
prudhomme	66	14	text	text	NN
prudhomme	66	15	combinations	combination	NNS
prudhomme	66	16	from	from	IN
prudhomme	66	17	the	the	DT
prudhomme	66	18	archives	archive	NNS
prudhomme	66	19	to	to	TO
prudhomme	66	20	extract	extract	VB
prudhomme	66	21	printed	print	VBN
prudhomme	66	22	text	text	NN
prudhomme	66	23	from	from	IN
prudhomme	66	24	those	those	DT
prudhomme	66	25	images	image	NNS
prudhomme	66	26	and	and	CC
prudhomme	66	27	photographs	photograph	NNS
prudhomme	66	28	.	.	.
prudhomme	67	1	“	"	``
prudhomme	67	2	Tesseract	tesseract	NN
prudhomme	67	3	4	4	CD
prudhomme	67	4	adds	add	VBZ
prudhomme	67	5	a	a	DT
prudhomme	67	6	new	new	JJ
prudhomme	67	7	neural	neural	JJ
prudhomme	67	8	net	net	NN
prudhomme	67	9	(	(	-LRB-
prudhomme	67	10	LSTM	LSTM	NNP
prudhomme	67	11	)	)	-RRB-
prudhomme	67	12	[	[	-LRB-
prudhomme	67	13	Long	Long	NNP
prudhomme	67	14	Short	Short	NNP
prudhomme	67	15	-	-	HYPH
prudhomme	67	16	Term	Term	NNP
prudhomme	67	17	Memory	Memory	NNP
prudhomme	67	18	]	]	-RRB-
prudhomme	67	19	based	base	VBN
prudhomme	67	20	OCR	ocr	NN
prudhomme	67	21	engine	engine	NN
prudhomme	67	22	which	which	WDT
prudhomme	67	23	is	be	VBZ
prudhomme	67	24	focused	focus	VBN
prudhomme	67	25	on	on	IN
prudhomme	67	26	line	line	NN
prudhomme	67	27	recognition	recognition	NN
prudhomme	67	28	,	,	,
prudhomme	67	29	”	"	''
prudhomme	67	30	while	while	IN
prudhomme	67	31	also	also	RB
prudhomme	67	32	recognizing	recognize	VBG
prudhomme	67	33	character	character	NN
prudhomme	67	34	patterns	pattern	NNS
prudhomme	67	35	(	(	-LRB-
prudhomme	67	36	“	"	``
prudhomme	67	37	Tesseract	Tesseract	NNP
prudhomme	67	38	”	"	''
prudhomme	67	39	n.d	n.d	NNP
prudhomme	67	40	.	.	NNP
prudhomme	67	41	)	)	-RRB-
prudhomme	67	42	.	.	.
prudhomme	68	1	We	-PRON-	PRP
prudhomme	68	2	were	be	VBD
prudhomme	68	3	able	able	JJ
prudhomme	68	4	to	to	TO
prudhomme	68	5	obtain	obtain	VB
prudhomme	68	6	successful	successful	JJ
prudhomme	68	7	output	output	NN
prudhomme	68	8	for	for	IN
prudhomme	68	9	the	the	DT
prudhomme	68	10	most	most	JJS
prudhomme	68	11	part	part	NN
prudhomme	68	12	,	,	,
prudhomme	68	13	with	with	IN
prudhomme	68	14	the	the	DT
prudhomme	68	15	exception	exception	NN
prudhomme	68	16	of	of	IN
prudhomme	68	17	a	a	DT
prudhomme	68	18	few	few	JJ
prudhomme	68	19	characters	character	NNS
prudhomme	68	20	that	that	WDT
prudhomme	68	21	were	be	VBD
prudhomme	68	22	hard	hard	JJ
prudhomme	68	23	to	to	TO
prudhomme	68	24	detect	detect	VB
prudhomme	68	25	due	due	IN
prudhomme	68	26	to	to	IN
prudhomme	68	27	pixelation	pixelation	NN
prudhomme	68	28	and	and	CC
prudhomme	68	29	font	font	NN
prudhomme	68	30	types	type	NNS
prudhomme	68	31	.	.	.
prudhomme	69	1	Fourth	fourth	JJ
prudhomme	69	2	,	,	,
prudhomme	69	3	we	-PRON-	PRP
prudhomme	69	4	looked	look	VBD
prudhomme	69	5	into	into	IN
prudhomme	69	6	object	object	NN
prudhomme	69	7	identifiers	identifier	NNS
prudhomme	69	8	,	,	,
prudhomme	69	9	keeping	keep	VBG
prudhomme	69	10	in	in	IN
prudhomme	69	11	mind	mind	NN
prudhomme	69	12	that	that	IN
prudhomme	69	13	“	"	``
prudhomme	69	14	When	when	WRB
prudhomme	69	15	there	there	EX
prudhomme	69	16	are	be	VBP
prudhomme	69	17	scarce	scarce	JJ
prudhomme	69	18	or	or	CC
prudhomme	69	19	insufficient	insufficient	JJ
prudhomme	69	20	labeled	label	VBN
prudhomme	69	21	data	datum	NNS
prudhomme	69	22	,	,	,
prudhomme	69	23	pre	pre	JJ
prudhomme	69	24	-	-	JJ
prudhomme	69	25	training	training	NN
prudhomme	69	26	is	be	VBZ
prudhomme	69	27	usually	usually	RB
prudhomme	69	28	conducted	conduct	VBN
prudhomme	69	29	”	"	''
prudhomme	69	30	(	(	-LRB-
prudhomme	69	31	Zhao	Zhao	NNP
prudhomme	69	32	2019	2019	CD
prudhomme	69	33	,	,	,
prudhomme	69	34	3215	3215	CD
prudhomme	69	35	)	)	-RRB-
prudhomme	69	36	.	.	.
prudhomme	70	1	Working	work	VBG
prudhomme	70	2	through	through	IN
prudhomme	70	3	the	the	DT
prudhomme	70	4	inventory	inventory	NN
prudhomme	70	5	process	process	NN
prudhomme	70	6	,	,	,
prudhomme	70	7	we	-PRON-	PRP
prudhomme	70	8	knew	know	VBD
prudhomme	70	9	that	that	IN
prudhomme	70	10	we	-PRON-	PRP
prudhomme	70	11	would	would	MD
prudhomme	70	12	also	also	RB
prudhomme	70	13	need	need	VB
prudhomme	70	14	to	to	TO
prudhomme	70	15	label	label	VB
prudhomme	70	16	more	more	JJR
prudhomme	70	17	data	datum	NNS
prudhomme	70	18	to	to	TO
prudhomme	70	19	grow	grow	VB
prudhomme	70	20	our	-PRON-	PRP$
prudhomme	70	21	capacity	capacity	NN
prudhomme	70	22	.	.	.
prudhomme	71	1	We	-PRON-	PRP
prudhomme	71	2	chose	choose	VBD
prudhomme	71	3	to	to	TO
prudhomme	71	4	use	use	VB
prudhomme	71	5	ResNet-50	ResNet-50	NNS
prudhomme	71	6	,	,	,
prudhomme	71	7	a	a	DT
prudhomme	71	8	smaller	small	JJR
prudhomme	71	9	version	version	NN
prudhomme	71	10	backbone	backbone	NN
prudhomme	71	11	of	of	IN
prudhomme	71	12	Keras	Keras	NNP
prudhomme	71	13	-	-	HYPH
prudhomme	71	14	Retinanet	Retinanet	NNP
prudhomme	71	15	,	,	,
prudhomme	71	16	frequently	frequently	RB
prudhomme	71	17	used	use	VBN
prudhomme	71	18	as	as	IN
prudhomme	71	19	a	a	DT
prudhomme	71	20	starting	starting	NN
prudhomme	71	21	point	point	NN
prudhomme	71	22	for	for	IN
prudhomme	71	23	transfer	transfer	NN
prudhomme	71	24	learning	learning	NN
prudhomme	71	25	.	.	.
prudhomme	72	1	Keras	Keras	NNP
prudhomme	72	2	is	be	VBZ
prudhomme	72	3	a	a	DT
prudhomme	72	4	deep	deep	JJ
prudhomme	72	5	learning	learning	NN
prudhomme	72	6	network	network	NN
prudhomme	72	7	API	api	NN
prudhomme	72	8	(	(	-LRB-
prudhomme	72	9	Application	Application	NNP
prudhomme	72	10	Programming	Programming	NNP
prudhomme	72	11	Interface	Interface	NNP
prudhomme	72	12	)	)	-RRB-
prudhomme	72	13	that	that	WDT
prudhomme	72	14	supports	support	VBZ
prudhomme	72	15	multiple	multiple	JJ
prudhomme	72	16	back	back	JJ
prudhomme	72	17	-	-	HYPH
prudhomme	72	18	end	end	NN
prudhomme	72	19	neural	neural	JJ
prudhomme	72	20	network	network	NN
prudhomme	72	21	computation	computation	NN
prudhomme	72	22	engines	engine	NNS
prudhomme	72	23	(	(	-LRB-
prudhomme	72	24	Heller	Heller	NNP
prudhomme	72	25	2019	2019	CD
prudhomme	72	26	)	)	-RRB-
prudhomme	72	27	and	and	CC
prudhomme	72	28	RetinaNet	RetinaNet	NNP
prudhomme	72	29	is	be	VBZ
prudhomme	72	30	a	a	DT
prudhomme	72	31	single	single	JJ
prudhomme	72	32	,	,	,
prudhomme	72	33	unified	unified	JJ
prudhomme	72	34	network	network	NN
prudhomme	72	35	consisting	consist	VBG
prudhomme	72	36	of	of	IN
prudhomme	72	37	a	a	DT
prudhomme	72	38	backbone	backbone	NN
prudhomme	72	39	network	network	NN
prudhomme	72	40	and	and	CC
prudhomme	72	41	two	two	CD
prudhomme	72	42	task	task	NN
prudhomme	72	43	-	-	HYPH
prudhomme	72	44	specific	specific	JJ
prudhomme	72	45	subnetworks	subnetwork	NNS
prudhomme	72	46	used	use	VBN
prudhomme	72	47	for	for	IN
prudhomme	72	48	object	object	NN
prudhomme	72	49	detection	detection	NN
prudhomme	72	50	(	(	-LRB-
prudhomme	72	51	Karaka	Karaka	NNP
prudhomme	72	52	2019	2019	CD
prudhomme	72	53	)	)	-RRB-
prudhomme	72	54	.	.	.
prudhomme	73	1	We	-PRON-	PRP
prudhomme	73	2	proceeded	proceed	VBD
prudhomme	73	3	by	by	IN
prudhomme	73	4	first	first	RB
prudhomme	73	5	dumping	dump	VBG
prudhomme	73	6	a	a	DT
prudhomme	73	7	lot	lot	NN
prudhomme	73	8	of	of	IN
prudhomme	73	9	pre	pre	JJ
prudhomme	73	10	-	-	JJ
prudhomme	73	11	tagged	tagged	JJ
prudhomme	73	12	information	information	NN
prudhomme	73	13	from	from	IN
prudhomme	73	14	pre	pre	JJ
prudhomme	73	15	-	-	JJ
prudhomme	73	16	existing	existing	JJ
prudhomme	73	17	datasets	dataset	NNS
prudhomme	73	18	into	into	IN
prudhomme	73	19	this	this	DT
prudhomme	73	20	neural	neural	JJ
prudhomme	73	21	network	network	NN
prudhomme	73	22	.	.	.
prudhomme	74	1	We	-PRON-	PRP
prudhomme	74	2	experimented	experiment	VBD
prudhomme	74	3	with	with	IN
prudhomme	74	4	three	three	CD
prudhomme	74	5	open	open	JJ
prudhomme	74	6	source	source	NN
prudhomme	74	7	datasets	dataset	NNS
prudhomme	74	8	:	:	:
prudhomme	74	9	PASCAL	PASCAL	NNP
prudhomme	74	10	VOC	VOC	NNP
prudhomme	74	11	2012	2012	CD
prudhomme	74	12	,	,	,
prudhomme	74	13	a	a	DT
prudhomme	74	14	set	set	NN
prudhomme	74	15	including	include	VBG
prudhomme	74	16	20	20	CD
prudhomme	74	17	object	object	NN
prudhomme	74	18	categories	category	NNS
prudhomme	74	19	;	;	:
prudhomme	74	20	Open	Open	NNP
prudhomme	74	21	Images	image	NNS
prudhomme	74	22	Database	Database	NNP
prudhomme	74	23	(	(	-LRB-
prudhomme	74	24	OID	OID	NNP
prudhomme	74	25	)	)	-RRB-
prudhomme	74	26	,	,	,
prudhomme	74	27	a	a	DT
prudhomme	74	28	very	very	RB
prudhomme	74	29	large	large	JJ
prudhomme	74	30	dataset	dataset	NN
prudhomme	74	31	annotated	annotate	VBD
prudhomme	74	32	with	with	IN
prudhomme	74	33	image	image	NN
prudhomme	74	34	-	-	HYPH
prudhomme	74	35	level	level	NN
prudhomme	74	36	labels	label	NNS
prudhomme	74	37	and	and	CC
prudhomme	74	38	object	object	NN
prudhomme	74	39	bounding	bounding	NN
prudhomme	74	40	boxes	box	NNS
prudhomme	74	41	;	;	:
prudhomme	74	42	and	and	CC
prudhomme	74	43	Microsoft	Microsoft	NNP
prudhomme	74	44	COCO	COCO	NNP
prudhomme	74	45	,	,	,
prudhomme	74	46	a	a	DT
prudhomme	74	47	large	large	JJ
prudhomme	74	48	-	-	HYPH
prudhomme	74	49	scale	scale	NN
prudhomme	74	50	object	object	NN
prudhomme	74	51	detection	detection	NN
prudhomme	74	52	,	,	,
prudhomme	74	53	segmentation	segmentation	NN
prudhomme	74	54	,	,	,
prudhomme	74	55	and	and	CC
prudhomme	74	56	captioning	caption	VBG
prudhomme	74	57	dataset	dataset	NN
prudhomme	74	58	.	.	.
prudhomme	75	1	With	with	IN
prudhomme	75	2	a	a	DT
prudhomme	75	3	few	few	JJ
prudhomme	75	4	faces	face	NNS
prudhomme	75	5	from	from	IN
prudhomme	75	6	the	the	DT
prudhomme	75	7	OID	OID	NNP
prudhomme	75	8	dataset	dataset	NN
prudhomme	75	9	,	,	,
prudhomme	75	10	we	-PRON-	PRP
prudhomme	75	11	could	could	MD
prudhomme	75	12	compare	compare	VB
prudhomme	75	13	and	and	CC
prudhomme	75	14	see	see	VB
prudhomme	75	15	if	if	IN
prudhomme	75	16	a	a	DT
prudhomme	75	17	face	face	NN
prudhomme	75	18	was	be	VBD
prudhomme	75	19	previously	previously	RB
prudhomme	75	20	recognized	recognize	VBN
prudhomme	75	21	.	.	.
prudhomme	76	1	Expanding	expand	VBG
prudhomme	76	2	our	-PRON-	PRP$
prudhomme	76	3	process	process	NN
prudhomme	76	4	to	to	IN
prudhomme	76	5	data	datum	NNS
prudhomme	76	6	known	know	VBN
prudhomme	76	7	from	from	IN
prudhomme	76	8	the	the	DT
prudhomme	76	9	archives	archive	NNS
prudhomme	76	10	collection	collection	NN
prudhomme	76	11	,	,	,
prudhomme	76	12	we	-PRON-	PRP
prudhomme	76	13	determined	determine	VBD
prudhomme	76	14	facial	facial	JJ
prudhomme	76	15	areas	area	NNS
prudhomme	76	16	,	,	,
prudhomme	76	17	and	and	CC
prudhomme	76	18	more	more	RBR
prudhomme	76	19	specifically	specifically	RB
prudhomme	76	20	,	,	,
prudhomme	76	21	assigned	assign	VBN
prudhomme	76	22	bounding	bound	VBG
prudhomme	76	23	box	box	NN
prudhomme	76	24	regressions	regression	NNS
prudhomme	76	25	to	to	TO
prudhomme	76	26	feed	feed	VB
prudhomme	76	27	into	into	IN
prudhomme	76	28	the	the	DT
prudhomme	76	29	facial	facial	JJ
prudhomme	76	30	recognition	recognition	NN
prudhomme	76	31	API	API	NNP
prudhomme	76	32	,	,	,
prudhomme	76	33	based	base	VBN
prudhomme	76	34	on	on	IN
prudhomme	76	35	Keras	Keras	NNP
prudhomme	76	36	code	code	NN
prudhomme	76	37	written	write	VBN
prudhomme	76	38	in	in	IN
prudhomme	76	39	Python	Python	NNP
prudhomme	76	40	.	.	.
prudhomme	77	1	The	the	DT
prudhomme	77	2	face	face	NN
prudhomme	77	3	recognition	recognition	NN
prudhomme	77	4	API	API	NNP
prudhomme	77	5	is	be	VBZ
prudhomme	77	6	available	available	JJ
prudhomme	77	7	via	via	IN
prudhomme	77	8	GitHub	GitHub	NNP
prudhomme	77	9	.	.	.
prudhomme	78	1	[	[	-LRB-
prudhomme	78	2	footnoteRef:1	footnoteRef:1	NNP
prudhomme	78	3	]	]	-RRB-
prudhomme	78	4	It	-PRON-	PRP
prudhomme	78	5	uses	use	VBZ
prudhomme	78	6	a	a	DT
prudhomme	78	7	method	method	NN
prudhomme	78	8	called	call	VBN
prudhomme	78	9	Histogram	Histogram	NNP
prudhomme	78	10	of	of	IN
prudhomme	78	11	Oriented	Oriented	NNP
prudhomme	78	12	Gradient	Gradient	NNP
prudhomme	78	13	(	(	-LRB-
prudhomme	78	14	HOG	HOG	NNP
prudhomme	78	15	)	)	-RRB-
prudhomme	78	16	encoding	encode	VBG
prudhomme	78	17	that	that	WDT
prudhomme	78	18	makes	make	VBZ
prudhomme	78	19	the	the	DT
prudhomme	78	20	actual	actual	JJ
prudhomme	78	21	face	face	NN
prudhomme	78	22	recognition	recognition	NN
prudhomme	78	23	process	process	NN
prudhomme	78	24	much	much	RB
prudhomme	78	25	easier	easy	JJR
prudhomme	78	26	to	to	TO
prudhomme	78	27	implement	implement	VB
prudhomme	78	28	for	for	IN
prudhomme	78	29	individuals	individual	NNS
prudhomme	78	30	because	because	IN
prudhomme	78	31	the	the	DT
prudhomme	78	32	encodings	encoding	NNS
prudhomme	78	33	are	be	VBP
prudhomme	78	34	fairly	fairly	RB
prudhomme	78	35	unique	unique	JJ
prudhomme	78	36	for	for	IN
prudhomme	78	37	every	every	DT
prudhomme	78	38	person	person	NN
prudhomme	78	39	,	,	,
prudhomme	78	40	as	as	IN
prudhomme	78	41	opposed	oppose	VBN
prudhomme	78	42	to	to	IN
prudhomme	78	43	encoding	encode	VBG
prudhomme	78	44	images	image	NNS
prudhomme	78	45	and	and	CC
prudhomme	78	46	trying	try	VBG
prudhomme	78	47	to	to	TO
prudhomme	78	48	blindly	blindly	RB
prudhomme	78	49	figure	figure	VB
prudhomme	78	50	out	out	RP
prudhomme	78	51	which	which	WDT
prudhomme	78	52	parts	part	NNS
prudhomme	78	53	are	be	VBP
prudhomme	78	54	faces	face	NNS
prudhomme	78	55	based	base	VBN
prudhomme	78	56	on	on	IN
prudhomme	78	57	our	-PRON-	PRP$
prudhomme	78	58	label	label	NN
prudhomme	78	59	boxes	box	NNS
prudhomme	78	60	.	.	.
prudhomme	79	1	Figure	figure	NN
prudhomme	79	2	1	1	CD
prudhomme	79	3	illustrates	illustrate	VBZ
prudhomme	79	4	our	-PRON-	PRP$
prudhomme	79	5	test	test	NN
prudhomme	79	6	,	,	,
prudhomme	79	7	confirming	confirm	VBG
prudhomme	79	8	from	from	IN
prudhomme	79	9	two	two	CD
prudhomme	79	10	very	very	RB
prudhomme	79	11	different	different	JJ
prudhomme	79	12	photographs	photograph	NNS
prudhomme	79	13	the	the	DT
prudhomme	79	14	presence	presence	NN
prudhomme	79	15	of	of	IN
prudhomme	79	16	Jessie	Jessie	NNP
prudhomme	79	17	Thatcher	Thatcher	NNP
prudhomme	79	18	Bost	Bost	NNP
prudhomme	79	19	,	,	,
prudhomme	79	20	the	the	DT
prudhomme	79	21	first	first	JJ
prudhomme	79	22	female	female	JJ
prudhomme	79	23	graduate	graduate	NN
prudhomme	79	24	from	from	IN
prudhomme	79	25	Oklahoma	Oklahoma	NNP
prudhomme	79	26	A&M	A&M	NNP
prudhomme	79	27	College	College	NNP
prudhomme	79	28	in	in	IN
prudhomme	79	29	1897	1897	CD
prudhomme	79	30	.	.	.
prudhomme	80	1	Comment	comment	NN
prudhomme	80	2	by	by	IN
prudhomme	80	3	Daniel	Daniel	NNP
prudhomme	80	4	Johnson	Johnson	NNP
prudhomme	80	5	:	:	:
prudhomme	80	6	This	this	DT
prudhomme	80	7	is	be	VBZ
prudhomme	80	8	a	a	DT
prudhomme	80	9	really	really	RB
prudhomme	80	10	good	good	JJ
prudhomme	80	11	example	example	NN
prudhomme	80	12	.	.	.
prudhomme	81	1	[	[	-LRB-
prudhomme	81	2	1	1	CD
prudhomme	81	3	:	:	:
prudhomme	81	4	See	see	VB
prudhomme	81	5	https://github.com/ageitgey/face_recognition	https://github.com/ageitgey/face_recognition	NN
prudhomme	81	6	.	.	.
prudhomme	81	7	]	]	-RRB-
prudhomme	82	1	Figure	figure	NN
prudhomme	82	2	1	1	CD
prudhomme	82	3	:	:	:
prudhomme	82	4	Face	face	NN
prudhomme	82	5	recognition	recognition	NN
prudhomme	82	6	test	test	NN
prudhomme	82	7	Ren	Ren	NNP
prudhomme	82	8	et	et	NNP
prudhomme	82	9	al	al	NNP
prudhomme	82	10	.	.	.
prudhomme	83	1	stated	state	VBN
prudhomme	83	2	that	that	IN
prudhomme	83	3	it	-PRON-	PRP
prudhomme	83	4	is	be	VBZ
prudhomme	83	5	important	important	JJ
prudhomme	83	6	to	to	TO
prudhomme	83	7	construct	construct	VB
prudhomme	83	8	a	a	DT
prudhomme	83	9	deep	deep	JJ
prudhomme	83	10	and	and	CC
prudhomme	83	11	convolutional	convolutional	JJ
prudhomme	83	12	per	per	IN
prudhomme	83	13	-	-	HYPH
prudhomme	83	14	region	region	NN
prudhomme	83	15	object	object	NN
prudhomme	83	16	classifier	classifier	NN
prudhomme	83	17	to	to	TO
prudhomme	83	18	obtain	obtain	VB
prudhomme	83	19	good	good	JJ
prudhomme	83	20	accuracy	accuracy	NN
prudhomme	83	21	using	use	VBG
prudhomme	83	22	ResNets	ResNets	NNP
prudhomme	83	23	(	(	-LRB-
prudhomme	83	24	2015	2015	CD
prudhomme	83	25	)	)	-RRB-
prudhomme	83	26	.	.	.
prudhomme	84	1	Going	go	VBG
prudhomme	84	2	forward	forward	RB
prudhomme	84	3	,	,	,
prudhomme	84	4	we	-PRON-	PRP
prudhomme	84	5	could	could	MD
prudhomme	84	6	use	use	VB
prudhomme	84	7	the	the	DT
prudhomme	84	8	tool	tool	NN
prudhomme	84	9	“	"	``
prudhomme	84	10	as	as	IN
prudhomme	84	11	is	be	VBZ
prudhomme	84	12	”	"	''
prudhomme	84	13	despite	despite	IN
prudhomme	84	14	the	the	DT
prudhomme	84	15	low	low	JJ
prudhomme	84	16	tolerance	tolerance	NN
prudhomme	84	17	for	for	IN
prudhomme	84	18	accuracy	accuracy	NN
prudhomme	84	19	,	,	,
prudhomme	84	20	or	or	CC
prudhomme	84	21	instead	instead	RB
prudhomme	84	22	try	try	VB
prudhomme	84	23	to	to	TO
prudhomme	84	24	establish	establish	VB
prudhomme	84	25	large	large	JJ
prudhomme	84	26	datasets	dataset	NNS
prudhomme	84	27	of	of	IN
prudhomme	84	28	faces	face	NNS
prudhomme	84	29	by	by	IN
prudhomme	84	30	training	train	VBG
prudhomme	84	31	on	on	IN
prudhomme	84	32	our	-PRON-	PRP$
prudhomme	84	33	own	own	JJ
prudhomme	84	34	collections	collection	NNS
prudhomme	84	35	in	in	IN
prudhomme	84	36	hopes	hope	NNS
prudhomme	84	37	of	of	IN
prudhomme	84	38	improving	improve	VBG
prudhomme	84	39	accuracy	accuracy	NN
prudhomme	84	40	.	.	.
prudhomme	85	1	We	-PRON-	PRP
prudhomme	85	2	proceeded	proceed	VBD
prudhomme	85	3	with	with	IN
prudhomme	85	4	utilizing	utilize	VBG
prudhomme	85	5	the	the	DT
prudhomme	85	6	Oklahoma	Oklahoma	NNP
prudhomme	85	7	State	State	NNP
prudhomme	85	8	University	University	NNP
prudhomme	85	9	Yearbook	Yearbook	NNP
prudhomme	85	10	collections	collection	NNS
prudhomme	85	11	,	,	,
prudhomme	85	12	comparing	compare	VBG
prudhomme	85	13	image	image	NN
prudhomme	85	14	sets	set	NNS
prudhomme	85	15	with	with	IN
prudhomme	85	16	other	other	JJ
prudhomme	85	17	photographs	photograph	NNS
prudhomme	85	18	that	that	WDT
prudhomme	85	19	may	may	MD
prudhomme	85	20	include	include	VB
prudhomme	85	21	these	these	DT
prudhomme	85	22	faces	face	NNS
prudhomme	85	23	.	.	.
prudhomme	86	1	We	-PRON-	PRP
prudhomme	86	2	look	look	VBP
prudhomme	86	3	forward	forward	RB
prudhomme	86	4	to	to	IN
prudhomme	86	5	automating	automate	VBG
prudhomme	86	6	more	more	JJR
prudhomme	86	7	of	of	IN
prudhomme	86	8	these	these	DT
prudhomme	86	9	processes	process	NNS
prudhomme	86	10	.	.	.
prudhomme	87	1	A	a	DT
prudhomme	87	2	Conclusive	conclusive	JJ
prudhomme	87	3	First	first	JJ
prudhomme	87	4	Experiment	experiment	NN
prudhomme	87	5	We	-PRON-	PRP
prudhomme	87	6	can	can	MD
prudhomme	87	7	say	say	VB
prudhomme	87	8	that	that	IN
prudhomme	87	9	our	-PRON-	PRP$
prudhomme	87	10	first	first	JJ
prudhomme	87	11	experiment	experiment	NN
prudhomme	87	12	developing	develop	VBG
prudhomme	87	13	a	a	DT
prudhomme	87	14	machine	machine	NN
prudhomme	87	15	learning	learn	VBG
prudhomme	87	16	solution	solution	NN
prudhomme	87	17	on	on	IN
prudhomme	87	18	a	a	DT
prudhomme	87	19	known	know	VBN
prudhomme	87	20	set	set	NN
prudhomme	87	21	of	of	IN
prudhomme	87	22	archival	archival	NN
prudhomme	87	23	data	datum	NNS
prudhomme	87	24	resulted	result	VBD
prudhomme	87	25	in	in	IN
prudhomme	87	26	positive	positive	JJ
prudhomme	87	27	output	output	NN
prudhomme	87	28	,	,	,
prudhomme	87	29	while	while	IN
prudhomme	87	30	recognizing	recognize	VBG
prudhomme	87	31	that	that	IN
prudhomme	87	32	it	-PRON-	PRP
prudhomme	87	33	is	be	VBZ
prudhomme	87	34	still	still	RB
prudhomme	87	35	a	a	DT
prudhomme	87	36	work	work	NN
prudhomme	87	37	in	in	IN
prudhomme	87	38	progress	progress	NN
prudhomme	87	39	.	.	.
prudhomme	88	1	For	for	IN
prudhomme	88	2	example	example	NN
prudhomme	88	3	,	,	,
prudhomme	88	4	the	the	DT
prudhomme	88	5	model	model	NN
prudhomme	88	6	we	-PRON-	PRP
prudhomme	88	7	ran	run	VBD
prudhomme	88	8	for	for	IN
prudhomme	88	9	the	the	DT
prudhomme	88	10	pilot	pilot	NN
prudhomme	88	11	is	be	VBZ
prudhomme	88	12	not	not	RB
prudhomme	88	13	natively	natively	RB
prudhomme	88	14	supported	support	VBN
prudhomme	88	15	on	on	IN
prudhomme	88	16	Windows	Windows	NNP
prudhomme	88	17	,	,	,
prudhomme	88	18	which	which	WDT
prudhomme	88	19	hindered	hinder	VBD
prudhomme	88	20	team	team	NN
prudhomme	88	21	collaboration	collaboration	NN
prudhomme	88	22	.	.	.
prudhomme	89	1	In	in	IN
prudhomme	89	2	light	light	NN
prudhomme	89	3	of	of	IN
prudhomme	89	4	these	these	DT
prudhomme	89	5	challenges	challenge	NNS
prudhomme	89	6	,	,	,
prudhomme	89	7	we	-PRON-	PRP
prudhomme	89	8	think	think	VBP
prudhomme	89	9	that	that	IN
prudhomme	89	10	our	-PRON-	PRP$
prudhomme	89	11	experiment	experiment	NN
prudhomme	89	12	was	be	VBD
prudhomme	89	13	a	a	DT
prudhomme	89	14	step	step	NN
prudhomme	89	15	in	in	IN
prudhomme	89	16	the	the	DT
prudhomme	89	17	right	right	JJ
prudhomme	89	18	direction	direction	NN
prudhomme	89	19	of	of	IN
prudhomme	89	20	adding	add	VBG
prudhomme	89	21	value	value	NN
prudhomme	89	22	to	to	IN
prudhomme	89	23	collections	collection	NNS
prudhomme	89	24	by	by	IN
prudhomme	89	25	bringing	bring	VBG
prudhomme	89	26	in	in	RP
prudhomme	89	27	a	a	DT
prudhomme	89	28	new	new	JJ
prudhomme	89	29	layer	layer	NN
prudhomme	89	30	of	of	IN
prudhomme	89	31	discovery	discovery	NN
prudhomme	89	32	for	for	IN
prudhomme	89	33	hidden	hidden	JJ
prudhomme	89	34	or	or	CC
prudhomme	89	35	unidentified	unidentified	JJ
prudhomme	89	36	content	content	NN
prudhomme	89	37	.	.	.
prudhomme	90	1	Above	above	IN
prudhomme	90	2	all	all	DT
prudhomme	90	3	,	,	,
prudhomme	90	4	this	this	DT
prudhomme	90	5	type	type	NN
prudhomme	90	6	of	of	IN
prudhomme	90	7	work	work	NN
prudhomme	90	8	relies	rely	VBZ
prudhomme	90	9	greatly	greatly	RB
prudhomme	90	10	on	on	IN
prudhomme	90	11	transparency	transparency	NN
prudhomme	90	12	.	.	.
prudhomme	91	1	As	as	IN
prudhomme	91	2	Schweikert	Schweikert	NNP
prudhomme	91	3	notes	note	VBZ
prudhomme	91	4	,	,	,
prudhomme	91	5	“	"	``
prudhomme	91	6	Transparency	transparency	NN
prudhomme	91	7	is	be	VBZ
prudhomme	91	8	not	not	RB
prudhomme	91	9	a	a	DT
prudhomme	91	10	perk	perk	NN
prudhomme	91	11	,	,	,
prudhomme	91	12	but	but	CC
prudhomme	91	13	a	a	DT
prudhomme	91	14	key	key	NN
prudhomme	91	15	to	to	IN
prudhomme	91	16	the	the	DT
prudhomme	91	17	responsible	responsible	JJ
prudhomme	91	18	adoption	adoption	NN
prudhomme	91	19	of	of	IN
prudhomme	91	20	machine	machine	NN
prudhomme	91	21	learning	learning	NN
prudhomme	91	22	solutions	solution	NNS
prudhomme	91	23	”	"	''
prudhomme	91	24	(	(	-LRB-
prudhomme	91	25	2019	2019	CD
prudhomme	91	26	,	,	,
prudhomme	91	27	72	72	CD
prudhomme	91	28	)	)	-RRB-
prudhomme	91	29	.	.	.
prudhomme	92	1	More	more	RBR
prudhomme	92	2	broadly	broadly	RB
prudhomme	92	3	,	,	,
prudhomme	92	4	issues	issue	NNS
prudhomme	92	5	in	in	IN
prudhomme	92	6	transparency	transparency	NN
prudhomme	92	7	and	and	CC
prudhomme	92	8	ethics	ethic	NNS
prudhomme	92	9	in	in	IN
prudhomme	92	10	machine	machine	NN
prudhomme	92	11	learning	learning	NN
prudhomme	92	12	are	be	VBP
prudhomme	92	13	important	important	JJ
prudhomme	92	14	concerns	concern	NNS
prudhomme	92	15	in	in	IN
prudhomme	92	16	the	the	DT
prudhomme	92	17	collecting	collecting	NN
prudhomme	92	18	and	and	CC
prudhomme	92	19	handling	handling	NN
prudhomme	92	20	of	of	IN
prudhomme	92	21	data	datum	NNS
prudhomme	92	22	.	.	.
prudhomme	93	1	In	in	IN
prudhomme	93	2	order	order	NN
prudhomme	93	3	to	to	TO
prudhomme	93	4	boost	boost	VB
prudhomme	93	5	adoption	adoption	NN
prudhomme	93	6	and	and	CC
prudhomme	93	7	get	get	VB
prudhomme	93	8	more	more	JJR
prudhomme	93	9	buy	buy	NN
prudhomme	93	10	-	-	HYPH
prudhomme	93	11	in	in	NN
prudhomme	93	12	with	with	IN
prudhomme	93	13	this	this	DT
prudhomme	93	14	new	new	JJ
prudhomme	93	15	type	type	NN
prudhomme	93	16	of	of	IN
prudhomme	93	17	discovery	discovery	NN
prudhomme	93	18	layer	layer	NN
prudhomme	93	19	,	,	,
prudhomme	93	20	our	-PRON-	PRP$
prudhomme	93	21	team	team	NN
prudhomme	93	22	shared	share	VBD
prudhomme	93	23	information	information	NN
prudhomme	93	24	intentionally	intentionally	RB
prudhomme	93	25	about	about	IN
prudhomme	93	26	the	the	DT
prudhomme	93	27	process	process	NN
prudhomme	93	28	to	to	TO
prudhomme	93	29	help	help	VB
prudhomme	93	30	add	add	VB
prudhomme	93	31	credibility	credibility	NN
prudhomme	93	32	to	to	IN
prudhomme	93	33	the	the	DT
prudhomme	93	34	work	work	NN
prudhomme	93	35	and	and	CC
prudhomme	93	36	foster	foster	VB
prudhomme	93	37	a	a	DT
prudhomme	93	38	more	more	RBR
prudhomme	93	39	collaborative	collaborative	JJ
prudhomme	93	40	environment	environment	NN
prudhomme	93	41	within	within	IN
prudhomme	93	42	the	the	DT
prudhomme	93	43	library	library	NN
prudhomme	93	44	.	.	.
prudhomme	94	1	Also	also	RB
prudhomme	94	2	,	,	,
prudhomme	94	3	the	the	DT
prudhomme	94	4	team	team	NN
prudhomme	94	5	developed	develop	VBD
prudhomme	94	6	a	a	DT
prudhomme	94	7	Graphic	Graphic	NNP
prudhomme	94	8	User	User	NNP
prudhomme	94	9	Interface	Interface	NNP
prudhomme	94	10	(	(	-LRB-
prudhomme	94	11	GUI	GUI	NNP
prudhomme	94	12	)	)	-RRB-
prudhomme	94	13	to	to	TO
prudhomme	94	14	search	search	VB
prudhomme	94	15	the	the	DT
prudhomme	94	16	inventory	inventory	NN
prudhomme	94	17	within	within	IN
prudhomme	94	18	the	the	DT
prudhomme	94	19	archives	archive	NNS
prudhomme	94	20	and	and	CC
prudhomme	94	21	ultimately	ultimately	RB
prudhomme	94	22	grow	grow	VBP
prudhomme	94	23	the	the	DT
prudhomme	94	24	solution	solution	NN
prudhomme	94	25	beyond	beyond	IN
prudhomme	94	26	the	the	DT
prudhomme	94	27	department	department	NN
prudhomme	94	28	.	.	.
prudhomme	95	1	Challenges	Challenges	NNPS
prudhomme	95	2	and	and	CC
prudhomme	95	3	Opportunities	Opportunities	NNPS
prudhomme	95	4	of	of	IN
prudhomme	95	5	Machine	Machine	NNP
prudhomme	95	6	Learning	Learning	NNP
prudhomme	95	7	Challenges	Challenges	NNPS
prudhomme	95	8	In	in	IN
prudhomme	95	9	a	a	DT
prudhomme	95	10	National	National	NNP
prudhomme	95	11	Library	Library	NNP
prudhomme	95	12	of	of	IN
prudhomme	95	13	Medicine	Medicine	NNP
prudhomme	95	14	blog	blog	NN
prudhomme	95	15	post	post	NN
prudhomme	95	16	,	,	,
prudhomme	95	17	Patti	Patti	NNP
prudhomme	95	18	Brennan	Brennan	NNP
prudhomme	95	19	points	point	VBZ
prudhomme	95	20	out	out	RP
prudhomme	95	21	“	"	``
prudhomme	95	22	that	that	IN
prudhomme	95	23	AI	AI	NNP
prudhomme	95	24	applications	application	NNS
prudhomme	95	25	are	be	VBP
prudhomme	95	26	only	only	RB
prudhomme	95	27	as	as	RB
prudhomme	95	28	good	good	JJ
prudhomme	95	29	as	as	IN
prudhomme	95	30	the	the	DT
prudhomme	95	31	data	datum	NNS
prudhomme	95	32	upon	upon	IN
prudhomme	95	33	which	which	WDT
prudhomme	95	34	they	-PRON-	PRP
prudhomme	95	35	are	be	VBP
prudhomme	95	36	trained	train	VBN
prudhomme	95	37	and	and	CC
prudhomme	95	38	built”(2019	built”(2019	NNP
prudhomme	95	39	)	)	-RRB-
prudhomme	95	40	,	,	,
prudhomme	95	41	and	and	CC
prudhomme	95	42	having	have	VBG
prudhomme	95	43	these	these	DT
prudhomme	95	44	data	datum	NNS
prudhomme	95	45	ready	ready	JJ
prudhomme	95	46	for	for	IN
prudhomme	95	47	analysis	analysis	NN
prudhomme	95	48	is	be	VBZ
prudhomme	95	49	a	a	DT
prudhomme	95	50	must	must	NN
prudhomme	95	51	in	in	IN
prudhomme	95	52	order	order	NN
prudhomme	95	53	to	to	TO
prudhomme	95	54	yield	yield	VB
prudhomme	95	55	accurate	accurate	JJ
prudhomme	95	56	results	result	NNS
prudhomme	95	57	.	.	.
prudhomme	96	1	Scaling	scale	VBG
prudhomme	96	2	of	of	IN
prudhomme	96	3	input	input	NN
prudhomme	96	4	and	and	CC
prudhomme	96	5	output	output	NN
prudhomme	96	6	variables	variable	NNS
prudhomme	96	7	also	also	RB
prudhomme	96	8	plays	play	VBZ
prudhomme	96	9	an	an	DT
prudhomme	96	10	important	important	JJ
prudhomme	96	11	role	role	NN
prudhomme	96	12	in	in	IN
prudhomme	96	13	the	the	DT
prudhomme	96	14	performance	performance	NN
prudhomme	96	15	improvement	improvement	NN
prudhomme	96	16	when	when	WRB
prudhomme	96	17	using	use	VBG
prudhomme	96	18	neural	neural	JJ
prudhomme	96	19	network	network	NN
prudhomme	96	20	models	model	NNS
prudhomme	96	21	.	.	.
prudhomme	97	1	Jerome	Jerome	NNP
prudhomme	97	2	Pesenti	Pesenti	NNP
prudhomme	97	3	,	,	,
prudhomme	97	4	Head	Head	NNP
prudhomme	97	5	of	of	IN
prudhomme	97	6	AI	AI	NNP
prudhomme	97	7	at	at	IN
prudhomme	97	8	Facebook	Facebook	NNP
prudhomme	97	9	,	,	,
prudhomme	97	10	states	state	VBZ
prudhomme	97	11	that	that	IN
prudhomme	97	12	“	"	``
prudhomme	97	13	When	when	WRB
prudhomme	97	14	you	-PRON-	PRP
prudhomme	97	15	scale	scale	VBP
prudhomme	97	16	deep	deep	JJ
prudhomme	97	17	learning	learning	NN
prudhomme	97	18	,	,	,
prudhomme	97	19	it	-PRON-	PRP
prudhomme	97	20	tends	tend	VBZ
prudhomme	97	21	to	to	TO
prudhomme	97	22	behave	behave	VB
prudhomme	97	23	better	well	JJR
prudhomme	97	24	and	and	CC
prudhomme	97	25	to	to	TO
prudhomme	97	26	be	be	VB
prudhomme	97	27	able	able	JJ
prudhomme	97	28	to	to	TO
prudhomme	97	29	solve	solve	VB
prudhomme	97	30	a	a	DT
prudhomme	97	31	broader	broad	JJR
prudhomme	97	32	task	task	NN
prudhomme	97	33	in	in	IN
prudhomme	97	34	a	a	DT
prudhomme	97	35	better	well	JJR
prudhomme	97	36	way	way	NN
prudhomme	97	37	”	"	''
prudhomme	97	38	(	(	-LRB-
prudhomme	97	39	2019	2019	CD
prudhomme	97	40	)	)	-RRB-
prudhomme	97	41	.	.	.
prudhomme	98	1	Clifford	Clifford	NNP
prudhomme	98	2	Lynch	Lynch	NNP
prudhomme	98	3	affirms	affirm	VBZ
prudhomme	98	4	,	,	,
prudhomme	98	5	“	"	``
prudhomme	98	6	machine	machine	NN
prudhomme	98	7	learning	learning	NN
prudhomme	98	8	applications	application	NNS
prudhomme	98	9	could	could	MD
prudhomme	98	10	substantially	substantially	RB
prudhomme	98	11	help	help	VB
prudhomme	98	12	archives	archive	NNS
prudhomme	98	13	make	make	VB
prudhomme	98	14	their	-PRON-	PRP$
prudhomme	98	15	collections	collection	NNS
prudhomme	98	16	more	more	RBR
prudhomme	98	17	discoverable	discoverable	JJ
prudhomme	98	18	to	to	IN
prudhomme	98	19	the	the	DT
prudhomme	98	20	public	public	NN
prudhomme	98	21	,	,	,
prudhomme	98	22	to	to	IN
prudhomme	98	23	the	the	DT
prudhomme	98	24	extent	extent	NN
prudhomme	98	25	that	that	IN
prudhomme	98	26	memory	memory	NN
prudhomme	98	27	organizations	organization	NNS
prudhomme	98	28	can	can	MD
prudhomme	98	29	develop	develop	VB
prudhomme	98	30	the	the	DT
prudhomme	98	31	skills	skill	NNS
prudhomme	98	32	and	and	CC
prudhomme	98	33	workflows	workflow	NNS
prudhomme	98	34	to	to	TO
prudhomme	98	35	apply	apply	VB
prudhomme	98	36	them	-PRON-	PRP
prudhomme	98	37	”	"	''
prudhomme	98	38	(	(	-LRB-
prudhomme	98	39	2019	2019	CD
prudhomme	98	40	)	)	-RRB-
prudhomme	98	41	.	.	.
prudhomme	99	1	This	this	DT
prudhomme	99	2	raises	raise	VBZ
prudhomme	99	3	the	the	DT
prudhomme	99	4	question	question	NN
prudhomme	99	5	whether	whether	IN
prudhomme	99	6	archives	archive	NNS
prudhomme	99	7	can	can	MD
prudhomme	99	8	also	also	RB
prudhomme	99	9	afford	afford	VB
prudhomme	99	10	to	to	TO
prudhomme	99	11	create	create	VB
prudhomme	99	12	the	the	DT
prudhomme	99	13	large	large	JJ
prudhomme	99	14	amount	amount	NN
prudhomme	99	15	of	of	IN
prudhomme	99	16	data	datum	NNS
prudhomme	99	17	from	from	IN
prudhomme	99	18	print	print	NN
prudhomme	99	19	heritage	heritage	NN
prudhomme	99	20	materials	material	NNS
prudhomme	99	21	or	or	CC
prudhomme	99	22	refine	refine	VB
prudhomme	99	23	their	-PRON-	PRP$
prudhomme	99	24	born	bear	VBN
prudhomme	99	25	-	-	HYPH
prudhomme	99	26	digital	digital	JJ
prudhomme	99	27	collections	collection	NNS
prudhomme	99	28	in	in	IN
prudhomme	99	29	order	order	NN
prudhomme	99	30	to	to	TO
prudhomme	99	31	build	build	VB
prudhomme	99	32	the	the	DT
prudhomme	99	33	capacity	capacity	NN
prudhomme	99	34	to	to	TO
prudhomme	99	35	sustain	sustain	VB
prudhomme	99	36	the	the	DT
prudhomme	99	37	use	use	NN
prudhomme	99	38	of	of	IN
prudhomme	99	39	deep	deep	JJ
prudhomme	99	40	-	-	HYPH
prudhomme	99	41	learning	learn	VBG
prudhomme	99	42	applications	application	NNS
prudhomme	99	43	.	.	.
prudhomme	100	1	Granted	grant	VBN
prudhomme	100	2	,	,	,
prudhomme	100	3	the	the	DT
prudhomme	100	4	increasing	increase	VBG
prudhomme	100	5	volume	volume	NN
prudhomme	100	6	of	of	IN
prudhomme	100	7	born	bear	VBN
prudhomme	100	8	-	-	HYPH
prudhomme	100	9	digital	digital	JJ
prudhomme	100	10	materials	material	NNS
prudhomme	100	11	could	could	MD
prudhomme	100	12	help	help	VB
prudhomme	100	13	leverage	leverage	VB
prudhomme	100	14	this	this	DT
prudhomme	100	15	data	data	NN
prudhomme	100	16	capacity	capacity	NN
prudhomme	100	17	somehow	somehow	RB
prudhomme	100	18	;	;	:
prudhomme	100	19	it	-PRON-	PRP
prudhomme	100	20	does	do	VBZ
prudhomme	100	21	not	not	RB
prudhomme	100	22	exclude	exclude	VB
prudhomme	100	23	the	the	DT
prudhomme	100	24	fact	fact	NN
prudhomme	100	25	that	that	IN
prudhomme	100	26	all	all	DT
prudhomme	100	27	data	datum	NNS
prudhomme	100	28	will	will	MD
prudhomme	100	29	need	need	VB
prudhomme	100	30	to	to	TO
prudhomme	100	31	be	be	VB
prudhomme	100	32	ready	ready	JJ
prudhomme	100	33	prior	prior	RB
prudhomme	100	34	to	to	IN
prudhomme	100	35	using	use	VBG
prudhomme	100	36	deep	deep	JJ
prudhomme	100	37	learning	learning	NN
prudhomme	100	38	.	.	.
prudhomme	101	1	Since	since	IN
prudhomme	101	2	machine	machine	NN
prudhomme	101	3	learning	learning	NN
prudhomme	101	4	is	be	VBZ
prudhomme	101	5	only	only	RB
prudhomme	101	6	good	good	JJ
prudhomme	101	7	so	so	RB
prudhomme	101	8	long	long	RB
prudhomme	101	9	as	as	IN
prudhomme	101	10	value	value	NN
prudhomme	101	11	is	be	VBZ
prudhomme	101	12	added	add	VBN
prudhomme	101	13	,	,	,
prudhomme	101	14	archives	archive	NNS
prudhomme	101	15	and	and	CC
prudhomme	101	16	libraries	library	NNS
prudhomme	101	17	will	will	MD
prudhomme	101	18	need	need	VB
prudhomme	101	19	to	to	TO
prudhomme	101	20	think	think	VB
prudhomme	101	21	in	in	IN
prudhomme	101	22	terms	term	NNS
prudhomme	101	23	of	of	IN
prudhomme	101	24	optimization	optimization	NN
prudhomme	101	25	as	as	RB
prudhomme	101	26	well	well	RB
prudhomme	101	27	,	,	,
prudhomme	101	28	deciding	decide	VBG
prudhomme	101	29	when	when	WRB
prudhomme	101	30	value	value	NN
prudhomme	101	31	-	-	HYPH
prudhomme	101	32	generated	generate	VBN
prudhomme	101	33	output	output	NN
prudhomme	101	34	is	be	VBZ
prudhomme	101	35	justified	justify	VBN
prudhomme	101	36	compared	compare	VBN
prudhomme	101	37	to	to	IN
prudhomme	101	38	the	the	DT
prudhomme	101	39	cost	cost	NN
prudhomme	101	40	of	of	IN
prudhomme	101	41	computing	compute	VBG
prudhomme	101	42	infrastructure	infrastructure	NN
prudhomme	101	43	and	and	CC
prudhomme	101	44	skilled	skilled	JJ
prudhomme	101	45	labor	labor	NN
prudhomme	101	46	needs	need	NNS
prudhomme	101	47	.	.	.
prudhomme	102	1	Besides	besides	IN
prudhomme	102	2	value	value	NN
prudhomme	102	3	,	,	,
prudhomme	102	4	operations	operation	NNS
prudhomme	102	5	,	,	,
prudhomme	102	6	such	such	JJ
prudhomme	102	7	as	as	IN
prudhomme	102	8	storing	store	VBG
prudhomme	102	9	and	and	CC
prudhomme	102	10	ensuring	ensure	VBG
prudhomme	102	11	access	access	NN
prudhomme	102	12	to	to	IN
prudhomme	102	13	these	these	DT
prudhomme	102	14	data	datum	NNS
prudhomme	102	15	,	,	,
prudhomme	102	16	are	be	VBP
prudhomme	102	17	just	just	RB
prudhomme	102	18	as	as	RB
prudhomme	102	19	important	important	JJ
prudhomme	102	20	considerations	consideration	NNS
prudhomme	102	21	to	to	IN
prudhomme	102	22	making	make	VBG
prudhomme	102	23	machine	machine	NN
prudhomme	102	24	learning	learn	VBG
prudhomme	102	25	a	a	DT
prudhomme	102	26	feasible	feasible	JJ
prudhomme	102	27	endeavor	endeavor	NN
prudhomme	102	28	.	.	.
prudhomme	103	1	Comment	comment	NN
prudhomme	103	2	by	by	IN
prudhomme	103	3	Mark	Mark	NNP
prudhomme	103	4	Dehmlow	Dehmlow	NNP
prudhomme	103	5	:	:	:
prudhomme	103	6	Maybe	maybe	RB
prudhomme	103	7	use	use	VBP
prudhomme	103	8	active	active	JJ
prudhomme	103	9	voice	voice	NN
prudhomme	103	10	?	?	.
prudhomme	104	1	Comment	comment	NN
prudhomme	104	2	by	by	IN
prudhomme	104	3	Daniel	Daniel	NNP
prudhomme	104	4	Johnson	Johnson	NNP
prudhomme	104	5	:	:	:
prudhomme	104	6	I	-PRON-	PRP
prudhomme	104	7	'm	be	VBP
prudhomme	104	8	ok	ok	JJ
prudhomme	104	9	with	with	IN
prudhomme	104	10	this	this	DT
prudhomme	104	11	construction	construction	NN
prudhomme	104	12	,	,	,
prudhomme	104	13	in	in	IN
prudhomme	104	14	this	this	DT
prudhomme	104	15	case	case	NN
prudhomme	104	16	,	,	,
prudhomme	104	17	if	if	IN
prudhomme	104	18	you	-PRON-	PRP
prudhomme	104	19	wish	wish	VBP
prudhomme	104	20	to	to	TO
prudhomme	104	21	leave	leave	VB
prudhomme	104	22	as	as	IN
prudhomme	104	23	is	be	VBZ
prudhomme	104	24	.	.	.
prudhomme	105	1	Opportunities	Opportunities	NNPS
prudhomme	105	2	Investment	Investment	NNP
prudhomme	105	3	in	in	IN
prudhomme	105	4	resources	resource	NNS
prudhomme	105	5	is	be	VBZ
prudhomme	105	6	also	also	RB
prudhomme	105	7	needed	need	VBN
prudhomme	105	8	for	for	IN
prudhomme	105	9	interpreting	interpret	VBG
prudhomme	105	10	results	result	NNS
prudhomme	105	11	,	,	,
prudhomme	105	12	in	in	IN
prudhomme	105	13	that	that	DT
prudhomme	105	14	“	"	``
prudhomme	105	15	results	result	NNS
prudhomme	105	16	of	of	IN
prudhomme	105	17	an	an	DT
prudhomme	105	18	AI	ai	NN
prudhomme	105	19	-	-	HYPH
prudhomme	105	20	powered	power	VBN
prudhomme	105	21	analysis	analysis	NN
prudhomme	105	22	should	should	MD
prudhomme	105	23	only	only	RB
prudhomme	105	24	factor	factor	VB
prudhomme	105	25	into	into	IN
prudhomme	105	26	the	the	DT
prudhomme	105	27	final	final	JJ
prudhomme	105	28	decision	decision	NN
prudhomme	105	29	;	;	:
prudhomme	105	30	they	-PRON-	PRP
prudhomme	105	31	should	should	MD
prudhomme	105	32	not	not	RB
prudhomme	105	33	be	be	VB
prudhomme	105	34	the	the	DT
prudhomme	105	35	final	final	JJ
prudhomme	105	36	arbiter	arbiter	NN
prudhomme	105	37	of	of	IN
prudhomme	105	38	that	that	DT
prudhomme	105	39	decision	decision	NN
prudhomme	105	40	”	"	''
prudhomme	105	41	(	(	-LRB-
prudhomme	105	42	Brennan	Brennan	NNP
prudhomme	105	43	2019	2019	CD
prudhomme	105	44	)	)	-RRB-
prudhomme	105	45	.	.	.
prudhomme	106	1	While	while	IN
prudhomme	106	2	this	this	DT
prudhomme	106	3	could	could	MD
prudhomme	106	4	be	be	VB
prudhomme	106	5	a	a	DT
prudhomme	106	6	challenge	challenge	NN
prudhomme	106	7	in	in	IN
prudhomme	106	8	itself	-PRON-	PRP
prudhomme	106	9	,	,	,
prudhomme	106	10	it	-PRON-	PRP
prudhomme	106	11	can	can	MD
prudhomme	106	12	also	also	RB
prudhomme	106	13	be	be	VB
prudhomme	106	14	an	an	DT
prudhomme	106	15	opportunity	opportunity	NN
prudhomme	106	16	when	when	WRB
prudhomme	106	17	machine	machine	NN
prudhomme	106	18	learning	learning	NN
prudhomme	106	19	helps	help	VBZ
prudhomme	106	20	minimize	minimize	VB
prudhomme	106	21	institutional	institutional	JJ
prudhomme	106	22	memory	memory	NN
prudhomme	106	23	loss	loss	NN
prudhomme	106	24	in	in	IN
prudhomme	106	25	archives	archive	NNS
prudhomme	106	26	and	and	CC
prudhomme	106	27	libraries	library	NNS
prudhomme	106	28	(	(	-LRB-
prudhomme	106	29	e.g.	e.g.	RB
prudhomme	106	30	,	,	,
prudhomme	106	31	when	when	WRB
prudhomme	106	32	long	long	JJ
prudhomme	106	33	-	-	HYPH
prudhomme	106	34	time	time	NN
prudhomme	106	35	archivists	archivist	NNS
prudhomme	106	36	and	and	CC
prudhomme	106	37	librarians	librarian	NNS
prudhomme	106	38	leave	leave	VBP
prudhomme	106	39	the	the	DT
prudhomme	106	40	institution	institution	NN
prudhomme	106	41	)	)	-RRB-
prudhomme	106	42	.	.	.
prudhomme	107	1	Machine	machine	NN
prudhomme	107	2	learning	learning	NN
prudhomme	107	3	could	could	MD
prudhomme	107	4	supplement	supplement	VB
prudhomme	107	5	practices	practice	NNS
prudhomme	107	6	that	that	WDT
prudhomme	107	7	are	be	VBP
prudhomme	107	8	already	already	RB
prudhomme	107	9	in	in	IN
prudhomme	107	10	place	place	NN
prudhomme	107	11	–	–	:
prudhomme	107	12	it	-PRON-	PRP
prudhomme	107	13	may	may	MD
prudhomme	107	14	not	not	RB
prudhomme	107	15	necessarily	necessarily	RB
prudhomme	107	16	replace	replace	VB
prudhomme	107	17	people	people	NNS
prudhomme	107	18	–	–	:
prudhomme	107	19	and	and	CC
prudhomme	107	20	at	at	IN
prudhomme	107	21	the	the	DT
prudhomme	107	22	same	same	JJ
prudhomme	107	23	time	time	NN
prudhomme	107	24	generate	generate	VB
prudhomme	107	25	metadata	metadata	NN
prudhomme	107	26	for	for	IN
prudhomme	107	27	the	the	DT
prudhomme	107	28	access	access	NN
prudhomme	107	29	and	and	CC
prudhomme	107	30	discovery	discovery	NN
prudhomme	107	31	of	of	IN
prudhomme	107	32	collections	collection	NNS
prudhomme	107	33	that	that	WDT
prudhomme	107	34	people	people	NNS
prudhomme	107	35	may	may	MD
prudhomme	107	36	never	never	RB
prudhomme	107	37	have	have	VB
prudhomme	107	38	the	the	DT
prudhomme	107	39	time	time	NN
prudhomme	107	40	to	to	TO
prudhomme	107	41	get	get	VB
prudhomme	107	42	to	to	IN
prudhomme	107	43	otherwise	otherwise	RB
prudhomme	107	44	.	.	.
prudhomme	108	1	But	but	CC
prudhomme	108	2	we	-PRON-	PRP
prudhomme	108	3	will	will	MD
prudhomme	108	4	still	still	RB
prudhomme	108	5	need	need	VB
prudhomme	108	6	to	to	TO
prudhomme	108	7	determine	determine	VB
prudhomme	108	8	accuracy	accuracy	NN
prudhomme	108	9	in	in	IN
prudhomme	108	10	results	result	NNS
prudhomme	108	11	.	.	.
prudhomme	109	1	As	as	IN
prudhomme	109	2	deep	deep	JJ
prudhomme	109	3	learning	learning	NN
prudhomme	109	4	applications	application	NNS
prudhomme	109	5	will	will	MD
prudhomme	109	6	only	only	RB
prudhomme	109	7	be	be	VB
prudhomme	109	8	as	as	RB
prudhomme	109	9	effective	effective	JJ
prudhomme	109	10	as	as	IN
prudhomme	109	11	the	the	DT
prudhomme	109	12	data	datum	NNS
prudhomme	109	13	,	,	,
prudhomme	109	14	archives	archive	NNS
prudhomme	109	15	and	and	CC
prudhomme	109	16	libraries	library	NNS
prudhomme	109	17	should	should	MD
prudhomme	109	18	expand	expand	VB
prudhomme	109	19	their	-PRON-	PRP$
prudhomme	109	20	capacity	capacity	NN
prudhomme	109	21	by	by	IN
prudhomme	109	22	working	work	VBG
prudhomme	109	23	with	with	IN
prudhomme	109	24	academic	academic	JJ
prudhomme	109	25	departments	department	NNS
prudhomme	109	26	and	and	CC
prudhomme	109	27	partnering	partner	VBG
prudhomme	109	28	with	with	IN
prudhomme	109	29	university	university	NN
prudhomme	109	30	supercomputing	supercompute	VBG
prudhomme	109	31	centers	center	NNS
prudhomme	109	32	or	or	CC
prudhomme	109	33	other	other	JJ
prudhomme	109	34	highly	highly	RB
prudhomme	109	35	performant	performant	JJ
prudhomme	109	36	computing	computing	NN
prudhomme	109	37	environments	environment	NNS
prudhomme	109	38	across	across	IN
prudhomme	109	39	consortium	consortium	NN
prudhomme	109	40	aggregating	aggregate	VBG
prudhomme	109	41	networks	network	NNS
prudhomme	109	42	.	.	.
prudhomme	110	1	Such	such	JJ
prudhomme	110	2	networks	network	NNS
prudhomme	110	3	provide	provide	VBP
prudhomme	110	4	a	a	DT
prudhomme	110	5	computing	computing	NN
prudhomme	110	6	environment	environment	NN
prudhomme	110	7	with	with	IN
prudhomme	110	8	greater	great	JJR
prudhomme	110	9	data	data	NN
prudhomme	110	10	capacity	capacity	NN
prudhomme	110	11	and	and	CC
prudhomme	110	12	more	more	JJR
prudhomme	110	13	GPUs	gpu	NNS
prudhomme	110	14	.	.	.
prudhomme	111	1	Along	along	IN
prudhomme	111	2	similar	similar	JJ
prudhomme	111	3	lines	line	NNS
prudhomme	111	4	,	,	,
prudhomme	111	5	there	there	EX
prudhomme	111	6	are	be	VBP
prudhomme	111	7	opportunities	opportunity	NNS
prudhomme	111	8	to	to	TO
prudhomme	111	9	build	build	VB
prudhomme	111	10	upon	upon	IN
prudhomme	111	11	Carpentries	Carpentries	NNPS
prudhomme	111	12	workshops	workshop	NNS
prudhomme	111	13	and	and	CC
prudhomme	111	14	the	the	DT
prudhomme	111	15	communities	community	NNS
prudhomme	111	16	of	of	IN
prudhomme	111	17	practice	practice	NN
prudhomme	111	18	that	that	WDT
prudhomme	111	19	surround	surround	VBP
prudhomme	111	20	this	this	DT
prudhomme	111	21	type	type	NN
prudhomme	111	22	of	of	IN
prudhomme	111	23	interest	interest	NN
prudhomme	111	24	.	.	.
prudhomme	112	1	These	these	DT
prudhomme	112	2	growing	grow	VBG
prudhomme	112	3	opportunities	opportunity	NNS
prudhomme	112	4	could	could	MD
prudhomme	112	5	help	help	VB
prudhomme	112	6	boost	boost	VB
prudhomme	112	7	the	the	DT
prudhomme	112	8	use	use	NN
prudhomme	112	9	of	of	IN
prudhomme	112	10	machine	machine	NN
prudhomme	112	11	learning	learning	NN
prudhomme	112	12	and	and	CC
prudhomme	112	13	deep	deep	JJ
prudhomme	112	14	learning	learning	NN
prudhomme	112	15	applications	application	NNS
prudhomme	112	16	to	to	TO
prudhomme	112	17	minimize	minimize	VB
prudhomme	112	18	our	-PRON-	PRP$
prudhomme	112	19	knowledge	knowledge	NN
prudhomme	112	20	gaps	gap	NNS
prudhomme	112	21	about	about	IN
prudhomme	112	22	local	local	JJ
prudhomme	112	23	history	history	NN
prudhomme	112	24	and	and	CC
prudhomme	112	25	the	the	DT
prudhomme	112	26	surrounding	surround	VBG
prudhomme	112	27	community	community	NN
prudhomme	112	28	,	,	,
prudhomme	112	29	bringing	bring	VBG
prudhomme	112	30	together	together	RB
prudhomme	112	31	different	different	JJ
prudhomme	112	32	types	type	NNS
prudhomme	112	33	of	of	IN
prudhomme	112	34	data	datum	NNS
prudhomme	112	35	scattered	scatter	VBN
prudhomme	112	36	across	across	IN
prudhomme	112	37	organizations	organization	NNS
prudhomme	112	38	.	.	.
prudhomme	113	1	This	this	DT
prudhomme	113	2	increased	increase	VBN
prudhomme	113	3	capacity	capacity	NN
prudhomme	113	4	for	for	IN
prudhomme	113	5	knowledge	knowledge	NN
prudhomme	113	6	could	could	MD
prudhomme	113	7	grow	grow	VB
prudhomme	113	8	through	through	IN
prudhomme	113	9	collaborative	collaborative	JJ
prudhomme	113	10	partnerships	partnership	NNS
prudhomme	113	11	,	,	,
prudhomme	113	12	connecting	connect	VBG
prudhomme	113	13	people	people	NNS
prudhomme	113	14	,	,	,
prudhomme	113	15	scholars	scholar	NNS
prudhomme	113	16	,	,	,
prudhomme	113	17	computer	computer	NN
prudhomme	113	18	scientists	scientist	NNS
prudhomme	113	19	,	,	,
prudhomme	113	20	archivists	archivist	NNS
prudhomme	113	21	and	and	CC
prudhomme	113	22	librarians	librarian	NNS
prudhomme	113	23	,	,	,
prudhomme	113	24	to	to	TO
prudhomme	113	25	share	share	VB
prudhomme	113	26	their	-PRON-	PRP$
prudhomme	113	27	expertise	expertise	NN
prudhomme	113	28	through	through	IN
prudhomme	113	29	different	different	JJ
prudhomme	113	30	types	type	NNS
prudhomme	113	31	of	of	IN
prudhomme	113	32	projects	project	NNS
prudhomme	113	33	.	.	.
prudhomme	114	1	Such	such	JJ
prudhomme	114	2	projects	project	NNS
prudhomme	114	3	could	could	MD
prudhomme	114	4	emphasize	emphasize	VB
prudhomme	114	5	the	the	DT
prudhomme	114	6	multi-	multi-	JJ
prudhomme	114	7	and	and	CC
prudhomme	114	8	interdisciplinary	interdisciplinary	JJ
prudhomme	114	9	academic	academic	JJ
prudhomme	114	10	approach	approach	NN
prudhomme	114	11	to	to	IN
prudhomme	114	12	research	research	NN
prudhomme	114	13	,	,	,
prudhomme	114	14	including	include	VBG
prudhomme	114	15	digital	digital	JJ
prudhomme	114	16	humanities	humanity	NNS
prudhomme	114	17	and	and	CC
prudhomme	114	18	other	other	JJ
prudhomme	114	19	forms	form	NNS
prudhomme	114	20	or	or	CC
prudhomme	114	21	models	model	NNS
prudhomme	114	22	of	of	IN
prudhomme	114	23	digital	digital	JJ
prudhomme	114	24	scholarship	scholarship	NN
prudhomme	114	25	.	.	.
prudhomme	115	1	Conclusion	conclusion	NN
prudhomme	115	2	Along	along	IN
prudhomme	115	3	with	with	IN
prudhomme	115	4	greater	great	JJR
prudhomme	115	5	computing	computing	NN
prudhomme	115	6	capabilities	capability	NNS
prudhomme	115	7	,	,	,
prudhomme	115	8	artificial	artificial	JJ
prudhomme	115	9	intelligence	intelligence	NN
prudhomme	115	10	could	could	MD
prudhomme	115	11	be	be	VB
prudhomme	115	12	an	an	DT
prudhomme	115	13	opportunity	opportunity	NN
prudhomme	115	14	for	for	IN
prudhomme	115	15	libraries	library	NNS
prudhomme	115	16	and	and	CC
prudhomme	115	17	archives	archive	NNS
prudhomme	115	18	to	to	TO
prudhomme	115	19	boost	boost	VB
prudhomme	115	20	the	the	DT
prudhomme	115	21	discovery	discovery	NN
prudhomme	115	22	of	of	IN
prudhomme	115	23	their	-PRON-	PRP$
prudhomme	115	24	digital	digital	JJ
prudhomme	115	25	collections	collection	NNS
prudhomme	115	26	by	by	IN
prudhomme	115	27	pushing	push	VBG
prudhomme	115	28	text	text	NN
prudhomme	115	29	and	and	CC
prudhomme	115	30	image	image	NN
prudhomme	115	31	recognition	recognition	NN
prudhomme	115	32	machine	machine	NN
prudhomme	115	33	learning	learning	NN
prudhomme	115	34	techniques	technique	NNS
prudhomme	115	35	to	to	IN
prudhomme	115	36	new	new	JJ
prudhomme	115	37	limits	limit	NNS
prudhomme	115	38	.	.	.
prudhomme	116	1	Machine	machine	NN
prudhomme	116	2	learning	learning	NN
prudhomme	116	3	applications	application	NNS
prudhomme	116	4	could	could	MD
prudhomme	116	5	help	help	VB
prudhomme	116	6	increase	increase	VB
prudhomme	116	7	our	-PRON-	PRP$
prudhomme	116	8	knowledge	knowledge	NN
prudhomme	116	9	of	of	IN
prudhomme	116	10	texts	text	NNS
prudhomme	116	11	,	,	,
prudhomme	116	12	photographs	photograph	NNS
prudhomme	116	13	,	,	,
prudhomme	116	14	and	and	CC
prudhomme	116	15	more	more	RBR
prudhomme	116	16	,	,	,
prudhomme	116	17	and	and	CC
prudhomme	116	18	determine	determine	VB
prudhomme	116	19	their	-PRON-	PRP$
prudhomme	116	20	relevance	relevance	NN
prudhomme	116	21	within	within	IN
prudhomme	116	22	the	the	DT
prudhomme	116	23	context	context	NN
prudhomme	116	24	of	of	IN
prudhomme	116	25	research	research	NN
prudhomme	116	26	and	and	CC
prudhomme	116	27	education	education	NN
prudhomme	116	28	.	.	.
prudhomme	117	1	It	-PRON-	PRP
prudhomme	117	2	could	could	MD
prudhomme	117	3	minimize	minimize	VB
prudhomme	117	4	institutional	institutional	JJ
prudhomme	117	5	memory	memory	NN
prudhomme	117	6	loss	loss	NN
prudhomme	117	7	,	,	,
prudhomme	117	8	especially	especially	RB
prudhomme	117	9	as	as	IN
prudhomme	117	10	long	long	JJ
prudhomme	117	11	-	-	HYPH
prudhomme	117	12	time	time	NN
prudhomme	117	13	professionals	professional	NNS
prudhomme	117	14	are	be	VBP
prudhomme	117	15	leaving	leave	VBG
prudhomme	117	16	the	the	DT
prudhomme	117	17	profession	profession	NN
prudhomme	117	18	.	.	.
prudhomme	118	1	However	however	RB
prudhomme	118	2	,	,	,
prudhomme	118	3	these	these	DT
prudhomme	118	4	applications	application	NNS
prudhomme	118	5	will	will	MD
prudhomme	118	6	only	only	RB
prudhomme	118	7	be	be	VB
prudhomme	118	8	as	as	RB
prudhomme	118	9	effective	effective	JJ
prudhomme	118	10	as	as	IN
prudhomme	118	11	the	the	DT
prudhomme	118	12	data	datum	NNS
prudhomme	118	13	are	be	VBP
prudhomme	118	14	good	good	JJ
prudhomme	118	15	for	for	IN
prudhomme	118	16	training	training	NN
prudhomme	118	17	and	and	CC
prudhomme	118	18	for	for	IN
prudhomme	118	19	the	the	DT
prudhomme	118	20	added	add	VBN
prudhomme	118	21	value	value	NN
prudhomme	118	22	they	-PRON-	PRP
prudhomme	118	23	generate	generate	VBP
prudhomme	118	24	.	.	.
prudhomme	119	1	At	at	IN
prudhomme	119	2	Oklahoma	Oklahoma	NNP
prudhomme	119	3	State	State	NNP
prudhomme	119	4	University	University	NNP
prudhomme	119	5	,	,	,
prudhomme	119	6	we	-PRON-	PRP
prudhomme	119	7	took	take	VBD
prudhomme	119	8	a	a	DT
prudhomme	119	9	leap	leap	NN
prudhomme	119	10	forward	forward	RB
prudhomme	119	11	developing	develop	VBG
prudhomme	119	12	a	a	DT
prudhomme	119	13	machine	machine	NN
prudhomme	119	14	learning	learn	VBG
prudhomme	119	15	solution	solution	NN
prudhomme	119	16	to	to	TO
prudhomme	119	17	facilitate	facilitate	VB
prudhomme	119	18	metadata	metadata	NN
prudhomme	119	19	creation	creation	NN
prudhomme	119	20	in	in	IN
prudhomme	119	21	support	support	NN
prudhomme	119	22	of	of	IN
prudhomme	119	23	curation	curation	NN
prudhomme	119	24	,	,	,
prudhomme	119	25	preservation	preservation	NN
prudhomme	119	26	,	,	,
prudhomme	119	27	and	and	CC
prudhomme	119	28	discovery	discovery	NN
prudhomme	119	29	.	.	.
prudhomme	120	1	Our	-PRON-	PRP$
prudhomme	120	2	experiment	experiment	NN
prudhomme	120	3	with	with	IN
prudhomme	120	4	text	text	NN
prudhomme	120	5	extraction	extraction	NN
prudhomme	120	6	and	and	CC
prudhomme	120	7	face	face	VB
prudhomme	120	8	recognition	recognition	NN
prudhomme	120	9	models	model	NNS
prudhomme	120	10	generated	generate	VBD
prudhomme	120	11	conclusive	conclusive	JJ
prudhomme	120	12	results	result	NNS
prudhomme	120	13	within	within	IN
prudhomme	120	14	one	one	CD
prudhomme	120	15	academic	academic	JJ
prudhomme	120	16	year	year	NN
prudhomme	120	17	with	with	IN
prudhomme	120	18	two	two	CD
prudhomme	120	19	student	student	NN
prudhomme	120	20	interns	intern	NNS
prudhomme	120	21	.	.	.
prudhomme	121	1	The	the	DT
prudhomme	121	2	team	team	NN
prudhomme	121	3	was	be	VBD
prudhomme	121	4	satisfied	satisfied	JJ
prudhomme	121	5	with	with	IN
prudhomme	121	6	the	the	DT
prudhomme	121	7	final	final	JJ
prudhomme	121	8	output	output	NN
prudhomme	121	9	and	and	CC
prudhomme	121	10	so	so	RB
prudhomme	121	11	was	be	VBD
prudhomme	121	12	the	the	DT
prudhomme	121	13	library	library	NN
prudhomme	121	14	as	as	IN
prudhomme	121	15	we	-PRON-	PRP
prudhomme	121	16	reported	report	VBD
prudhomme	121	17	on	on	IN
prudhomme	121	18	our	-PRON-	PRP$
prudhomme	121	19	work	work	NN
prudhomme	121	20	.	.	.
prudhomme	122	1	Again	again	RB
prudhomme	122	2	,	,	,
prudhomme	122	3	it	-PRON-	PRP
prudhomme	122	4	is	be	VBZ
prudhomme	122	5	still	still	RB
prudhomme	122	6	a	a	DT
prudhomme	122	7	work	work	NN
prudhomme	122	8	in	in	IN
prudhomme	122	9	progress	progress	NN
prudhomme	122	10	and	and	CC
prudhomme	122	11	we	-PRON-	PRP
prudhomme	122	12	look	look	VBP
prudhomme	122	13	forward	forward	RB
prudhomme	122	14	to	to	IN
prudhomme	122	15	taking	take	VBG
prudhomme	122	16	another	another	DT
prudhomme	122	17	leap	leap	NN
prudhomme	122	18	forward	forward	RB
prudhomme	122	19	.	.	.
prudhomme	123	1	In	in	IN
prudhomme	123	2	sum	sum	NN
prudhomme	123	3	,	,	,
prudhomme	123	4	it	-PRON-	PRP
prudhomme	123	5	will	will	MD
prudhomme	123	6	be	be	VB
prudhomme	123	7	organizations	organization	NNS
prudhomme	123	8	’	’	POS
prudhomme	123	9	responsibility	responsibility	NN
prudhomme	123	10	to	to	TO
prudhomme	123	11	build	build	VB
prudhomme	123	12	their	-PRON-	PRP$
prudhomme	123	13	data	datum	NNS
prudhomme	123	14	capacity	capacity	NN
prudhomme	123	15	to	to	TO
prudhomme	123	16	sustain	sustain	VB
prudhomme	123	17	deep	deep	JJ
prudhomme	123	18	learning	learning	NN
prudhomme	123	19	applications	application	NNS
prudhomme	123	20	and	and	CC
prudhomme	123	21	justify	justify	VB
prudhomme	123	22	their	-PRON-	PRP$
prudhomme	123	23	commitment	commitment	NN
prudhomme	123	24	of	of	IN
prudhomme	123	25	resources	resource	NNS
prudhomme	123	26	.	.	.
prudhomme	124	1	Nonetheless	nonetheless	RB
prudhomme	124	2	,	,	,
prudhomme	124	3	as	as	IN
prudhomme	124	4	Oklahoma	Oklahoma	NNP
prudhomme	124	5	State	State	NNP
prudhomme	124	6	University	University	NNP
prudhomme	124	7	’s	’s	POS
prudhomme	124	8	face	face	NN
prudhomme	124	9	recognition	recognition	NN
prudhomme	124	10	initiative	initiative	NN
prudhomme	124	11	suggests	suggest	VBZ
prudhomme	124	12	,	,	,
prudhomme	124	13	these	these	DT
prudhomme	124	14	applications	application	NNS
prudhomme	124	15	can	can	MD
prudhomme	124	16	augment	augment	VB
prudhomme	124	17	archives	archive	NNS
prudhomme	124	18	’	'	''
prudhomme	124	19	and	and	CC
prudhomme	124	20	libraries	library	NNS
prudhomme	124	21	’	'	''
prudhomme	124	22	support	support	NN
prudhomme	124	23	for	for	IN
prudhomme	124	24	multi	multi	JJ
prudhomme	124	25	and	and	CC
prudhomme	124	26	interdisciplinary	interdisciplinary	JJ
prudhomme	124	27	research	research	NN
prudhomme	124	28	and	and	CC
prudhomme	124	29	scholarship	scholarship	NN
prudhomme	124	30	.	.	.
prudhomme	125	1	References	reference	NNS
prudhomme	125	2	Brennan	Brennan	NNP
prudhomme	125	3	,	,	,
prudhomme	125	4	Patti	Patti	NNP
prudhomme	125	5	.	.	.
prudhomme	126	1	2019	2019	CD
prudhomme	126	2	.	.	.
prudhomme	127	1	“	"	``
prudhomme	127	2	AI	AI	NNP
prudhomme	127	3	is	be	VBZ
prudhomme	127	4	Coming	come	VBG
prudhomme	127	5	.	.	.
prudhomme	128	1	Are	be	VBP
prudhomme	128	2	Data	datum	NNS
prudhomme	128	3	Ready	ready	JJ
prudhomme	128	4	?	?	.
prudhomme	128	5	”	"	''
prudhomme	128	6	NLM	NLM	NNP
prudhomme	128	7	Musings	Musings	NNPS
prudhomme	128	8	from	from	IN
prudhomme	128	9	the	the	DT
prudhomme	128	10	Mezzanine	Mezzanine	NNP
prudhomme	128	11	(	(	-LRB-
prudhomme	128	12	blog	blog	NN
prudhomme	128	13	)	)	-RRB-
prudhomme	128	14	.	.	.
prudhomme	129	1	March	March	NNP
prudhomme	129	2	26	26	CD
prudhomme	129	3	,	,	,
prudhomme	129	4	2019	2019	CD
prudhomme	129	5	.	.	.
prudhomme	129	6	https://nlmdirector.nlm.nih.gov/2019/03/26/ai-is-coming-are-the-data-ready/.	https://nlmdirector.nlm.nih.gov/2019/03/26/ai-is-coming-are-the-data-ready/.	ADD
prudhomme	130	1	Carmel	Carmel	NNP
prudhomme	130	2	,	,	,
prudhomme	130	3	Kent	Kent	NNP
prudhomme	130	4	.	.	.
prudhomme	131	1	2019	2019	CD
prudhomme	131	2	.	.	.
prudhomme	132	1	“	"	``
prudhomme	132	2	Evidence	Evidence	NNP
prudhomme	132	3	Summary	Summary	NNP
prudhomme	132	4	:	:	:
prudhomme	132	5	Artificial	Artificial	NNP
prudhomme	132	6	Intelligence	Intelligence	NNP
prudhomme	132	7	in	in	IN
prudhomme	132	8	Education	Education	NNP
prudhomme	132	9	.	.	.
prudhomme	132	10	”	"	''
prudhomme	132	11	European	European	NNP
prudhomme	132	12	EdTech	EdTech	NNP
prudhomme	132	13	Network	Network	NNP
prudhomme	132	14	.	.	.
prudhomme	133	1	https://eetn.eu/knowledge/detail/Evidence-Summary--Artificial-Intelligence-in-education	https://eetn.eu/knowledge/detail/evidence-summary--artificial-intelligence-in-education	ADD
prudhomme	133	2	.	.	.
prudhomme	134	1	Coleman	Coleman	NNP
prudhomme	134	2	,	,	,
prudhomme	134	3	Catherine	Catherine	NNP
prudhomme	134	4	Nicole	Nicole	NNP
prudhomme	134	5	.	.	.
prudhomme	135	1	2017	2017	CD
prudhomme	135	2	.	.	.
prudhomme	136	1	“	"	``
prudhomme	136	2	Artificial	Artificial	NNP
prudhomme	136	3	Intelligence	Intelligence	NNP
prudhomme	136	4	and	and	CC
prudhomme	136	5	the	the	DT
prudhomme	136	6	Library	Library	NNP
prudhomme	136	7	of	of	IN
prudhomme	136	8	the	the	DT
prudhomme	136	9	Future	Future	NNP
prudhomme	136	10	,	,	,
prudhomme	136	11	Revisited	Revisited	NNP
prudhomme	136	12	.	.	.
prudhomme	136	13	”	"	''
prudhomme	136	14	Stanford	Stanford	NNP
prudhomme	136	15	Libraries	Libraries	NNP
prudhomme	136	16	(	(	-LRB-
prudhomme	136	17	blog	blog	NN
prudhomme	136	18	)	)	-RRB-
prudhomme	136	19	.	.	.
prudhomme	137	1	November	November	NNP
prudhomme	137	2	3	3	CD
prudhomme	137	3	,	,	,
prudhomme	137	4	2017	2017	CD
prudhomme	137	5	.	.	.
prudhomme	137	6	https://library.stanford.edu/blogs/digital-library-blog/2017/11/artificial-intelligence-and-library-future-revisited	https://library.stanford.edu/blogs/digital-library-blog/2017/11/artificial-intelligence-and-library-future-revisite	VBN
prudhomme	137	7	.	.	.
prudhomme	138	1	“	"	``
prudhomme	138	2	Face	Face	NNP
prudhomme	138	3	Recognition	Recognition	NNP
prudhomme	138	4	.	.	.
prudhomme	138	5	”	"	''
prudhomme	138	6	n.d	n.d	NNP
prudhomme	138	7	.	.	NNP
prudhomme	138	8	Accessed	Accessed	NNP
prudhomme	138	9	November	November	NNP
prudhomme	138	10	30	30	CD
prudhomme	138	11	,	,	,
prudhomme	138	12	2019	2019	CD
prudhomme	138	13	.	.	.
prudhomme	138	14	https://github.com/ageitgey/face_recognition	https://github.com/ageitgey/face_recognition	NN
prudhomme	138	15	.	.	.
prudhomme	139	1	Griffey	Griffey	NNP
prudhomme	139	2	,	,	,
prudhomme	139	3	Jason	Jason	NNP
prudhomme	139	4	,	,	,
prudhomme	139	5	ed	ed	NNP
prudhomme	139	6	..	..	NFP
prudhomme	139	7	2019	2019	CD
prudhomme	139	8	.	.	.
prudhomme	140	1	“	"	``
prudhomme	140	2	Artificial	Artificial	NNP
prudhomme	140	3	Intelligence	Intelligence	NNP
prudhomme	140	4	and	and	CC
prudhomme	140	5	Machine	Machine	NNP
prudhomme	140	6	Learning	Learning	NNP
prudhomme	140	7	in	in	IN
prudhomme	140	8	Libraries	library	NNS
prudhomme	140	9	.	.	.
prudhomme	140	10	”	"	''
prudhomme	140	11	Special	special	JJ
prudhomme	140	12	issue	issue	NN
prudhomme	140	13	,	,	,
prudhomme	140	14	Library	Library	NNP
prudhomme	140	15	Technology	Technology	NNP
prudhomme	140	16	Reports	report	VBZ
prudhomme	140	17	55	55	CD
prudhomme	140	18	,	,	,
prudhomme	140	19	no	no	UH
prudhomme	140	20	.	.	.
prudhomme	141	1	1	1	LS
prudhomme	141	2	(	(	-LRB-
prudhomme	141	3	January	January	NNP
prudhomme	141	4	)	)	-RRB-
prudhomme	141	5	.	.	.
prudhomme	142	1	https://journals.ala.org/index.php/ltr/issue/viewIssue/709/471	https://journals.ala.org/index.php/ltr/issue/viewIssue/709/471	NNP
prudhomme	142	2	.	.	.
prudhomme	143	1	Harper	Harper	NNP
prudhomme	143	2	,	,	,
prudhomme	143	3	Charlie	Charlie	NNP
prudhomme	143	4	.	.	.
prudhomme	144	1	2018	2018	CD
prudhomme	144	2	.	.	.
prudhomme	145	1	“	"	``
prudhomme	145	2	Machine	Machine	NNP
prudhomme	145	3	Learning	Learning	NNP
prudhomme	145	4	and	and	CC
prudhomme	145	5	the	the	DT
prudhomme	145	6	Library	library	NN
prudhomme	145	7	or	or	CC
prudhomme	145	8	:	:	:
prudhomme	145	9	How	how	WRB
prudhomme	145	10	I	-PRON-	PRP
prudhomme	145	11	Learned	learn	VBD
prudhomme	145	12	to	to	TO
prudhomme	145	13	Stop	stop	VB
prudhomme	145	14	Worrying	worry	VBG
prudhomme	145	15	and	and	CC
prudhomme	145	16	Love	love	VB
prudhomme	145	17	My	-PRON-	PRP$
prudhomme	145	18	Robot	Robot	NNP
prudhomme	145	19	Overlords	Overlords	NNPS
prudhomme	145	20	.	.	.
prudhomme	145	21	”	"	''
prudhomme	145	22	Code4Lib	Code4Lib	NNP
prudhomme	145	23	,	,	,
prudhomme	145	24	no	no	UH
prudhomme	145	25	.	.	.
prudhomme	146	1	41	41	CD
prudhomme	146	2	(	(	-LRB-
prudhomme	146	3	August	August	NNP
prudhomme	146	4	)	)	-RRB-
prudhomme	146	5	.	.	.
prudhomme	147	1	https://journal.code4lib.org/articles/13671	https://journal.code4lib.org/articles/13671	NNP
prudhomme	147	2	.	.	.
prudhomme	148	1	Heller	Heller	NNP
prudhomme	148	2	,	,	,
prudhomme	148	3	Martin	Martin	NNP
prudhomme	148	4	.	.	.
prudhomme	149	1	2019	2019	CD
prudhomme	149	2	.	.	.
prudhomme	150	1	“	"	``
prudhomme	150	2	What	what	WP
prudhomme	150	3	is	be	VBZ
prudhomme	150	4	Keras	Keras	NNP
prudhomme	150	5	?	?	.
prudhomme	151	1	The	the	DT
prudhomme	151	2	Deep	Deep	NNP
prudhomme	151	3	Neural	Neural	NNP
prudhomme	151	4	Network	Network	NNP
prudhomme	151	5	API	API	NNP
prudhomme	151	6	Explained	Explained	NNP
prudhomme	151	7	.	.	.
prudhomme	151	8	”	"	''
prudhomme	151	9	InfoWorld	InfoWorld	NNP
prudhomme	151	10	(	(	-LRB-
prudhomme	151	11	website	website	NN
prudhomme	151	12	)	)	-RRB-
prudhomme	151	13	.	.	.
prudhomme	152	1	January	January	NNP
prudhomme	152	2	28	28	CD
prudhomme	152	3	,	,	,
prudhomme	152	4	2019	2019	CD
prudhomme	152	5	.	.	.
prudhomme	152	6	https://www.infoworld.com/article/3336192/what-is-keras-the-deep-neural-network-api-explained.html	https://www.infoworld.com/article/3336192/what-is-keras-the-deep-neural-network-api-explained.html	ADD
prudhomme	152	7	.	.	.
prudhomme	153	1	Karaka	Karaka	NNP
prudhomme	153	2	,	,	,
prudhomme	153	3	Anil	Anil	NNP
prudhomme	153	4	.	.	.
prudhomme	154	1	2019	2019	CD
prudhomme	154	2	.	.	.
prudhomme	155	1	“	"	``
prudhomme	155	2	Object	object	VB
prudhomme	155	3	Detection	detection	NN
prudhomme	155	4	with	with	IN
prudhomme	155	5	RetinaNet	RetinaNet	NNP
prudhomme	155	6	.	.	.
prudhomme	155	7	”	"	''
prudhomme	155	8	Weights	Weights	NNP
prudhomme	155	9	&	&	CC
prudhomme	155	10	Biases	Biases	NNPS
prudhomme	155	11	(	(	-LRB-
prudhomme	155	12	website	website	NN
prudhomme	155	13	)	)	-RRB-
prudhomme	155	14	.	.	.
prudhomme	156	1	July	July	NNP
prudhomme	156	2	18	18	CD
prudhomme	156	3	,	,	,
prudhomme	156	4	2019	2019	CD
prudhomme	156	5	.	.	.
prudhomme	156	6	https://www.wandb.com/articles/object-detection-with-retinanet	https://www.wandb.com/articles/object-detection-with-retinanet	ADD
prudhomme	156	7	.	.	.
prudhomme	157	1	Lynch	Lynch	NNP
prudhomme	157	2	,	,	,
prudhomme	157	3	Clifford	Clifford	NNP
prudhomme	157	4	.	.	.
prudhomme	158	1	2019	2019	CD
prudhomme	158	2	.	.	.
prudhomme	159	1	“	"	``
prudhomme	159	2	Machine	Machine	NNP
prudhomme	159	3	Learning	Learning	NNP
prudhomme	159	4	,	,	,
prudhomme	159	5	Archives	Archives	NNPS
prudhomme	159	6	and	and	CC
prudhomme	159	7	Special	Special	NNP
prudhomme	159	8	Collections	Collections	NNPS
prudhomme	159	9	:	:	:
prudhomme	159	10	A	a	DT
prudhomme	159	11	High	High	NNP
prudhomme	159	12	Level	Level	NNP
prudhomme	159	13	View	view	NN
prudhomme	159	14	.	.	.
prudhomme	159	15	”	"	''
prudhomme	159	16	International	International	NNP
prudhomme	159	17	Council	Council	NNP
prudhomme	159	18	on	on	IN
prudhomme	159	19	Archives	Archives	NNP
prudhomme	159	20	Blog	Blog	NNP
prudhomme	159	21	.	.	.
prudhomme	160	1	October	October	NNP
prudhomme	160	2	1	1	CD
prudhomme	160	3	,	,	,
prudhomme	160	4	2019	2019	CD
prudhomme	160	5	.	.	.
prudhomme	160	6	https://blog-ica.org/2019/10/02/machine-learning-archives-and-special-collections-a-high-level-view/.	https://blog-ica.org/2019/10/02/machine-learning-archives-and-special-collections-a-high-level-view/.	JJ
prudhomme	161	1	Mahapatra	Mahapatra	NNP
prudhomme	161	2	,	,	,
prudhomme	161	3	Sambit	Sambit	NNP
prudhomme	161	4	.	.	.
prudhomme	162	1	“	"	``
prudhomme	162	2	Why	why	WRB
prudhomme	162	3	Deep	deep	RB
prudhomme	162	4	Learning	learn	VBG
prudhomme	162	5	over	over	IN
prudhomme	162	6	Traditional	traditional	JJ
prudhomme	162	7	Machine	machine	NN
prudhomme	162	8	Learning	Learning	NNP
prudhomme	162	9	?	?	.
prudhomme	162	10	”	"	''
prudhomme	162	11	Towards	towards	IN
prudhomme	162	12	Data	Data	NNP
prudhomme	162	13	Science	Science	NNP
prudhomme	162	14	(	(	-LRB-
prudhomme	162	15	website).March	website).March	NNP
prudhomme	162	16	21	21	CD
prudhomme	162	17	,	,	,
prudhomme	162	18	2018	2018	CD
prudhomme	162	19	.	.	.
prudhomme	162	20	https://towardsdatascience.com/why-deep-learning-is-needed-over-traditional-machine-learning-1b6a99177063	https://towardsdatascience.com/why-deep-learning-is-needed-over-traditional-machine-learning-1b6a99177063	NNP
prudhomme	162	21	.	.	.
prudhomme	163	1	McCarthy	McCarthy	NNP
prudhomme	163	2	,	,	,
prudhomme	163	3	John	John	NNP
prudhomme	163	4	.	.	.
prudhomme	164	1	“	"	``
prudhomme	164	2	What	what	WP
prudhomme	164	3	is	be	VBZ
prudhomme	164	4	Artificial	Artificial	NNP
prudhomme	164	5	Intelligence	Intelligence	NNP
prudhomme	164	6	?	?	.
prudhomme	164	7	”	"	''
prudhomme	164	8	Professor	Professor	NNP
prudhomme	164	9	John	John	NNP
prudhomme	164	10	McCarthy	McCarthy	NNP
prudhomme	164	11	(	(	-LRB-
prudhomme	164	12	website	website	NN
prudhomme	164	13	)	)	-RRB-
prudhomme	164	14	.	.	.
prudhomme	165	1	Revised	Revised	NNP
prudhomme	165	2	November	November	NNP
prudhomme	165	3	12	12	CD
prudhomme	165	4	,	,	,
prudhomme	165	5	2007	2007	CD
prudhomme	165	6	.	.	.
prudhomme	165	7	http://jmc.stanford.edu/articles/whatisai/whatisai.pdf	http://jmc.stanford.edu/articles/whatisai/whatisai.pdf	NNP
prudhomme	165	8	.	.	.
prudhomme	166	1	Padilla	Padilla	NNP
prudhomme	166	2	,	,	,
prudhomme	166	3	Thomas	Thomas	NNP
prudhomme	166	4	.	.	.
prudhomme	167	1	2019	2019	CD
prudhomme	167	2	.	.	.
prudhomme	168	1	Responsible	responsible	JJ
prudhomme	168	2	Operations	operation	NNS
prudhomme	168	3	:	:	:
prudhomme	168	4	Data	Data	NNP
prudhomme	168	5	Science	Science	NNP
prudhomme	168	6	,	,	,
prudhomme	168	7	Machine	Machine	NNP
prudhomme	168	8	Learning	Learning	NNP
prudhomme	168	9	,	,	,
prudhomme	168	10	and	and	CC
prudhomme	168	11	AI	AI	NNP
prudhomme	168	12	in	in	IN
prudhomme	168	13	Libraries	Libraries	NNP
prudhomme	168	14	.	.	.
prudhomme	169	1	Dublin	Dublin	NNP
prudhomme	169	2	,	,	,
prudhomme	169	3	OH	oh	NN
prudhomme	169	4	:	:	:
prudhomme	169	5	OCLC	OCLC	NNP
prudhomme	169	6	Research	Research	NNP
prudhomme	169	7	.	.	.
prudhomme	170	1	https://doi.org/10.25333/xk7z-9g97	https://doi.org/10.25333/xk7z-9g97	NNP
prudhomme	170	2	.	.	.
prudhomme	171	1	Pesenti	Pesenti	NNP
prudhomme	171	2	,	,	,
prudhomme	171	3	Jerome	Jerome	NNP
prudhomme	171	4	.	.	.
prudhomme	172	1	2019	2019	CD
prudhomme	172	2	.	.	.
prudhomme	173	1	“	"	``
prudhomme	173	2	Facebook	Facebook	NNP
prudhomme	173	3	's	's	POS
prudhomme	173	4	Head	Head	NNP
prudhomme	173	5	of	of	IN
prudhomme	173	6	AI	AI	NNP
prudhomme	173	7	Says	say	VBZ
prudhomme	173	8	the	the	DT
prudhomme	173	9	Field	field	NN
prudhomme	173	10	Will	Will	MD
prudhomme	173	11	Soon	soon	RB
prudhomme	173	12	‘	'	``
prudhomme	173	13	Hit	hit	VB
prudhomme	173	14	the	the	DT
prudhomme	173	15	Wall	Wall	NNP
prudhomme	173	16	.	.	.
prudhomme	173	17	’	'	''
prudhomme	173	18	”	"	''
prudhomme	173	19	Interview	interview	NN
prudhomme	173	20	by	by	IN
prudhomme	173	21	Will	Will	NNP
prudhomme	173	22	Knight	Knight	NNP
prudhomme	173	23	.	.	.
prudhomme	174	1	Wired	Wired	NNP
prudhomme	174	2	(	(	-LRB-
prudhomme	174	3	website	website	NN
prudhomme	174	4	)	)	-RRB-
prudhomme	174	5	.	.	.
prudhomme	175	1	December	December	NNP
prudhomme	175	2	4	4	CD
prudhomme	175	3	,	,	,
prudhomme	175	4	2019	2019	CD
prudhomme	175	5	.	.	.
prudhomme	175	6	https://www.wired.com/story/facebooks-ai-says-field-hit-wall/.	https://www.wired.com/story/facebooks-ai-says-field-hit-wall/.	ADD
prudhomme	176	1	Ren	Ren	NNP
prudhomme	176	2	,	,	,
prudhomme	176	3	Shaoqing	shaoqe	VBG
prudhomme	176	4	,	,	,
prudhomme	176	5	Kaiming	Kaiming	NNP
prudhomme	176	6	He	He	NNP
prudhomme	176	7	,	,	,
prudhomme	176	8	Ross	Ross	NNP
prudhomme	176	9	Girshick	Girshick	NNP
prudhomme	176	10	,	,	,
prudhomme	176	11	Xiangyu	Xiangyu	NNP
prudhomme	176	12	Zhang	Zhang	NNP
prudhomme	176	13	,	,	,
prudhomme	176	14	and	and	CC
prudhomme	176	15	Jian	Jian	NNP
prudhomme	176	16	Sun	Sun	NNP
prudhomme	176	17	.	.	.
prudhomme	177	1	2015	2015	CD
prudhomme	177	2	.	.	.
prudhomme	178	1	“	"	``
prudhomme	178	2	Object	object	VB
prudhomme	178	3	Detection	detection	NN
prudhomme	178	4	Networks	Networks	NNPS
prudhomme	178	5	on	on	IN
prudhomme	178	6	Convolutional	Convolutional	NNP
prudhomme	178	7	Feature	Feature	NNP
prudhomme	178	8	Maps	Maps	NNPS
prudhomme	178	9	.	.	.
prudhomme	178	10	”	"	''
prudhomme	178	11	IEEE	IEEE	NNP
prudhomme	178	12	Transactions	transaction	NNS
prudhomme	178	13	on	on	IN
prudhomme	178	14	Pattern	Pattern	NNP
prudhomme	178	15	Analysis	Analysis	NNP
prudhomme	178	16	and	and	CC
prudhomme	178	17	Machine	Machine	NNP
prudhomme	178	18	Intelligence	Intelligence	NNP
prudhomme	178	19	39	39	CD
prudhomme	178	20	,	,	,
prudhomme	178	21	no	no	UH
prudhomme	178	22	.	.	.
prudhomme	179	1	7	7	LS
prudhomme	179	2	(	(	-LRB-
prudhomme	179	3	April	April	NNP
prudhomme	179	4	)	)	-RRB-
prudhomme	179	5	.	.	.
prudhomme	180	1	SAS	SAS	NNP
prudhomme	180	2	.	.	.
prudhomme	181	1	n.d	n.d	NNP
prudhomme	181	2	.	.	.
prudhomme	181	3	“	"	``
prudhomme	181	4	Machine	Machine	NNP
prudhomme	181	5	Learning	Learning	NNP
prudhomme	181	6	:	:	:
prudhomme	181	7	What	what	WP
prudhomme	181	8	It	-PRON-	PRP
prudhomme	181	9	Is	be	VBZ
prudhomme	181	10	and	and	CC
prudhomme	181	11	Why	why	WRB
prudhomme	181	12	It	-PRON-	PRP
prudhomme	181	13	Matters	matter	VBZ
prudhomme	181	14	.	.	.
prudhomme	181	15	”	"	''
prudhomme	181	16	Accessed	access	VBN
prudhomme	181	17	December	December	NNP
prudhomme	181	18	17	17	CD
prudhomme	181	19	,	,	,
prudhomme	181	20	2019	2019	CD
prudhomme	181	21	.	.	.
prudhomme	181	22	https://www.sas.com/en_us/insights/analytics/machine-learning.html	https://www.sas.com/en_us/insights/analytics/machine-learning.html	CD
prudhomme	181	23	.	.	.
prudhomme	182	1	Schweikert	Schweikert	NNP
prudhomme	182	2	,	,	,
prudhomme	182	3	Annie	Annie	NNP
prudhomme	182	4	.	.	.
prudhomme	183	1	2019	2019	CD
prudhomme	183	2	.	.	.
prudhomme	184	1	“	"	``
prudhomme	184	2	Audiovisual	Audiovisual	NNP
prudhomme	184	3	Algorithms	Algorithms	NNP
prudhomme	184	4	,	,	,
prudhomme	184	5	New	New	NNP
prudhomme	184	6	Techniques	Techniques	NNP
prudhomme	184	7	for	for	IN
prudhomme	184	8	Digital	Digital	NNP
prudhomme	184	9	Processing	Processing	NNP
prudhomme	184	10	.	.	.
prudhomme	184	11	”	"	''
prudhomme	184	12	Master	Master	NNP
prudhomme	184	13	’s	’s	POS
prudhomme	184	14	Thesis	Thesis	NNP
prudhomme	184	15	,	,	,
prudhomme	184	16	New	New	NNP
prudhomme	184	17	York	York	NNP
prudhomme	184	18	University	University	NNP
prudhomme	184	19	.	.	.
prudhomme	185	1	https://www.nyu.edu/tisch/preservation/program/student_work/2019spring/19s_thesis_Schweikert.pdf	https://www.nyu.edu/tisch/preservation/program/student_work/2019spring/19s_thesis_Schweikert.pdf	NNP
prudhomme	185	2	.	.	.
prudhomme	186	1	“	"	``
prudhomme	186	2	Tesseract	Tesseract	NNP
prudhomme	186	3	OCR	OCR	NNP
prudhomme	186	4	.	.	.
prudhomme	186	5	”	"	''
prudhomme	186	6	n.d	n.d	NNP
prudhomme	186	7	.	.	NNP
prudhomme	186	8	Accessed	access	VBN
prudhomme	186	9	December	December	NNP
prudhomme	186	10	11	11	CD
prudhomme	186	11	,	,	,
prudhomme	186	12	2019	2019	CD
prudhomme	186	13	.	.	.
prudhomme	186	14	https://github.com/tesseract-ocr/tesseract	https://github.com/tesseract-ocr/tesseract	NNP
prudhomme	186	15	.	.	.
prudhomme	187	1	Zhao	Zhao	NNP
prudhomme	187	2	,	,	,
prudhomme	187	3	Zhong	Zhong	NNP
prudhomme	187	4	-	-	HYPH
prudhomme	187	5	Qiu	Qiu	NNP
prudhomme	187	6	,	,	,
prudhomme	187	7	Peng	Peng	NNP
prudhomme	187	8	Zheng	Zheng	NNP
prudhomme	187	9	,	,	,
prudhomme	187	10	Shou	Shou	NNP
prudhomme	187	11	-	-	HYPH
prudhomme	187	12	tao	tao	NNP
prudhomme	187	13	Xu	Xu	NNP
prudhomme	187	14	,	,	,
prudhomme	187	15	and	and	CC
prudhomme	187	16	Xindong	Xindong	NNP
prudhomme	187	17	Wu	Wu	NNP
prudhomme	187	18	.	.	.
prudhomme	188	1	2017	2017	CD
prudhomme	188	2	“	"	``
prudhomme	188	3	Object	object	VB
prudhomme	188	4	Detection	detection	NN
prudhomme	188	5	with	with	IN
prudhomme	188	6	Deep	Deep	NNP
prudhomme	188	7	Learning	learning	NN
prudhomme	188	8	:	:	:
prudhomme	188	9	A	a	DT
prudhomme	188	10	Review	Review	NNP
prudhomme	188	11	.	.	.
prudhomme	188	12	”	"	''
prudhomme	188	13	IEEE	IEEE	NNP
prudhomme	188	14	Transactions	transaction	NNS
prudhomme	188	15	on	on	IN
prudhomme	188	16	Neural	Neural	NNP
prudhomme	188	17	Networks	Networks	NNPS
prudhomme	188	18	and	and	CC
prudhomme	188	19	Learning	Learning	NNP
prudhomme	188	20	Systems	Systems	NNPS
prudhomme	188	21	30	30	CD
prudhomme	188	22	,	,	,
prudhomme	188	23	no	no	UH
prudhomme	188	24	.	.	.
prudhomme	189	1	11	11	CD
prudhomme	189	2	(	(	-LRB-
prudhomme	189	3	2019	2019	CD
prudhomme	189	4	)	)	-RRB-
prudhomme	189	5	:	:	:
prudhomme	189	6	3212	3212	CD
prudhomme	189	7	-	-	SYM
prudhomme	189	8	3232	3232	CD
prudhomme	189	9	.	.	.
prudhomme	190	1	Comment	comment	NN
prudhomme	190	2	by	by	IN
prudhomme	190	3	Daniel	Daniel	NNP
prudhomme	190	4	Johnson	Johnson	NNP
prudhomme	190	5	:	:	:
prudhomme	190	6	It	-PRON-	PRP
prudhomme	190	7	appears	appear	VBZ
prudhomme	190	8	you	-PRON-	PRP
prudhomme	190	9	were	be	VBD
prudhomme	190	10	using	use	VBG
prudhomme	190	11	a	a	DT
prudhomme	190	12	preprint	preprint	NN
prudhomme	190	13	.	.	.
prudhomme	191	1	This	this	DT
prudhomme	191	2	information	information	NN
prudhomme	191	3	is	be	VBZ
prudhomme	191	4	for	for	IN
prudhomme	191	5	the	the	DT
prudhomme	191	6	official	official	JJ
prudhomme	191	7	published	publish	VBN
prudhomme	191	8	version	version	NN
prudhomme	191	9	.	.	.