id	sid	tid	token	lemma	pos
hansen	1	1	Can	Can	MD
hansen	1	2	a	a	DT
hansen	1	3	Hammer	Hammer	NNP
hansen	1	4	Categorize	Categorize	NNP
hansen	1	5	Highly	highly	RB
hansen	1	6	Technical	technical	JJ
hansen	1	7	Articles	article	NNS
hansen	1	8	?	?	.
hansen	2	1	Samuel	Samuel	NNP
hansen	2	2	Hansen	Hansen	NNP
hansen	2	3	University	University	NNP
hansen	2	4	of	of	IN
hansen	2	5	Michigan	Michigan	NNP
hansen	2	6	hansensm@umich.edu	hansensm@umich.edu	NNP
hansen	2	7	When	when	WRB
hansen	2	8	everything	everything	NN
hansen	2	9	looks	look	VBZ
hansen	2	10	like	like	IN
hansen	2	11	a	a	DT
hansen	2	12	nail	nail	NN
hansen	2	13	...	...	.
hansen	3	1	I	-PRON-	PRP
hansen	3	2	was	be	VBD
hansen	3	3	sure	sure	JJ
hansen	3	4	I	-PRON-	PRP
hansen	3	5	had	have	VBD
hansen	3	6	the	the	DT
hansen	3	7	most	most	RBS
hansen	3	8	brilliant	brilliant	JJ
hansen	3	9	research	research	NN
hansen	3	10	project	project	NN
hansen	3	11	idea	idea	NN
hansen	3	12	for	for	IN
hansen	3	13	my	-PRON-	PRP$
hansen	3	14	course	course	NN
hansen	3	15	in	in	IN
hansen	3	16	Digital	Digital	NNP
hansen	3	17	Scholarship	Scholarship	NNP
hansen	3	18	techniques	technique	NNS
hansen	3	19	.	.	.
hansen	4	1	I	-PRON-	PRP
hansen	4	2	would	would	MD
hansen	4	3	use	use	VB
hansen	4	4	the	the	DT
hansen	4	5	Mathematical	Mathematical	NNP
hansen	4	6	Subject	Subject	NNP
hansen	4	7	Classification	Classification	NNP
hansen	4	8	(	(	-LRB-
hansen	4	9	MSC	MSC	NNP
hansen	4	10	)	)	-RRB-
hansen	4	11	values	value	NNS
hansen	4	12	assigned	assign	VBN
hansen	4	13	to	to	IN
hansen	4	14	the	the	DT
hansen	4	15	publications	publication	NNS
hansen	4	16	in	in	IN
hansen	4	17	MathSciNet[footnoteRef:0	MathSciNet[footnoteRef:0	NNP
hansen	4	18	]	]	-RRB-
hansen	4	19	to	to	TO
hansen	4	20	create	create	VB
hansen	4	21	a	a	DT
hansen	4	22	temporal	temporal	JJ
hansen	4	23	citation	citation	NN
hansen	4	24	network	network	NN
hansen	4	25	which	which	WDT
hansen	4	26	would	would	MD
hansen	4	27	allow	allow	VB
hansen	4	28	me	-PRON-	PRP
hansen	4	29	to	to	TO
hansen	4	30	visualize	visualize	VB
hansen	4	31	how	how	WRB
hansen	4	32	new	new	JJ
hansen	4	33	subfields	subfield	NNS
hansen	4	34	were	be	VBD
hansen	4	35	created	create	VBN
hansen	4	36	and	and	CC
hansen	4	37	perhaps	perhaps	RB
hansen	4	38	even	even	RB
hansen	4	39	predict	predict	VBP
hansen	4	40	them	-PRON-	PRP
hansen	4	41	while	while	IN
hansen	4	42	they	-PRON-	PRP
hansen	4	43	were	be	VBD
hansen	4	44	still	still	RB
hansen	4	45	in	in	IN
hansen	4	46	their	-PRON-	PRP$
hansen	4	47	infancy	infancy	NN
hansen	4	48	.	.	.
hansen	5	1	I	-PRON-	PRP
hansen	5	2	thought	think	VBD
hansen	5	3	it	-PRON-	PRP
hansen	5	4	would	would	MD
hansen	5	5	be	be	VB
hansen	5	6	an	an	DT
hansen	5	7	easy	easy	JJ
hansen	5	8	enough	enough	JJ
hansen	5	9	project	project	NN
hansen	5	10	.	.	.
hansen	6	1	I	-PRON-	PRP
hansen	6	2	already	already	RB
hansen	6	3	knew	know	VBD
hansen	6	4	how	how	WRB
hansen	6	5	to	to	TO
hansen	6	6	analyze	analyze	VB
hansen	6	7	network	network	NN
hansen	6	8	data	datum	NNS
hansen	6	9	and	and	CC
hansen	6	10	the	the	DT
hansen	6	11	data	datum	NNS
hansen	6	12	I	-PRON-	PRP
hansen	6	13	needed	need	VBD
hansen	6	14	already	already	RB
hansen	6	15	existed	exist	VBD
hansen	6	16	,	,	,
hansen	6	17	I	-PRON-	PRP
hansen	6	18	just	just	RB
hansen	6	19	had	have	VBD
hansen	6	20	to	to	TO
hansen	6	21	get	get	VB
hansen	6	22	my	-PRON-	PRP$
hansen	6	23	hands	hand	NNS
hansen	6	24	on	on	IN
hansen	6	25	it	-PRON-	PRP
hansen	6	26	.	.	.
hansen	7	1	I	-PRON-	PRP
hansen	7	2	even	even	RB
hansen	7	3	sold	sell	VBD
hansen	7	4	a	a	DT
hansen	7	5	couple	couple	NN
hansen	7	6	of	of	IN
hansen	7	7	my	-PRON-	PRP$
hansen	7	8	fellow	fellow	JJ
hansen	7	9	coursemates	coursemate	NNS
hansen	7	10	on	on	IN
hansen	7	11	the	the	DT
hansen	7	12	idea	idea	NN
hansen	7	13	and	and	CC
hansen	7	14	they	-PRON-	PRP
hansen	7	15	agreed	agree	VBD
hansen	7	16	to	to	TO
hansen	7	17	work	work	VB
hansen	7	18	with	with	IN
hansen	7	19	me	-PRON-	PRP
hansen	7	20	.	.	.
hansen	8	1	Of	of	RB
hansen	8	2	course	course	RB
hansen	8	3	nothing	nothing	NN
hansen	8	4	is	be	VBZ
hansen	8	5	as	as	RB
hansen	8	6	easy	easy	JJ
hansen	8	7	as	as	IN
hansen	8	8	that	that	DT
hansen	8	9	,	,	,
hansen	8	10	and	and	CC
hansen	8	11	numerous	numerous	JJ
hansen	8	12	requests	request	NNS
hansen	8	13	for	for	IN
hansen	8	14	data	datum	NNS
hansen	8	15	went	go	VBD
hansen	8	16	without	without	IN
hansen	8	17	response	response	NN
hansen	8	18	.	.	.
hansen	9	1	Even	even	RB
hansen	9	2	after	after	IN
hansen	9	3	I	-PRON-	PRP
hansen	9	4	reached	reach	VBD
hansen	9	5	out	out	RP
hansen	9	6	to	to	IN
hansen	9	7	personal	personal	JJ
hansen	9	8	contacts	contact	NNS
hansen	9	9	at	at	IN
hansen	9	10	MathSciNet	MathSciNet	NNP
hansen	9	11	,	,	,
hansen	9	12	we	-PRON-	PRP
hansen	9	13	came	come	VBD
hansen	9	14	to	to	TO
hansen	9	15	understand	understand	VB
hansen	9	16	we	-PRON-	PRP
hansen	9	17	would	would	MD
hansen	9	18	not	not	RB
hansen	9	19	be	be	VB
hansen	9	20	getting	get	VBG
hansen	9	21	the	the	DT
hansen	9	22	MSC	MSC	NNP
hansen	9	23	data	datum	NNS
hansen	9	24	the	the	DT
hansen	9	25	entire	entire	JJ
hansen	9	26	project	project	NN
hansen	9	27	relied	rely	VBN
hansen	9	28	upon	upon	IN
hansen	9	29	.	.	.
hansen	10	1	Not	not	RB
hansen	10	2	that	that	IN
hansen	10	3	we	-PRON-	PRP
hansen	10	4	were	be	VBD
hansen	10	5	going	go	VBG
hansen	10	6	to	to	TO
hansen	10	7	let	let	VB
hansen	10	8	a	a	DT
hansen	10	9	little	little	JJ
hansen	10	10	setback	setback	NN
hansen	10	11	like	like	UH
hansen	10	12	not	not	RB
hansen	10	13	having	have	VBG
hansen	10	14	the	the	DT
hansen	10	15	necessary	necessary	JJ
hansen	10	16	data	datum	NNS
hansen	10	17	stop	stop	VB
hansen	10	18	us	-PRON-	PRP
hansen	10	19	.	.	.
hansen	11	1	Comment	comment	NN
hansen	11	2	by	by	IN
hansen	11	3	Mark	Mark	NNP
hansen	11	4	Dehmlow	Dehmlow	NNP
hansen	11	5	:	:	:
hansen	11	6	Is	be	VBZ
hansen	11	7	this	this	DT
hansen	11	8	meant	mean	VBN
hansen	11	9	to	to	TO
hansen	11	10	be	be	VB
hansen	11	11	MARC	MARC	NNP
hansen	11	12	subfields	subfield	NNS
hansen	11	13	or	or	CC
hansen	11	14	professional	professional	JJ
hansen	11	15	subfields	subfield	NNS
hansen	11	16	?	?	.
hansen	12	1	It	-PRON-	PRP
hansen	12	2	might	may	MD
hansen	12	3	be	be	VB
hansen	12	4	good	good	JJ
hansen	12	5	to	to	TO
hansen	12	6	disambiguate	disambiguate	VB
hansen	12	7	.	.	.
hansen	13	1	[	[	-LRB-
hansen	13	2	0	0	CD
hansen	13	3	:	:	:
hansen	13	4	https://mathscinet.ams.org/	https://mathscinet.ams.org/	NNP
hansen	13	5	]	]	-RRB-
hansen	13	6	After	after	RB
hansen	13	7	all	all	RB
hansen	13	8	,	,	,
hansen	13	9	this	this	DT
hansen	13	10	was	be	VBD
hansen	13	11	early	early	JJ
hansen	13	12	2018	2018	CD
hansen	13	13	and	and	CC
hansen	13	14	there	there	EX
hansen	13	15	had	have	VBD
hansen	13	16	already	already	RB
hansen	13	17	been	be	VBN
hansen	13	18	years	year	NNS
hansen	13	19	of	of	IN
hansen	13	20	stories	story	NNS
hansen	13	21	about	about	IN
hansen	13	22	how	how	WRB
hansen	13	23	artificial	artificial	JJ
hansen	13	24	intelligence	intelligence	NN
hansen	13	25	,	,	,
hansen	13	26	machine	machine	NN
hansen	13	27	learning	learning	NN
hansen	13	28	in	in	IN
hansen	13	29	particular	particular	JJ
hansen	13	30	,	,	,
hansen	13	31	was	be	VBD
hansen	13	32	going	go	VBG
hansen	13	33	to	to	TO
hansen	13	34	revolutionize	revolutionize	VB
hansen	13	35	every	every	DT
hansen	13	36	aspect	aspect	NN
hansen	13	37	of	of	IN
hansen	13	38	our	-PRON-	PRP$
hansen	13	39	world	world	NN
hansen	13	40	(	(	-LRB-
hansen	13	41	Kelly	Kelly	NNP
hansen	13	42	2014	2014	CD
hansen	13	43	;	;	:
hansen	13	44	Clark	Clark	NNP
hansen	13	45	2015	2015	CD
hansen	13	46	;	;	:
hansen	13	47	Parloff	Parloff	NNP
hansen	13	48	2016	2016	CD
hansen	13	49	;	;	:
hansen	13	50	Sangwani	Sangwani	NNP
hansen	13	51	2017	2017	CD
hansen	13	52	;	;	:
hansen	13	53	Tank	tank	NN
hansen	13	54	2017	2017	CD
hansen	13	55	)	)	-RRB-
hansen	13	56	.	.	.
hansen	14	1	All	all	PDT
hansen	14	2	the	the	DT
hansen	14	3	coverage	coverage	NN
hansen	14	4	made	make	VBD
hansen	14	5	it	-PRON-	PRP
hansen	14	6	seem	seem	VB
hansen	14	7	like	like	IN
hansen	14	8	AI	AI	NNP
hansen	14	9	was	be	VBD
hansen	14	10	not	not	RB
hansen	14	11	only	only	RB
hansen	14	12	a	a	DT
hansen	14	13	tool	tool	NN
hansen	14	14	with	with	IN
hansen	14	15	as	as	RB
hansen	14	16	many	many	JJ
hansen	14	17	applications	application	NNS
hansen	14	18	as	as	IN
hansen	14	19	a	a	DT
hansen	14	20	hammer	hammer	NN
hansen	14	21	,	,	,
hansen	14	22	but	but	CC
hansen	14	23	that	that	IN
hansen	14	24	it	-PRON-	PRP
hansen	14	25	also	also	RB
hansen	14	26	magically	magically	RB
hansen	14	27	turned	turn	VBD
hansen	14	28	all	all	DT
hansen	14	29	problems	problem	NNS
hansen	14	30	into	into	IN
hansen	14	31	nails	nail	NNS
hansen	14	32	.	.	.
hansen	15	1	While	while	IN
hansen	15	2	none	none	NN
hansen	15	3	of	of	IN
hansen	15	4	us	-PRON-	PRP
hansen	15	5	were	be	VBD
hansen	15	6	AI	AI	NNP
hansen	15	7	experts	expert	NNS
hansen	15	8	,	,	,
hansen	15	9	we	-PRON-	PRP
hansen	15	10	knew	know	VBD
hansen	15	11	that	that	DT
hansen	15	12	machine	machine	NN
hansen	15	13	learning	learning	NN
hansen	15	14	was	be	VBD
hansen	15	15	supposed	suppose	VBN
hansen	15	16	to	to	TO
hansen	15	17	be	be	VB
hansen	15	18	good	good	JJ
hansen	15	19	at	at	IN
hansen	15	20	classification	classification	NN
hansen	15	21	and	and	CC
hansen	15	22	categorization	categorization	NN
hansen	15	23	.	.	.
hansen	16	1	The	the	DT
hansen	16	2	promise	promise	NN
hansen	16	3	seemed	seem	VBD
hansen	16	4	to	to	TO
hansen	16	5	be	be	VB
hansen	16	6	that	that	IN
hansen	16	7	if	if	IN
hansen	16	8	you	-PRON-	PRP
hansen	16	9	had	have	VBD
hansen	16	10	stacks	stack	NNS
hansen	16	11	of	of	IN
hansen	16	12	data	datum	NNS
hansen	16	13	,	,	,
hansen	16	14	a	a	DT
hansen	16	15	machine	machine	NN
hansen	16	16	learning	learn	VBG
hansen	16	17	algorithm	algorithm	NN
hansen	16	18	could	could	MD
hansen	16	19	dive	dive	VB
hansen	16	20	in	in	RP
hansen	16	21	,	,	,
hansen	16	22	find	find	VB
hansen	16	23	the	the	DT
hansen	16	24	needles	needle	NNS
hansen	16	25	,	,	,
hansen	16	26	and	and	CC
hansen	16	27	arrange	arrange	VB
hansen	16	28	them	-PRON-	PRP
hansen	16	29	into	into	IN
hansen	16	30	neatly	neatly	RB
hansen	16	31	divided	divide	VBN
hansen	16	32	piles	pile	NNS
hansen	16	33	of	of	IN
hansen	16	34	similar	similar	JJ
hansen	16	35	sharpness	sharpness	NN
hansen	16	36	and	and	CC
hansen	16	37	length	length	NN
hansen	16	38	.	.	.
hansen	17	1	Not	not	RB
hansen	17	2	only	only	RB
hansen	17	3	that	that	DT
hansen	17	4	,	,	,
hansen	17	5	but	but	CC
hansen	17	6	there	there	EX
hansen	17	7	were	be	VBD
hansen	17	8	pre	pre	VBN
hansen	17	9	-	-	JJ
hansen	17	10	built	build	VBN
hansen	17	11	tools	tool	NNS
hansen	17	12	that	that	WDT
hansen	17	13	made	make	VBD
hansen	17	14	it	-PRON-	PRP
hansen	17	15	so	so	RB
hansen	17	16	almost	almost	RB
hansen	17	17	anyone	anyone	NN
hansen	17	18	could	could	MD
hansen	17	19	do	do	VB
hansen	17	20	it	-PRON-	PRP
hansen	17	21	.	.	.
hansen	18	1	For	for	IN
hansen	18	2	a	a	DT
hansen	18	3	group	group	NN
hansen	18	4	of	of	IN
hansen	18	5	people	people	NNS
hansen	18	6	whose	whose	WP$
hansen	18	7	project	project	NN
hansen	18	8	was	be	VBD
hansen	18	9	on	on	IN
hansen	18	10	life	life	NN
hansen	18	11	support	support	NN
hansen	18	12	because	because	IN
hansen	18	13	we	-PRON-	PRP
hansen	18	14	could	could	MD
hansen	18	15	not	not	RB
hansen	18	16	get	get	VB
hansen	18	17	the	the	DT
hansen	18	18	categorization	categorization	NN
hansen	18	19	data	datum	NNS
hansen	18	20	we	-PRON-	PRP
hansen	18	21	needed	need	VBD
hansen	18	22	,	,	,
hansen	18	23	machine	machine	NN
hansen	18	24	learning	learning	NN
hansen	18	25	began	begin	VBD
hansen	18	26	to	to	TO
hansen	18	27	look	look	VB
hansen	18	28	like	like	IN
hansen	18	29	our	-PRON-	PRP$
hansen	18	30	only	only	JJ
hansen	18	31	potential	potential	JJ
hansen	18	32	savior	savior	NN
hansen	18	33	.	.	.
hansen	19	1	So	so	RB
hansen	19	2	,	,	,
hansen	19	3	machine	machine	NN
hansen	19	4	learning	learning	NN
hansen	19	5	is	be	VBZ
hansen	19	6	what	what	WP
hansen	19	7	we	-PRON-	PRP
hansen	19	8	did	do	VBD
hansen	19	9	.	.	.
hansen	20	1	Comment	comment	NN
hansen	20	2	by	by	IN
hansen	20	3	Mark	Mark	NNP
hansen	20	4	Dehmlow	Dehmlow	NNP
hansen	20	5	:	:	:
hansen	20	6	Maybe	maybe	RB
hansen	20	7	pursued	pursue	VBN
hansen	20	8	?	?	.
hansen	21	1	Is	be	VBZ
hansen	21	2	machine	machine	NN
hansen	21	3	learning	learn	VBG
hansen	21	4	an	an	DT
hansen	21	5	action	action	NN
hansen	21	6	or	or	CC
hansen	21	7	a	a	DT
hansen	21	8	technique	technique	NN
hansen	21	9	?	?	.
hansen	22	1	Comment	comment	NN
hansen	22	2	by	by	IN
hansen	22	3	Daniel	Daniel	NNP
hansen	22	4	Johnson	Johnson	NNP
hansen	22	5	:	:	:
hansen	22	6	"	"	``
hansen	22	7	used	use	VBN
hansen	22	8	"	"	''
hansen	22	9	would	would	MD
hansen	22	10	work	work	VB
hansen	22	11	too	too	RB
hansen	22	12	.	.	.
hansen	23	1	I	-PRON-	PRP
hansen	23	2	will	will	MD
hansen	23	3	not	not	RB
hansen	23	4	go	go	VB
hansen	23	5	too	too	RB
hansen	23	6	deep	deep	RB
hansen	23	7	into	into	IN
hansen	23	8	the	the	DT
hansen	23	9	actual	actual	JJ
hansen	23	10	process	process	NN
hansen	23	11	,	,	,
hansen	23	12	but	but	CC
hansen	23	13	I	-PRON-	PRP
hansen	23	14	will	will	MD
hansen	23	15	give	give	VB
hansen	23	16	a	a	DT
hansen	23	17	brief	brief	JJ
hansen	23	18	outline	outline	NN
hansen	23	19	of	of	IN
hansen	23	20	the	the	DT
hansen	23	21	techniques	technique	NNS
hansen	23	22	we	-PRON-	PRP
hansen	23	23	employed	employ	VBD
hansen	23	24	.	.	.
hansen	24	1	Machine	machine	NN
hansen	24	2	-	-	HYPH
hansen	24	3	learning	learning	NN
hansen	24	4	-	-	HYPH
hansen	24	5	based	base	VBN
hansen	24	6	categorization	categorization	NN
hansen	24	7	needs	need	VBZ
hansen	24	8	data	datum	NNS
hansen	24	9	to	to	TO
hansen	24	10	classify	classify	VB
hansen	24	11	,	,	,
hansen	24	12	which	which	WDT
hansen	24	13	in	in	IN
hansen	24	14	our	-PRON-	PRP$
hansen	24	15	case	case	NN
hansen	24	16	were	be	VBD
hansen	24	17	mathematics	mathematic	NNS
hansen	24	18	publications	publication	NNS
hansen	24	19	.	.	.
hansen	25	1	While	while	IN
hansen	25	2	this	this	DT
hansen	25	3	can	can	MD
hansen	25	4	be	be	VB
hansen	25	5	done	do	VBN
hansen	25	6	with	with	IN
hansen	25	7	titles	title	NNS
hansen	25	8	and	and	CC
hansen	25	9	abstracts	abstract	NNS
hansen	25	10	we	-PRON-	PRP
hansen	25	11	wanted	want	VBD
hansen	25	12	to	to	TO
hansen	25	13	provide	provide	VB
hansen	25	14	the	the	DT
hansen	25	15	machine	machine	NN
hansen	25	16	with	with	IN
hansen	25	17	as	as	RB
hansen	25	18	much	much	JJ
hansen	25	19	data	datum	NNS
hansen	25	20	as	as	IN
hansen	25	21	we	-PRON-	PRP
hansen	25	22	could	could	MD
hansen	25	23	,	,	,
hansen	25	24	so	so	RB
hansen	25	25	we	-PRON-	PRP
hansen	25	26	decided	decide	VBD
hansen	25	27	to	to	TO
hansen	25	28	work	work	VB
hansen	25	29	with	with	IN
hansen	25	30	full	full	JJ
hansen	25	31	-	-	HYPH
hansen	25	32	text	text	NN
hansen	25	33	articles	article	NNS
hansen	25	34	.	.	.
hansen	26	1	Since	since	IN
hansen	26	2	we	-PRON-	PRP
hansen	26	3	were	be	VBD
hansen	26	4	at	at	IN
hansen	26	5	the	the	DT
hansen	26	6	University	University	NNP
hansen	26	7	of	of	IN
hansen	26	8	Wisconsin	Wisconsin	NNP
hansen	26	9	at	at	IN
hansen	26	10	the	the	DT
hansen	26	11	time	time	NN
hansen	26	12	,	,	,
hansen	26	13	we	-PRON-	PRP
hansen	26	14	were	be	VBD
hansen	26	15	able	able	JJ
hansen	26	16	to	to	TO
hansen	26	17	connect	connect	VB
hansen	26	18	with	with	IN
hansen	26	19	the	the	DT
hansen	26	20	team	team	NN
hansen	26	21	behind	behind	IN
hansen	26	22	GeoDeepDive[footnoteRef:1	GeoDeepDive[footnoteRef:1	NNS
hansen	26	23	]	]	-RRB-
hansen	26	24	who	who	WP
hansen	26	25	have	have	VBP
hansen	26	26	agreements	agreement	NNS
hansen	26	27	with	with	IN
hansen	26	28	many	many	JJ
hansen	26	29	publishers	publisher	NNS
hansen	26	30	to	to	TO
hansen	26	31	provide	provide	VB
hansen	26	32	the	the	DT
hansen	26	33	full	full	JJ
hansen	26	34	text	text	NN
hansen	26	35	of	of	IN
hansen	26	36	articles	article	NNS
hansen	26	37	for	for	IN
hansen	26	38	text	text	NN
hansen	26	39	and	and	CC
hansen	26	40	data	datum	NNS
hansen	26	41	mining	mining	NN
hansen	26	42	research	research	NN
hansen	26	43	(	(	-LRB-
hansen	26	44	“	"	``
hansen	26	45	GeoDeepDive	geodeepdive	NN
hansen	26	46	:	:	:
hansen	26	47	Project	Project	NNP
hansen	26	48	Overview	Overview	NNP
hansen	26	49	”	"	''
hansen	26	50	n.d	n.d	NNP
hansen	26	51	.	.	NNP
hansen	26	52	)	)	-RRB-
hansen	26	53	.	.	.
hansen	27	1	GeoDeepDive	geodeepdive	NN
hansen	27	2	provided	provide	VBD
hansen	27	3	us	-PRON-	PRP
hansen	27	4	with	with	IN
hansen	27	5	the	the	DT
hansen	27	6	full	full	JJ
hansen	27	7	text	text	NN
hansen	27	8	of	of	IN
hansen	27	9	22,397	22,397	CD
hansen	27	10	mathematics	mathematic	NNS
hansen	27	11	articles	article	NNS
hansen	27	12	which	which	WDT
hansen	27	13	we	-PRON-	PRP
hansen	27	14	used	use	VBD
hansen	27	15	as	as	IN
hansen	27	16	our	-PRON-	PRP$
hansen	27	17	corpus	corpus	NN
hansen	27	18	.	.	.
hansen	28	1	In	in	IN
hansen	28	2	order	order	NN
hansen	28	3	to	to	TO
hansen	28	4	classify	classify	VB
hansen	28	5	these	these	DT
hansen	28	6	articles	article	NNS
hansen	28	7	,	,	,
hansen	28	8	which	which	WDT
hansen	28	9	were	be	VBD
hansen	28	10	already	already	RB
hansen	28	11	pre	pre	VBN
hansen	28	12	-	-	VBN
hansen	28	13	processed	process	VBN
hansen	28	14	by	by	IN
hansen	28	15	GeoDeepDive	GeoDeepDive	NNP
hansen	28	16	with	with	IN
hansen	28	17	CoreNLP,[footnoteRef:2	CoreNLP,[footnoteRef:2	NNPS
hansen	28	18	]	]	-RRB-
hansen	28	19	we	-PRON-	PRP
hansen	28	20	first	first	RB
hansen	28	21	used	use	VBD
hansen	28	22	the	the	DT
hansen	28	23	Python	Python	NNP
hansen	28	24	package	package	NN
hansen	28	25	Gensim[footnoteRef:3	Gensim[footnoteRef:3	NNP
hansen	28	26	]	]	-RRB-
hansen	28	27	to	to	TO
hansen	28	28	process	process	VB
hansen	28	29	the	the	DT
hansen	28	30	articles	article	NNS
hansen	28	31	into	into	IN
hansen	28	32	a	a	DT
hansen	28	33	Python	Python	NNP
hansen	28	34	-	-	HYPH
hansen	28	35	friendly	friendly	JJ
hansen	28	36	format	format	NN
hansen	28	37	and	and	CC
hansen	28	38	to	to	TO
hansen	28	39	remove	remove	VB
hansen	28	40	stopwords	stopword	NNS
hansen	28	41	.	.	.
hansen	29	1	Then	then	RB
hansen	29	2	we	-PRON-	PRP
hansen	29	3	randomly	randomly	RB
hansen	29	4	sampled	sample	VBD
hansen	29	5	⅓	⅓	NNS
hansen	29	6	of	of	IN
hansen	29	7	the	the	DT
hansen	29	8	corpus	corpus	NNP
hansen	29	9	to	to	TO
hansen	29	10	create	create	VB
hansen	29	11	a	a	DT
hansen	29	12	topic	topic	JJ
hansen	29	13	model	model	NN
hansen	29	14	using	use	VBG
hansen	29	15	the	the	DT
hansen	29	16	MALLET[footnoteRef:4	mallet[footnoteref:4	NN
hansen	29	17	]	]	-RRB-
hansen	29	18	topic	topic	NN
hansen	29	19	modeling	modeling	NN
hansen	29	20	tool	tool	NN
hansen	29	21	.	.	.
hansen	30	1	Finally	finally	RB
hansen	30	2	,	,	,
hansen	30	3	we	-PRON-	PRP
hansen	30	4	applied	apply	VBD
hansen	30	5	the	the	DT
hansen	30	6	model	model	NN
hansen	30	7	to	to	IN
hansen	30	8	the	the	DT
hansen	30	9	remaining	remain	VBG
hansen	30	10	articles	article	NNS
hansen	30	11	in	in	IN
hansen	30	12	our	-PRON-	PRP$
hansen	30	13	corpus	corpus	NN
hansen	30	14	.	.	.
hansen	31	1	We	-PRON-	PRP
hansen	31	2	then	then	RB
hansen	31	3	coded	code	VBD
hansen	31	4	the	the	DT
hansen	31	5	words	word	NNS
hansen	31	6	within	within	IN
hansen	31	7	the	the	DT
hansen	31	8	generated	generate	VBN
hansen	31	9	topics	topic	NNS
hansen	31	10	to	to	IN
hansen	31	11	subfields	subfield	NNS
hansen	31	12	within	within	IN
hansen	31	13	mathematics	mathematic	NNS
hansen	31	14	and	and	CC
hansen	31	15	used	use	VBD
hansen	31	16	those	those	DT
hansen	31	17	codes	code	NNS
hansen	31	18	to	to	IN
hansen	31	19	assign	assign	VB
hansen	31	20	articles	article	NNS
hansen	31	21	a	a	DT
hansen	31	22	subfield	subfield	JJ
hansen	31	23	category	category	NN
hansen	31	24	.	.	.
hansen	32	1	In	in	IN
hansen	32	2	order	order	NN
hansen	32	3	to	to	TO
hansen	32	4	make	make	VB
hansen	32	5	sure	sure	JJ
hansen	32	6	our	-PRON-	PRP$
hansen	32	7	results	result	NNS
hansen	32	8	were	be	VBD
hansen	32	9	not	not	RB
hansen	32	10	just	just	RB
hansen	32	11	a	a	DT
hansen	32	12	one	one	CD
hansen	32	13	-	-	HYPH
hansen	32	14	off	off	NN
hansen	32	15	,	,	,
hansen	32	16	we	-PRON-	PRP
hansen	32	17	repeated	repeat	VBD
hansen	32	18	this	this	DT
hansen	32	19	process	process	NN
hansen	32	20	multiple	multiple	JJ
hansen	32	21	times	time	NNS
hansen	32	22	and	and	CC
hansen	32	23	checked	check	VBD
hansen	32	24	for	for	IN
hansen	32	25	variance	variance	NN
hansen	32	26	in	in	IN
hansen	32	27	the	the	DT
hansen	32	28	results	result	NNS
hansen	32	29	.	.	.
hansen	33	1	There	there	EX
hansen	33	2	was	be	VBD
hansen	33	3	none	none	NN
hansen	33	4	,	,	,
hansen	33	5	the	the	DT
hansen	33	6	results	result	NNS
hansen	33	7	were	be	VBD
hansen	33	8	uniformly	uniformly	RB
hansen	33	9	poor	poor	JJ
hansen	33	10	.	.	.
hansen	34	1	[	[	-LRB-
hansen	34	2	1	1	CD
hansen	34	3	:	:	:
hansen	34	4	https://geodeepdive.org/	https://geodeepdive.org/	LS
hansen	34	5	]	]	-RRB-
hansen	34	6	[	[	-LRB-
hansen	34	7	2	2	CD
hansen	34	8	:	:	:
hansen	34	9	https://stanfordnlp.github.io/CoreNLP/	https://stanfordnlp.github.io/CoreNLP/	NNS
hansen	34	10	]	]	-RRB-
hansen	34	11	[	[	-LRB-
hansen	34	12	3	3	CD
hansen	34	13	:	:	:
hansen	34	14	https://radimrehurek.com/gensim/	https://radimrehurek.com/gensim/	NN
hansen	34	15	]	]	-RRB-
hansen	34	16	[	[	-LRB-
hansen	34	17	4	4	CD
hansen	34	18	:	:	SYM
hansen	34	19	http://mallet.cs.umass.edu/topics.php	http://mallet.cs.umass.edu/topics.php	LS
hansen	34	20	]	]	-RRB-
hansen	34	21	That	that	DT
hansen	34	22	might	may	MD
hansen	34	23	not	not	RB
hansen	34	24	be	be	VB
hansen	34	25	entirely	entirely	RB
hansen	34	26	fair	fair	JJ
hansen	34	27	.	.	.
hansen	35	1	There	there	EX
hansen	35	2	were	be	VBD
hansen	35	3	interesting	interesting	JJ
hansen	35	4	aspects	aspect	NNS
hansen	35	5	to	to	IN
hansen	35	6	the	the	DT
hansen	35	7	results	result	NNS
hansen	35	8	of	of	IN
hansen	35	9	the	the	DT
hansen	35	10	topic	topic	NN
hansen	35	11	modeling	modeling	NN
hansen	35	12	,	,	,
hansen	35	13	but	but	CC
hansen	35	14	when	when	WRB
hansen	35	15	it	-PRON-	PRP
hansen	35	16	came	come	VBD
hansen	35	17	to	to	IN
hansen	35	18	categorization	categorization	NN
hansen	35	19	they	-PRON-	PRP
hansen	35	20	were	be	VBD
hansen	35	21	useless	useless	JJ
hansen	35	22	.	.	.
hansen	36	1	Of	of	IN
hansen	36	2	the	the	DT
hansen	36	3	subfield	subfield	NN
hansen	36	4	codes	code	NNS
hansen	36	5	assigned	assign	VBN
hansen	36	6	to	to	IN
hansen	36	7	articles	article	NNS
hansen	36	8	,	,	,
hansen	36	9	only	only	RB
hansen	36	10	two	two	CD
hansen	36	11	were	be	VBD
hansen	36	12	ever	ever	RB
hansen	36	13	the	the	DT
hansen	36	14	dominant	dominant	JJ
hansen	36	15	result	result	NN
hansen	36	16	for	for	IN
hansen	36	17	any	any	DT
hansen	36	18	given	give	VBN
hansen	36	19	article	article	NN
hansen	36	20	:	:	:
hansen	36	21	Graph	Graph	NNP
hansen	36	22	Theory	Theory	NNP
hansen	36	23	and	and	CC
hansen	36	24	Undefined	Undefined	NNP
hansen	36	25	,	,	,
hansen	36	26	which	which	WDT
hansen	36	27	does	do	VBZ
hansen	36	28	not	not	RB
hansen	36	29	really	really	RB
hansen	36	30	tell	tell	VB
hansen	36	31	the	the	DT
hansen	36	32	whole	whole	JJ
hansen	36	33	story	story	NN
hansen	36	34	as	as	IN
hansen	36	35	Undefined	Undefined	NNP
hansen	36	36	was	be	VBD
hansen	36	37	the	the	DT
hansen	36	38	run	run	NN
hansen	36	39	-	-	HYPH
hansen	36	40	away	away	RP
hansen	36	41	winner	winner	NN
hansen	36	42	in	in	IN
hansen	36	43	the	the	DT
hansen	36	44	article	article	NN
hansen	36	45	classification	classification	NN
hansen	36	46	race	race	NN
hansen	36	47	with	with	IN
hansen	36	48	more	more	JJR
hansen	36	49	than	than	IN
hansen	36	50	70	70	CD
hansen	36	51	%	%	NN
hansen	36	52	of	of	IN
hansen	36	53	articles	article	NNS
hansen	36	54	classified	classify	VBN
hansen	36	55	as	as	IN
hansen	36	56	Undefined	undefined	JJ
hansen	36	57	in	in	IN
hansen	36	58	each	each	DT
hansen	36	59	run	run	NN
hansen	36	60	,	,	,
hansen	36	61	including	include	VBG
hansen	36	62	one	one	CD
hansen	36	63	for	for	IN
hansen	36	64	which	which	WDT
hansen	36	65	it	-PRON-	PRP
hansen	36	66	hit	hit	VBD
hansen	36	67	95	95	CD
hansen	36	68	%	%	NN
hansen	36	69	.	.	.
hansen	37	1	The	the	DT
hansen	37	2	topics	topic	NNS
hansen	37	3	generated	generate	VBN
hansen	37	4	by	by	IN
hansen	37	5	MALLET	MALLET	NNP
hansen	37	6	were	be	VBD
hansen	37	7	often	often	RB
hansen	37	8	plagued	plague	VBN
hansen	37	9	by	by	IN
hansen	37	10	gibberish	gibberish	NN
hansen	37	11	caused	cause	VBN
hansen	37	12	by	by	IN
hansen	37	13	equations	equation	NNS
hansen	37	14	in	in	IN
hansen	37	15	the	the	DT
hansen	37	16	mathematics	mathematic	NNS
hansen	37	17	articles	article	NNS
hansen	37	18	and	and	CC
hansen	37	19	there	there	EX
hansen	37	20	was	be	VBD
hansen	37	21	at	at	RB
hansen	37	22	least	least	RBS
hansen	37	23	one	one	CD
hansen	37	24	topic	topic	NN
hansen	37	25	in	in	IN
hansen	37	26	each	each	DT
hansen	37	27	run	run	NN
hansen	37	28	that	that	WDT
hansen	37	29	was	be	VBD
hansen	37	30	filled	fill	VBN
hansen	37	31	with	with	IN
hansen	37	32	the	the	DT
hansen	37	33	names	name	NNS
hansen	37	34	of	of	IN
hansen	37	35	months	month	NNS
hansen	37	36	and	and	CC
hansen	37	37	locations	location	NNS
hansen	37	38	.	.	.
hansen	38	1	Add	add	VB
hansen	38	2	how	how	WRB
hansen	38	3	the	the	DT
hansen	38	4	technical	technical	JJ
hansen	38	5	language	language	NN
hansen	38	6	of	of	IN
hansen	38	7	mathematics	mathematic	NNS
hansen	38	8	is	be	VBZ
hansen	38	9	filled	fill	VBN
hansen	38	10	with	with	IN
hansen	38	11	words	word	NNS
hansen	38	12	that	that	WDT
hansen	38	13	have	have	VBP
hansen	38	14	non	non	JJ
hansen	38	15	-	-	JJ
hansen	38	16	technical	technical	JJ
hansen	38	17	definitions	definition	NNS
hansen	38	18	(	(	-LRB-
hansen	38	19	for	for	IN
hansen	38	20	example	example	NN
hansen	38	21	,	,	,
hansen	38	22	map	map	NN
hansen	38	23	or	or	CC
hansen	38	24	space	space	NN
hansen	38	25	)	)	-RRB-
hansen	38	26	,	,	,
hansen	38	27	or	or	CC
hansen	38	28	words	word	NNS
hansen	38	29	which	which	WDT
hansen	38	30	have	have	VBP
hansen	38	31	their	-PRON-	PRP$
hansen	38	32	own	own	JJ
hansen	38	33	subfield	subfield	NN
hansen	38	34	-	-	HYPH
hansen	38	35	specific	specific	JJ
hansen	38	36	meanings	meaning	NNS
hansen	38	37	(	(	-LRB-
hansen	38	38	such	such	JJ
hansen	38	39	as	as	IN
hansen	38	40	homomorphism	homomorphism	NN
hansen	38	41	or	or	CC
hansen	38	42	degree	degree	NN
hansen	38	43	)	)	-RRB-
hansen	38	44	,	,	,
hansen	38	45	both	both	DT
hansen	38	46	of	of	IN
hansen	38	47	which	which	WDT
hansen	38	48	frustrate	frustrate	VBP
hansen	38	49	attempts	attempt	NNS
hansen	38	50	to	to	TO
hansen	38	51	code	code	VB
hansen	38	52	a	a	DT
hansen	38	53	subfield	subfield	NN
hansen	38	54	.	.	.
hansen	39	1	These	these	DT
hansen	39	2	issues	issue	NNS
hansen	39	3	help	help	VBP
hansen	39	4	make	make	VB
hansen	39	5	it	-PRON-	PRP
hansen	39	6	clear	clear	JJ
hansen	39	7	why	why	WRB
hansen	39	8	so	so	RB
hansen	39	9	many	many	JJ
hansen	39	10	articles	article	NNS
hansen	39	11	ended	end	VBD
hansen	39	12	up	up	RP
hansen	39	13	as	as	IN
hansen	39	14	“	"	``
hansen	39	15	Undefined	undefined	JJ
hansen	39	16	.	.	.
hansen	39	17	”	"	''
hansen	39	18	Even	even	RB
hansen	39	19	for	for	IN
hansen	39	20	the	the	DT
hansen	39	21	one	one	CD
hansen	39	22	subfield	subfield	NN
hansen	39	23	which	which	WDT
hansen	39	24	had	have	VBD
hansen	39	25	a	a	DT
hansen	39	26	unique	unique	JJ
hansen	39	27	enough	enough	JJ
hansen	39	28	vocabulary	vocabulary	NN
hansen	39	29	for	for	IN
hansen	39	30	our	-PRON-	PRP$
hansen	39	31	topic	topic	JJ
hansen	39	32	model	model	NN
hansen	39	33	to	to	TO
hansen	39	34	partially	partially	RB
hansen	39	35	be	be	VB
hansen	39	36	able	able	JJ
hansen	39	37	to	to	TO
hansen	39	38	identify	identify	VB
hansen	39	39	,	,	,
hansen	39	40	Graph	Graph	NNP
hansen	39	41	Theory	Theory	NNP
hansen	39	42	,	,	,
hansen	39	43	the	the	DT
hansen	39	44	results	result	NNS
hansen	39	45	were	be	VBD
hansen	39	46	marginally	marginally	RB
hansen	39	47	positive	positive	JJ
hansen	39	48	at	at	IN
hansen	39	49	best	good	JJS
hansen	39	50	.	.	.
hansen	40	1	We	-PRON-	PRP
hansen	40	2	were	be	VBD
hansen	40	3	able	able	JJ
hansen	40	4	to	to	TO
hansen	40	5	obtain	obtain	VB
hansen	40	6	Mathematical	Mathematical	NNP
hansen	40	7	Subject	Subject	NNP
hansen	40	8	Classification	Classification	NNP
hansen	40	9	(	(	-LRB-
hansen	40	10	MSC	MSC	NNP
hansen	40	11	)	)	-RRB-
hansen	40	12	values	value	VBZ
hansen	40	13	for	for	IN
hansen	40	14	around	around	IN
hansen	40	15	10	10	CD
hansen	40	16	%	%	NN
hansen	40	17	of	of	IN
hansen	40	18	our	-PRON-	PRP$
hansen	40	19	corpus	corpus	NN
hansen	40	20	.	.	.
hansen	41	1	When	when	WRB
hansen	41	2	we	-PRON-	PRP
hansen	41	3	compared	compare	VBD
hansen	41	4	the	the	DT
hansen	41	5	articles	article	NNS
hansen	41	6	we	-PRON-	PRP
hansen	41	7	categorized	categorize	VBD
hansen	41	8	as	as	IN
hansen	41	9	Graph	Graph	NNP
hansen	41	10	Theory	Theory	NNP
hansen	41	11	to	to	IN
hansen	41	12	the	the	DT
hansen	41	13	articles	article	NNS
hansen	41	14	which	which	WDT
hansen	41	15	had	have	VBD
hansen	41	16	been	be	VBN
hansen	41	17	assigned	assign	VBN
hansen	41	18	the	the	DT
hansen	41	19	MSC	MSC	NNP
hansen	41	20	value	value	NN
hansen	41	21	for	for	IN
hansen	41	22	Graph	Graph	NNP
hansen	41	23	Theory	Theory	NNP
hansen	41	24	(	(	-LRB-
hansen	41	25	05Cxx	05Cxx	NNP
hansen	41	26	)	)	-RRB-
hansen	41	27	,	,	,
hansen	41	28	we	-PRON-	PRP
hansen	41	29	found	find	VBD
hansen	41	30	we	-PRON-	PRP
hansen	41	31	had	have	VBD
hansen	41	32	a	a	DT
hansen	41	33	textbook	textbook	NN
hansen	41	34	recall	recall	NN
hansen	41	35	-	-	HYPH
hansen	41	36	versus	versus	IN
hansen	41	37	-	-	HYPH
hansen	41	38	precision	precision	NN
hansen	41	39	problem	problem	NN
hansen	41	40	.	.	.
hansen	42	1	We	-PRON-	PRP
hansen	42	2	could	could	MD
hansen	42	3	either	either	RB
hansen	42	4	correctly	correctly	RB
hansen	42	5	categorize	categorize	VB
hansen	42	6	nearly	nearly	RB
hansen	42	7	all	all	DT
hansen	42	8	of	of	IN
hansen	42	9	the	the	DT
hansen	42	10	Graph	Graph	NNP
hansen	42	11	Theory	Theory	NNP
hansen	42	12	articles	article	VBZ
hansen	42	13	with	with	IN
hansen	42	14	a	a	DT
hansen	42	15	very	very	RB
hansen	42	16	high	high	JJ
hansen	42	17	rate	rate	NN
hansen	42	18	of	of	IN
hansen	42	19	false	false	JJ
hansen	42	20	positives	positive	NNS
hansen	42	21	(	(	-LRB-
hansen	42	22	high	high	JJ
hansen	42	23	recall	recall	NN
hansen	42	24	and	and	CC
hansen	42	25	low	low	JJ
hansen	42	26	precision	precision	NN
hansen	42	27	)	)	-RRB-
hansen	42	28	or	or	CC
hansen	42	29	we	-PRON-	PRP
hansen	42	30	could	could	MD
hansen	42	31	almost	almost	RB
hansen	42	32	never	never	RB
hansen	42	33	incorrectly	incorrectly	RB
hansen	42	34	categorize	categorize	VB
hansen	42	35	an	an	DT
hansen	42	36	article	article	NN
hansen	42	37	as	as	IN
hansen	42	38	Graph	Graph	NNP
hansen	42	39	Theory	Theory	NNP
hansen	42	40	,	,	,
hansen	42	41	but	but	CC
hansen	42	42	miss	miss	VB
hansen	42	43	over	over	IN
hansen	42	44	30	30	CD
hansen	42	45	%	%	NN
hansen	42	46	that	that	IN
hansen	42	47	we	-PRON-	PRP
hansen	42	48	should	should	MD
hansen	42	49	have	have	VB
hansen	42	50	categorized	categorize	VBN
hansen	42	51	as	as	IN
hansen	42	52	Graph	Graph	NNP
hansen	42	53	Theory	Theory	NNP
hansen	42	54	(	(	-LRB-
hansen	42	55	high	high	JJ
hansen	42	56	precision	precision	NN
hansen	42	57	and	and	CC
hansen	42	58	low	low	JJ
hansen	42	59	recall	recall	NN
hansen	42	60	)	)	-RRB-
hansen	42	61	.	.	.
hansen	43	1	Needless	needless	JJ
hansen	43	2	to	to	TO
hansen	43	3	say	say	VB
hansen	43	4	,	,	,
hansen	43	5	we	-PRON-	PRP
hansen	43	6	were	be	VBD
hansen	43	7	not	not	RB
hansen	43	8	able	able	JJ
hansen	43	9	to	to	TO
hansen	43	10	create	create	VB
hansen	43	11	the	the	DT
hansen	43	12	temporal	temporal	JJ
hansen	43	13	subfield	subfield	NN
hansen	43	14	network	network	NN
hansen	43	15	I	-PRON-	PRP
hansen	43	16	had	have	VBD
hansen	43	17	imagined	imagine	VBN
hansen	43	18	.	.	.
hansen	44	1	While	while	IN
hansen	44	2	we	-PRON-	PRP
hansen	44	3	could	could	MD
hansen	44	4	reasonably	reasonably	RB
hansen	44	5	claim	claim	VB
hansen	44	6	that	that	IN
hansen	44	7	we	-PRON-	PRP
hansen	44	8	learned	learn	VBD
hansen	44	9	very	very	RB
hansen	44	10	interesting	interesting	JJ
hansen	44	11	things	thing	NNS
hansen	44	12	about	about	IN
hansen	44	13	the	the	DT
hansen	44	14	language	language	NN
hansen	44	15	of	of	IN
hansen	44	16	mathematics	mathematic	NNS
hansen	44	17	and	and	CC
hansen	44	18	its	-PRON-	PRP$
hansen	44	19	subfields	subfield	NNS
hansen	44	20	,	,	,
hansen	44	21	we	-PRON-	PRP
hansen	44	22	could	could	MD
hansen	44	23	not	not	RB
hansen	44	24	claim	claim	VB
hansen	44	25	we	-PRON-	PRP
hansen	44	26	even	even	RB
hansen	44	27	came	come	VBD
hansen	44	28	close	close	RB
hansen	44	29	to	to	IN
hansen	44	30	automatically	automatically	RB
hansen	44	31	categorizing	categorize	VBG
hansen	44	32	mathematics	mathematic	NNS
hansen	44	33	articles	article	NNS
hansen	44	34	.	.	.
hansen	45	1	When	when	WRB
hansen	45	2	we	-PRON-	PRP
hansen	45	3	had	have	VBD
hansen	45	4	to	to	TO
hansen	45	5	report	report	VB
hansen	45	6	back	back	RP
hansen	45	7	on	on	IN
hansen	45	8	our	-PRON-	PRP$
hansen	45	9	work	work	NN
hansen	45	10	at	at	IN
hansen	45	11	the	the	DT
hansen	45	12	end	end	NN
hansen	45	13	of	of	IN
hansen	45	14	the	the	DT
hansen	45	15	course	course	NN
hansen	45	16	,	,	,
hansen	45	17	our	-PRON-	PRP$
hansen	45	18	main	main	JJ
hansen	45	19	result	result	NN
hansen	45	20	was	be	VBD
hansen	45	21	that	that	IN
hansen	45	22	basic	basic	JJ
hansen	45	23	,	,	,
hansen	45	24	off	off	IN
hansen	45	25	-	-	HYPH
hansen	45	26	the	the	DT
hansen	45	27	-	-	HYPH
hansen	45	28	shelf	shelf	NN
hansen	45	29	topic	topic	NN
hansen	45	30	modelling	modelling	NN
hansen	45	31	does	do	VBZ
hansen	45	32	not	not	RB
hansen	45	33	work	work	VB
hansen	45	34	well	well	RB
hansen	45	35	when	when	WRB
hansen	45	36	it	-PRON-	PRP
hansen	45	37	comes	come	VBZ
hansen	45	38	to	to	IN
hansen	45	39	highly	highly	RB
hansen	45	40	technical	technical	JJ
hansen	45	41	articles	article	NNS
hansen	45	42	from	from	IN
hansen	45	43	subjects	subject	NNS
hansen	45	44	like	like	IN
hansen	45	45	mathematics	mathematic	NNS
hansen	45	46	.	.	.
hansen	46	1	It	-PRON-	PRP
hansen	46	2	was	be	VBD
hansen	46	3	also	also	RB
hansen	46	4	a	a	DT
hansen	46	5	welcome	welcome	JJ
hansen	46	6	lesson	lesson	NN
hansen	46	7	in	in	IN
hansen	46	8	not	not	RB
hansen	46	9	believing	believe	VBG
hansen	46	10	the	the	DT
hansen	46	11	hype	hype	NN
hansen	46	12	of	of	IN
hansen	46	13	machine	machine	NN
hansen	46	14	learning	learning	NN
hansen	46	15	,	,	,
hansen	46	16	even	even	RB
hansen	46	17	when	when	WRB
hansen	46	18	a	a	DT
hansen	46	19	project	project	NN
hansen	46	20	looks	look	VBZ
hansen	46	21	exactly	exactly	RB
hansen	46	22	like	like	IN
hansen	46	23	the	the	DT
hansen	46	24	kind	kind	JJ
hansen	46	25	machine	machine	NN
hansen	46	26	learning	learning	NN
hansen	46	27	was	be	VBD
hansen	46	28	supposed	suppose	VBN
hansen	46	29	to	to	TO
hansen	46	30	excel	excel	VB
hansen	46	31	at	at	IN
hansen	46	32	solving	solve	VBG
hansen	46	33	.	.	.
hansen	47	1	While	while	IN
hansen	47	2	we	-PRON-	PRP
hansen	47	3	had	have	VBD
hansen	47	4	a	a	DT
hansen	47	5	hammer	hammer	NN
hansen	47	6	and	and	CC
hansen	47	7	our	-PRON-	PRP$
hansen	47	8	problem	problem	NN
hansen	47	9	looked	look	VBD
hansen	47	10	like	like	IN
hansen	47	11	a	a	DT
hansen	47	12	nail	nail	NN
hansen	47	13	,	,	,
hansen	47	14	it	-PRON-	PRP
hansen	47	15	seemed	seem	VBD
hansen	47	16	that	that	IN
hansen	47	17	the	the	DT
hansen	47	18	former	former	JJ
hansen	47	19	was	be	VBD
hansen	47	20	a	a	DT
hansen	47	21	ball	ball	NN
hansen	47	22	peen	peen	NN
hansen	47	23	and	and	CC
hansen	47	24	the	the	DT
hansen	47	25	latter	latter	JJ
hansen	47	26	a	a	DT
hansen	47	27	railroad	railroad	NN
hansen	47	28	tie	tie	NN
hansen	47	29	.	.	.
hansen	48	1	In	in	IN
hansen	48	2	the	the	DT
hansen	48	3	end	end	NN
hansen	48	4	,	,	,
hansen	48	5	even	even	RB
hansen	48	6	in	in	IN
hansen	48	7	the	the	DT
hansen	48	8	land	land	NN
hansen	48	9	of	of	IN
hansen	48	10	hammers	hammer	NNS
hansen	48	11	and	and	CC
hansen	48	12	nails	nail	NNS
hansen	48	13	,	,	,
hansen	48	14	the	the	DT
hansen	48	15	tool	tool	NN
hansen	48	16	has	have	VBZ
hansen	48	17	to	to	TO
hansen	48	18	match	match	VB
hansen	48	19	the	the	DT
hansen	48	20	task	task	NN
hansen	48	21	.	.	.
hansen	49	1	Though	though	IN
hansen	49	2	we	-PRON-	PRP
hansen	49	3	failed	fail	VBD
hansen	49	4	to	to	TO
hansen	49	5	accomplish	accomplish	VB
hansen	49	6	automated	automate	VBN
hansen	49	7	categorization	categorization	NN
hansen	49	8	of	of	IN
hansen	49	9	mathematics	mathematic	NNS
hansen	49	10	,	,	,
hansen	49	11	we	-PRON-	PRP
hansen	49	12	were	be	VBD
hansen	49	13	dilettantes	dilettante	NNS
hansen	49	14	in	in	IN
hansen	49	15	the	the	DT
hansen	49	16	world	world	NN
hansen	49	17	of	of	IN
hansen	49	18	machine	machine	NN
hansen	49	19	learning	learning	NN
hansen	49	20	.	.	.
hansen	50	1	I	-PRON-	PRP
hansen	50	2	believe	believe	VBP
hansen	50	3	our	-PRON-	PRP$
hansen	50	4	project	project	NN
hansen	50	5	is	be	VBZ
hansen	50	6	a	a	DT
hansen	50	7	good	good	JJ
hansen	50	8	example	example	NN
hansen	50	9	of	of	IN
hansen	50	10	how	how	WRB
hansen	50	11	machine	machine	NN
hansen	50	12	learning	learning	NN
hansen	50	13	is	be	VBZ
hansen	50	14	still	still	RB
hansen	50	15	a	a	DT
hansen	50	16	long	long	JJ
hansen	50	17	way	way	NN
hansen	50	18	from	from	IN
hansen	50	19	being	be	VBG
hansen	50	20	the	the	DT
hansen	50	21	magic	magic	JJ
hansen	50	22	tool	tool	NN
hansen	50	23	as	as	IN
hansen	50	24	some	some	DT
hansen	50	25	,	,	,
hansen	50	26	though	though	IN
hansen	50	27	not	not	RB
hansen	50	28	all	all	DT
hansen	50	29	(	(	-LRB-
hansen	50	30	Rahimi	Rahimi	NNP
hansen	50	31	and	and	CC
hansen	50	32	Recht	Recht	NNP
hansen	50	33	2017	2017	CD
hansen	50	34	)	)	-RRB-
hansen	50	35	,	,	,
hansen	50	36	have	have	VBP
hansen	50	37	portrayed	portray	VBN
hansen	50	38	it	-PRON-	PRP
hansen	50	39	.	.	.
hansen	51	1	Let	let	VB
hansen	51	2	us	-PRON-	PRP
hansen	51	3	look	look	VB
hansen	51	4	at	at	IN
hansen	51	5	what	what	WP
hansen	51	6	happens	happen	VBZ
hansen	51	7	when	when	WRB
hansen	51	8	smarter	smart	JJR
hansen	51	9	and	and	CC
hansen	51	10	more	more	RBR
hansen	51	11	capable	capable	JJ
hansen	51	12	minds	mind	NNS
hansen	51	13	tackle	tackle	VBP
hansen	51	14	the	the	DT
hansen	51	15	problem	problem	NN
hansen	51	16	of	of	IN
hansen	51	17	classifying	classify	VBG
hansen	51	18	mathematics	mathematic	NNS
hansen	51	19	and	and	CC
hansen	51	20	other	other	JJ
hansen	51	21	highly	highly	RB
hansen	51	22	technical	technical	JJ
hansen	51	23	subjects	subject	NNS
hansen	51	24	using	use	VBG
hansen	51	25	advanced	advanced	JJ
hansen	51	26	machine	machine	NN
hansen	51	27	learning	learning	NN
hansen	51	28	techniques	technique	NNS
hansen	51	29	.	.	.
hansen	52	1	Comment	comment	NN
hansen	52	2	by	by	IN
hansen	52	3	Mark	Mark	NNP
hansen	52	4	Dehmlow	Dehmlow	NNP
hansen	52	5	:	:	:
hansen	52	6	Perhaps	perhaps	RB
hansen	52	7	a	a	DT
hansen	52	8	different	different	JJ
hansen	52	9	word	word	NN
hansen	52	10	choice	choice	NN
hansen	52	11	?	?	.
hansen	53	1	Solving	solving	NN
hansen	53	2	refers	refer	VBZ
hansen	53	3	to	to	IN
hansen	53	4	the	the	DT
hansen	53	5	word	word	NN
hansen	53	6	project	project	NN
hansen	53	7	here	here	RB
hansen	53	8	.	.	.
hansen	54	1	Maybe	maybe	RB
hansen	54	2	something	something	NN
hansen	54	3	like	like	IN
hansen	54	4	problem	problem	NN
hansen	54	5	instead	instead	RB
hansen	54	6	of	of	IN
hansen	54	7	project	project	NN
hansen	54	8	OR	or	CC
hansen	54	9	supporting	support	VBG
hansen	54	10	instead	instead	RB
hansen	54	11	of	of	IN
hansen	54	12	solving	solve	VBG
hansen	54	13	?	?	.
hansen	55	1	Comment	comment	NN
hansen	55	2	by	by	IN
hansen	55	3	Daniel	Daniel	NNP
hansen	55	4	Johnson	Johnson	NNP
hansen	55	5	:	:	:
hansen	55	6	+1	+1	:
hansen	55	7	Finding	find	VBG
hansen	55	8	the	the	DT
hansen	55	9	Right	Right	NNP
hansen	55	10	Hammer	Hammer	NNP
hansen	55	11	To	to	TO
hansen	55	12	illustrate	illustrate	VB
hansen	55	13	the	the	DT
hansen	55	14	quest	quest	NN
hansen	55	15	to	to	TO
hansen	55	16	find	find	VB
hansen	55	17	the	the	DT
hansen	55	18	right	right	JJ
hansen	55	19	hammer	hammer	NN
hansen	55	20	I	-PRON-	PRP
hansen	55	21	am	be	VBP
hansen	55	22	going	go	VBG
hansen	55	23	to	to	TO
hansen	55	24	focus	focus	VB
hansen	55	25	on	on	IN
hansen	55	26	three	three	CD
hansen	55	27	different	different	JJ
hansen	55	28	projects	project	NNS
hansen	55	29	that	that	WDT
hansen	55	30	tackled	tackle	VBD
hansen	55	31	the	the	DT
hansen	55	32	automated	automate	VBN
hansen	55	33	categorization	categorization	NN
hansen	55	34	of	of	IN
hansen	55	35	highly	highly	RB
hansen	55	36	technical	technical	JJ
hansen	55	37	content	content	NN
hansen	55	38	,	,	,
hansen	55	39	two	two	CD
hansen	55	40	of	of	IN
hansen	55	41	which	which	WDT
hansen	55	42	also	also	RB
hansen	55	43	attempted	attempt	VBD
hansen	55	44	to	to	TO
hansen	55	45	categorize	categorize	VB
hansen	55	46	mathematical	mathematical	JJ
hansen	55	47	content	content	NN
hansen	55	48	and	and	CC
hansen	55	49	one	one	CD
hansen	55	50	that	that	WDT
hansen	55	51	looked	look	VBD
hansen	55	52	to	to	TO
hansen	55	53	categorize	categorize	VB
hansen	55	54	scholarly	scholarly	NNP
hansen	55	55	works	work	NNS
hansen	55	56	in	in	IN
hansen	55	57	general	general	JJ
hansen	55	58	.	.	.
hansen	56	1	These	these	DT
hansen	56	2	three	three	CD
hansen	56	3	projects	project	NNS
hansen	56	4	provide	provide	VBP
hansen	56	5	examples	example	NNS
hansen	56	6	of	of	IN
hansen	56	7	many	many	JJ
hansen	56	8	of	of	IN
hansen	56	9	the	the	DT
hansen	56	10	approaches	approach	NNS
hansen	56	11	and	and	CC
hansen	56	12	practices	practice	NNS
hansen	56	13	employed	employ	VBN
hansen	56	14	by	by	IN
hansen	56	15	experts	expert	NNS
hansen	56	16	in	in	IN
hansen	56	17	automated	automated	JJ
hansen	56	18	classification	classification	NN
hansen	56	19	and	and	CC
hansen	56	20	demonstrate	demonstrate	VB
hansen	56	21	the	the	DT
hansen	56	22	two	two	CD
hansen	56	23	main	main	JJ
hansen	56	24	paths	path	NNS
hansen	56	25	that	that	WDT
hansen	56	26	these	these	DT
hansen	56	27	types	type	NNS
hansen	56	28	of	of	IN
hansen	56	29	projects	project	NNS
hansen	56	30	follow	follow	VBP
hansen	56	31	to	to	TO
hansen	56	32	accomplish	accomplish	VB
hansen	56	33	their	-PRON-	PRP$
hansen	56	34	goals	goal	NNS
hansen	56	35	.	.	.
hansen	57	1	Since	since	IN
hansen	57	2	we	-PRON-	PRP
hansen	57	3	have	have	VBP
hansen	57	4	been	be	VBN
hansen	57	5	discussing	discuss	VBG
hansen	57	6	mathematics	mathematic	NNS
hansen	57	7	,	,	,
hansen	57	8	let	let	VB
hansen	57	9	us	-PRON-	PRP
hansen	57	10	start	start	VB
hansen	57	11	with	with	IN
hansen	57	12	those	those	DT
hansen	57	13	two	two	CD
hansen	57	14	projects	project	NNS
hansen	57	15	.	.	.
hansen	58	1	Both	both	DT
hansen	58	2	projects	project	NNS
hansen	58	3	began	begin	VBD
hansen	58	4	because	because	IN
hansen	58	5	the	the	DT
hansen	58	6	participants	participant	NNS
hansen	58	7	were	be	VBD
hansen	58	8	struggling	struggle	VBG
hansen	58	9	to	to	TO
hansen	58	10	categorize	categorize	VB
hansen	58	11	mathematics	mathematic	NNS
hansen	58	12	publications	publication	NNS
hansen	58	13	so	so	IN
hansen	58	14	they	-PRON-	PRP
hansen	58	15	would	would	MD
hansen	58	16	be	be	VB
hansen	58	17	properly	properly	RB
hansen	58	18	indexed	index	VBN
hansen	58	19	and	and	CC
hansen	58	20	searchable	searchable	JJ
hansen	58	21	in	in	IN
hansen	58	22	digital	digital	JJ
hansen	58	23	mathematics	mathematic	NNS
hansen	58	24	databases	database	NNS
hansen	58	25	:	:	:
hansen	58	26	the	the	DT
hansen	58	27	Czech	Czech	NNP
hansen	58	28	Digital	Digital	NNP
hansen	58	29	Mathematics	Mathematics	NNP
hansen	58	30	Library	Library	NNP
hansen	58	31	(	(	-LRB-
hansen	58	32	DML	DML	NNP
hansen	58	33	-	-	HYPH
hansen	58	34	CZ)[footnoteRef:5	CZ)[footnoteRef:5	NNP
hansen	58	35	]	]	-RRB-
hansen	58	36	and	and	CC
hansen	58	37	NUMDAM[footnoteRef:6	numdam[footnoteref:6	RB
hansen	58	38	]	]	-RRB-
hansen	58	39	in	in	IN
hansen	58	40	the	the	DT
hansen	58	41	case	case	NN
hansen	58	42	of	of	IN
hansen	58	43	Radim	Radim	NNP
hansen	58	44	Řehůřek	řehůřek	CD
hansen	58	45	and	and	CC
hansen	58	46	Petr	Petr	VBN
hansen	58	47	Sojka	sojka	RB
hansen	58	48	(	(	-LRB-
hansen	58	49	Řehůřek	Řehůřek	NNP
hansen	58	50	and	and	CC
hansen	58	51	Sojka	sojka	RB
hansen	58	52	2008	2008	CD
hansen	58	53	)	)	-RRB-
hansen	58	54	,	,	,
hansen	58	55	and	and	CC
hansen	58	56	Zentralblatt	Zentralblatt	NNP
hansen	58	57	MATH	MATH	NNP
hansen	58	58	(	(	-LRB-
hansen	58	59	zbMath)[footnoteRef:7	zbmath)[footnoteref:7	FW
hansen	58	60	]	]	-RRB-
hansen	58	61	in	in	IN
hansen	58	62	the	the	DT
hansen	58	63	case	case	NN
hansen	58	64	of	of	IN
hansen	58	65	Simon	Simon	NNP
hansen	58	66	Barthel	Barthel	NNP
hansen	58	67	,	,	,
hansen	58	68	Sascha	Sascha	NNP
hansen	58	69	Tönnies	tönnie	NNS
hansen	58	70	,	,	,
hansen	58	71	and	and	CC
hansen	58	72	Wolf	Wolf	NNP
hansen	58	73	-	-	HYPH
hansen	58	74	Tilo	Tilo	NNP
hansen	58	75	Balke	Balke	NNP
hansen	58	76	(	(	-LRB-
hansen	58	77	Barthel	Barthel	NNP
hansen	58	78	,	,	,
hansen	58	79	Tönnies	Tönnies	NNP
hansen	58	80	,	,	,
hansen	58	81	and	and	CC
hansen	58	82	Balke	Balke	NNP
hansen	58	83	2013	2013	CD
hansen	58	84	)	)	-RRB-
hansen	58	85	.	.	.
hansen	59	1	All	all	DT
hansen	59	2	of	of	IN
hansen	59	3	these	these	DT
hansen	59	4	databases	database	NNS
hansen	59	5	rely	rely	VBP
hansen	59	6	on	on	IN
hansen	59	7	the	the	DT
hansen	59	8	aforementioned	aforementioned	JJ
hansen	59	9	MSC[footnoteRef:8	MSC[footnoteRef:8	NNP
hansen	59	10	]	]	-RRB-
hansen	59	11	to	to	TO
hansen	59	12	aid	aid	VB
hansen	59	13	in	in	IN
hansen	59	14	indexing	indexing	NN
hansen	59	15	and	and	CC
hansen	59	16	retrieval	retrieval	NN
hansen	59	17	,	,	,
hansen	59	18	and	and	CC
hansen	59	19	so	so	RB
hansen	59	20	their	-PRON-	PRP$
hansen	59	21	goal	goal	NN
hansen	59	22	was	be	VBD
hansen	59	23	to	to	TO
hansen	59	24	automate	automate	VB
hansen	59	25	the	the	DT
hansen	59	26	assignment	assignment	NN
hansen	59	27	of	of	IN
hansen	59	28	MSC	MSC	NNP
hansen	59	29	values	value	NNS
hansen	59	30	to	to	TO
hansen	59	31	lower	lower	VB
hansen	59	32	the	the	DT
hansen	59	33	time	time	NN
hansen	59	34	and	and	CC
hansen	59	35	labor	labor	NN
hansen	59	36	cost	cost	NN
hansen	59	37	of	of	IN
hansen	59	38	requiring	require	VBG
hansen	59	39	humans	human	NNS
hansen	59	40	to	to	TO
hansen	59	41	do	do	VB
hansen	59	42	this	this	DT
hansen	59	43	task	task	NN
hansen	59	44	.	.	.
hansen	60	1	The	the	DT
hansen	60	2	main	main	JJ
hansen	60	3	differences	difference	NNS
hansen	60	4	between	between	IN
hansen	60	5	their	-PRON-	PRP$
hansen	60	6	tasks	task	NNS
hansen	60	7	related	relate	VBN
hansen	60	8	to	to	IN
hansen	60	9	the	the	DT
hansen	60	10	number	number	NN
hansen	60	11	of	of	IN
hansen	60	12	documents	document	NNS
hansen	60	13	they	-PRON-	PRP
hansen	60	14	were	be	VBD
hansen	60	15	working	work	VBG
hansen	60	16	with	with	IN
hansen	60	17	(	(	-LRB-
hansen	60	18	thousands	thousand	NNS
hansen	60	19	for	for	IN
hansen	60	20	Řehůřek	Řehůřek	NNP
hansen	60	21	and	and	CC
hansen	60	22	Sojka	sojka	JJ
hansen	60	23	and	and	CC
hansen	60	24	millions	million	NNS
hansen	60	25	for	for	IN
hansen	60	26	Barthel	Barthel	NNP
hansen	60	27	,	,	,
hansen	60	28	Tönnies	Tönnies	NNP
hansen	60	29	,	,	,
hansen	60	30	and	and	CC
hansen	60	31	Balke	Balke	NNP
hansen	60	32	)	)	-RRB-
hansen	60	33	,	,	,
hansen	60	34	the	the	DT
hansen	60	35	amount	amount	NN
hansen	60	36	of	of	IN
hansen	60	37	the	the	DT
hansen	60	38	works	work	NNS
hansen	60	39	available	available	JJ
hansen	60	40	(	(	-LRB-
hansen	60	41	full	full	JJ
hansen	60	42	text	text	NN
hansen	60	43	for	for	IN
hansen	60	44	Řehůřek	Řehůřek	NNP
hansen	60	45	and	and	CC
hansen	60	46	Sojka	sojka	RB
hansen	60	47	,	,	,
hansen	60	48	and	and	CC
hansen	60	49	titles	title	NNS
hansen	60	50	,	,	,
hansen	60	51	authors	author	NNS
hansen	60	52	,	,	,
hansen	60	53	and	and	CC
hansen	60	54	abstracts	abstract	NNS
hansen	60	55	for	for	IN
hansen	60	56	Barthel	Barthel	NNP
hansen	60	57	,	,	,
hansen	60	58	Tönnies	Tönnies	NNP
hansen	60	59	,	,	,
hansen	60	60	and	and	CC
hansen	60	61	Balke	Balke	NNP
hansen	60	62	)	)	-RRB-
hansen	60	63	,	,	,
hansen	60	64	and	and	CC
hansen	60	65	the	the	DT
hansen	60	66	quality	quality	NN
hansen	60	67	of	of	IN
hansen	60	68	the	the	DT
hansen	60	69	data	datum	NNS
hansen	60	70	(	(	-LRB-
hansen	60	71	mostly	mostly	RB
hansen	60	72	OCR	ocr	NN
hansen	60	73	scans	scan	NNS
hansen	60	74	for	for	IN
hansen	60	75	Řehůřek	Řehůřek	NNP
hansen	60	76	and	and	CC
hansen	60	77	Sojka	sojka	RB
hansen	60	78	and	and	CC
hansen	60	79	mostly	mostly	RB
hansen	60	80	TeX	TeX	VBD
hansen	60	81	for	for	IN
hansen	60	82	Barthel	Barthel	NNP
hansen	60	83	,	,	,
hansen	60	84	Tönnies	Tönnies	NNP
hansen	60	85	,	,	,
hansen	60	86	and	and	CC
hansen	60	87	Balke	Balke	NNP
hansen	60	88	)	)	-RRB-
hansen	60	89	.	.	.
hansen	61	1	Even	even	RB
hansen	61	2	with	with	IN
hansen	61	3	these	these	DT
hansen	61	4	differences	difference	NNS
hansen	61	5	,	,	,
hansen	61	6	both	both	DT
hansen	61	7	projects	project	NNS
hansen	61	8	took	take	VBD
hansen	61	9	a	a	DT
hansen	61	10	similar	similar	JJ
hansen	61	11	approach	approach	NN
hansen	61	12	,	,	,
hansen	61	13	and	and	CC
hansen	61	14	it	-PRON-	PRP
hansen	61	15	is	be	VBZ
hansen	61	16	the	the	DT
hansen	61	17	first	first	JJ
hansen	61	18	of	of	IN
hansen	61	19	the	the	DT
hansen	61	20	two	two	CD
hansen	61	21	main	main	JJ
hansen	61	22	pathways	pathway	NNS
hansen	61	23	toward	toward	IN
hansen	61	24	classification	classification	NN
hansen	61	25	I	-PRON-	PRP
hansen	61	26	spoke	speak	VBD
hansen	61	27	of	of	IN
hansen	61	28	earlier	early	RBR
hansen	61	29	:	:	:
hansen	61	30	using	use	VBG
hansen	61	31	a	a	DT
hansen	61	32	predetermined	predetermined	JJ
hansen	61	33	taxonomy	taxonomy	NN
hansen	61	34	and	and	CC
hansen	61	35	a	a	DT
hansen	61	36	set	set	NN
hansen	61	37	of	of	IN
hansen	61	38	pre	pre	JJ
hansen	61	39	-	-	JJ
hansen	61	40	categorized	categorized	JJ
hansen	61	41	data	datum	NNS
hansen	61	42	to	to	TO
hansen	61	43	build	build	VB
hansen	61	44	a	a	DT
hansen	61	45	machine	machine	NN
hansen	61	46	learning	learn	VBG
hansen	61	47	categorizer	categorizer	NN
hansen	61	48	.	.	.
hansen	62	1	[	[	-LRB-
hansen	62	2	5	5	CD
hansen	62	3	:	:	:
hansen	62	4	https://dml.cz/	https://dml.cz/	NNP
hansen	62	5	]	]	-RRB-
hansen	62	6	[	[	-LRB-
hansen	62	7	6	6	CD
hansen	62	8	:	:	:
hansen	62	9	http://www.numdam.org/	http://www.numdam.org/	NN
hansen	62	10	]	]	-RRB-
hansen	62	11	[	[	-LRB-
hansen	62	12	7	7	CD
hansen	62	13	:	:	:
hansen	62	14	https://zbmath.org/	https://zbmath.org/	LS
hansen	62	15	]	]	-RRB-
hansen	62	16	[	[	-LRB-
hansen	62	17	8	8	CD
hansen	62	18	:	:	:
hansen	62	19	Mathematical	Mathematical	NNP
hansen	62	20	Subject	Subject	NNP
hansen	62	21	Classification	Classification	NNP
hansen	62	22	(	(	-LRB-
hansen	62	23	MSC	MSC	NNP
hansen	62	24	)	)	-RRB-
hansen	62	25	values	value	NNS
hansen	62	26	in	in	IN
hansen	62	27	MathSciNet	MathSciNet	NNP
hansen	62	28	and	and	CC
hansen	62	29	zbMath	zbMath	NNP
hansen	62	30	are	be	VBP
hansen	62	31	a	a	DT
hansen	62	32	particularly	particularly	RB
hansen	62	33	interesting	interesting	JJ
hansen	62	34	categorization	categorization	NN
hansen	62	35	set	set	VBN
hansen	62	36	to	to	TO
hansen	62	37	work	work	VB
hansen	62	38	with	with	IN
hansen	62	39	as	as	IN
hansen	62	40	they	-PRON-	PRP
hansen	62	41	are	be	VBP
hansen	62	42	assigned	assign	VBN
hansen	62	43	and	and	CC
hansen	62	44	reviewed	review	VBN
hansen	62	45	by	by	IN
hansen	62	46	a	a	DT
hansen	62	47	subject	subject	JJ
hansen	62	48	area	area	NN
hansen	62	49	expert	expert	NN
hansen	62	50	editor	editor	NN
hansen	62	51	and	and	CC
hansen	62	52	an	an	DT
hansen	62	53	active	active	JJ
hansen	62	54	researcher	researcher	NN
hansen	62	55	in	in	IN
hansen	62	56	the	the	DT
hansen	62	57	same	same	JJ
hansen	62	58	,	,	,
hansen	62	59	or	or	CC
hansen	62	60	closely	closely	RB
hansen	62	61	related	relate	VBN
hansen	62	62	,	,	,
hansen	62	63	subfield	subfield	JJ
hansen	62	64	as	as	IN
hansen	62	65	the	the	DT
hansen	62	66	article	article	NN
hansen	62	67	’s	’s	POS
hansen	62	68	content	content	NN
hansen	62	69	before	before	IN
hansen	62	70	they	-PRON-	PRP
hansen	62	71	are	be	VBP
hansen	62	72	published	publish	VBN
hansen	62	73	.	.	.
hansen	63	1	This	this	DT
hansen	63	2	multi	multi	JJ
hansen	63	3	-	-	JJ
hansen	63	4	step	step	JJ
hansen	63	5	process	process	NN
hansen	63	6	of	of	IN
hansen	63	7	review	review	NN
hansen	63	8	yields	yield	NNS
hansen	63	9	a	a	DT
hansen	63	10	built	build	VBN
hansen	63	11	-	-	HYPH
hansen	63	12	in	in	RP
hansen	63	13	accuracy	accuracy	NN
hansen	63	14	check	check	NN
hansen	63	15	for	for	IN
hansen	63	16	the	the	DT
hansen	63	17	categorization	categorization	NN
hansen	63	18	.	.	.
hansen	63	19	]	]	-RRB-
hansen	64	1	In	in	IN
hansen	64	2	the	the	DT
hansen	64	3	end	end	NN
hansen	64	4	,	,	,
hansen	64	5	while	while	IN
hansen	64	6	both	both	DT
hansen	64	7	projects	project	NNS
hansen	64	8	determined	determine	VBD
hansen	64	9	that	that	IN
hansen	64	10	the	the	DT
hansen	64	11	use	use	NN
hansen	64	12	of	of	IN
hansen	64	13	Support	Support	NNP
hansen	64	14	Vector	Vector	NNP
hansen	64	15	Machines	Machines	NNPS
hansen	64	16	(	(	-LRB-
hansen	64	17	Gandhi	Gandhi	NNP
hansen	64	18	2018)[footnoteRef:9	2018)[footnoteRef:9	NNP
hansen	64	19	]	]	-RRB-
hansen	64	20	provided	provide	VBD
hansen	64	21	the	the	DT
hansen	64	22	best	good	JJS
hansen	64	23	categorization	categorization	NN
hansen	64	24	results	result	NNS
hansen	64	25	,	,	,
hansen	64	26	their	-PRON-	PRP$
hansen	64	27	implementations	implementation	NNS
hansen	64	28	were	be	VBD
hansen	64	29	different	different	JJ
hansen	64	30	.	.	.
hansen	65	1	The	the	DT
hansen	65	2	Řehůřek	Řehůřek	NNP
hansen	65	3	and	and	CC
hansen	65	4	Sojka	sojka	RB
hansen	65	5	SVMs	svm	NNS
hansen	65	6	were	be	VBD
hansen	65	7	trained	train	VBN
hansen	65	8	with	with	IN
hansen	65	9	terms	term	NNS
hansen	65	10	weighted	weight	VBN
hansen	65	11	using	use	VBG
hansen	65	12	augmented	augment	VBN
hansen	65	13	term	term	NN
hansen	65	14	frequency[footnoteRef:10	frequency[footnoteref:10	NN
hansen	65	15	]	]	-RRB-
hansen	65	16	and	and	CC
hansen	65	17	dynamic	dynamic	JJ
hansen	65	18	decision	decision	NN
hansen	65	19	threshold[footnoteRef:11	threshold[footnoteref:11	NN
hansen	65	20	]	]	-RRB-
hansen	65	21	selection	selection	NN
hansen	65	22	using	use	VBG
hansen	65	23	s	s	NNPS
hansen	65	24	-	-	HYPH
hansen	65	25	cut[footnoteRef:12	cut[footnoteRef:12	NNP
hansen	65	26	]	]	-RRB-
hansen	65	27	(	(	-LRB-
hansen	65	28	Řehůřek	Řehůřek	NNP
hansen	65	29	and	and	CC
hansen	65	30	Sojka	sojka	RB
hansen	65	31	2008	2008	CD
hansen	65	32	,	,	,
hansen	65	33	549	549	CD
hansen	65	34	)	)	-RRB-
hansen	65	35	and	and	CC
hansen	65	36	Barthel	Barthel	NNP
hansen	65	37	,	,	,
hansen	65	38	Tönnies	Tönnies	NNP
hansen	65	39	,	,	,
hansen	65	40	and	and	CC
hansen	65	41	Balke	Balke	NNP
hansen	65	42	's	's	POS
hansen	65	43	with	with	IN
hansen	65	44	term	term	NN
hansen	65	45	weighting	weight	VBG
hansen	65	46	using	use	VBG
hansen	65	47	term	term	NN
hansen	65	48	frequency	frequency	NN
hansen	65	49	–	–	:
hansen	65	50	inverse	inverse	NN
hansen	65	51	document	document	NN
hansen	65	52	frequency[footnoteRef:13	frequency[footnoteRef:13	NNP
hansen	65	53	]	]	-RRB-
hansen	65	54	and	and	CC
hansen	65	55	Euclidean	Euclidean	NNP
hansen	65	56	normalization[footnoteRef:14	normalization[footnoteRef:14	''
hansen	65	57	]	]	-RRB-
hansen	65	58	(	(	-LRB-
hansen	65	59	Barthel	Barthel	NNP
hansen	65	60	,	,	,
hansen	65	61	Tönnies	Tönnies	NNP
hansen	65	62	,	,	,
hansen	65	63	and	and	CC
hansen	65	64	Balke	Balke	NNP
hansen	65	65	2013	2013	CD
hansen	65	66	,	,	,
hansen	65	67	88	88	CD
hansen	65	68	)	)	-RRB-
hansen	65	69	,	,	,
hansen	65	70	but	but	CC
hansen	65	71	the	the	DT
hansen	65	72	main	main	JJ
hansen	65	73	difference	difference	NN
hansen	65	74	was	be	VBD
hansen	65	75	how	how	WRB
hansen	65	76	they	-PRON-	PRP
hansen	65	77	handled	handle	VBD
hansen	65	78	formulae	formulae	NNP
hansen	65	79	.	.	.
hansen	66	1	In	in	IN
hansen	66	2	particular	particular	JJ
hansen	66	3	the	the	DT
hansen	66	4	Barthel	Barthel	NNP
hansen	66	5	,	,	,
hansen	66	6	Tönnies	Tönnies	NNP
hansen	66	7	,	,	,
hansen	66	8	and	and	CC
hansen	66	9	Balke	Balke	NNP
hansen	66	10	group	group	NN
hansen	66	11	split	split	VBD
hansen	66	12	their	-PRON-	PRP$
hansen	66	13	corpus	corpus	NN
hansen	66	14	into	into	IN
hansen	66	15	words	word	NNS
hansen	66	16	and	and	CC
hansen	66	17	formulae	formula	NNS
hansen	66	18	and	and	CC
hansen	66	19	mapped	map	VBD
hansen	66	20	them	-PRON-	PRP
hansen	66	21	to	to	TO
hansen	66	22	separate	separate	JJ
hansen	66	23	vectors	vector	NNS
hansen	66	24	which	which	WDT
hansen	66	25	were	be	VBD
hansen	66	26	then	then	RB
hansen	66	27	merged	merge	VBN
hansen	66	28	together	together	RB
hansen	66	29	for	for	IN
hansen	66	30	a	a	DT
hansen	66	31	combined	combine	VBN
hansen	66	32	vector	vector	NN
hansen	66	33	used	use	VBN
hansen	66	34	for	for	IN
hansen	66	35	categorization	categorization	NN
hansen	66	36	.	.	.
hansen	67	1	Řehůřek	Řehůřek	NNP
hansen	67	2	and	and	CC
hansen	67	3	Sojka	Sojka	NNP
hansen	67	4	did	do	VBD
hansen	67	5	not	not	RB
hansen	67	6	differentiate	differentiate	VB
hansen	67	7	between	between	IN
hansen	67	8	words	word	NNS
hansen	67	9	and	and	CC
hansen	67	10	formulae	formula	NNS
hansen	67	11	in	in	IN
hansen	67	12	their	-PRON-	PRP$
hansen	67	13	corpus	corpus	NN
hansen	67	14	,	,	,
hansen	67	15	and	and	CC
hansen	67	16	they	-PRON-	PRP
hansen	67	17	did	do	VBD
hansen	67	18	note	note	VB
hansen	67	19	that	that	IN
hansen	67	20	their	-PRON-	PRP$
hansen	67	21	OCR	ocr	NN
hansen	67	22	scans	scan	NNS
hansen	67	23	’	’	POS
hansen	67	24	poor	poor	JJ
hansen	67	25	handling	handling	NN
hansen	67	26	of	of	IN
hansen	67	27	formulae	formula	NNS
hansen	67	28	could	could	MD
hansen	67	29	have	have	VB
hansen	67	30	hindered	hinder	VBN
hansen	67	31	their	-PRON-	PRP$
hansen	67	32	results	result	NNS
hansen	67	33	(	(	-LRB-
hansen	67	34	Řehůřek	Řehůřek	NNP
hansen	67	35	and	and	CC
hansen	67	36	Sojka	sojka	RB
hansen	67	37	2008	2008	CD
hansen	67	38	,	,	,
hansen	67	39	555	555	CD
hansen	67	40	)	)	-RRB-
hansen	67	41	.	.	.
hansen	68	1	In	in	IN
hansen	68	2	the	the	DT
hansen	68	3	end	end	NN
hansen	68	4	,	,	,
hansen	68	5	not	not	RB
hansen	68	6	having	have	VBG
hansen	68	7	the	the	DT
hansen	68	8	ability	ability	NN
hansen	68	9	to	to	TO
hansen	68	10	handle	handle	VB
hansen	68	11	formulae	formulae	NNP
hansen	68	12	separately	separately	RB
hansen	68	13	did	do	VBD
hansen	68	14	not	not	RB
hansen	68	15	seem	seem	VB
hansen	68	16	to	to	TO
hansen	68	17	matter	matter	VB
hansen	68	18	as	as	IN
hansen	68	19	Řehůřek	Řehůřek	NNP
hansen	68	20	and	and	CC
hansen	68	21	Sojka	Sojka	NNP
hansen	68	22	claimed	claim	VBD
hansen	68	23	microaveraged	microaverage	VBD
hansen	68	24	F₁	F₁	NNP
hansen	68	25	scores	score	NNS
hansen	68	26	of	of	IN
hansen	68	27	89.03	89.03	CD
hansen	68	28	%	%	NN
hansen	68	29	(	(	-LRB-
hansen	68	30	Řehůřek	Řehůřek	NNP
hansen	68	31	and	and	CC
hansen	68	32	Sojka	sojka	RB
hansen	68	33	2008	2008	CD
hansen	68	34	,	,	,
hansen	68	35	549	549	CD
hansen	68	36	)	)	-RRB-
hansen	68	37	when	when	WRB
hansen	68	38	classifying	classify	VBG
hansen	68	39	the	the	DT
hansen	68	40	top	top	JJ
hansen	68	41	level	level	NN
hansen	68	42	MSC	MSC	NNP
hansen	68	43	category	category	NN
hansen	68	44	with	with	IN
hansen	68	45	their	-PRON-	PRP$
hansen	68	46	best	good	JJS
hansen	68	47	performing	perform	VBG
hansen	68	48	SVM	SVM	NNP
hansen	68	49	.	.	.
hansen	69	1	When	when	WRB
hansen	69	2	this	this	DT
hansen	69	3	is	be	VBZ
hansen	69	4	compared	compare	VBN
hansen	69	5	to	to	IN
hansen	69	6	the	the	DT
hansen	69	7	microaveraged	microaverage	VBN
hansen	69	8	F₁	F₁	NNP
hansen	69	9	of	of	IN
hansen	69	10	67.3	67.3	CD
hansen	69	11	%	%	NN
hansen	69	12	obtained	obtain	VBN
hansen	69	13	by	by	IN
hansen	69	14	Barthel	Barthel	NNP
hansen	69	15	,	,	,
hansen	69	16	Tönnies	Tönnies	NNP
hansen	69	17	,	,	,
hansen	69	18	and	and	CC
hansen	69	19	Balke	Balke	NNP
hansen	69	20	(	(	-LRB-
hansen	69	21	Barthel	Barthel	NNP
hansen	69	22	,	,	,
hansen	69	23	Tönnies	Tönnies	NNP
hansen	69	24	,	,	,
hansen	69	25	and	and	CC
hansen	69	26	Balke	Balke	NNP
hansen	69	27	2013	2013	CD
hansen	69	28	,	,	,
hansen	69	29	88	88	CD
hansen	69	30	)	)	-RRB-
hansen	69	31	,	,	,
hansen	69	32	it	-PRON-	PRP
hansen	69	33	would	would	MD
hansen	69	34	seem	seem	VB
hansen	69	35	that	that	IN
hansen	69	36	either	either	CC
hansen	69	37	Řehůřek	Řehůřek	NNP
hansen	69	38	’s	’s	,
hansen	69	39	and	and	CC
hansen	69	40	Sojka	sojka	RB
hansen	69	41	’s	’	VBZ
hansen	69	42	implementation	implementation	NN
hansen	69	43	of	of	IN
hansen	69	44	SVMs	svm	NNS
hansen	69	45	or	or	CC
hansen	69	46	their	-PRON-	PRP$
hansen	69	47	access	access	NN
hansen	69	48	to	to	IN
hansen	69	49	full	full	JJ
hansen	69	50	-	-	HYPH
hansen	69	51	text	text	NN
hansen	69	52	led	lead	VBD
hansen	69	53	to	to	IN
hansen	69	54	a	a	DT
hansen	69	55	clear	clear	JJ
hansen	69	56	advantage	advantage	NN
hansen	69	57	.	.	.
hansen	70	1	This	this	DT
hansen	70	2	advantage	advantage	NN
hansen	70	3	becomes	become	VBZ
hansen	70	4	less	less	RBR
hansen	70	5	clear	clear	JJ
hansen	70	6	when	when	WRB
hansen	70	7	one	one	PRP
hansen	70	8	takes	take	VBZ
hansen	70	9	into	into	IN
hansen	70	10	account	account	NN
hansen	70	11	that	that	IN
hansen	70	12	Řehůřek	Řehůřek	NNP
hansen	70	13	and	and	CC
hansen	70	14	Sojka	Sojka	NNP
hansen	70	15	were	be	VBD
hansen	70	16	only	only	RB
hansen	70	17	working	work	VBG
hansen	70	18	with	with	IN
hansen	70	19	top	top	JJ
hansen	70	20	level	level	NN
hansen	70	21	MSCs	msc	NNS
hansen	70	22	where	where	WRB
hansen	70	23	they	-PRON-	PRP
hansen	70	24	had	have	VBD
hansen	70	25	at	at	RB
hansen	70	26	least	least	RBS
hansen	70	27	30	30	CD
hansen	70	28	(	(	-LRB-
hansen	70	29	60	60	CD
hansen	70	30	in	in	IN
hansen	70	31	the	the	DT
hansen	70	32	case	case	NN
hansen	70	33	of	of	IN
hansen	70	34	their	-PRON-	PRP$
hansen	70	35	best	good	JJS
hansen	70	36	result	result	NN
hansen	70	37	)	)	-RRB-
hansen	70	38	articles	article	NNS
hansen	70	39	,	,	,
hansen	70	40	and	and	CC
hansen	70	41	their	-PRON-	PRP$
hansen	70	42	limited	limited	JJ
hansen	70	43	corpus	corpus	NN
hansen	70	44	meant	mean	VBD
hansen	70	45	that	that	IN
hansen	70	46	many	many	JJ
hansen	70	47	top	top	JJ
hansen	70	48	-	-	HYPH
hansen	70	49	level	level	NN
hansen	70	50	MSC	MSC	NNP
hansen	70	51	categories	category	NNS
hansen	70	52	would	would	MD
hansen	70	53	not	not	RB
hansen	70	54	have	have	VB
hansen	70	55	been	be	VBN
hansen	70	56	included	include	VBN
hansen	70	57	.	.	.
hansen	71	1	Looking	look	VBG
hansen	71	2	at	at	IN
hansen	71	3	the	the	DT
hansen	71	4	work	work	NN
hansen	71	5	done	do	VBN
hansen	71	6	by	by	IN
hansen	71	7	Barthel	Barthel	NNP
hansen	71	8	,	,	,
hansen	71	9	Tönnies	Tönnies	NNP
hansen	71	10	,	,	,
hansen	71	11	and	and	CC
hansen	71	12	Balke	Balke	NNP
hansen	71	13	makes	make	VBZ
hansen	71	14	it	-PRON-	PRP
hansen	71	15	clear	clear	JJ
hansen	71	16	that	that	IN
hansen	71	17	these	these	DT
hansen	71	18	less	less	RBR
hansen	71	19	common	common	JJ
hansen	71	20	MSC	MSC	NNP
hansen	71	21	categories	category	NNS
hansen	71	22	such	such	JJ
hansen	71	23	as	as	IN
hansen	71	24	K	K	NNP
hansen	71	25	-	-	HYPH
hansen	71	26	Theory	Theory	NNP
hansen	71	27	or	or	CC
hansen	71	28	Potential	Potential	NNP
hansen	71	29	Theory	Theory	NNP
hansen	71	30	,	,	,
hansen	71	31	for	for	IN
hansen	71	32	which	which	WDT
hansen	71	33	Barthel	Barthel	NNP
hansen	71	34	,	,	,
hansen	71	35	Tönnies	Tönnies	NNP
hansen	71	36	,	,	,
hansen	71	37	and	and	CC
hansen	71	38	Balke	Balke	NNP
hansen	71	39	achieved	achieve	VBD
hansen	71	40	microaveraged	microaverage	VBD
hansen	71	41	F₁	F₁	NNP
hansen	71	42	measures	measure	NNS
hansen	71	43	of	of	IN
hansen	71	44	18.2	18.2	CD
hansen	71	45	%	%	NN
hansen	71	46	and	and	CC
hansen	71	47	24	24	CD
hansen	71	48	%	%	NN
hansen	71	49	respectively	respectively	RB
hansen	71	50	,	,	,
hansen	71	51	have	have	VBP
hansen	71	52	a	a	DT
hansen	71	53	large	large	JJ
hansen	71	54	impact	impact	NN
hansen	71	55	on	on	IN
hansen	71	56	the	the	DT
hansen	71	57	overall	overall	JJ
hansen	71	58	effectiveness	effectiveness	NN
hansen	71	59	of	of	IN
hansen	71	60	the	the	DT
hansen	71	61	automated	automated	JJ
hansen	71	62	categorization	categorization	NN
hansen	71	63	.	.	.
hansen	72	1	Remember	remember	VB
hansen	72	2	,	,	,
hansen	72	3	this	this	DT
hansen	72	4	is	be	VBZ
hansen	72	5	only	only	RB
hansen	72	6	for	for	IN
hansen	72	7	the	the	DT
hansen	72	8	top	top	JJ
hansen	72	9	level	level	NN
hansen	72	10	of	of	IN
hansen	72	11	MSC	MSC	NNP
hansen	72	12	codes	code	NNS
hansen	72	13	,	,	,
hansen	72	14	and	and	CC
hansen	72	15	the	the	DT
hansen	72	16	work	work	NN
hansen	72	17	of	of	IN
hansen	72	18	Barthel	Barthel	NNP
hansen	72	19	,	,	,
hansen	72	20	Tönnies	Tönnies	NNP
hansen	72	21	,	,	,
hansen	72	22	and	and	CC
hansen	72	23	Balke	Balke	NNP
hansen	72	24	suggests	suggest	VBZ
hansen	72	25	it	-PRON-	PRP
hansen	72	26	would	would	MD
hansen	72	27	get	get	VB
hansen	72	28	worse	bad	JJR
hansen	72	29	when	when	WRB
hansen	72	30	trying	try	VBG
hansen	72	31	to	to	TO
hansen	72	32	apply	apply	VB
hansen	72	33	the	the	DT
hansen	72	34	second	second	JJ
hansen	72	35	and	and	CC
hansen	72	36	third	third	JJ
hansen	72	37	level	level	NN
hansen	72	38	for	for	IN
hansen	72	39	full	full	JJ
hansen	72	40	MSC	MSC	NNP
hansen	72	41	categorization	categorization	NN
hansen	72	42	to	to	IN
hansen	72	43	these	these	DT
hansen	72	44	less	less	RBR
hansen	72	45	-	-	HYPH
hansen	72	46	common	common	JJ
hansen	72	47	categories	category	NNS
hansen	72	48	.	.	.
hansen	73	1	This	this	DT
hansen	73	2	leads	lead	VBZ
hansen	73	3	me	-PRON-	PRP
hansen	73	4	to	to	TO
hansen	73	5	believe	believe	VB
hansen	73	6	that	that	IN
hansen	73	7	in	in	IN
hansen	73	8	the	the	DT
hansen	73	9	case	case	NN
hansen	73	10	of	of	IN
hansen	73	11	categorizing	categorize	VBG
hansen	73	12	highly	highly	RB
hansen	73	13	technical	technical	JJ
hansen	73	14	mathematical	mathematical	JJ
hansen	73	15	works	work	NNS
hansen	73	16	to	to	IN
hansen	73	17	an	an	DT
hansen	73	18	existing	exist	VBG
hansen	73	19	taxonomy	taxonomy	NN
hansen	73	20	,	,	,
hansen	73	21	people	people	NNS
hansen	73	22	have	have	VBP
hansen	73	23	come	come	VBN
hansen	73	24	close	close	RB
hansen	73	25	to	to	IN
hansen	73	26	identifying	identify	VBG
hansen	73	27	the	the	DT
hansen	73	28	overall	overall	JJ
hansen	73	29	size	size	NN
hansen	73	30	of	of	IN
hansen	73	31	the	the	DT
hansen	73	32	machine	machine	NN
hansen	73	33	learning	learning	NN
hansen	73	34	hammer	hammer	NNP
hansen	73	35	,	,	,
hansen	73	36	but	but	CC
hansen	73	37	are	be	VBP
hansen	73	38	still	still	RB
hansen	73	39	a	a	DT
hansen	73	40	long	long	JJ
hansen	73	41	way	way	NN
hansen	73	42	away	away	RB
hansen	73	43	from	from	IN
hansen	73	44	finding	find	VBG
hansen	73	45	the	the	DT
hansen	73	46	right	right	JJ
hansen	73	47	match	match	NN
hansen	73	48	for	for	IN
hansen	73	49	the	the	DT
hansen	73	50	categorization	categorization	NN
hansen	73	51	nail	nail	NN
hansen	73	52	.	.	.
hansen	74	1	[	[	-LRB-
hansen	74	2	9	9	CD
hansen	74	3	:	:	:
hansen	74	4	Support	Support	NNP
hansen	74	5	Vector	Vector	NNP
hansen	74	6	Machines	Machines	NNPS
hansen	74	7	(	(	-LRB-
hansen	74	8	SVMs	svm	NNS
hansen	74	9	)	)	-RRB-
hansen	74	10	are	be	VBP
hansen	74	11	machine	machine	NN
hansen	74	12	learning	learning	NN
hansen	74	13	models	model	NNS
hansen	74	14	which	which	WDT
hansen	74	15	are	be	VBP
hansen	74	16	trained	train	VBN
hansen	74	17	using	use	VBG
hansen	74	18	a	a	DT
hansen	74	19	pre	pre	JJ
hansen	74	20	-	-	JJ
hansen	74	21	classified	classified	JJ
hansen	74	22	corpus	corpus	NN
hansen	74	23	to	to	TO
hansen	74	24	split	split	VB
hansen	74	25	a	a	DT
hansen	74	26	vector	vector	NN
hansen	74	27	space	space	NN
hansen	74	28	into	into	IN
hansen	74	29	a	a	DT
hansen	74	30	set	set	NN
hansen	74	31	of	of	IN
hansen	74	32	differentiated	differentiate	VBN
hansen	74	33	areas	area	NNS
hansen	74	34	(	(	-LRB-
hansen	74	35	or	or	CC
hansen	74	36	categories	category	NNS
hansen	74	37	)	)	-RRB-
hansen	74	38	and	and	CC
hansen	74	39	then	then	RB
hansen	74	40	attempt	attempt	VB
hansen	74	41	to	to	TO
hansen	74	42	classify	classify	VB
hansen	74	43	new	new	JJ
hansen	74	44	items	item	NNS
hansen	74	45	through	through	IN
hansen	74	46	where	where	WRB
hansen	74	47	in	in	IN
hansen	74	48	the	the	DT
hansen	74	49	vector	vector	NN
hansen	74	50	space	space	NN
hansen	74	51	the	the	DT
hansen	74	52	trained	train	VBN
hansen	74	53	model	model	NN
hansen	74	54	places	place	VBZ
hansen	74	55	them	-PRON-	PRP
hansen	74	56	.	.	.
hansen	75	1	For	for	IN
hansen	75	2	a	a	DT
hansen	75	3	more	more	JJR
hansen	75	4	in	in	IN
hansen	75	5	-	-	HYPH
hansen	75	6	depth	depth	NN
hansen	75	7	,	,	,
hansen	75	8	technical	technical	JJ
hansen	75	9	explanation	explanation	NN
hansen	75	10	,	,	,
hansen	75	11	see	see	VBP
hansen	75	12	:	:	:
hansen	75	13	https://towardsdatascience.com/support-vector-machine-introduction-to-machine-learning-algorithms-934a444fca47	https://towardsdatascience.com/support-vector-machine-introduction-to-machine-learning-algorithms-934a444fca47	NNP
hansen	75	14	]	]	-RRB-
hansen	75	15	[	[	-LRB-
hansen	75	16	10	10	CD
hansen	75	17	:	:	:
hansen	75	18	Augmented	augment	VBN
hansen	75	19	term	term	NN
hansen	75	20	frequency	frequency	NN
hansen	75	21	refers	refer	VBZ
hansen	75	22	to	to	IN
hansen	75	23	the	the	DT
hansen	75	24	number	number	NN
hansen	75	25	of	of	IN
hansen	75	26	times	time	NNS
hansen	75	27	a	a	DT
hansen	75	28	term	term	NN
hansen	75	29	occurs	occur	VBZ
hansen	75	30	in	in	IN
hansen	75	31	the	the	DT
hansen	75	32	document	document	NN
hansen	75	33	divided	divide	VBN
hansen	75	34	by	by	IN
hansen	75	35	the	the	DT
hansen	75	36	number	number	NN
hansen	75	37	of	of	IN
hansen	75	38	times	time	NNS
hansen	75	39	the	the	DT
hansen	75	40	most	most	RBS
hansen	75	41	frequent	frequent	JJ
hansen	75	42	occurring	occurring	NN
hansen	75	43	term	term	NN
hansen	75	44	appears	appear	VBZ
hansen	75	45	in	in	IN
hansen	75	46	the	the	DT
hansen	75	47	document	document	NN
hansen	75	48	.	.	.
hansen	75	49	]	]	-RRB-
hansen	76	1	[	[	-LRB-
hansen	76	2	11	11	CD
hansen	76	3	:	:	:
hansen	76	4	The	the	DT
hansen	76	5	decision	decision	NN
hansen	76	6	threshold	threshold	NN
hansen	76	7	is	be	VBZ
hansen	76	8	the	the	DT
hansen	76	9	cut	cut	NN
hansen	76	10	-	-	HYPH
hansen	76	11	off	off	RP
hansen	76	12	for	for	IN
hansen	76	13	how	how	WRB
hansen	76	14	close	close	RB
hansen	76	15	to	to	IN
hansen	76	16	a	a	DT
hansen	76	17	category	category	NN
hansen	76	18	the	the	DT
hansen	76	19	SVM	SVM	NNP
hansen	76	20	must	must	MD
hansen	76	21	determine	determine	VB
hansen	76	22	an	an	DT
hansen	76	23	item	item	NN
hansen	76	24	to	to	TO
hansen	76	25	be	be	VB
hansen	76	26	in	in	IN
hansen	76	27	order	order	NN
hansen	76	28	for	for	IN
hansen	76	29	it	-PRON-	PRP
hansen	76	30	to	to	TO
hansen	76	31	be	be	VB
hansen	76	32	assigned	assign	VBN
hansen	76	33	that	that	DT
hansen	76	34	category	category	NN
hansen	76	35	.	.	.
hansen	77	1	Řehůřek	Řehůřek	NNP
hansen	77	2	and	and	CC
hansen	77	3	Sojka	Sojka	NNP
hansen	77	4	’s	’s	POS
hansen	77	5	work	work	NN
hansen	77	6	varied	vary	VBD
hansen	77	7	this	this	DT
hansen	77	8	threshold	threshold	NN
hansen	77	9	dynamically	dynamically	RB
hansen	77	10	.	.	.
hansen	77	11	]	]	-RRB-
hansen	78	1	[	[	-LRB-
hansen	78	2	12	12	CD
hansen	78	3	:	:	:
hansen	78	4	Score	score	RB
hansen	78	5	-	-	HYPH
hansen	78	6	based	base	VBN
hansen	78	7	local	local	JJ
hansen	78	8	optimization	optimization	NN
hansen	78	9	,	,	,
hansen	78	10	or	or	CC
hansen	78	11	s	s	NN
hansen	78	12	-	-	HYPH
hansen	78	13	cut	cut	VBN
hansen	78	14	,	,	,
hansen	78	15	allows	allow	VBZ
hansen	78	16	a	a	DT
hansen	78	17	machine	machine	NN
hansen	78	18	-	-	HYPH
hansen	78	19	learning	learn	VBG
hansen	78	20	model	model	NN
hansen	78	21	to	to	TO
hansen	78	22	set	set	VB
hansen	78	23	different	different	JJ
hansen	78	24	thresholds	threshold	NNS
hansen	78	25	for	for	IN
hansen	78	26	each	each	DT
hansen	78	27	category	category	NN
hansen	78	28	with	with	IN
hansen	78	29	an	an	DT
hansen	78	30	emphasis	emphasis	NN
hansen	78	31	on	on	IN
hansen	78	32	local	local	JJ
hansen	78	33	,	,	,
hansen	78	34	or	or	CC
hansen	78	35	category	category	NN
hansen	78	36	,	,	,
hansen	78	37	instead	instead	RB
hansen	78	38	of	of	IN
hansen	78	39	global	global	JJ
hansen	78	40	performance	performance	NN
hansen	78	41	.	.	.
hansen	78	42	]	]	-RRB-
hansen	79	1	[	[	-LRB-
hansen	79	2	13	13	CD
hansen	79	3	:	:	:
hansen	79	4	Term	term	NN
hansen	79	5	frequency	frequency	NN
hansen	79	6	–	–	:
hansen	79	7	inverse	inverse	NN
hansen	79	8	document	document	NN
hansen	79	9	frequency	frequency	NN
hansen	79	10	provides	provide	VBZ
hansen	79	11	a	a	DT
hansen	79	12	weight	weight	NN
hansen	79	13	for	for	IN
hansen	79	14	terms	term	NNS
hansen	79	15	depending	depend	VBG
hansen	79	16	on	on	IN
hansen	79	17	how	how	WRB
hansen	79	18	frequently	frequently	RB
hansen	79	19	it	-PRON-	PRP
hansen	79	20	occurs	occur	VBZ
hansen	79	21	across	across	IN
hansen	79	22	the	the	DT
hansen	79	23	corpus	corpus	NN
hansen	79	24	.	.	.
hansen	80	1	A	a	DT
hansen	80	2	term	term	NN
hansen	80	3	which	which	WDT
hansen	80	4	occurs	occur	VBZ
hansen	80	5	rarely	rarely	RB
hansen	80	6	across	across	IN
hansen	80	7	the	the	DT
hansen	80	8	corpus	corpus	NN
hansen	80	9	but	but	CC
hansen	80	10	with	with	IN
hansen	80	11	a	a	DT
hansen	80	12	high	high	JJ
hansen	80	13	frequency	frequency	NN
hansen	80	14	within	within	IN
hansen	80	15	a	a	DT
hansen	80	16	single	single	JJ
hansen	80	17	document	document	NN
hansen	80	18	will	will	MD
hansen	80	19	have	have	VB
hansen	80	20	a	a	DT
hansen	80	21	higher	high	JJR
hansen	80	22	weight	weight	NN
hansen	80	23	when	when	WRB
hansen	80	24	classifying	classify	VBG
hansen	80	25	the	the	DT
hansen	80	26	document	document	NN
hansen	80	27	in	in	IN
hansen	80	28	question	question	NN
hansen	80	29	.	.	.
hansen	80	30	]	]	-RRB-
hansen	81	1	[	[	-LRB-
hansen	81	2	14	14	CD
hansen	81	3	:	:	:
hansen	81	4	A	a	DT
hansen	81	5	Euclidean	euclidean	JJ
hansen	81	6	norm	norm	NN
hansen	81	7	provides	provide	VBZ
hansen	81	8	the	the	DT
hansen	81	9	distance	distance	NN
hansen	81	10	from	from	IN
hansen	81	11	the	the	DT
hansen	81	12	origin	origin	NN
hansen	81	13	to	to	IN
hansen	81	14	a	a	DT
hansen	81	15	point	point	NN
hansen	81	16	in	in	IN
hansen	81	17	an	an	DT
hansen	81	18	n	n	JJ
hansen	81	19	-	-	HYPH
hansen	81	20	dimensional	dimensional	JJ
hansen	81	21	space	space	NN
hansen	81	22	.	.	.
hansen	82	1	It	-PRON-	PRP
hansen	82	2	is	be	VBZ
hansen	82	3	calculated	calculate	VBN
hansen	82	4	by	by	IN
hansen	82	5	taking	take	VBG
hansen	82	6	the	the	DT
hansen	82	7	square	square	JJ
hansen	82	8	root	root	NN
hansen	82	9	of	of	IN
hansen	82	10	the	the	DT
hansen	82	11	sum	sum	NN
hansen	82	12	of	of	IN
hansen	82	13	the	the	DT
hansen	82	14	squares	square	NNS
hansen	82	15	of	of	IN
hansen	82	16	all	all	DT
hansen	82	17	coordinate	coordinate	JJ
hansen	82	18	values	value	NNS
hansen	82	19	.	.	.
hansen	82	20	]	]	-RRB-
hansen	83	1	Now	now	RB
hansen	83	2	let	let	VB
hansen	83	3	us	-PRON-	PRP
hansen	83	4	shift	shift	VB
hansen	83	5	from	from	IN
hansen	83	6	mathematics	mathematics	NN
hansen	83	7	-	-	HYPH
hansen	83	8	specific	specific	JJ
hansen	83	9	categorization	categorization	NN
hansen	83	10	to	to	TO
hansen	83	11	subject	subject	JJ
hansen	83	12	categorization	categorization	NN
hansen	83	13	in	in	IN
hansen	83	14	general	general	JJ
hansen	83	15	and	and	CC
hansen	83	16	look	look	VB
hansen	83	17	at	at	IN
hansen	83	18	the	the	DT
hansen	83	19	work	work	NN
hansen	83	20	Microsoft	Microsoft	NNP
hansen	83	21	has	have	VBZ
hansen	83	22	done	do	VBN
hansen	83	23	assigning	assigning	NN
hansen	83	24	Fields	Fields	NNP
hansen	83	25	of	of	IN
hansen	83	26	Study	Study	NNP
hansen	83	27	(	(	-LRB-
hansen	83	28	FoS	FoS	NNP
hansen	83	29	)	)	-RRB-
hansen	83	30	in	in	IN
hansen	83	31	the	the	DT
hansen	83	32	Microsoft	Microsoft	NNP
hansen	83	33	Academic	Academic	NNP
hansen	83	34	Graph	Graph	NNP
hansen	83	35	(	(	-LRB-
hansen	83	36	MAG	MAG	NNP
hansen	83	37	)	)	-RRB-
hansen	83	38	which	which	WDT
hansen	83	39	is	be	VBZ
hansen	83	40	used	use	VBN
hansen	83	41	to	to	TO
hansen	83	42	create	create	VB
hansen	83	43	their	-PRON-	PRP$
hansen	83	44	Microsoft	Microsoft	NNP
hansen	83	45	Academic	Academic	NNP
hansen	83	46	article	article	NN
hansen	83	47	search	search	NN
hansen	83	48	product	product	NN
hansen	83	49	.	.	.
hansen	84	1	[	[	-LRB-
hansen	84	2	footnoteRef:15	footnoteref:15	CD
hansen	84	3	]	]	-RRB-
hansen	84	4	While	while	IN
hansen	84	5	the	the	DT
hansen	84	6	MAG	MAG	NNP
hansen	84	7	FoS	fos	NN
hansen	84	8	project	project	NN
hansen	84	9	is	be	VBZ
hansen	84	10	also	also	RB
hansen	84	11	attempting	attempt	VBG
hansen	84	12	to	to	TO
hansen	84	13	categorize	categorize	VB
hansen	84	14	articles	article	NNS
hansen	84	15	for	for	IN
hansen	84	16	proper	proper	JJ
hansen	84	17	indexing	indexing	NN
hansen	84	18	and	and	CC
hansen	84	19	search	search	NN
hansen	84	20	,	,	,
hansen	84	21	it	-PRON-	PRP
hansen	84	22	represents	represent	VBZ
hansen	84	23	the	the	DT
hansen	84	24	second	second	JJ
hansen	84	25	path	path	NN
hansen	84	26	which	which	WDT
hansen	84	27	is	be	VBZ
hansen	84	28	taken	take	VBN
hansen	84	29	by	by	IN
hansen	84	30	automated	automate	VBN
hansen	84	31	categorization	categorization	NN
hansen	84	32	projects	project	NNS
hansen	84	33	:	:	:
hansen	84	34	using	use	VBG
hansen	84	35	machine	machine	NN
hansen	84	36	learning	learning	NN
hansen	84	37	techniques	technique	NNS
hansen	84	38	to	to	TO
hansen	84	39	both	both	DT
hansen	84	40	create	create	VB
hansen	84	41	the	the	DT
hansen	84	42	taxonomy	taxonomy	NN
hansen	84	43	and	and	CC
hansen	84	44	to	to	TO
hansen	84	45	classify	classify	VB
hansen	84	46	.	.	.
hansen	85	1	[	[	-LRB-
hansen	85	2	15	15	CD
hansen	85	3	:	:	:
hansen	85	4	https://academic.microsoft.com/	https://academic.microsoft.com/	LS
hansen	85	5	]	]	-RRB-
hansen	85	6	Microsoft	Microsoft	NNP
hansen	85	7	took	take	VBD
hansen	85	8	a	a	DT
hansen	85	9	unique	unique	JJ
hansen	85	10	approach	approach	NN
hansen	85	11	in	in	IN
hansen	85	12	the	the	DT
hansen	85	13	development	development	NN
hansen	85	14	of	of	IN
hansen	85	15	their	-PRON-	PRP$
hansen	85	16	taxonomy	taxonomy	NN
hansen	85	17	.	.	.
hansen	86	1	Instead	instead	RB
hansen	86	2	of	of	IN
hansen	86	3	relying	rely	VBG
hansen	86	4	on	on	IN
hansen	86	5	the	the	DT
hansen	86	6	corpus	corpus	NN
hansen	86	7	of	of	IN
hansen	86	8	articles	article	NNS
hansen	86	9	in	in	IN
hansen	86	10	the	the	DT
hansen	86	11	MAG	MAG	NNP
hansen	86	12	to	to	TO
hansen	86	13	develop	develop	VB
hansen	86	14	it	-PRON-	PRP
hansen	86	15	,	,	,
hansen	86	16	they	-PRON-	PRP
hansen	86	17	relied	rely	VBD
hansen	86	18	primarily	primarily	RB
hansen	86	19	on	on	IN
hansen	86	20	Wikipedia	Wikipedia	NNP
hansen	86	21	for	for	IN
hansen	86	22	its	-PRON-	PRP$
hansen	86	23	creation	creation	NN
hansen	86	24	.	.	.
hansen	87	1	They	-PRON-	PRP
hansen	87	2	generated	generate	VBD
hansen	87	3	an	an	DT
hansen	87	4	initial	initial	JJ
hansen	87	5	seed	seed	NN
hansen	87	6	by	by	IN
hansen	87	7	referencing	reference	VBG
hansen	87	8	the	the	DT
hansen	87	9	Science	Science	NNP
hansen	87	10	Metrix	Metrix	NNP
hansen	87	11	classification	classification	NN
hansen	87	12	scheme[footnoteRef:16	scheme[footnoteref:16	NN
hansen	87	13	]	]	-RRB-
hansen	87	14	and	and	CC
hansen	87	15	a	a	DT
hansen	87	16	couple	couple	NN
hansen	87	17	thousand	thousand	CD
hansen	87	18	FoS	FoS	NNP
hansen	87	19	Wikipedia	Wikipedia	NNP
hansen	87	20	articles	article	NNS
hansen	87	21	they	-PRON-	PRP
hansen	87	22	identified	identify	VBD
hansen	87	23	internally	internally	RB
hansen	87	24	.	.	.
hansen	88	1	They	-PRON-	PRP
hansen	88	2	then	then	RB
hansen	88	3	used	use	VBD
hansen	88	4	an	an	DT
hansen	88	5	iterative	iterative	JJ
hansen	88	6	process	process	NN
hansen	88	7	to	to	TO
hansen	88	8	identify	identify	VB
hansen	88	9	more	more	JJR
hansen	88	10	FoS	fos	NN
hansen	88	11	in	in	IN
hansen	88	12	Wikipedia	Wikipedia	NNP
hansen	88	13	based	base	VBN
hansen	88	14	on	on	IN
hansen	88	15	whether	whether	IN
hansen	88	16	they	-PRON-	PRP
hansen	88	17	were	be	VBD
hansen	88	18	linked	link	VBN
hansen	88	19	to	to	IN
hansen	88	20	Wikipedia	Wikipedia	NNP
hansen	88	21	articles	article	NNS
hansen	88	22	that	that	WDT
hansen	88	23	were	be	VBD
hansen	88	24	already	already	RB
hansen	88	25	identified	identify	VBN
hansen	88	26	as	as	IN
hansen	88	27	FoS	fos	NN
hansen	88	28	and	and	CC
hansen	88	29	whether	whether	IN
hansen	88	30	the	the	DT
hansen	88	31	new	new	JJ
hansen	88	32	articles	article	NNS
hansen	88	33	represented	represent	VBD
hansen	88	34	valid	valid	JJ
hansen	88	35	entity	entity	NN
hansen	88	36	types	type	NNS
hansen	88	37	–	–	:
hansen	88	38	e.g.	e.g.	RB
hansen	89	1	an	an	DT
hansen	89	2	entity	entity	NN
hansen	89	3	type	type	NN
hansen	89	4	of	of	IN
hansen	89	5	protein	protein	NN
hansen	89	6	would	would	MD
hansen	89	7	be	be	VB
hansen	89	8	added	add	VBN
hansen	89	9	and	and	CC
hansen	89	10	an	an	DT
hansen	89	11	entity	entity	NN
hansen	89	12	type	type	NN
hansen	89	13	of	of	IN
hansen	89	14	person	person	NN
hansen	89	15	would	would	MD
hansen	89	16	be	be	VB
hansen	89	17	excluded	exclude	VBN
hansen	89	18	(	(	-LRB-
hansen	89	19	Shen	Shen	NNP
hansen	89	20	,	,	,
hansen	89	21	Ma	Ma	NNP
hansen	89	22	,	,	,
hansen	89	23	and	and	CC
hansen	89	24	Wang	Wang	NNP
hansen	89	25	2018	2018	CD
hansen	89	26	,	,	,
hansen	89	27	3	3	CD
hansen	89	28	)	)	-RRB-
hansen	89	29	.	.	.
hansen	90	1	This	this	DT
hansen	90	2	work	work	NN
hansen	90	3	allowed	allow	VBD
hansen	90	4	Microsoft	Microsoft	NNP
hansen	90	5	to	to	TO
hansen	90	6	develop	develop	VB
hansen	90	7	a	a	DT
hansen	90	8	list	list	NN
hansen	90	9	of	of	IN
hansen	90	10	more	more	JJR
hansen	90	11	than	than	IN
hansen	90	12	200,000	200,000	CD
hansen	90	13	Fields	field	NNS
hansen	90	14	of	of	IN
hansen	90	15	Study	Study	NNP
hansen	90	16	for	for	IN
hansen	90	17	use	use	NN
hansen	90	18	as	as	IN
hansen	90	19	categories	category	NNS
hansen	90	20	in	in	IN
hansen	90	21	the	the	DT
hansen	90	22	MAG	MAG	NNP
hansen	90	23	.	.	.
hansen	91	1	[	[	-LRB-
hansen	91	2	16	16	CD
hansen	91	3	:	:	:
hansen	91	4	http://science-metrix.com/?q=en/classification	http://science-metrix.com/?q=en/classification	NNP
hansen	91	5	]	]	-RRB-
hansen	91	6	Microsoft	Microsoft	NNP
hansen	91	7	then	then	RB
hansen	91	8	used	use	VBD
hansen	91	9	machine	machine	NN
hansen	91	10	learning	learning	NN
hansen	91	11	techniques	technique	NNS
hansen	91	12	to	to	TO
hansen	91	13	apply	apply	VB
hansen	91	14	these	these	DT
hansen	91	15	FoS	fos	NN
hansen	91	16	to	to	IN
hansen	91	17	their	-PRON-	PRP$
hansen	91	18	corpus	corpus	NN
hansen	91	19	of	of	IN
hansen	91	20	over	over	IN
hansen	91	21	140	140	CD
hansen	91	22	million	million	CD
hansen	91	23	academic	academic	JJ
hansen	91	24	articles	article	NNS
hansen	91	25	.	.	.
hansen	92	1	The	the	DT
hansen	92	2	specific	specific	JJ
hansen	92	3	techniques	technique	NNS
hansen	92	4	are	be	VBP
hansen	92	5	not	not	RB
hansen	92	6	as	as	RB
hansen	92	7	clear	clear	JJ
hansen	92	8	as	as	IN
hansen	92	9	they	-PRON-	PRP
hansen	92	10	were	be	VBD
hansen	92	11	with	with	IN
hansen	92	12	the	the	DT
hansen	92	13	previous	previous	JJ
hansen	92	14	examples	example	NNS
hansen	92	15	,	,	,
hansen	92	16	likely	likely	RB
hansen	92	17	due	due	IN
hansen	92	18	to	to	IN
hansen	92	19	Microsoft	Microsoft	NNP
hansen	92	20	protecting	protect	VBG
hansen	92	21	their	-PRON-	PRP$
hansen	92	22	specific	specific	JJ
hansen	92	23	methods	method	NNS
hansen	92	24	from	from	IN
hansen	92	25	competitors	competitor	NNS
hansen	92	26	,	,	,
hansen	92	27	but	but	CC
hansen	92	28	the	the	DT
hansen	92	29	article	article	NN
hansen	92	30	published	publish	VBN
hansen	92	31	to	to	IN
hansen	92	32	the	the	DT
hansen	92	33	arXiv	arXiv	NNP
hansen	92	34	by	by	IN
hansen	92	35	their	-PRON-	PRP$
hansen	92	36	researchers	researcher	NNS
hansen	92	37	(	(	-LRB-
hansen	92	38	Shen	Shen	NNP
hansen	92	39	,	,	,
hansen	92	40	Ma	Ma	NNP
hansen	92	41	,	,	,
hansen	92	42	and	and	CC
hansen	92	43	Wang	Wang	NNP
hansen	92	44	2018	2018	CD
hansen	92	45	)	)	-RRB-
hansen	92	46	and	and	CC
hansen	92	47	the	the	DT
hansen	92	48	write	write	NN
hansen	92	49	up	up	RP
hansen	92	50	on	on	IN
hansen	92	51	the	the	DT
hansen	92	52	MAG	MAG	NNP
hansen	92	53	website	website	NN
hansen	92	54	does	do	VBZ
hansen	92	55	make	make	VB
hansen	92	56	it	-PRON-	PRP
hansen	92	57	clear	clear	JJ
hansen	92	58	they	-PRON-	PRP
hansen	92	59	used	use	VBD
hansen	92	60	vector	vector	NN
hansen	92	61	based	base	VBN
hansen	92	62	convolutional	convolutional	JJ
hansen	92	63	neural	neural	JJ
hansen	92	64	networks	network	NNS
hansen	92	65	which	which	WDT
hansen	92	66	relied	rely	VBD
hansen	92	67	on	on	IN
hansen	92	68	Skip	Skip	NNP
hansen	92	69	-	-	HYPH
hansen	92	70	gram	gram	NNP
hansen	92	71	(	(	-LRB-
hansen	92	72	Mikolov	Mikolov	NNP
hansen	92	73	et	et	FW
hansen	92	74	al	al	NNP
hansen	92	75	.	.	.
hansen	93	1	2013	2013	LS
hansen	93	2	)	)	-RRB-
hansen	93	3	embeddings	embedding	NNS
hansen	93	4	and	and	CC
hansen	93	5	bag	bag	NN
hansen	93	6	-	-	HYPH
hansen	93	7	of	of	IN
hansen	93	8	-	-	HYPH
hansen	93	9	words	word	NNS
hansen	93	10	/	/	SYM
hansen	93	11	entities	entity	NNS
hansen	93	12	features	feature	VBZ
hansen	93	13	to	to	TO
hansen	93	14	create	create	VB
hansen	93	15	their	-PRON-	PRP$
hansen	93	16	vectors	vector	NNS
hansen	93	17	(	(	-LRB-
hansen	93	18	“	"	``
hansen	93	19	Microsoft	Microsoft	NNP
hansen	93	20	Academic	Academic	NNP
hansen	93	21	Increases	Increases	NNPS
hansen	93	22	Power	Power	NNP
hansen	93	23	of	of	IN
hansen	93	24	Semantic	Semantic	NNP
hansen	93	25	Search	Search	NNP
hansen	93	26	by	by	IN
hansen	93	27	Adding	add	VBG
hansen	93	28	More	More	JJR
hansen	93	29	Fields	field	NNS
hansen	93	30	of	of	IN
hansen	93	31	Study	Study	NNP
hansen	93	32	-	-	HYPH
hansen	93	33	Microsoft	Microsoft	NNP
hansen	93	34	Research	Research	NNP
hansen	93	35	”	"	''
hansen	93	36	2018	2018	CD
hansen	93	37	)	)	-RRB-
hansen	93	38	.	.	.
hansen	94	1	One	one	CD
hansen	94	2	really	really	RB
hansen	94	3	interesting	interesting	JJ
hansen	94	4	part	part	NN
hansen	94	5	of	of	IN
hansen	94	6	the	the	DT
hansen	94	7	machine	machine	NN
hansen	94	8	learning	learning	NN
hansen	94	9	method	method	NN
hansen	94	10	used	use	VBN
hansen	94	11	by	by	IN
hansen	94	12	Microsoft	Microsoft	NNP
hansen	94	13	was	be	VBD
hansen	94	14	that	that	IN
hansen	94	15	it	-PRON-	PRP
hansen	94	16	did	do	VBD
hansen	94	17	not	not	RB
hansen	94	18	rely	rely	VB
hansen	94	19	only	only	RB
hansen	94	20	on	on	IN
hansen	94	21	information	information	NN
hansen	94	22	from	from	IN
hansen	94	23	the	the	DT
hansen	94	24	article	article	NN
hansen	94	25	being	be	VBG
hansen	94	26	categorized	categorize	VBN
hansen	94	27	.	.	.
hansen	95	1	It	-PRON-	PRP
hansen	95	2	also	also	RB
hansen	95	3	utilized	utilize	VBD
hansen	95	4	the	the	DT
hansen	95	5	citations	citation	NNS
hansen	95	6	to	to	IN
hansen	95	7	and	and	CC
hansen	95	8	references	reference	NNS
hansen	95	9	from	from	IN
hansen	95	10	information	information	NN
hansen	95	11	about	about	IN
hansen	95	12	the	the	DT
hansen	95	13	article	article	NN
hansen	95	14	in	in	IN
hansen	95	15	the	the	DT
hansen	95	16	MAG	MAG	NNP
hansen	95	17	,	,	,
hansen	95	18	and	and	CC
hansen	95	19	used	use	VBD
hansen	95	20	the	the	DT
hansen	95	21	FoS	fos	NN
hansen	95	22	the	the	DT
hansen	95	23	citations	citation	NNS
hansen	95	24	and	and	CC
hansen	95	25	references	reference	NNS
hansen	95	26	were	be	VBD
hansen	95	27	assigned	assign	VBN
hansen	95	28	in	in	IN
hansen	95	29	order	order	NN
hansen	95	30	to	to	TO
hansen	95	31	influence	influence	VB
hansen	95	32	the	the	DT
hansen	95	33	FoS	fos	NN
hansen	95	34	of	of	IN
hansen	95	35	the	the	DT
hansen	95	36	original	original	JJ
hansen	95	37	article	article	NN
hansen	95	38	.	.	.
hansen	96	1	The	the	DT
hansen	96	2	identification	identification	NN
hansen	96	3	of	of	IN
hansen	96	4	potential	potential	JJ
hansen	96	5	FoS	fos	NN
hansen	96	6	and	and	CC
hansen	96	7	their	-PRON-	PRP$
hansen	96	8	assignment	assignment	NN
hansen	96	9	to	to	IN
hansen	96	10	articles	article	NNS
hansen	96	11	was	be	VBD
hansen	96	12	only	only	RB
hansen	96	13	a	a	DT
hansen	96	14	part	part	NN
hansen	96	15	of	of	IN
hansen	96	16	Microsoft	Microsoft	NNP
hansen	96	17	's	's	POS
hansen	96	18	purpose	purpose	NN
hansen	96	19	.	.	.
hansen	97	1	In	in	IN
hansen	97	2	order	order	NN
hansen	97	3	to	to	TO
hansen	97	4	fully	fully	RB
hansen	97	5	index	index	VB
hansen	97	6	the	the	DT
hansen	97	7	MAG	MAG	NNP
hansen	97	8	and	and	CC
hansen	97	9	make	make	VB
hansen	97	10	it	-PRON-	PRP
hansen	97	11	searchable	searchable	JJ
hansen	97	12	they	-PRON-	PRP
hansen	97	13	also	also	RB
hansen	97	14	wished	wish	VBD
hansen	97	15	to	to	TO
hansen	97	16	determine	determine	VB
hansen	97	17	the	the	DT
hansen	97	18	relationships	relationship	NNS
hansen	97	19	between	between	IN
hansen	97	20	the	the	DT
hansen	97	21	FoS	fos	NN
hansen	97	22	;	;	:
hansen	97	23	in	in	IN
hansen	97	24	other	other	JJ
hansen	97	25	words	word	NNS
hansen	97	26	they	-PRON-	PRP
hansen	97	27	wanted	want	VBD
hansen	97	28	to	to	TO
hansen	97	29	build	build	VB
hansen	97	30	a	a	DT
hansen	97	31	hierarchical	hierarchical	JJ
hansen	97	32	taxonomy	taxonomy	NN
hansen	97	33	.	.	.
hansen	98	1	To	to	TO
hansen	98	2	achieve	achieve	VB
hansen	98	3	this	this	DT
hansen	98	4	they	-PRON-	PRP
hansen	98	5	used	use	VBD
hansen	98	6	the	the	DT
hansen	98	7	article	article	NN
hansen	98	8	categorizations	categorization	NNS
hansen	98	9	and	and	CC
hansen	98	10	defined	define	VBD
hansen	98	11	a	a	DT
hansen	98	12	Field	Field	NNP
hansen	98	13	of	of	IN
hansen	98	14	Study	Study	NNP
hansen	98	15	A	A	NNP
hansen	98	16	as	as	IN
hansen	98	17	the	the	DT
hansen	98	18	parent	parent	NN
hansen	98	19	of	of	IN
hansen	98	20	B	b	NN
hansen	98	21	if	if	IN
hansen	98	22	the	the	DT
hansen	98	23	articles	article	NNS
hansen	98	24	categorized	categorize	VBN
hansen	98	25	as	as	IN
hansen	98	26	B	B	NNP
hansen	98	27	were	be	VBD
hansen	98	28	close	close	JJ
hansen	98	29	to	to	IN
hansen	98	30	a	a	DT
hansen	98	31	subset	subset	NN
hansen	98	32	of	of	IN
hansen	98	33	the	the	DT
hansen	98	34	articles	article	NNS
hansen	98	35	categorized	categorize	VBN
hansen	98	36	as	as	IN
hansen	98	37	A	a	DT
hansen	98	38	(	(	-LRB-
hansen	98	39	a	a	DT
hansen	98	40	more	more	RBR
hansen	98	41	formal	formal	JJ
hansen	98	42	definition	definition	NN
hansen	98	43	can	can	MD
hansen	98	44	be	be	VB
hansen	98	45	found	find	VBN
hansen	98	46	in	in	IN
hansen	98	47	Shen	Shen	NNP
hansen	98	48	,	,	,
hansen	98	49	Ma	Ma	NNP
hansen	98	50	,	,	,
hansen	98	51	and	and	CC
hansen	98	52	Wang	Wang	NNP
hansen	98	53	2018	2018	CD
hansen	98	54	,	,	,
hansen	98	55	4	4	CD
hansen	98	56	)	)	-RRB-
hansen	98	57	.	.	.
hansen	99	1	This	this	DT
hansen	99	2	work	work	NN
hansen	99	3	,	,	,
hansen	99	4	which	which	WDT
hansen	99	5	created	create	VBD
hansen	99	6	a	a	DT
hansen	99	7	six	six	CD
hansen	99	8	-	-	HYPH
hansen	99	9	level	level	NN
hansen	99	10	hierarchy	hierarchy	NN
hansen	99	11	,	,	,
hansen	99	12	was	be	VBD
hansen	99	13	mostly	mostly	RB
hansen	99	14	automated	automate	VBN
hansen	99	15	,	,	,
hansen	99	16	but	but	CC
hansen	99	17	Microsoft	Microsoft	NNP
hansen	99	18	did	do	VBD
hansen	99	19	inspect	inspect	VB
hansen	99	20	and	and	CC
hansen	99	21	manually	manually	RB
hansen	99	22	adjust	adjust	VB
hansen	99	23	the	the	DT
hansen	99	24	relationships	relationship	NNS
hansen	99	25	between	between	IN
hansen	99	26	FoS	FoS	NNP
hansen	99	27	on	on	IN
hansen	99	28	the	the	DT
hansen	99	29	highest	high	JJS
hansen	99	30	two	two	CD
hansen	99	31	levels	level	NNS
hansen	99	32	.	.	.
hansen	100	1	To	to	TO
hansen	100	2	evaluate	evaluate	VB
hansen	100	3	the	the	DT
hansen	100	4	quality	quality	NN
hansen	100	5	of	of	IN
hansen	100	6	their	-PRON-	PRP$
hansen	100	7	FoS	FoS	NNP
hansen	100	8	taxonomy	taxonomy	NN
hansen	100	9	and	and	CC
hansen	100	10	categorization	categorization	NN
hansen	100	11	work	work	NN
hansen	100	12	,	,	,
hansen	100	13	Microsoft	Microsoft	NNP
hansen	100	14	randomly	randomly	RB
hansen	100	15	sampled	sample	VBD
hansen	100	16	data	datum	NNS
hansen	100	17	at	at	IN
hansen	100	18	each	each	DT
hansen	100	19	of	of	IN
hansen	100	20	the	the	DT
hansen	100	21	three	three	CD
hansen	100	22	steps	step	NNS
hansen	100	23	of	of	IN
hansen	100	24	the	the	DT
hansen	100	25	project	project	NN
hansen	100	26	and	and	CC
hansen	100	27	used	use	VBD
hansen	100	28	human	human	JJ
hansen	100	29	judges	judge	NNS
hansen	100	30	to	to	TO
hansen	100	31	assess	assess	VB
hansen	100	32	their	-PRON-	PRP$
hansen	100	33	accuracy	accuracy	NN
hansen	100	34	.	.	.
hansen	101	1	The	the	DT
hansen	101	2	accuracy	accuracy	NN
hansen	101	3	assessments	assessment	NNS
hansen	101	4	of	of	IN
hansen	101	5	the	the	DT
hansen	101	6	three	three	CD
hansen	101	7	steps	step	NNS
hansen	101	8	were	be	VBD
hansen	101	9	not	not	RB
hansen	101	10	as	as	RB
hansen	101	11	complete	complete	JJ
hansen	101	12	as	as	IN
hansen	101	13	they	-PRON-	PRP
hansen	101	14	would	would	MD
hansen	101	15	be	be	VB
hansen	101	16	with	with	IN
hansen	101	17	the	the	DT
hansen	101	18	mathematics	mathematic	NNS
hansen	101	19	categorization	categorization	NN
hansen	101	20	,	,	,
hansen	101	21	as	as	IN
hansen	101	22	that	that	DT
hansen	101	23	approach	approach	NN
hansen	101	24	would	would	MD
hansen	101	25	evaluate	evaluate	VB
hansen	101	26	terms	term	NNS
hansen	101	27	across	across	IN
hansen	101	28	the	the	DT
hansen	101	29	whole	whole	NN
hansen	101	30	of	of	IN
hansen	101	31	their	-PRON-	PRP$
hansen	101	32	data	data	NN
hansen	101	33	sets	set	NNS
hansen	101	34	,	,	,
hansen	101	35	but	but	CC
hansen	101	36	the	the	DT
hansen	101	37	projects	project	NNS
hansen	101	38	are	be	VBP
hansen	101	39	of	of	IN
hansen	101	40	very	very	RB
hansen	101	41	different	different	JJ
hansen	101	42	scales	scale	NNS
hansen	101	43	so	so	RB
hansen	101	44	different	different	JJ
hansen	101	45	methods	method	NNS
hansen	101	46	are	be	VBP
hansen	101	47	appropriate	appropriate	JJ
hansen	101	48	.	.	.
hansen	102	1	In	in	IN
hansen	102	2	the	the	DT
hansen	102	3	end	end	NN
hansen	102	4	Microsoft	Microsoft	NNP
hansen	102	5	estimates	estimate	VBZ
hansen	102	6	the	the	DT
hansen	102	7	accuracy	accuracy	NN
hansen	102	8	of	of	IN
hansen	102	9	the	the	DT
hansen	102	10	FoS	fos	NN
hansen	102	11	at	at	IN
hansen	102	12	94.75	94.75	CD
hansen	102	13	%	%	NN
hansen	102	14	,	,	,
hansen	102	15	the	the	DT
hansen	102	16	article	article	NN
hansen	102	17	categorization	categorization	NN
hansen	102	18	at	at	IN
hansen	102	19	81.2	81.2	CD
hansen	102	20	%	%	NN
hansen	102	21	,	,	,
hansen	102	22	and	and	CC
hansen	102	23	the	the	DT
hansen	102	24	hierarchy	hierarchy	NN
hansen	102	25	at	at	IN
hansen	102	26	78	78	CD
hansen	102	27	%	%	NN
hansen	102	28	(	(	-LRB-
hansen	102	29	Shen	Shen	NNP
hansen	102	30	,	,	,
hansen	102	31	Ma	Ma	NNP
hansen	102	32	,	,	,
hansen	102	33	and	and	CC
hansen	102	34	Wang	Wang	NNP
hansen	102	35	2018	2018	CD
hansen	102	36	,	,	,
hansen	102	37	5	5	CD
hansen	102	38	)	)	-RRB-
hansen	102	39	.	.	.
hansen	103	1	Since	since	IN
hansen	103	2	MSC	MSC	NNP
hansen	103	3	was	be	VBD
hansen	103	4	created	create	VBN
hansen	103	5	by	by	IN
hansen	103	6	humans	human	NNS
hansen	103	7	there	there	EX
hansen	103	8	is	be	VBZ
hansen	103	9	no	no	DT
hansen	103	10	meaningful	meaningful	JJ
hansen	103	11	way	way	NN
hansen	103	12	to	to	TO
hansen	103	13	compare	compare	VB
hansen	103	14	the	the	DT
hansen	103	15	FoS	FoS	NNP
hansen	103	16	accuracy	accuracy	NN
hansen	103	17	measurements	measurement	NNS
hansen	103	18	,	,	,
hansen	103	19	but	but	CC
hansen	103	20	the	the	DT
hansen	103	21	categorization	categorization	NN
hansen	103	22	accuracy	accuracy	NN
hansen	103	23	falls	fall	VBZ
hansen	103	24	somewhere	somewhere	RB
hansen	103	25	between	between	IN
hansen	103	26	that	that	DT
hansen	103	27	of	of	IN
hansen	103	28	the	the	DT
hansen	103	29	two	two	CD
hansen	103	30	mathematics	mathematic	NNS
hansen	103	31	projects	project	NNS
hansen	103	32	.	.	.
hansen	104	1	This	this	DT
hansen	104	2	is	be	VBZ
hansen	104	3	a	a	DT
hansen	104	4	very	very	RB
hansen	104	5	impressive	impressive	JJ
hansen	104	6	result	result	NN
hansen	104	7	,	,	,
hansen	104	8	especially	especially	RB
hansen	104	9	when	when	WRB
hansen	104	10	the	the	DT
hansen	104	11	aforementioned	aforementioned	JJ
hansen	104	12	scale	scale	NN
hansen	104	13	is	be	VBZ
hansen	104	14	taken	take	VBN
hansen	104	15	into	into	IN
hansen	104	16	account	account	NN
hansen	104	17	.	.	.
hansen	105	1	Instead	instead	RB
hansen	105	2	of	of	IN
hansen	105	3	trying	try	VBG
hansen	105	4	to	to	TO
hansen	105	5	replace	replace	VB
hansen	105	6	the	the	DT
hansen	105	7	work	work	NN
hansen	105	8	of	of	IN
hansen	105	9	humans	human	NNS
hansen	105	10	categorizing	categorize	VBG
hansen	105	11	mathematics	mathematic	NNS
hansen	105	12	articles	article	NNS
hansen	105	13	indexed	index	VBN
hansen	105	14	in	in	IN
hansen	105	15	a	a	DT
hansen	105	16	database	database	NN
hansen	105	17	,	,	,
hansen	105	18	which	which	WDT
hansen	105	19	for	for	IN
hansen	105	20	2018	2018	CD
hansen	105	21	was	be	VBD
hansen	105	22	120,324	120,324	CD
hansen	105	23	items	item	NNS
hansen	105	24	in	in	IN
hansen	105	25	MathSciNet[footnoteRef:17	mathscinet[footnoteref:17	CD
hansen	105	26	]	]	-RRB-
hansen	105	27	and	and	CC
hansen	105	28	97,819	97,819	CD
hansen	105	29	in	in	IN
hansen	105	30	zbMath[footnoteRef:18	zbMath[footnoteRef:18	NNP
hansen	105	31	]	]	-RRB-
hansen	105	32	,	,	,
hansen	105	33	it	-PRON-	PRP
hansen	105	34	is	be	VBZ
hansen	105	35	trying	try	VBG
hansen	105	36	to	to	TO
hansen	105	37	replace	replace	VB
hansen	105	38	the	the	DT
hansen	105	39	human	human	JJ
hansen	105	40	categorization	categorization	NN
hansen	105	41	of	of	IN
hansen	105	42	all	all	DT
hansen	105	43	items	item	NNS
hansen	105	44	indexed	index	VBN
hansen	105	45	in	in	IN
hansen	105	46	MAG	MAG	NNP
hansen	105	47	,	,	,
hansen	105	48	which	which	WDT
hansen	105	49	was	be	VBD
hansen	105	50	10,616,601	10,616,601	CD
hansen	105	51	in	in	IN
hansen	105	52	2018	2018	CD
hansen	105	53	.	.	.
hansen	106	1	[	[	-LRB-
hansen	106	2	footnoteRef:19	footnoteRef:19	NNP
hansen	106	3	]	]	-RRB-
hansen	106	4	Both	both	CC
hansen	106	5	zbMath	zbMath	NNP
hansen	106	6	and	and	CC
hansen	106	7	MathSciNet	MathSciNet	NNP
hansen	106	8	were	be	VBD
hansen	106	9	capable	capable	JJ
hansen	106	10	of	of	IN
hansen	106	11	providing	provide	VBG
hansen	106	12	the	the	DT
hansen	106	13	human	human	JJ
hansen	106	14	labor	labor	NN
hansen	106	15	to	to	TO
hansen	106	16	do	do	VB
hansen	106	17	the	the	DT
hansen	106	18	work	work	NN
hansen	106	19	of	of	IN
hansen	106	20	assigning	assign	VBG
hansen	106	21	MSC	MSC	NNP
hansen	106	22	values	value	NNS
hansen	106	23	to	to	IN
hansen	106	24	the	the	DT
hansen	106	25	mathematics	mathematic	NNS
hansen	106	26	articles	article	NNS
hansen	106	27	they	-PRON-	PRP
hansen	106	28	indexed	index	VBD
hansen	106	29	in	in	IN
hansen	106	30	2018	2018	CD
hansen	106	31	.	.	.
hansen	107	1	Therefore	therefore	RB
hansen	107	2	using	use	VBG
hansen	107	3	an	an	DT
hansen	107	4	automated	automate	VBN
hansen	107	5	categorization	categorization	NN
hansen	107	6	,	,	,
hansen	107	7	which	which	WDT
hansen	107	8	at	at	IN
hansen	107	9	best	good	JJS
hansen	107	10	could	could	MD
hansen	107	11	only	only	RB
hansen	107	12	get	get	VB
hansen	107	13	the	the	DT
hansen	107	14	top	top	JJ
hansen	107	15	level	level	NN
hansen	107	16	right	right	RB
hansen	107	17	with	with	IN
hansen	107	18	90	90	CD
hansen	107	19	%	%	NN
hansen	107	20	accuracy	accuracy	NN
hansen	107	21	,	,	,
hansen	107	22	was	be	VBD
hansen	107	23	not	not	RB
hansen	107	24	the	the	DT
hansen	107	25	right	right	JJ
hansen	107	26	path	path	NN
hansen	107	27	.	.	.
hansen	108	1	On	on	IN
hansen	108	2	the	the	DT
hansen	108	3	other	other	JJ
hansen	108	4	hand	hand	NN
hansen	108	5	,	,	,
hansen	108	6	it	-PRON-	PRP
hansen	108	7	seems	seem	VBZ
hansen	108	8	clear	clear	JJ
hansen	108	9	that	that	IN
hansen	108	10	no	no	DT
hansen	108	11	one	one	PRP
hansen	108	12	could	could	MD
hansen	108	13	feasibly	feasibly	RB
hansen	108	14	provide	provide	VB
hansen	108	15	the	the	DT
hansen	108	16	human	human	JJ
hansen	108	17	labor	labor	NN
hansen	108	18	to	to	TO
hansen	108	19	categorize	categorize	VB
hansen	108	20	all	all	DT
hansen	108	21	articles	article	NNS
hansen	108	22	indexed	index	VBN
hansen	108	23	by	by	IN
hansen	108	24	MAG	MAG	NNP
hansen	108	25	in	in	IN
hansen	108	26	2018	2018	CD
hansen	108	27	so	so	IN
hansen	108	28	an	an	DT
hansen	108	29	80	80	CD
hansen	108	30	%	%	NN
hansen	108	31	accurate	accurate	JJ
hansen	108	32	categorization	categorization	NN
hansen	108	33	is	be	VBZ
hansen	108	34	a	a	DT
hansen	108	35	significant	significant	JJ
hansen	108	36	accomplishment	accomplishment	NN
hansen	108	37	.	.	.
hansen	109	1	To	to	TO
hansen	109	2	go	go	VB
hansen	109	3	back	back	RB
hansen	109	4	to	to	IN
hansen	109	5	the	the	DT
hansen	109	6	nail	nail	NN
hansen	109	7	and	and	CC
hansen	109	8	hammer	hammer	NN
hansen	109	9	analogy	analogy	NN
hansen	109	10	,	,	,
hansen	109	11	Microsoft	Microsoft	NNP
hansen	109	12	may	may	MD
hansen	109	13	have	have	VB
hansen	109	14	used	use	VBN
hansen	109	15	a	a	DT
hansen	109	16	sledgehammer	sledgehammer	NN
hansen	109	17	but	but	CC
hansen	109	18	they	-PRON-	PRP
hansen	109	19	were	be	VBD
hansen	109	20	hammering	hammer	VBG
hansen	109	21	a	a	DT
hansen	109	22	rather	rather	RB
hansen	109	23	giant	giant	JJ
hansen	109	24	nail	nail	NN
hansen	109	25	.	.	.
hansen	110	1	Comment	comment	NN
hansen	110	2	by	by	IN
hansen	110	3	Mark	Mark	NNP
hansen	110	4	Dehmlow	Dehmlow	NNP
hansen	110	5	:	:	:
hansen	110	6	Maybe	maybe	RB
hansen	110	7	use	use	VB
hansen	110	8	"	"	``
hansen	110	9	the	the	DT
hansen	110	10	project	project	NN
hansen	110	11	?	?	.
hansen	110	12	?	?	.
hansen	111	1	Comment	comment	NN
hansen	111	2	by	by	IN
hansen	111	3	Mark	Mark	NNP
hansen	111	4	Dehmlow	Dehmlow	NNP
hansen	111	5	:	:	:
hansen	111	6	It	-PRON-	PRP
hansen	111	7	would	would	MD
hansen	111	8	be	be	VB
hansen	111	9	interesting	interesting	JJ
hansen	111	10	to	to	TO
hansen	111	11	know	know	VB
hansen	111	12	the	the	DT
hansen	111	13	accuracy	accuracy	NN
hansen	111	14	of	of	IN
hansen	111	15	human	human	JJ
hansen	111	16	indexing	indexing	NN
hansen	111	17	in	in	IN
hansen	111	18	MSN	MSN	NNP
hansen	111	19	or	or	CC
hansen	111	20	ZBM	ZBM	NNP
hansen	111	21	.	.	.
hansen	112	1	Comment	comment	NN
hansen	112	2	by	by	IN
hansen	112	3	Samuel	Samuel	NNP
hansen	112	4	Hansen	Hansen	NNP
hansen	112	5	:	:	:
hansen	112	6	Sadly	sadly	RB
hansen	112	7	no	no	DT
hansen	112	8	one	one	NN
hansen	112	9	seems	seem	VBZ
hansen	112	10	to	to	TO
hansen	112	11	have	have	VB
hansen	112	12	studied	study	VBN
hansen	112	13	this	this	DT
hansen	112	14	,	,	,
hansen	112	15	but	but	CC
hansen	112	16	both	both	DT
hansen	112	17	databases	database	NNS
hansen	112	18	are	be	VBP
hansen	112	19	considered	consider	VBN
hansen	112	20	Gold	Gold	NNP
hansen	112	21	Standard	Standard	NNP
hansen	112	22	when	when	WRB
hansen	112	23	it	-PRON-	PRP
hansen	112	24	comes	come	VBZ
hansen	112	25	to	to	IN
hansen	112	26	classification	classification	NN
hansen	112	27	.	.	.
hansen	113	1	They	-PRON-	PRP
hansen	113	2	do	do	VBP
hansen	113	3	not	not	RB
hansen	113	4	rely	rely	VB
hansen	113	5	on	on	IN
hansen	113	6	authors	author	NNS
hansen	113	7	or	or	CC
hansen	113	8	publishers	publisher	NNS
hansen	113	9	providing	provide	VBG
hansen	113	10	classification	classification	NN
hansen	113	11	information	information	NN
hansen	113	12	and	and	CC
hansen	113	13	each	each	DT
hansen	113	14	article	article	NN
hansen	113	15	is	be	VBZ
hansen	113	16	classified	classify	VBN
hansen	113	17	not	not	RB
hansen	113	18	only	only	RB
hansen	113	19	by	by	IN
hansen	113	20	the	the	DT
hansen	113	21	subject	subject	JJ
hansen	113	22	area	area	NN
hansen	113	23	expert	expert	NN
hansen	113	24	reviewer	reviewer	NN
hansen	113	25	they	-PRON-	PRP
hansen	113	26	send	send	VBP
hansen	113	27	it	-PRON-	PRP
hansen	113	28	to	to	IN
hansen	113	29	it	-PRON-	PRP
hansen	113	30	is	be	VBZ
hansen	113	31	also	also	RB
hansen	113	32	identified	identify	VBN
hansen	113	33	before	before	IN
hansen	113	34	the	the	DT
hansen	113	35	reviewer	reviewer	NN
hansen	113	36	receives	receive	VBZ
hansen	113	37	it	-PRON-	PRP
hansen	113	38	by	by	IN
hansen	113	39	an	an	DT
hansen	113	40	editor	editor	NN
hansen	113	41	who	who	WP
hansen	113	42	is	be	VBZ
hansen	113	43	trying	try	VBG
hansen	113	44	to	to	TO
hansen	113	45	find	find	VB
hansen	113	46	a	a	DT
hansen	113	47	reviewer	reviewer	NN
hansen	113	48	,	,	,
hansen	113	49	and	and	CC
hansen	113	50	again	again	RB
hansen	113	51	after	after	IN
hansen	113	52	the	the	DT
hansen	113	53	paper	paper	NN
hansen	113	54	is	be	VBZ
hansen	113	55	reviewed	review	VBN
hansen	113	56	as	as	IN
hansen	113	57	a	a	DT
hansen	113	58	check	check	NN
hansen	113	59	.	.	.
hansen	114	1	There	there	EX
hansen	114	2	is	be	VBZ
hansen	114	3	rarely	rarely	RB
hansen	114	4	much	much	JJ
hansen	114	5	disagreement	disagreement	NN
hansen	114	6	over	over	IN
hansen	114	7	MSC	MSC	NNP
hansen	114	8	values	value	NNS
hansen	114	9	between	between	IN
hansen	114	10	MSN	MSN	NNP
hansen	114	11	and	and	CC
hansen	114	12	zbMath	zbMath	NNP
hansen	114	13	and	and	CC
hansen	114	14	when	when	WRB
hansen	114	15	it	-PRON-	PRP
hansen	114	16	does	do	VBZ
hansen	114	17	occur	occur	VB
hansen	114	18	,	,	,
hansen	114	19	from	from	IN
hansen	114	20	what	what	WP
hansen	114	21	I	-PRON-	PRP
hansen	114	22	am	be	VBP
hansen	114	23	trying	try	VBG
hansen	114	24	to	to	TO
hansen	114	25	remember	remember	VB
hansen	114	26	from	from	IN
hansen	114	27	a	a	DT
hansen	114	28	presentation	presentation	NN
hansen	114	29	,	,	,
hansen	114	30	it	-PRON-	PRP
hansen	114	31	is	be	VBZ
hansen	114	32	in	in	IN
hansen	114	33	the	the	DT
hansen	114	34	last	last	JJ
hansen	114	35	part	part	NN
hansen	114	36	so	so	IN
hansen	114	37	we	-PRON-	PRP
hansen	114	38	are	be	VBP
hansen	114	39	taking	take	VBG
hansen	114	40	just	just	RB
hansen	114	41	tiny	tiny	JJ
hansen	114	42	differences	difference	NNS
hansen	114	43	given	give	VBN
hansen	114	44	how	how	WRB
hansen	114	45	fine	fine	JJ
hansen	114	46	a	a	DT
hansen	114	47	classification	classification	NN
hansen	114	48	MSN	MSN	NNP
hansen	114	49	is	be	VBZ
hansen	114	50	.	.	.
hansen	115	1	Comment	comment	NN
hansen	115	2	by	by	IN
hansen	115	3	Mark	Mark	NNP
hansen	115	4	Dehmlow	Dehmlow	NNP
hansen	115	5	:	:	:
hansen	115	6	If	if	IN
hansen	115	7	you	-PRON-	PRP
hansen	115	8	can	can	MD
hansen	115	9	find	find	VB
hansen	115	10	it	-PRON-	PRP
hansen	115	11	,	,	,
hansen	115	12	it	-PRON-	PRP
hansen	115	13	may	may	MD
hansen	115	14	be	be	VB
hansen	115	15	worth	worth	JJ
hansen	115	16	citing	cite	VBG
hansen	115	17	OR	or	CC
hansen	115	18	you	-PRON-	PRP
hansen	115	19	could	could	MD
hansen	115	20	explain	explain	VB
hansen	115	21	their	-PRON-	PRP$
hansen	115	22	methodology	methodology	NN
hansen	115	23	.	.	.
hansen	116	1	It	-PRON-	PRP
hansen	116	2	helps	help	VBZ
hansen	116	3	validate	validate	VB
hansen	116	4	the	the	DT
hansen	116	5	claim	claim	NN
hansen	116	6	that	that	IN
hansen	116	7	human	human	JJ
hansen	116	8	categorization	categorization	NN
hansen	116	9	is	be	VBZ
hansen	116	10	better	well	JJR
hansen	116	11	than	than	IN
hansen	116	12	machine	machine	NN
hansen	116	13	.	.	.
hansen	117	1	Comment	comment	NN
hansen	117	2	by	by	IN
hansen	117	3	Daniel	Daniel	NNP
hansen	117	4	Johnson	Johnson	NNP
hansen	117	5	:	:	:
hansen	117	6	This	this	DT
hansen	117	7	is	be	VBZ
hansen	117	8	kind	kind	RB
hansen	117	9	of	of	RB
hansen	117	10	interesting	interesting	JJ
hansen	117	11	,	,	,
hansen	117	12	as	as	IN
hansen	117	13	it	-PRON-	PRP
hansen	117	14	might	may	MD
hansen	117	15	be	be	VB
hansen	117	16	a	a	DT
hansen	117	17	method	method	NN
hansen	117	18	of	of	IN
hansen	117	19	working	work	VBG
hansen	117	20	other	other	JJ
hansen	117	21	disciplines	discipline	NNS
hansen	117	22	have	have	VBP
hansen	117	23	never	never	RB
hansen	117	24	heard	hear	VBN
hansen	117	25	of	of	IN
hansen	117	26	.	.	.
hansen	118	1	Perhaps	perhaps	RB
hansen	118	2	a	a	DT
hansen	118	3	short	short	JJ
hansen	118	4	footnote	footnote	NN
hansen	118	5	?	?	.
hansen	119	1	Comment	comment	NN
hansen	119	2	by	by	IN
hansen	119	3	Mark	Mark	NNP
hansen	119	4	Dehmlow	Dehmlow	NNP
hansen	119	5	:	:	:
hansen	119	6	Maybe	maybe	RB
hansen	119	7	"	"	``
hansen	119	8	approach	approach	NN
hansen	119	9	"	"	''
hansen	119	10	?	?	.
hansen	120	1	[	[	-LRB-
hansen	120	2	17	17	CD
hansen	120	3	:	:	:
hansen	120	4	https://mathscinet.ams.org/mathscinet/search/publications.html?dr=pubyear&yrop=eq&arg3=2018	https://mathscinet.ams.org/mathscinet/search/publications.html?dr=pubyear&yrop=eq&arg3=2018	NNP
hansen	120	5	]	]	-RRB-
hansen	120	6	[	[	-LRB-
hansen	120	7	18	18	CD
hansen	120	8	:	:	:
hansen	120	9	https://zbmath.org/?q=py%3A2018	https://zbmath.org/?q=py%3a2018	LS
hansen	120	10	]	]	-RRB-
hansen	120	11	[	[	-LRB-
hansen	120	12	19	19	CD
hansen	120	13	:	:	:
hansen	120	14	https://academic.microsoft.com/publications/33923547	https://academic.microsoft.com/publications/33923547	LS
hansen	120	15	]	]	-RRB-
hansen	120	16	Are	be	VBP
hansen	120	17	You	-PRON-	PRP
hansen	120	18	Sure	sure	JJ
hansen	120	19	it	-PRON-	PRP
hansen	120	20	’s	’	VBZ
hansen	120	21	a	a	DT
hansen	120	22	Nail	nail	NN
hansen	120	23	?	?	.
hansen	121	1	I	-PRON-	PRP
hansen	121	2	started	start	VBD
hansen	121	3	this	this	DT
hansen	121	4	chapter	chapter	NN
hansen	121	5	talking	talk	VBG
hansen	121	6	about	about	IN
hansen	121	7	how	how	WRB
hansen	121	8	we	-PRON-	PRP
hansen	121	9	have	have	VBP
hansen	121	10	all	all	DT
hansen	121	11	been	be	VBN
hansen	121	12	told	tell	VBN
hansen	121	13	that	that	IN
hansen	121	14	AI	AI	NNP
hansen	121	15	and	and	CC
hansen	121	16	machine	machine	NN
hansen	121	17	learning	learning	NN
hansen	121	18	were	be	VBD
hansen	121	19	going	go	VBG
hansen	121	20	to	to	TO
hansen	121	21	revolutionize	revolutionize	VB
hansen	121	22	everything	everything	NN
hansen	121	23	in	in	IN
hansen	121	24	the	the	DT
hansen	121	25	world	world	NN
hansen	121	26	.	.	.
hansen	122	1	That	that	IN
hansen	122	2	they	-PRON-	PRP
hansen	122	3	were	be	VBD
hansen	122	4	the	the	DT
hansen	122	5	hammers	hammer	NNS
hansen	122	6	and	and	CC
hansen	122	7	all	all	PDT
hansen	122	8	the	the	DT
hansen	122	9	world	world	NN
hansen	122	10	's	's	POS
hansen	122	11	problems	problem	NNS
hansen	122	12	were	be	VBD
hansen	122	13	nails	nail	NNS
hansen	122	14	.	.	.
hansen	123	1	I	-PRON-	PRP
hansen	123	2	found	find	VBD
hansen	123	3	that	that	IN
hansen	123	4	this	this	DT
hansen	123	5	was	be	VBD
hansen	123	6	not	not	RB
hansen	123	7	the	the	DT
hansen	123	8	case	case	NN
hansen	123	9	when	when	WRB
hansen	123	10	we	-PRON-	PRP
hansen	123	11	tried	try	VBD
hansen	123	12	to	to	TO
hansen	123	13	employ	employ	VB
hansen	123	14	it	-PRON-	PRP
hansen	123	15	,	,	,
hansen	123	16	in	in	IN
hansen	123	17	an	an	DT
hansen	123	18	admittedly	admittedly	RB
hansen	123	19	rather	rather	RB
hansen	123	20	naive	naive	JJ
hansen	123	21	fashion	fashion	NN
hansen	123	22	,	,	,
hansen	123	23	to	to	TO
hansen	123	24	automatically	automatically	RB
hansen	123	25	categorize	categorize	VB
hansen	123	26	mathematical	mathematical	JJ
hansen	123	27	articles	article	NNS
hansen	123	28	.	.	.
hansen	124	1	From	from	IN
hansen	124	2	the	the	DT
hansen	124	3	other	other	JJ
hansen	124	4	examples	example	NNS
hansen	124	5	I	-PRON-	PRP
hansen	124	6	included	include	VBD
hansen	124	7	,	,	,
hansen	124	8	it	-PRON-	PRP
hansen	124	9	is	be	VBZ
hansen	124	10	also	also	RB
hansen	124	11	clear	clear	JJ
hansen	124	12	computational	computational	JJ
hansen	124	13	experts	expert	NNS
hansen	124	14	find	find	VBP
hansen	124	15	the	the	DT
hansen	124	16	automatic	automatic	JJ
hansen	124	17	categorization	categorization	NN
hansen	124	18	of	of	IN
hansen	124	19	highly	highly	RB
hansen	124	20	technical	technical	JJ
hansen	124	21	content	content	NN
hansen	124	22	a	a	DT
hansen	124	23	hard	hard	JJ
hansen	124	24	problem	problem	NN
hansen	124	25	to	to	TO
hansen	124	26	tackle	tackle	VB
hansen	124	27	,	,	,
hansen	124	28	one	one	CD
hansen	124	29	where	where	WRB
hansen	124	30	success	success	NN
hansen	124	31	is	be	VBZ
hansen	124	32	very	very	RB
hansen	124	33	much	much	RB
hansen	124	34	dependent	dependent	JJ
hansen	124	35	on	on	IN
hansen	124	36	what	what	WP
hansen	124	37	it	-PRON-	PRP
hansen	124	38	is	be	VBZ
hansen	124	39	being	be	VBG
hansen	124	40	measured	measure	VBN
hansen	124	41	against	against	IN
hansen	124	42	.	.	.
hansen	125	1	In	in	IN
hansen	125	2	the	the	DT
hansen	125	3	case	case	NN
hansen	125	4	of	of	IN
hansen	125	5	classifying	classify	VBG
hansen	125	6	mathematics	mathematic	NNS
hansen	125	7	,	,	,
hansen	125	8	machine	machine	NN
hansen	125	9	learning	learning	NN
hansen	125	10	can	can	MD
hansen	125	11	do	do	VB
hansen	125	12	a	a	DT
hansen	125	13	decent	decent	JJ
hansen	125	14	job	job	NN
hansen	125	15	but	but	CC
hansen	125	16	not	not	RB
hansen	125	17	enough	enough	JJ
hansen	125	18	to	to	TO
hansen	125	19	compete	compete	VB
hansen	125	20	with	with	IN
hansen	125	21	humans	human	NNS
hansen	125	22	.	.	.
hansen	126	1	In	in	IN
hansen	126	2	the	the	DT
hansen	126	3	case	case	NN
hansen	126	4	of	of	IN
hansen	126	5	classifying	classify	VBG
hansen	126	6	everything	everything	NN
hansen	126	7	,	,	,
hansen	126	8	scale	scale	NN
hansen	126	9	gives	give	VBZ
hansen	126	10	machines	machine	NNS
hansen	126	11	an	an	DT
hansen	126	12	edge	edge	NN
hansen	126	13	,	,	,
hansen	126	14	as	as	RB
hansen	126	15	long	long	RB
hansen	126	16	as	as	IN
hansen	126	17	you	-PRON-	PRP
hansen	126	18	have	have	VBP
hansen	126	19	the	the	DT
hansen	126	20	computational	computational	JJ
hansen	126	21	power	power	NN
hansen	126	22	and	and	CC
hansen	126	23	knowledge	knowledge	NN
hansen	126	24	wielded	wield	VBN
hansen	126	25	by	by	IN
hansen	126	26	a	a	DT
hansen	126	27	company	company	NN
hansen	126	28	like	like	IN
hansen	126	29	Microsoft	Microsoft	NNP
hansen	126	30	.	.	.
hansen	127	1	Comment	comment	NN
hansen	127	2	by	by	IN
hansen	127	3	Mark	Mark	NNP
hansen	127	4	Dehmlow	Dehmlow	NNP
hansen	127	5	:	:	:
hansen	127	6	This	this	DT
hansen	127	7	is	be	VBZ
hansen	127	8	a	a	DT
hansen	127	9	bit	bit	NN
hansen	127	10	hyperbolic	hyperbolic	JJ
hansen	127	11	-	-	:
hansen	127	12	which	which	WDT
hansen	127	13	is	be	VBZ
hansen	127	14	okay	okay	JJ
hansen	127	15	if	if	IN
hansen	127	16	the	the	DT
hansen	127	17	tone	tone	NN
hansen	127	18	is	be	VBZ
hansen	127	19	intended	intend	VBN
hansen	127	20	.	.	.
hansen	128	1	Certainly	certainly	RB
hansen	128	2	,	,	,
hansen	128	3	it	-PRON-	PRP
hansen	128	4	would	would	MD
hansen	128	5	be	be	VB
hansen	128	6	easy	easy	JJ
hansen	128	7	to	to	TO
hansen	128	8	claim	claim	VB
hansen	128	9	that	that	IN
hansen	128	10	ML	ML	NNP
hansen	128	11	intended	intend	VBD
hansen	128	12	to	to	TO
hansen	128	13	revolutionize	revolutionize	VB
hansen	128	14	classification	classification	NN
hansen	128	15	.	.	.
hansen	129	1	This	this	DT
hansen	129	2	collection	collection	NN
hansen	129	3	is	be	VBZ
hansen	129	4	about	about	IN
hansen	129	5	the	the	DT
hansen	129	6	intersection	intersection	NN
hansen	129	7	of	of	IN
hansen	129	8	AI	AI	NNP
hansen	129	9	,	,	,
hansen	129	10	machine	machine	NN
hansen	129	11	learning	learning	NN
hansen	129	12	,	,	,
hansen	129	13	deep	deep	JJ
hansen	129	14	learning	learning	NN
hansen	129	15	,	,	,
hansen	129	16	and	and	CC
hansen	129	17	libraries	library	NNS
hansen	129	18	.	.	.
hansen	130	1	While	while	IN
hansen	130	2	there	there	EX
hansen	130	3	are	be	VBP
hansen	130	4	definitely	definitely	RB
hansen	130	5	problems	problem	NNS
hansen	130	6	in	in	IN
hansen	130	7	libraries	library	NNS
hansen	130	8	where	where	WRB
hansen	130	9	these	these	DT
hansen	130	10	techniques	technique	NNS
hansen	130	11	will	will	MD
hansen	130	12	be	be	VB
hansen	130	13	the	the	DT
hansen	130	14	answer	answer	NN
hansen	130	15	,	,	,
hansen	130	16	I	-PRON-	PRP
hansen	130	17	think	think	VBP
hansen	130	18	it	-PRON-	PRP
hansen	130	19	is	be	VBZ
hansen	130	20	important	important	JJ
hansen	130	21	to	to	TO
hansen	130	22	pause	pause	VB
hansen	130	23	,	,	,
hansen	130	24	think	think	VB
hansen	130	25	,	,	,
hansen	130	26	and	and	CC
hansen	130	27	be	be	VB
hansen	130	28	sure	sure	JJ
hansen	130	29	artificial	artificial	JJ
hansen	130	30	intelligence	intelligence	NN
hansen	130	31	techniques	technique	NNS
hansen	130	32	are	be	VBP
hansen	130	33	the	the	DT
hansen	130	34	most	most	RBS
hansen	130	35	effective	effective	JJ
hansen	130	36	approach	approach	NN
hansen	130	37	before	before	IN
hansen	130	38	trying	try	VBG
hansen	130	39	to	to	TO
hansen	130	40	use	use	VB
hansen	130	41	them	-PRON-	PRP
hansen	130	42	.	.	.
hansen	131	1	Libraries	library	NNS
hansen	131	2	,	,	,
hansen	131	3	even	even	RB
hansen	131	4	those	those	DT
hansen	131	5	like	like	IN
hansen	131	6	the	the	DT
hansen	131	7	one	one	NN
hansen	131	8	I	-PRON-	PRP
hansen	131	9	work	work	VBP
hansen	131	10	in	in	RP
hansen	131	11	,	,	,
hansen	131	12	which	which	WDT
hansen	131	13	are	be	VBP
hansen	131	14	lucky	lucky	JJ
hansen	131	15	enough	enough	RB
hansen	131	16	to	to	TO
hansen	131	17	boast	boast	VB
hansen	131	18	of	of	IN
hansen	131	19	incredibly	incredibly	RB
hansen	131	20	talented	talented	JJ
hansen	131	21	IT	IT	NNP
hansen	131	22	departments	department	NNS
hansen	131	23	,	,	,
hansen	131	24	do	do	VB
hansen	131	25	not	not	RB
hansen	131	26	tend	tend	VB
hansen	131	27	to	to	TO
hansen	131	28	have	have	VB
hansen	131	29	access	access	NN
hansen	131	30	to	to	IN
hansen	131	31	a	a	DT
hansen	131	32	large	large	JJ
hansen	131	33	amount	amount	NN
hansen	131	34	of	of	IN
hansen	131	35	unused	unused	JJ
hansen	131	36	computational	computational	JJ
hansen	131	37	power	power	NN
hansen	131	38	or	or	CC
hansen	131	39	numerous	numerous	JJ
hansen	131	40	experts	expert	NNS
hansen	131	41	in	in	IN
hansen	131	42	bleeding	bleed	VBG
hansen	131	43	-	-	HYPH
hansen	131	44	edge	edge	NN
hansen	131	45	AI	ai	NN
hansen	131	46	.	.	.
hansen	132	1	They	-PRON-	PRP
hansen	132	2	are	be	VBP
hansen	132	3	also	also	RB
hansen	132	4	rather	rather	RB
hansen	132	5	notoriously	notoriously	RB
hansen	132	6	limited	limited	JJ
hansen	132	7	budget	budget	NN
hansen	132	8	-	-	HYPH
hansen	132	9	wise	wise	JJ
hansen	132	10	and	and	CC
hansen	132	11	would	would	MD
hansen	132	12	likely	likely	RB
hansen	132	13	have	have	VB
hansen	132	14	to	to	TO
hansen	132	15	decide	decide	VB
hansen	132	16	between	between	IN
hansen	132	17	existing	exist	VBG
hansen	132	18	budget	budget	NN
hansen	132	19	items	item	NNS
hansen	132	20	and	and	CC
hansen	132	21	developing	develop	VBG
hansen	132	22	an	an	DT
hansen	132	23	in	in	IN
hansen	132	24	-	-	HYPH
hansen	132	25	house	house	NN
hansen	132	26	machine	machine	NN
hansen	132	27	learning	learning	NN
hansen	132	28	program	program	NN
hansen	132	29	.	.	.
hansen	133	1	Those	those	DT
hansen	133	2	realities	reality	NNS
hansen	133	3	combined	combine	VBN
hansen	133	4	with	with	IN
hansen	133	5	the	the	DT
hansen	133	6	legitimate	legitimate	JJ
hansen	133	7	questions	question	NNS
hansen	133	8	which	which	WDT
hansen	133	9	can	can	MD
hansen	133	10	be	be	VB
hansen	133	11	raised	raise	VBN
hansen	133	12	about	about	IN
hansen	133	13	the	the	DT
hansen	133	14	efficacy	efficacy	NN
hansen	133	15	of	of	IN
hansen	133	16	machine	machine	NN
hansen	133	17	learning	learning	NN
hansen	133	18	and	and	CC
hansen	133	19	AI	AI	NNP
hansen	133	20	with	with	IN
hansen	133	21	respect	respect	NN
hansen	133	22	to	to	IN
hansen	133	23	the	the	DT
hansen	133	24	types	type	NNS
hansen	133	25	of	of	IN
hansen	133	26	problems	problem	NNS
hansen	133	27	a	a	DT
hansen	133	28	library	library	NN
hansen	133	29	may	may	MD
hansen	133	30	encounter	encounter	VB
hansen	133	31	,	,	,
hansen	133	32	such	such	JJ
hansen	133	33	as	as	IN
hansen	133	34	categorizing	categorize	VBG
hansen	133	35	the	the	DT
hansen	133	36	contents	content	NNS
hansen	133	37	of	of	IN
hansen	133	38	highly	highly	RB
hansen	133	39	technical	technical	JJ
hansen	133	40	articles	article	NNS
hansen	133	41	,	,	,
hansen	133	42	make	make	VB
hansen	133	43	me	-PRON-	PRP
hansen	133	44	worry	worry	VB
hansen	133	45	.	.	.
hansen	134	1	While	while	IN
hansen	134	2	there	there	EX
hansen	134	3	will	will	MD
hansen	134	4	be	be	VB
hansen	134	5	many	many	JJ
hansen	134	6	cases	case	NNS
hansen	134	7	where	where	WRB
hansen	134	8	using	use	VBG
hansen	134	9	AI	AI	NNP
hansen	134	10	makes	make	VBZ
hansen	134	11	sense	sense	NN
hansen	134	12	,	,	,
hansen	134	13	I	-PRON-	PRP
hansen	134	14	want	want	VBP
hansen	134	15	to	to	TO
hansen	134	16	be	be	VB
hansen	134	17	sure	sure	JJ
hansen	134	18	libraries	library	NNS
hansen	134	19	are	be	VBP
hansen	134	20	asking	ask	VBG
hansen	134	21	themselves	-PRON-	PRP
hansen	134	22	a	a	DT
hansen	134	23	lot	lot	NN
hansen	134	24	of	of	IN
hansen	134	25	questions	question	NNS
hansen	134	26	before	before	IN
hansen	134	27	starting	start	VBG
hansen	134	28	to	to	TO
hansen	134	29	use	use	VB
hansen	134	30	it	-PRON-	PRP
hansen	134	31	.	.	.
hansen	135	1	Questions	question	NNS
hansen	135	2	like	like	IN
hansen	135	3	:	:	:
hansen	135	4	is	be	VBZ
hansen	135	5	this	this	DT
hansen	135	6	problem	problem	NN
hansen	135	7	large	large	JJ
hansen	135	8	enough	enough	RB
hansen	135	9	in	in	IN
hansen	135	10	scale	scale	NN
hansen	135	11	to	to	IN
hansen	135	12	substitute	substitute	NN
hansen	135	13	machines	machine	NNS
hansen	135	14	for	for	IN
hansen	135	15	human	human	JJ
hansen	135	16	labor	labor	NN
hansen	135	17	given	give	VBN
hansen	135	18	that	that	IN
hansen	135	19	machines	machine	NNS
hansen	135	20	will	will	MD
hansen	135	21	likely	likely	RB
hansen	135	22	be	be	VB
hansen	135	23	less	less	RBR
hansen	135	24	accurate	accurate	JJ
hansen	135	25	?	?	.
hansen	136	1	Or	or	CC
hansen	136	2	:	:	:
hansen	136	3	will	will	MD
hansen	136	4	using	use	VBG
hansen	136	5	machines	machine	NNS
hansen	136	6	to	to	TO
hansen	136	7	solve	solve	VB
hansen	136	8	this	this	DT
hansen	136	9	problem	problem	NN
hansen	136	10	cost	cost	VB
hansen	136	11	us	-PRON-	PRP
hansen	136	12	more	more	RBR
hansen	136	13	in	in	IN
hansen	136	14	equipment	equipment	NN
hansen	136	15	and	and	CC
hansen	136	16	highly	highly	RB
hansen	136	17	technical	technical	JJ
hansen	136	18	staff	staff	NN
hansen	136	19	than	than	IN
hansen	136	20	our	-PRON-	PRP$
hansen	136	21	current	current	JJ
hansen	136	22	solution	solution	NN
hansen	136	23	,	,	,
hansen	136	24	and	and	CC
hansen	136	25	has	have	VBZ
hansen	136	26	that	that	DT
hansen	136	27	factored	factor	VBN
hansen	136	28	in	in	IN
hansen	136	29	the	the	DT
hansen	136	30	people	people	NNS
hansen	136	31	and	and	CC
hansen	136	32	services	service	NNS
hansen	136	33	a	a	DT
hansen	136	34	library	library	NN
hansen	136	35	may	may	MD
hansen	136	36	need	need	VB
hansen	136	37	to	to	TO
hansen	136	38	cut	cut	VB
hansen	136	39	to	to	TO
hansen	136	40	afford	afford	VB
hansen	136	41	them	-PRON-	PRP
hansen	136	42	?	?	.
hansen	137	1	Not	not	RB
hansen	137	2	to	to	TO
hansen	137	3	mention	mention	VB
hansen	137	4	:	:	:
hansen	137	5	is	be	VBZ
hansen	137	6	this	this	DT
hansen	137	7	really	really	RB
hansen	137	8	a	a	DT
hansen	137	9	problem	problem	NN
hansen	137	10	or	or	CC
hansen	137	11	are	be	VBP
hansen	137	12	we	-PRON-	PRP
hansen	137	13	just	just	RB
hansen	137	14	looking	look	VBG
hansen	137	15	for	for	IN
hansen	137	16	a	a	DT
hansen	137	17	way	way	NN
hansen	137	18	to	to	TO
hansen	137	19	employ	employ	VB
hansen	137	20	machine	machine	NN
hansen	137	21	learning	learn	VBG
hansen	137	22	to	to	TO
hansen	137	23	say	say	VB
hansen	137	24	that	that	IN
hansen	137	25	we	-PRON-	PRP
hansen	137	26	did	do	VBD
hansen	137	27	?	?	.
hansen	138	1	In	in	IN
hansen	138	2	the	the	DT
hansen	138	3	cases	case	NNS
hansen	138	4	where	where	WRB
hansen	138	5	the	the	DT
hansen	138	6	answers	answer	NNS
hansen	138	7	to	to	IN
hansen	138	8	these	these	DT
hansen	138	9	questions	question	NNS
hansen	138	10	are	be	VBP
hansen	138	11	yes	yes	UH
hansen	138	12	,	,	,
hansen	138	13	it	-PRON-	PRP
hansen	138	14	will	will	MD
hansen	138	15	make	make	VB
hansen	138	16	sense	sense	NN
hansen	138	17	for	for	IN
hansen	138	18	libraries	library	NNS
hansen	138	19	to	to	TO
hansen	138	20	employ	employ	VB
hansen	138	21	machine	machine	NN
hansen	138	22	learning	learning	NN
hansen	138	23	.	.	.
hansen	139	1	I	-PRON-	PRP
hansen	139	2	just	just	RB
hansen	139	3	want	want	VBP
hansen	139	4	libraries	library	NNS
hansen	139	5	to	to	TO
hansen	139	6	look	look	VB
hansen	139	7	really	really	RB
hansen	139	8	carefully	carefully	RB
hansen	139	9	at	at	IN
hansen	139	10	how	how	WRB
hansen	139	11	they	-PRON-	PRP
hansen	139	12	approach	approach	VBP
hansen	139	13	problems	problem	NNS
hansen	139	14	and	and	CC
hansen	139	15	solutions	solution	NNS
hansen	139	16	,	,	,
hansen	139	17	to	to	TO
hansen	139	18	make	make	VB
hansen	139	19	sure	sure	JJ
hansen	139	20	that	that	IN
hansen	139	21	their	-PRON-	PRP$
hansen	139	22	problem	problem	NN
hansen	139	23	is	be	VBZ
hansen	139	24	,	,	,
hansen	139	25	in	in	IN
hansen	139	26	fact	fact	NN
hansen	139	27	,	,	,
hansen	139	28	a	a	DT
hansen	139	29	nail	nail	NN
hansen	139	30	,	,	,
hansen	139	31	and	and	CC
hansen	139	32	then	then	RB
hansen	139	33	to	to	TO
hansen	139	34	look	look	VB
hansen	139	35	even	even	RB
hansen	139	36	closer	close	RBR
hansen	139	37	and	and	CC
hansen	139	38	make	make	VB
hansen	139	39	sure	sure	JJ
hansen	139	40	it	-PRON-	PRP
hansen	139	41	is	be	VBZ
hansen	139	42	the	the	DT
hansen	139	43	type	type	NN
hansen	139	44	of	of	IN
hansen	139	45	nail	nail	NN
hansen	139	46	a	a	DT
hansen	139	47	machine	machine	NN
hansen	139	48	-	-	HYPH
hansen	139	49	learning	learn	VBG
hansen	139	50	hammer	hammer	NN
hansen	139	51	can	can	MD
hansen	139	52	hit	hit	VB
hansen	139	53	.	.	.
hansen	140	1	Comment	comment	NN
hansen	140	2	by	by	IN
hansen	140	3	Mark	Mark	NNP
hansen	140	4	Dehmlow	Dehmlow	NNP
hansen	140	5	:	:	:
hansen	140	6	Maybe	maybe	RB
hansen	140	7	"	"	``
hansen	140	8	think	think	VBP
hansen	140	9	analytically	analytically	RB
hansen	140	10	"	"	''
hansen	140	11	?	?	.
hansen	141	1	References	References	NNPS
hansen	141	2	Barthel	Barthel	NNP
hansen	141	3	,	,	,
hansen	141	4	Simon	Simon	NNP
hansen	141	5	,	,	,
hansen	141	6	Sascha	Sascha	NNP
hansen	141	7	Tönnies	Tönnies	NNP
hansen	141	8	,	,	,
hansen	141	9	and	and	CC
hansen	141	10	Wolf	Wolf	NNP
hansen	141	11	-	-	HYPH
hansen	141	12	Tilo	Tilo	NNP
hansen	141	13	Balke	Balke	NNP
hansen	141	14	.	.	.
hansen	142	1	2013	2013	CD
hansen	142	2	.	.	.
hansen	143	1	“	"	``
hansen	143	2	Large	large	JJ
hansen	143	3	-	-	HYPH
hansen	143	4	Scale	scale	NN
hansen	143	5	Experiments	experiment	NNS
hansen	143	6	for	for	IN
hansen	143	7	Mathematical	Mathematical	NNP
hansen	143	8	Document	Document	NNP
hansen	143	9	Classification	Classification	NNP
hansen	143	10	.	.	.
hansen	143	11	”	"	''
hansen	143	12	In	in	IN
hansen	143	13	Digital	Digital	NNP
hansen	143	14	Libraries	Libraries	NNPS
hansen	143	15	:	:	:
hansen	143	16	Social	Social	NNP
hansen	143	17	Media	Media	NNPS
hansen	143	18	and	and	CC
hansen	143	19	Community	Community	NNP
hansen	143	20	Networks	Networks	NNP
hansen	143	21	,	,	,
hansen	143	22	edited	edit	VBN
hansen	143	23	by	by	IN
hansen	143	24	Shalini	Shalini	NNP
hansen	143	25	R.	R.	NNP
hansen	143	26	Urs	Urs	NNP
hansen	143	27	,	,	,
hansen	143	28	Jin	Jin	NNP
hansen	143	29	-	-	HYPH
hansen	143	30	Cheon	Cheon	NNP
hansen	143	31	Na	Na	NNP
hansen	143	32	,	,	,
hansen	143	33	and	and	CC
hansen	143	34	George	George	NNP
hansen	143	35	Buchanan	Buchanan	NNP
hansen	143	36	,	,	,
hansen	143	37	83–92	83–92	NN
hansen	143	38	.	.	.
hansen	144	1	Cham	Cham	NNP
hansen	144	2	:	:	:
hansen	144	3	Springer	Springer	NNP
hansen	144	4	International	International	NNP
hansen	144	5	Publishing	Publishing	NNP
hansen	144	6	.	.	.
hansen	145	1	Clark	Clark	NNP
hansen	145	2	,	,	,
hansen	145	3	Jack	Jack	NNP
hansen	145	4	.	.	.
hansen	146	1	2015	2015	CD
hansen	146	2	.	.	.
hansen	147	1	“	"	``
hansen	147	2	Why	why	WRB
hansen	147	3	2015	2015	CD
hansen	147	4	Was	be	VBD
hansen	147	5	a	a	DT
hansen	147	6	Breakthrough	Breakthrough	NNP
hansen	147	7	Year	Year	NNP
hansen	147	8	in	in	IN
hansen	147	9	Artificial	Artificial	NNP
hansen	147	10	Intelligence	Intelligence	NNP
hansen	147	11	.	.	.
hansen	147	12	”	"	''
hansen	147	13	Bloomberg	Bloomberg	NNP
hansen	147	14	,	,	,
hansen	147	15	December	December	NNP
hansen	147	16	8	8	CD
hansen	147	17	,	,	,
hansen	147	18	2015	2015	CD
hansen	147	19	.	.	.
hansen	147	20	https://www.bloomberg.com/news/articles/2015-12-08/why-2015-was-a-breakthrough-year-in-artificial-intelligence	https://www.bloomberg.com/news/articles/2015-12-08/why-2015-was-a-breakthrough-year-in-artificial-intelligence	NNP
hansen	147	21	.	.	.
hansen	148	1	Gandhi	Gandhi	NNP
hansen	148	2	,	,	,
hansen	148	3	Rohith	Rohith	NNP
hansen	148	4	.	.	.
hansen	149	1	2018	2018	CD
hansen	149	2	.	.	.
hansen	150	1	“	"	``
hansen	150	2	Support	Support	NNP
hansen	150	3	Vector	Vector	NNP
hansen	150	4	Machine	Machine	NNP
hansen	150	5	—	—	:
hansen	150	6	Introduction	introduction	NN
hansen	150	7	to	to	IN
hansen	150	8	Machine	Machine	NNP
hansen	150	9	Learning	Learning	NNP
hansen	150	10	Algorithms	Algorithms	NNP
hansen	150	11	.	.	.
hansen	150	12	”	"	''
hansen	150	13	Medium	medium	NN
hansen	150	14	.	.	.
hansen	151	1	July	July	NNP
hansen	151	2	5	5	CD
hansen	151	3	,	,	,
hansen	151	4	2018	2018	CD
hansen	151	5	.	.	.
hansen	151	6	https://towardsdatascience.com/support-vector-machine-introduction-to-machine-learning-algorithms-934a444fca47	https://towardsdatascience.com/support-vector-machine-introduction-to-machine-learning-algorithms-934a444fca47	NNP
hansen	151	7	.	.	.
hansen	152	1	“	"	``
hansen	152	2	GeoDeepDive	geodeepdive	NN
hansen	152	3	:	:	:
hansen	152	4	Project	Project	NNP
hansen	152	5	Overview	Overview	NNP
hansen	152	6	.	.	.
hansen	152	7	”	"	''
hansen	152	8	n.d	n.d	NNP
hansen	152	9	.	.	NNP
hansen	152	10	Accessed	Accessed	NNP
hansen	152	11	May	May	NNP
hansen	152	12	7	7	CD
hansen	152	13	,	,	,
hansen	152	14	2018	2018	CD
hansen	152	15	.	.	.
hansen	152	16	https://geodeepdive.org/about.html	https://geodeepdive.org/about.html	ADD
hansen	152	17	.	.	.
hansen	153	1	Kelly	Kelly	NNP
hansen	153	2	,	,	,
hansen	153	3	Kevin	Kevin	NNP
hansen	153	4	.	.	.
hansen	154	1	2014	2014	CD
hansen	154	2	.	.	.
hansen	155	1	“	"	``
hansen	155	2	The	the	DT
hansen	155	3	Three	three	CD
hansen	155	4	Breakthroughs	Breakthroughs	NNPS
hansen	155	5	That	that	WDT
hansen	155	6	Have	have	VBP
hansen	155	7	Finally	finally	RB
hansen	155	8	Unleashed	unleash	VBN
hansen	155	9	AI	AI	NNP
hansen	155	10	on	on	IN
hansen	155	11	the	the	DT
hansen	155	12	World	World	NNP
hansen	155	13	.	.	.
hansen	155	14	”	"	''
hansen	155	15	Wired	wire	VBN
hansen	155	16	,	,	,
hansen	155	17	October	October	NNP
hansen	155	18	27	27	CD
hansen	155	19	,	,	,
hansen	155	20	2014	2014	CD
hansen	155	21	.	.	.
hansen	155	22	https://www.wired.com/2014/10/future-of-artificial-intelligence/.	https://www.wired.com/2014/10/future-of-artificial-intelligence/.	ADD
hansen	156	1	“	"	``
hansen	156	2	Microsoft	Microsoft	NNP
hansen	156	3	Academic	Academic	NNP
hansen	156	4	Increases	Increases	NNPS
hansen	156	5	Power	Power	NNP
hansen	156	6	of	of	IN
hansen	156	7	Semantic	Semantic	NNP
hansen	156	8	Search	Search	NNP
hansen	156	9	by	by	IN
hansen	156	10	Adding	add	VBG
hansen	156	11	More	More	JJR
hansen	156	12	Fields	field	NNS
hansen	156	13	of	of	IN
hansen	156	14	Study	Study	NNP
hansen	156	15	.	.	.
hansen	156	16	”	"	''
hansen	156	17	2018	2018	CD
hansen	156	18	.	.	.
hansen	157	1	Microsoft	Microsoft	NNP
hansen	157	2	Academic	Academic	NNP
hansen	157	3	(	(	-LRB-
hansen	157	4	blog	blog	NN
hansen	157	5	)	)	-RRB-
hansen	157	6	.	.	.
hansen	158	1	February	February	NNP
hansen	158	2	15	15	CD
hansen	158	3	,	,	,
hansen	158	4	2018	2018	CD
hansen	158	5	.	.	.
hansen	158	6	https://www.microsoft.com/en-us/research/project/academic/articles/microsoft-academic-increases-power-semantic-search-adding-fields-study/.	https://www.microsoft.com/en-us/research/project/academic/articles/microsoft-academic-increases-power-semantic-search-adding-fields-study/.	NNP
hansen	159	1	Mikolov	Mikolov	NNP
hansen	159	2	,	,	,
hansen	159	3	Tomas	Tomas	NNP
hansen	159	4	,	,	,
hansen	159	5	Ilya	Ilya	NNP
hansen	159	6	Sutskever	Sutskever	NNP
hansen	159	7	,	,	,
hansen	159	8	Kai	Kai	NNP
hansen	159	9	Chen	Chen	NNP
hansen	159	10	,	,	,
hansen	159	11	Greg	Greg	NNP
hansen	159	12	S	S	NNP
hansen	159	13	Corrado	Corrado	NNP
hansen	159	14	,	,	,
hansen	159	15	and	and	CC
hansen	159	16	Jeff	Jeff	NNP
hansen	159	17	Dean	Dean	NNP
hansen	159	18	.	.	.
hansen	160	1	2013	2013	CD
hansen	160	2	.	.	.
hansen	161	1	“	"	``
hansen	161	2	Distributed	distribute	VBN
hansen	161	3	Representations	Representations	NNPS
hansen	161	4	of	of	IN
hansen	161	5	Words	Words	NNPS
hansen	161	6	and	and	CC
hansen	161	7	Phrases	Phrases	NNPS
hansen	161	8	and	and	CC
hansen	161	9	Their	-PRON-	PRP$
hansen	161	10	Compositionality	compositionality	NN
hansen	161	11	.	.	.
hansen	161	12	”	"	''
hansen	161	13	In	in	IN
hansen	161	14	Advances	advance	NNS
hansen	161	15	in	in	IN
hansen	161	16	Neural	Neural	NNP
hansen	161	17	Information	Information	NNP
hansen	161	18	Processing	Processing	NNP
hansen	161	19	Systems	Systems	NNP
hansen	161	20	26	26	CD
hansen	161	21	,	,	,
hansen	161	22	edited	edit	VBN
hansen	161	23	by	by	IN
hansen	161	24	C.	C.	NNP
hansen	161	25	J.	J.	NNP
hansen	161	26	C.	C.	NNP
hansen	161	27	Burges	Burges	NNP
hansen	161	28	,	,	,
hansen	161	29	L.	L.	NNP
hansen	161	30	Bottou	Bottou	NNP
hansen	161	31	,	,	,
hansen	161	32	M.	M.	NNP
hansen	161	33	Welling	Welling	NNP
hansen	161	34	,	,	,
hansen	161	35	Z.	Z.	NNP
hansen	161	36	Ghahramani	Ghahramani	NNP
hansen	161	37	,	,	,
hansen	161	38	and	and	CC
hansen	161	39	K.	K.	NNP
hansen	161	40	Q.	Q.	NNP
hansen	161	41	Weinberger	Weinberger	NNP
hansen	161	42	,	,	,
hansen	161	43	3111–3119	3111–3119	CD
hansen	161	44	.	.	.
hansen	162	1	Curran	Curran	NNP
hansen	162	2	Associates	Associates	NNPS
hansen	162	3	,	,	,
hansen	162	4	Inc.	Inc.	NNP
hansen	162	5	http://papers.nips.cc/paper/5021-distributed-representations-of-words-and-phrases-and-their-compositionality.pdf	http://papers.nips.cc/paper/5021-distributed-representations-of-words-and-phrases-and-their-compositionality.pdf	NNP
hansen	162	6	.	.	.
hansen	163	1	Parloff	Parloff	NNP
hansen	163	2	,	,	,
hansen	163	3	Roger	Roger	NNP
hansen	163	4	.	.	.
hansen	164	1	2016	2016	CD
hansen	164	2	.	.	.
hansen	165	1	“	"	``
hansen	165	2	From	from	IN
hansen	165	3	2016	2016	CD
hansen	165	4	:	:	:
hansen	165	5	Why	why	WRB
hansen	165	6	Deep	deep	JJ
hansen	165	7	Learning	learning	NN
hansen	165	8	Is	be	VBZ
hansen	165	9	Suddenly	suddenly	RB
hansen	165	10	Changing	change	VBG
hansen	165	11	Your	-PRON-	PRP$
hansen	165	12	Life	life	NN
hansen	165	13	.	.	.
hansen	165	14	”	"	''
hansen	165	15	Fortune	Fortune	NNP
hansen	165	16	.	.	.
hansen	166	1	September	September	NNP
hansen	166	2	28	28	CD
hansen	166	3	,	,	,
hansen	166	4	2016	2016	CD
hansen	166	5	.	.	.
hansen	166	6	https://fortune.com/longform/ai-artificial-intelligence-deep-machine-learning/.	https://fortune.com/longform/ai-artificial-intelligence-deep-machine-learning/.	NNP
hansen	167	1	Rahimi	Rahimi	NNP
hansen	167	2	,	,	,
hansen	167	3	Ali	Ali	NNP
hansen	167	4	,	,	,
hansen	167	5	and	and	CC
hansen	167	6	Benjamin	Benjamin	NNP
hansen	167	7	Recht	Recht	NNP
hansen	167	8	.	.	.
hansen	168	1	2017	2017	CD
hansen	168	2	.	.	.
hansen	169	1	“	"	``
hansen	169	2	Back	back	RB
hansen	169	3	When	when	WRB
hansen	169	4	We	-PRON-	PRP
hansen	169	5	Were	be	VBD
hansen	169	6	Kids	kid	NNS
hansen	169	7	.	.	.
hansen	169	8	”	"	''
hansen	169	9	Presentation	presentation	NN
hansen	169	10	at	at	IN
hansen	169	11	the	the	DT
hansen	169	12	NIPS	NIPS	NNP
hansen	169	13	2017	2017	CD
hansen	169	14	Conference	Conference	NNP
hansen	169	15	.	.	.
hansen	170	1	https://www.youtube.com/watch?v=Qi1Yry33TQE	https://www.youtube.com/watch?v=Qi1Yry33TQE	NNP
hansen	170	2	.	.	.
hansen	171	1	Řehůřek	Řehůřek	NNP
hansen	171	2	,	,	,
hansen	171	3	Radim	Radim	NNP
hansen	171	4	,	,	,
hansen	171	5	and	and	CC
hansen	171	6	Petr	Petr	NNP
hansen	171	7	Sojka	sojka	RB
hansen	171	8	.	.	.
hansen	172	1	2008	2008	CD
hansen	172	2	.	.	.
hansen	173	1	“	"	``
hansen	173	2	Automated	Automated	NNP
hansen	173	3	Classification	Classification	NNP
hansen	173	4	and	and	CC
hansen	173	5	Categorization	Categorization	NNP
hansen	173	6	of	of	IN
hansen	173	7	 	  	_SP
hansen	173	8	Mathematical	mathematical	JJ
hansen	173	9	 	  	_SP
hansen	173	10	Knowledge	knowledge	NN
hansen	173	11	.	.	.
hansen	173	12	”	"	''
hansen	173	13	In	in	IN
hansen	173	14	Intelligent	Intelligent	NNP
hansen	173	15	Computer	Computer	NNP
hansen	173	16	Mathematics	Mathematics	NNP
hansen	173	17	,	,	,
hansen	173	18	edited	edit	VBN
hansen	173	19	by	by	IN
hansen	173	20	Serge	Serge	NNP
hansen	173	21	Autexier	Autexier	NNP
hansen	173	22	,	,	,
hansen	173	23	John	John	NNP
hansen	173	24	Campbell	Campbell	NNP
hansen	173	25	,	,	,
hansen	173	26	Julio	Julio	NNP
hansen	173	27	Rubio	Rubio	NNP
hansen	173	28	,	,	,
hansen	173	29	Volker	Volker	NNP
hansen	173	30	Sorge	Sorge	NNP
hansen	173	31	,	,	,
hansen	173	32	Masakazu	Masakazu	NNP
hansen	173	33	Suzuki	Suzuki	NNP
hansen	173	34	,	,	,
hansen	173	35	and	and	CC
hansen	173	36	Freek	Freek	NNP
hansen	173	37	Wiedijk	Wiedijk	NNP
hansen	173	38	,	,	,
hansen	173	39	543–57	543–57	CD
hansen	173	40	.	.	.
hansen	174	1	Berlin	Berlin	NNP
hansen	174	2	:	:	:
hansen	174	3	Springer	Springer	NNP
hansen	174	4	-	-	HYPH
hansen	174	5	Verlag	Verlag	NNP
hansen	174	6	.	.	.
hansen	175	1	Sangwani	Sangwani	NNP
hansen	175	2	,	,	,
hansen	175	3	Gaurav	Gaurav	NNP
hansen	175	4	.	.	.
hansen	176	1	2017	2017	CD
hansen	176	2	.	.	.
hansen	177	1	“	"	``
hansen	177	2	2017	2017	CD
hansen	177	3	Is	be	VBZ
hansen	177	4	the	the	DT
hansen	177	5	Year	Year	NNP
hansen	177	6	of	of	IN
hansen	177	7	Machine	Machine	NNP
hansen	177	8	Learning	Learning	NNP
hansen	177	9	.	.	.
hansen	178	1	Here	here	RB
hansen	178	2	’s	’	VBZ
hansen	178	3	Why	why	WRB
hansen	178	4	.	.	.
hansen	178	5	”	"	''
hansen	178	6	Business	Business	NNP
hansen	178	7	Insider	Insider	NNP
hansen	178	8	,	,	,
hansen	178	9	January	January	NNP
hansen	178	10	13	13	CD
hansen	178	11	,	,	,
hansen	178	12	2017	2017	CD
hansen	178	13	.	.	.
hansen	178	14	https://www.businessinsider.in/2017-is-the-year-of-machine-learning-heres-why/articleshow/56514535.cms	https://www.businessinsider.in/2017-is-the-year-of-machine-learning-heres-why/articleshow/56514535.cms	NNP
hansen	178	15	.	.	.
hansen	179	1	Shen	Shen	NNP
hansen	179	2	,	,	,
hansen	179	3	Zhihong	Zhihong	NNP
hansen	179	4	,	,	,
hansen	179	5	Hao	Hao	NNP
hansen	179	6	Ma	Ma	NNP
hansen	179	7	,	,	,
hansen	179	8	and	and	CC
hansen	179	9	Kuansan	Kuansan	NNP
hansen	179	10	Wang	Wang	NNP
hansen	179	11	.	.	.
hansen	180	1	2018	2018	CD
hansen	180	2	.	.	.
hansen	181	1	“	"	``
hansen	181	2	A	a	DT
hansen	181	3	Web	web	NN
hansen	181	4	-	-	HYPH
hansen	181	5	Scale	scale	NN
hansen	181	6	System	system	NN
hansen	181	7	for	for	IN
hansen	181	8	Scientific	Scientific	NNP
hansen	181	9	Knowledge	Knowledge	NNP
hansen	181	10	Exploration	Exploration	NNP
hansen	181	11	.	.	.
hansen	181	12	”	"	''
hansen	181	13	Paper	paper	NN
hansen	181	14	presented	present	VBD
hansen	181	15	at	at	IN
hansen	181	16	the	the	DT
hansen	181	17	56th	56th	JJ
hansen	181	18	Annual	Annual	NNP
hansen	181	19	Meeting	Meeting	NNP
hansen	181	20	of	of	IN
hansen	181	21	the	the	DT
hansen	181	22	Association	Association	NNP
hansen	181	23	for	for	IN
hansen	181	24	Computational	Computational	NNP
hansen	181	25	Linguistics	Linguistics	NNPS
hansen	181	26	,	,	,
hansen	181	27	Melbourne	Melbourne	NNP
hansen	181	28	,	,	,
hansen	181	29	July	July	NNP
hansen	181	30	2018	2018	CD
hansen	181	31	.	.	.
hansen	181	32	http://arxiv.org/abs/1805.12216	http://arxiv.org/abs/1805.12216	NNP
hansen	181	33	.	.	.
hansen	182	1	Tank	tank	NN
hansen	182	2	,	,	,
hansen	182	3	Aytekin	Aytekin	NNP
hansen	182	4	.	.	.
hansen	183	1	2017	2017	CD
hansen	183	2	.	.	.
hansen	184	1	“	"	``
hansen	184	2	This	this	DT
hansen	184	3	Is	be	VBZ
hansen	184	4	the	the	DT
hansen	184	5	Year	Year	NNP
hansen	184	6	of	of	IN
hansen	184	7	the	the	DT
hansen	184	8	Machine	Machine	NNP
hansen	184	9	Learning	Learning	NNP
hansen	184	10	Revolution	Revolution	NNP
hansen	184	11	.	.	.
hansen	184	12	”	"	''
hansen	184	13	Entrepreneur	entrepreneur	NN
hansen	184	14	,	,	,
hansen	184	15	January	January	NNP
hansen	184	16	12	12	CD
hansen	184	17	,	,	,
hansen	184	18	2017	2017	CD
hansen	184	19	.	.	.
hansen	184	20	https://www.entrepreneur.com/article/287324	https://www.entrepreneur.com/article/287324	UH
hansen	184	21	.	.	.
hansen	185	1	Editor	editor	NN
hansen	185	2	Notes	Notes	NNP
hansen	185	3	There	there	EX
hansen	185	4	are	be	VBP
hansen	185	5	some	some	DT
hansen	185	6	additional	additional	JJ
hansen	185	7	stylistic	stylistic	JJ
hansen	185	8	fixes	fix	NNS
hansen	185	9	Mark	Mark	NNP
hansen	185	10	and	and	CC
hansen	185	11	I	-PRON-	PRP
hansen	185	12	picked	pick	VBD
hansen	185	13	up	up	RP
hansen	185	14	on	on	IN
hansen	185	15	,	,	,
hansen	185	16	but	but	CC
hansen	185	17	those	those	DT
hansen	185	18	were	be	VBD
hansen	185	19	easily	easily	RB
hansen	185	20	made	make	VBN
hansen	185	21	,	,	,
hansen	185	22	and	and	CC
hansen	185	23	there	there	EX
hansen	185	24	's	be	VBZ
hansen	185	25	only	only	RB
hansen	185	26	a	a	DT
hansen	185	27	handful	handful	NN
hansen	185	28	of	of	IN
hansen	185	29	very	very	RB
hansen	185	30	small	small	JJ
hansen	185	31	issues	issue	NNS
hansen	185	32	for	for	IN
hansen	185	33	Samuel	Samuel	NNP
hansen	185	34	to	to	TO
hansen	185	35	look	look	VB
hansen	185	36	at	at	IN
hansen	185	37	.	.	.
hansen	186	1	The	the	DT
hansen	186	2	author	author	NN
hansen	186	3	addressed	address	VBD
hansen	186	4	the	the	DT
hansen	186	5	concerns	concern	NNS
hansen	186	6	we	-PRON-	PRP
hansen	186	7	had	have	VBD
hansen	186	8	,	,	,
hansen	186	9	especially	especially	RB
hansen	186	10	about	about	IN
hansen	186	11	the	the	DT
hansen	186	12	conclusion	conclusion	NN
hansen	186	13	,	,	,
hansen	186	14	which	which	WDT
hansen	186	15	is	be	VBZ
hansen	186	16	nicely	nicely	RB
hansen	186	17	moderated	moderate	VBN
hansen	186	18	and	and	CC
hansen	186	19	melds	meld	VBZ
hansen	186	20	together	together	RB
hansen	186	21	the	the	DT
hansen	186	22	implications	implication	NNS
hansen	186	23	of	of	IN
hansen	186	24	the	the	DT
hansen	186	25	strengths	strength	NNS
hansen	186	26	/	/	SYM
hansen	186	27	weaknesses	weakness	NNS
hansen	186	28	of	of	IN
hansen	186	29	ML	ML	NNP
hansen	186	30	classifying	classify	VBG
hansen	186	31	articulated	articulate	VBN
hansen	186	32	earlier	early	RBR
hansen	186	33	in	in	IN
hansen	186	34	the	the	DT
hansen	186	35	article	article	NN
hansen	186	36	.	.	.