id	sid	tid	token	lemma	pos
cril-1051	1	1	microsoft	microsoft	PROPN
cril-1051	1	2	word	word	PROPN
cril-1051	1	3	moeller-cril2021-proof_final.docx	moeller-cril2021-proof_final.docx	VERB
cril-1051	1	4	1	1	NUM
cril-1051	1	5	computational	computational	ADJ
cril-1051	1	6	morphology	morphology	NOUN
cril-1051	1	7	for	for	ADP
cril-1051	1	8	language	language	NOUN
cril-1051	1	9	description	description	NOUN
cril-1051	1	10	and	and	CCONJ
cril-1051	1	11	documentation	documentation	NOUN
cril-1051	1	12	sarah	sarah	PROPN
cril-1051	1	13	moeller	moeller	PROPN
cril-1051	1	14	university	university	PROPN
cril-1051	1	15	of	of	ADP
cril-1051	1	16	colorado	colorado	PROPN
cril-1051	1	17	boulder	boulder	PROPN
cril-1051	1	18	while	while	SCONJ
cril-1051	1	19	the	the	DET
cril-1051	1	20	field	field	NOUN
cril-1051	1	21	of	of	ADP
cril-1051	1	22	linguistics	linguistic	NOUN
cril-1051	1	23	has	have	AUX
cril-1051	1	24	slowly	slowly	ADV
cril-1051	1	25	but	but	CCONJ
cril-1051	1	26	surely	surely	ADV
cril-1051	1	27	widened	widen	VERB
cril-1051	1	28	the	the	DET
cril-1051	1	29	world	world	NOUN
cril-1051	1	30	’s	’s	PART
cril-1051	1	31	knowledge	knowledge	NOUN
cril-1051	1	32	about	about	ADP
cril-1051	1	33	human	human	ADJ
cril-1051	1	34	language	language	NOUN
cril-1051	1	35	,	,	PUNCT
cril-1051	1	36	thanks	thank	NOUN
cril-1051	1	37	in	in	ADP
cril-1051	1	38	part	part	NOUN
cril-1051	1	39	to	to	ADP
cril-1051	1	40	the	the	DET
cril-1051	1	41	recent	recent	ADJ
cril-1051	1	42	emphasis	emphasis	NOUN
cril-1051	1	43	on	on	ADP
cril-1051	1	44	language	language	NOUN
cril-1051	1	45	documentation	documentation	NOUN
cril-1051	1	46	and	and	CCONJ
cril-1051	1	47	description	description	NOUN
cril-1051	1	48	of	of	ADP
cril-1051	1	49	underdocumented	underdocumented	ADJ
cril-1051	1	50	languages	language	NOUN
cril-1051	1	51	,	,	PUNCT
cril-1051	1	52	computational	computational	ADJ
cril-1051	1	53	linguistics	linguistic	NOUN
cril-1051	1	54	has	have	AUX
cril-1051	1	55	barely	barely	ADV
cril-1051	1	56	expanded	expand	VERB
cril-1051	1	57	beyond	beyond	ADP
cril-1051	1	58	a	a	DET
cril-1051	1	59	handful	handful	NOUN
cril-1051	1	60	of	of	ADP
cril-1051	1	61	economically	economically	ADV
cril-1051	1	62	or	or	CCONJ
cril-1051	1	63	politically	politically	ADV
cril-1051	1	64	powerful	powerful	ADJ
cril-1051	1	65	languages	language	NOUN
cril-1051	1	66	.	.	PUNCT
cril-1051	2	1	this	this	DET
cril-1051	2	2	paper	paper	NOUN
cril-1051	2	3	is	be	AUX
cril-1051	2	4	a	a	DET
cril-1051	2	5	synthesis	synthesis	NOUN
cril-1051	2	6	of	of	ADP
cril-1051	2	7	natural	natural	ADJ
cril-1051	2	8	language	language	NOUN
cril-1051	2	9	processing	processing	NOUN
cril-1051	2	10	(	(	PUNCT
cril-1051	2	11	nlp	nlp	NOUN
cril-1051	2	12	)	)	PUNCT
cril-1051	2	13	models	model	NOUN
cril-1051	2	14	and	and	CCONJ
cril-1051	2	15	methods	method	NOUN
cril-1051	2	16	and	and	CCONJ
cril-1051	2	17	a	a	DET
cril-1051	2	18	history	history	NOUN
cril-1051	2	19	about	about	ADP
cril-1051	2	20	how	how	SCONJ
cril-1051	2	21	the	the	DET
cril-1051	2	22	models	model	NOUN
cril-1051	2	23	and	and	CCONJ
cril-1051	2	24	methods	method	NOUN
cril-1051	2	25	have	have	AUX
cril-1051	2	26	been	be	AUX
cril-1051	2	27	applied	apply	VERB
cril-1051	2	28	to	to	ADP
cril-1051	2	29	the	the	DET
cril-1051	2	30	study	study	NOUN
cril-1051	2	31	of	of	ADP
cril-1051	2	32	morphological	morphological	ADJ
cril-1051	2	33	structure	structure	NOUN
cril-1051	2	34	,	,	PUNCT
cril-1051	2	35	particularly	particularly	ADV
cril-1051	2	36	in	in	ADP
cril-1051	2	37	low	low	ADJ
cril-1051	2	38	-	-	PUNCT
cril-1051	2	39	resource	resource	NOUN
cril-1051	2	40	languages	language	NOUN
cril-1051	2	41	(	(	PUNCT
cril-1051	2	42	lrl	lrl	PROPN
cril-1051	2	43	)	)	PUNCT
cril-1051	2	44	.	.	PUNCT
cril-1051	3	1	the	the	DET
cril-1051	3	2	paper	paper	NOUN
cril-1051	3	3	assumes	assume	VERB
cril-1051	3	4	that	that	SCONJ
cril-1051	3	5	the	the	DET
cril-1051	3	6	study	study	NOUN
cril-1051	3	7	of	of	ADP
cril-1051	3	8	morphology	morphology	NOUN
cril-1051	3	9	has	have	VERB
cril-1051	3	10	an	an	DET
cril-1051	3	11	important	important	ADJ
cril-1051	3	12	role	role	NOUN
cril-1051	3	13	to	to	PART
cril-1051	3	14	play	play	VERB
cril-1051	3	15	in	in	ADP
cril-1051	3	16	both	both	CCONJ
cril-1051	3	17	nlp	nlp	NOUN
cril-1051	3	18	and	and	CCONJ
cril-1051	3	19	linguistics	linguistic	NOUN
cril-1051	3	20	.	.	PUNCT
cril-1051	4	1	it	it	PRON
cril-1051	4	2	explores	explore	VERB
cril-1051	4	3	the	the	DET
cril-1051	4	4	potential	potential	NOUN
cril-1051	4	5	for	for	ADP
cril-1051	4	6	discovering	discover	VERB
cril-1051	4	7	newer	new	ADJ
cril-1051	4	8	and	and	CCONJ
cril-1051	4	9	more	more	ADV
cril-1051	4	10	efficient	efficient	ADJ
cril-1051	4	11	methods	method	NOUN
cril-1051	4	12	while	while	SCONJ
cril-1051	4	13	training	train	VERB
cril-1051	4	14	computational	computational	ADJ
cril-1051	4	15	morphological	morphological	ADJ
cril-1051	4	16	models	model	NOUN
cril-1051	4	17	on	on	ADP
cril-1051	4	18	data	datum	NOUN
cril-1051	4	19	produced	produce	VERB
cril-1051	4	20	during	during	ADP
cril-1051	4	21	language	language	NOUN
cril-1051	4	22	documentation	documentation	NOUN
cril-1051	4	23	and	and	CCONJ
cril-1051	4	24	description	description	NOUN
cril-1051	4	25	(	(	PUNCT
cril-1051	4	26	ldd	ldd	PROPN
cril-1051	4	27	)	)	PUNCT
cril-1051	4	28	field	field	NOUN
cril-1051	4	29	projects.1	projects.1	ADJ
cril-1051	4	30	keywords	keyword	NOUN
cril-1051	4	31	:	:	PUNCT
cril-1051	4	32	language	language	NOUN
cril-1051	4	33	documentation	documentation	NOUN
cril-1051	4	34	,	,	PUNCT
cril-1051	4	35	natural	natural	ADJ
cril-1051	4	36	language	language	NOUN
cril-1051	4	37	processing	processing	NOUN
cril-1051	4	38	,	,	PUNCT
cril-1051	4	39	nlp	nlp	ADJ
cril-1051	4	40	,	,	PUNCT
cril-1051	4	41	low	low	ADJ
cril-1051	4	42	-	-	PUNCT
cril-1051	4	43	resource	resource	NOUN
cril-1051	4	44	languages	language	NOUN
cril-1051	4	45	,	,	PUNCT
cril-1051	4	46	machine	machine	NOUN
cril-1051	4	47	learning	learn	VERB
cril-1051	4	48	1	1	NUM
cril-1051	4	49	.	.	PUNCT
cril-1051	4	50	introduction	introduction	NOUN
cril-1051	4	51	morphology	morphology	NOUN
cril-1051	4	52	comprises	comprise	VERB
cril-1051	4	53	word	word	NOUN
cril-1051	4	54	-	-	PUNCT
cril-1051	4	55	building	building	NOUN
cril-1051	4	56	properties	property	NOUN
cril-1051	4	57	in	in	ADP
cril-1051	4	58	human	human	ADJ
cril-1051	4	59	languages	language	NOUN
cril-1051	4	60	and	and	CCONJ
cril-1051	4	61	their	their	PRON
cril-1051	4	62	accompanying	accompanying	ADJ
cril-1051	4	63	(	(	PUNCT
cril-1051	4	64	morpho-)syntactic	morpho-)syntactic	ADJ
cril-1051	4	65	phenomena	phenomenon	NOUN
cril-1051	4	66	.	.	PUNCT
cril-1051	5	1	historically	historically	ADV
cril-1051	5	2	,	,	PUNCT
cril-1051	5	3	computational	computational	ADJ
cril-1051	5	4	linguists	linguist	NOUN
cril-1051	5	5	and	and	CCONJ
cril-1051	5	6	“	"	PUNCT
cril-1051	5	7	paper	paper	NOUN
cril-1051	5	8	-	-	PUNCT
cril-1051	5	9	and	and	CCONJ
cril-1051	5	10	-	-	PUNCT
cril-1051	5	11	pencil	pencil	NOUN
cril-1051	5	12	linguists	linguist	NOUN
cril-1051	5	13	”	"	PUNCT
cril-1051	5	14	have	have	AUX
cril-1051	5	15	taken	take	VERB
cril-1051	5	16	different	different	ADJ
cril-1051	5	17	and	and	CCONJ
cril-1051	5	18	sometimes	sometimes	ADV
cril-1051	5	19	seemingly	seemingly	ADV
cril-1051	5	20	incompatible	incompatible	ADJ
cril-1051	5	21	approaches	approach	NOUN
cril-1051	5	22	to	to	ADP
cril-1051	5	23	morphology	morphology	NOUN
cril-1051	5	24	(	(	PUNCT
cril-1051	5	25	karttunen	karttunen	PROPN
cril-1051	5	26	&	&	CCONJ
cril-1051	5	27	beesley	beesley	PROPN
cril-1051	5	28	2005	2005	NUM
cril-1051	5	29	)	)	PUNCT
cril-1051	5	30	.	.	PUNCT
cril-1051	6	1	yet	yet	ADV
cril-1051	6	2	,	,	PUNCT
cril-1051	6	3	despite	despite	SCONJ
cril-1051	6	4	their	their	PRON
cril-1051	6	5	out	out	ADP
cril-1051	6	6	-	-	PUNCT
cril-1051	6	7	of	of	ADP
cril-1051	6	8	-	-	PUNCT
cril-1051	6	9	sync	sync	NOUN
cril-1051	6	10	approaches	approach	NOUN
cril-1051	6	11	,	,	PUNCT
cril-1051	6	12	both	both	PRON
cril-1051	6	13	computational	computational	ADJ
cril-1051	6	14	linguistics	linguistic	NOUN
cril-1051	6	15	and	and	CCONJ
cril-1051	6	16	“	"	PUNCT
cril-1051	6	17	traditional	traditional	ADJ
cril-1051	6	18	”	"	PUNCT
cril-1051	6	19	linguistics	linguistic	NOUN
cril-1051	6	20	benefit	benefit	VERB
cril-1051	6	21	from	from	ADP
cril-1051	6	22	morphological	morphological	ADJ
cril-1051	6	23	analysis	analysis	NOUN
cril-1051	6	24	(	(	PUNCT
cril-1051	6	25	cotterell	cotterell	NOUN
cril-1051	6	26	et	et	PROPN
cril-1051	6	27	al	al	PROPN
cril-1051	6	28	.	.	PROPN
cril-1051	6	29	2015	2015	NUM
cril-1051	6	30	)	)	PUNCT
cril-1051	6	31	.	.	PUNCT
cril-1051	7	1	for	for	ADP
cril-1051	7	2	natural	natural	ADJ
cril-1051	7	3	language	language	NOUN
cril-1051	7	4	processing	processing	NOUN
cril-1051	7	5	(	(	PUNCT
cril-1051	7	6	nlp	nlp	NOUN
cril-1051	7	7	)	)	PUNCT
cril-1051	7	8	,	,	PUNCT
cril-1051	7	9	work	work	VERB
cril-1051	7	10	with	with	ADP
cril-1051	7	11	low	low	ADJ
cril-1051	7	12	-	-	PUNCT
cril-1051	7	13	resource	resource	NOUN
cril-1051	7	14	languages	language	NOUN
cril-1051	7	15	(	(	PUNCT
cril-1051	7	16	lrl	lrl	PROPN
cril-1051	7	17	)	)	PUNCT
cril-1051	7	18	is	be	AUX
cril-1051	7	19	still	still	ADV
cril-1051	7	20	largely	largely	ADV
cril-1051	7	21	uncharted	uncharte	VERB
cril-1051	7	22	territory	territory	NOUN
cril-1051	7	23	.	.	PUNCT
cril-1051	8	1	this	this	DET
cril-1051	8	2	paper	paper	NOUN
cril-1051	8	3	explores	explore	VERB
cril-1051	8	4	the	the	DET
cril-1051	8	5	limited	limited	ADJ
cril-1051	8	6	work	work	NOUN
cril-1051	8	7	in	in	ADP
cril-1051	8	8	morphology	morphology	NOUN
cril-1051	8	9	by	by	ADP
cril-1051	8	10	asking	ask	VERB
cril-1051	8	11	this	this	DET
cril-1051	8	12	question	question	NOUN
cril-1051	8	13	:	:	PUNCT
cril-1051	8	14	“	"	PUNCT
cril-1051	8	15	what	what	PRON
cril-1051	8	16	[	[	X
cril-1051	8	17	computational	computational	ADJ
cril-1051	8	18	]	]	X
cril-1051	8	19	methods	method	NOUN
cril-1051	8	20	...	...	PUNCT
cril-1051	8	21	can	can	AUX
cril-1051	8	22	detect	detect	VERB
cril-1051	8	23	[	[	X
cril-1051	8	24	morphological	morphological	ADJ
cril-1051	8	25	]	]	PUNCT
cril-1051	8	26	structure	structure	NOUN
cril-1051	8	27	in	in	ADP
cril-1051	8	28	small	small	ADJ
cril-1051	8	29	,	,	PUNCT
cril-1051	8	30	noisy	noisy	ADJ
cril-1051	8	31	data	datum	NOUN
cril-1051	8	32	sets	set	NOUN
cril-1051	8	33	,	,	PUNCT
cril-1051	8	34	while	while	SCONJ
cril-1051	8	35	being	be	AUX
cril-1051	8	36	directly	directly	ADV
cril-1051	8	37	applicable	applicable	ADJ
cril-1051	8	38	to	to	ADP
cril-1051	8	39	a	a	DET
cril-1051	8	40	wide	wide	ADJ
cril-1051	8	41	variety	variety	NOUN
cril-1051	8	42	of	of	ADP
cril-1051	8	43	languages	language	NOUN
cril-1051	8	44	?	?	PUNCT
cril-1051	8	45	”	"	PUNCT
cril-1051	9	1	(	(	PUNCT
cril-1051	9	2	bird	bird	NOUN
cril-1051	9	3	2009	2009	NUM
cril-1051	9	4	)	)	PUNCT
cril-1051	9	5	.	.	PUNCT
cril-1051	10	1	the	the	DET
cril-1051	10	2	paper	paper	NOUN
cril-1051	10	3	is	be	AUX
cril-1051	10	4	organized	organize	VERB
cril-1051	10	5	as	as	SCONJ
cril-1051	10	6	follows	follow	VERB
cril-1051	10	7	.	.	PUNCT
cril-1051	11	1	section	section	NOUN
cril-1051	11	2	2	2	NUM
cril-1051	11	3	describes	describe	VERB
cril-1051	11	4	the	the	DET
cril-1051	11	5	workflow	workflow	NOUN
cril-1051	11	6	and	and	CCONJ
cril-1051	11	7	activities	activity	NOUN
cril-1051	11	8	of	of	ADP
cril-1051	11	9	ldd	ldd	PROPN
cril-1051	11	10	.	.	PUNCT
cril-1051	12	1	section	section	PROPN
cril-1051	12	2	3	3	NUM
cril-1051	12	3	sketches	sketch	VERB
cril-1051	12	4	the	the	DET
cril-1051	12	5	history	history	NOUN
cril-1051	12	6	of	of	ADP
cril-1051	12	7	nlp	nlp	ADJ
cril-1051	12	8	work	work	NOUN
cril-1051	12	9	with	with	ADP
cril-1051	12	10	lrl	lrl	PROPN
cril-1051	12	11	.	.	PUNCT
cril-1051	13	1	section	section	NOUN
cril-1051	13	2	4	4	NUM
cril-1051	13	3	defines	define	VERB
cril-1051	13	4	morphological	morphological	ADJ
cril-1051	13	5	analysis	analysis	NOUN
cril-1051	13	6	and	and	CCONJ
cril-1051	13	7	looks	look	VERB
cril-1051	13	8	specifically	specifically	ADV
cril-1051	13	9	at	at	ADP
cril-1051	13	10	nlp	nlp	NOUN
cril-1051	13	11	applied	apply	VERB
cril-1051	13	12	to	to	ADP
cril-1051	13	13	morpheme	morpheme	ADJ
cril-1051	13	14	segmentation	segmentation	NOUN
cril-1051	13	15	and	and	CCONJ
cril-1051	13	16	glossing	glossing	NOUN
cril-1051	13	17	and	and	CCONJ
cril-1051	13	18	section	section	NOUN
cril-1051	13	19	5	5	NUM
cril-1051	13	20	looks	look	NOUN
cril-1051	13	21	at	at	ADP
cril-1051	13	22	the	the	DET
cril-1051	13	23	application	application	NOUN
cril-1051	13	24	to	to	ADP
cril-1051	13	25	learning	learn	VERB
cril-1051	13	26	morphological	morphological	ADJ
cril-1051	13	27	inflectional	inflectional	ADJ
cril-1051	13	28	paradigmatic	paradigmatic	ADJ
cril-1051	13	29	patterns	pattern	NOUN
cril-1051	13	30	.	.	PUNCT
cril-1051	14	1	colorado	colorado	PROPN
cril-1051	14	2	research	research	PROPN
cril-1051	14	3	in	in	ADP
cril-1051	14	4	linguistics	linguistic	NOUN
cril-1051	14	5	,	,	PUNCT
cril-1051	14	6	volume	volume	NOUN
cril-1051	14	7	25	25	NUM
cril-1051	14	8	(	(	PUNCT
cril-1051	14	9	2021	2021	NUM
cril-1051	14	10	)	)	PUNCT
cril-1051	14	11	2	2	NUM
cril-1051	14	12	2	2	NUM
cril-1051	14	13	.	.	PUNCT
cril-1051	14	14	language	language	NOUN
cril-1051	14	15	documentation	documentation	NOUN
cril-1051	14	16	and	and	CCONJ
cril-1051	14	17	description	description	NOUN
cril-1051	14	18	in	in	ADP
cril-1051	14	19	linguistics	linguistic	NOUN
cril-1051	14	20	,	,	PUNCT
cril-1051	14	21	morphological	morphological	ADJ
cril-1051	14	22	description	description	NOUN
cril-1051	14	23	of	of	ADP
cril-1051	14	24	a	a	DET
cril-1051	14	25	broad	broad	ADJ
cril-1051	14	26	range	range	NOUN
cril-1051	14	27	of	of	ADP
cril-1051	14	28	languages	language	NOUN
cril-1051	14	29	is	be	AUX
cril-1051	14	30	a	a	DET
cril-1051	14	31	foundational	foundational	ADJ
cril-1051	14	32	step	step	NOUN
cril-1051	14	33	towards	towards	ADP
cril-1051	14	34	any	any	DET
cril-1051	14	35	reasonable	reasonable	ADJ
cril-1051	14	36	linguistic	linguistic	ADJ
cril-1051	14	37	theory	theory	NOUN
cril-1051	14	38	.	.	PUNCT
cril-1051	15	1	a	a	DET
cril-1051	15	2	focus	focus	NOUN
cril-1051	15	3	of	of	ADP
cril-1051	15	4	documenting	document	VERB
cril-1051	15	5	and	and	CCONJ
cril-1051	15	6	describing	describe	VERB
cril-1051	15	7	underdocumented	underdocumented	ADJ
cril-1051	15	8	languages	language	NOUN
cril-1051	15	9	,	,	PUNCT
cril-1051	15	10	including	include	VERB
cril-1051	15	11	their	their	PRON
cril-1051	15	12	morphological	morphological	ADJ
cril-1051	15	13	structure	structure	NOUN
cril-1051	15	14	,	,	PUNCT
cril-1051	15	15	has	have	AUX
cril-1051	15	16	been	be	AUX
cril-1051	15	17	emphasized	emphasize	VERB
cril-1051	15	18	since	since	SCONJ
cril-1051	15	19	the	the	DET
cril-1051	15	20	1990	1990	NUM
cril-1051	15	21	’s	’s	NOUN
cril-1051	15	22	along	along	ADP
cril-1051	15	23	with	with	ADP
cril-1051	15	24	the	the	DET
cril-1051	15	25	development	development	NOUN
cril-1051	15	26	of	of	ADP
cril-1051	15	27	language	language	NOUN
cril-1051	15	28	documentation	documentation	NOUN
cril-1051	15	29	as	as	ADP
cril-1051	15	30	a	a	DET
cril-1051	15	31	distinct	distinct	ADJ
cril-1051	15	32	subfield	subfield	NOUN
cril-1051	15	33	.	.	PUNCT
cril-1051	16	1	himmelmann	himmelmann	NOUN
cril-1051	16	2	(	(	PUNCT
cril-1051	16	3	1998	1998	NUM
cril-1051	16	4	)	)	PUNCT
cril-1051	16	5	defines	define	VERB
cril-1051	16	6	language	language	NOUN
cril-1051	16	7	documentation	documentation	NOUN
cril-1051	16	8	as	as	ADP
cril-1051	16	9	“	"	PUNCT
cril-1051	16	10	a	a	DET
cril-1051	16	11	comprehensive	comprehensive	ADJ
cril-1051	16	12	and	and	CCONJ
cril-1051	16	13	representative	representative	ADJ
cril-1051	16	14	sample	sample	NOUN
cril-1051	16	15	of	of	ADP
cril-1051	16	16	communicative	communicative	ADJ
cril-1051	16	17	events	event	NOUN
cril-1051	16	18	[	[	PUNCT
cril-1051	16	19	that	that	PRON
cril-1051	16	20	are	be	AUX
cril-1051	16	21	]	]	PUNCT
cril-1051	16	22	as	as	ADV
cril-1051	16	23	natural	natural	ADJ
cril-1051	16	24	as	as	ADP
cril-1051	16	25	possible	possible	ADJ
cril-1051	16	26	.	.	PUNCT
cril-1051	16	27	”	"	PUNCT
cril-1051	17	1	woodbury	woodbury	PROPN
cril-1051	17	2	(	(	PUNCT
cril-1051	17	3	2003	2003	NUM
cril-1051	17	4	)	)	PUNCT
cril-1051	17	5	defines	define	VERB
cril-1051	17	6	it	it	PRON
cril-1051	17	7	similarly	similarly	ADV
cril-1051	17	8	as	as	ADP
cril-1051	17	9	“	"	PUNCT
cril-1051	17	10	comprehensive	comprehensive	ADJ
cril-1051	17	11	and	and	CCONJ
cril-1051	17	12	transparent	transparent	ADJ
cril-1051	17	13	records	record	NOUN
cril-1051	17	14	supporting	support	VERB
cril-1051	17	15	wide	wide	ADJ
cril-1051	17	16	ranging	range	VERB
cril-1051	17	17	scientific	scientific	ADJ
cril-1051	17	18	investigations	investigation	NOUN
cril-1051	17	19	of	of	ADP
cril-1051	17	20	the	the	DET
cril-1051	17	21	language	language	NOUN
cril-1051	17	22	.	.	PUNCT
cril-1051	17	23	”	"	PUNCT
cril-1051	18	1	language	language	NOUN
cril-1051	18	2	description	description	NOUN
cril-1051	18	3	can	can	AUX
cril-1051	18	4	be	be	AUX
cril-1051	18	5	defined	define	VERB
cril-1051	18	6	as	as	ADP
cril-1051	18	7	work	work	NOUN
cril-1051	18	8	that	that	PRON
cril-1051	18	9	analyzes	analyze	VERB
cril-1051	18	10	language	language	NOUN
cril-1051	18	11	documentation	documentation	NOUN
cril-1051	18	12	to	to	PART
cril-1051	18	13	create	create	VERB
cril-1051	18	14	“	"	PUNCT
cril-1051	18	15	systematic	systematic	ADJ
cril-1051	18	16	presentations	presentation	NOUN
cril-1051	18	17	of	of	ADP
cril-1051	18	18	the	the	DET
cril-1051	18	19	phonology	phonology	NOUN
cril-1051	18	20	,	,	PUNCT
cril-1051	18	21	morphology	morphology	NOUN
cril-1051	18	22	,	,	PUNCT
cril-1051	18	23	syntax	syntax	NOUN
cril-1051	18	24	,	,	PUNCT
cril-1051	18	25	and	and	CCONJ
cril-1051	18	26	semantics	semantic	NOUN
cril-1051	18	27	of	of	ADP
cril-1051	18	28	the	the	DET
cril-1051	18	29	language	language	NOUN
cril-1051	18	30	”	"	PUNCT
cril-1051	18	31	(	(	PUNCT
cril-1051	18	32	bird	bird	NOUN
cril-1051	18	33	&	&	CCONJ
cril-1051	18	34	chiang	chiang	PROPN
cril-1051	18	35	2012	2012	NUM
cril-1051	18	36	)	)	PUNCT
cril-1051	18	37	.	.	PUNCT
cril-1051	19	1	the	the	DET
cril-1051	19	2	emphasis	emphasis	NOUN
cril-1051	19	3	on	on	ADP
cril-1051	19	4	endangered	endanger	VERB
cril-1051	19	5	languages	language	NOUN
cril-1051	19	6	over	over	ADP
cril-1051	19	7	the	the	DET
cril-1051	19	8	past	past	ADJ
cril-1051	19	9	three	three	NUM
cril-1051	19	10	decades	decade	NOUN
cril-1051	19	11	has	have	AUX
cril-1051	19	12	established	establish	VERB
cril-1051	19	13	best	good	ADJ
cril-1051	19	14	practices	practice	NOUN
cril-1051	19	15	documenting	document	VERB
cril-1051	19	16	a	a	DET
cril-1051	19	17	new	new	ADJ
cril-1051	19	18	language	language	NOUN
cril-1051	19	19	(	(	PUNCT
cril-1051	19	20	bowern	bowern	ADJ
cril-1051	19	21	2008	2008	NUM
cril-1051	19	22	;	;	PUNCT
cril-1051	19	23	czaykowskahiggins	czaykowskahiggin	NOUN
cril-1051	19	24	2009	2009	NUM
cril-1051	19	25	;	;	PUNCT
cril-1051	19	26	lupke	lupke	PROPN
cril-1051	19	27	2010	2010	NUM
cril-1051	19	28	;	;	PUNCT
cril-1051	19	29	vallejos	vallejos	NOUN
cril-1051	19	30	2014	2014	NUM
cril-1051	19	31	;	;	PUNCT
cril-1051	19	32	rice	rice	NOUN
cril-1051	19	33	&	&	CCONJ
cril-1051	19	34	thunder	thunder	NOUN
cril-1051	19	35	2017	2017	NUM
cril-1051	19	36	)	)	PUNCT
cril-1051	19	37	.	.	PUNCT
cril-1051	20	1	however	however	ADV
cril-1051	20	2	,	,	PUNCT
cril-1051	20	3	the	the	DET
cril-1051	20	4	specific	specific	ADJ
cril-1051	20	5	activities	activity	NOUN
cril-1051	20	6	that	that	PRON
cril-1051	20	7	divide	divide	VERB
cril-1051	20	8	the	the	DET
cril-1051	20	9	two	two	NUM
cril-1051	20	10	subfields	subfield	NOUN
cril-1051	20	11	are	be	AUX
cril-1051	20	12	not	not	PART
cril-1051	20	13	rigid	rigid	ADJ
cril-1051	20	14	.	.	PUNCT
cril-1051	21	1	therefore	therefore	ADV
cril-1051	21	2	,	,	PUNCT
cril-1051	21	3	the	the	DET
cril-1051	21	4	current	current	ADJ
cril-1051	21	5	work	work	NOUN
cril-1051	21	6	generally	generally	ADV
cril-1051	21	7	refers	refer	VERB
cril-1051	21	8	to	to	ADP
cril-1051	21	9	them	they	PRON
cril-1051	21	10	together	together	ADV
cril-1051	21	11	as	as	ADP
cril-1051	21	12	“	"	PUNCT
cril-1051	21	13	language	language	NOUN
cril-1051	21	14	documentation	documentation	NOUN
cril-1051	21	15	and	and	CCONJ
cril-1051	21	16	description	description	NOUN
cril-1051	21	17	(	(	PUNCT
cril-1051	21	18	ldd	ldd	PROPN
cril-1051	21	19	)	)	PUNCT
cril-1051	21	20	”	"	PUNCT
cril-1051	21	21	or	or	CCONJ
cril-1051	21	22	“	"	PUNCT
cril-1051	21	23	documentary	documentary	NOUN
cril-1051	21	24	and	and	CCONJ
cril-1051	21	25	descriptive	descriptive	ADJ
cril-1051	21	26	linguistics	linguistic	NOUN
cril-1051	21	27	”	"	PUNCT
cril-1051	21	28	.	.	PUNCT
cril-1051	22	1	the	the	DET
cril-1051	22	2	workflow	workflow	NOUN
cril-1051	22	3	of	of	ADP
cril-1051	22	4	ldd	ldd	PROPN
cril-1051	22	5	is	be	AUX
cril-1051	22	6	not	not	PART
cril-1051	22	7	standardized	standardize	VERB
cril-1051	22	8	,	,	PUNCT
cril-1051	22	9	although	although	SCONJ
cril-1051	22	10	most	most	ADJ
cril-1051	22	11	projects	project	NOUN
cril-1051	22	12	seem	seem	VERB
cril-1051	22	13	to	to	PART
cril-1051	22	14	follow	follow	VERB
cril-1051	22	15	a	a	DET
cril-1051	22	16	similar	similar	ADJ
cril-1051	22	17	sequence	sequence	NOUN
cril-1051	22	18	.	.	PUNCT
cril-1051	23	1	a	a	DET
cril-1051	23	2	version	version	NOUN
cril-1051	23	3	of	of	ADP
cril-1051	23	4	one	one	NUM
cril-1051	23	5	common	common	ADJ
cril-1051	23	6	sequence	sequence	NOUN
cril-1051	23	7	(	(	PUNCT
cril-1051	23	8	bird	bird	NOUN
cril-1051	23	9	&	&	CCONJ
cril-1051	23	10	chiang	chiang	PROPN
cril-1051	23	11	2012	2012	NUM
cril-1051	23	12	)	)	PUNCT
cril-1051	23	13	is	be	AUX
cril-1051	23	14	given	give	VERB
cril-1051	23	15	below	below	ADV
cril-1051	23	16	(	(	PUNCT
cril-1051	23	17	the	the	DET
cril-1051	23	18	numbers	number	NOUN
cril-1051	23	19	are	be	AUX
cril-1051	23	20	used	use	VERB
cril-1051	23	21	to	to	PART
cril-1051	23	22	refer	refer	VERB
cril-1051	23	23	to	to	ADP
cril-1051	23	24	each	each	DET
cril-1051	23	25	task	task	NOUN
cril-1051	23	26	,	,	PUNCT
cril-1051	23	27	e.g.	e.g.	ADV
cril-1051	23	28	,	,	PUNCT
cril-1051	23	29	“	"	PUNCT
cril-1051	23	30	task	task	NOUN
cril-1051	23	31	2a	2a	NUM
cril-1051	23	32	”	"	PUNCT
cril-1051	23	33	refers	refer	VERB
cril-1051	23	34	to	to	ADP
cril-1051	23	35	transcription	transcription	NOUN
cril-1051	23	36	)	)	PUNCT
cril-1051	23	37	.	.	PUNCT
cril-1051	24	1	each	each	DET
cril-1051	24	2	subsequent	subsequent	ADJ
cril-1051	24	3	task	task	NOUN
cril-1051	24	4	progressively	progressively	ADV
cril-1051	24	5	encompasses	encompass	VERB
cril-1051	24	6	more	more	ADJ
cril-1051	24	7	description	description	NOUN
cril-1051	24	8	than	than	ADP
cril-1051	24	9	documentation	documentation	NOUN
cril-1051	24	10	,	,	PUNCT
cril-1051	24	11	except	except	SCONJ
cril-1051	24	12	archiving	archive	VERB
cril-1051	24	13	,	,	PUNCT
cril-1051	24	14	which	which	PRON
cril-1051	24	15	comes	come	VERB
cril-1051	24	16	strictly	strictly	ADV
cril-1051	24	17	under	under	ADP
cril-1051	24	18	language	language	NOUN
cril-1051	24	19	documentation	documentation	NOUN
cril-1051	24	20	but	but	CCONJ
cril-1051	24	21	is	be	AUX
cril-1051	24	22	logically	logically	ADV
cril-1051	24	23	a	a	DET
cril-1051	24	24	last	last	ADJ
cril-1051	24	25	step	step	NOUN
cril-1051	24	26	.	.	PUNCT
cril-1051	25	1	(	(	PUNCT
cril-1051	25	2	1	1	X
cril-1051	25	3	)	)	PUNCT
cril-1051	25	4	collect	collect	NOUN
cril-1051	25	5	(	(	PUNCT
cril-1051	25	6	audio	audio	ADJ
cril-1051	25	7	/	/	SYM
cril-1051	25	8	video	video	NOUN
cril-1051	25	9	recordings	recording	NOUN
cril-1051	25	10	)	)	PUNCT
cril-1051	25	11	naturally	naturally	ADV
cril-1051	25	12	occurring	occur	VERB
cril-1051	25	13	speech	speech	NOUN
cril-1051	25	14	(	(	PUNCT
cril-1051	25	15	2	2	NUM
cril-1051	25	16	)	)	PUNCT
cril-1051	25	17	a	a	DET
cril-1051	25	18	)	)	PUNCT
cril-1051	25	19	transcribe	transcribe	NOUN
cril-1051	25	20	and	and	CCONJ
cril-1051	25	21	b	b	X
cril-1051	25	22	)	)	PUNCT
cril-1051	25	23	translate	translate	NOUN
cril-1051	25	24	(	(	PUNCT
cril-1051	25	25	3	3	X
cril-1051	25	26	)	)	PUNCT
cril-1051	25	27	perform	perform	VERB
cril-1051	25	28	basic	basic	ADJ
cril-1051	25	29	morphosyntactic	morphosyntactic	ADJ
cril-1051	25	30	analysis	analysis	NOUN
cril-1051	25	31	by	by	ADP
cril-1051	25	32	segmenting	segment	VERB
cril-1051	25	33	the	the	DET
cril-1051	25	34	morphemes	morpheme	NOUN
cril-1051	25	35	and	and	CCONJ
cril-1051	25	36	creating	create	VERB
cril-1051	25	37	morphological	morphological	ADJ
cril-1051	25	38	glosses	gloss	NOUN
cril-1051	25	39	and/or	and/or	CCONJ
cril-1051	25	40	a	a	DET
cril-1051	25	41	lexicon	lexicon	NOUN
cril-1051	25	42	(	(	PUNCT
cril-1051	25	43	4	4	NUM
cril-1051	25	44	)	)	PUNCT
cril-1051	25	45	elicit	elicit	ADJ
cril-1051	25	46	morphological	morphological	ADJ
cril-1051	25	47	paradigms	paradigm	NOUN
cril-1051	25	48	that	that	PRON
cril-1051	25	49	reveal	reveal	VERB
cril-1051	25	50	underlying	underlie	VERB
cril-1051	25	51	patterns	pattern	NOUN
cril-1051	25	52	(	(	PUNCT
cril-1051	25	53	5	5	X
cril-1051	25	54	)	)	PUNCT
cril-1051	25	55	prepare	prepare	VERB
cril-1051	25	56	descriptive	descriptive	ADJ
cril-1051	25	57	reports	report	NOUN
cril-1051	25	58	that	that	PRON
cril-1051	25	59	outline	outline	VERB
cril-1051	25	60	the	the	DET
cril-1051	25	61	language	language	NOUN
cril-1051	25	62	’s	’s	PART
cril-1051	25	63	structure	structure	NOUN
cril-1051	25	64	(	(	PUNCT
cril-1051	25	65	6	6	NUM
cril-1051	25	66	)	)	PUNCT
cril-1051	25	67	archive	archive	NOUN
cril-1051	25	68	data	datum	NOUN
cril-1051	25	69	in	in	ADP
cril-1051	25	70	a	a	DET
cril-1051	25	71	long	long	ADJ
cril-1051	25	72	-	-	PUNCT
cril-1051	25	73	term	term	NOUN
cril-1051	25	74	repository	repository	NOUN
cril-1051	25	75	computational	computational	ADJ
cril-1051	25	76	morphology	morphology	NOUN
cril-1051	25	77	for	for	ADP
cril-1051	25	78	language	language	NOUN
cril-1051	25	79	description	description	NOUN
cril-1051	25	80	and	and	CCONJ
cril-1051	25	81	documentation	documentation	NOUN
cril-1051	25	82	3	3	NUM
cril-1051	25	83	one	one	NUM
cril-1051	25	84	primary	primary	ADJ
cril-1051	25	85	output	output	NOUN
cril-1051	25	86	of	of	ADP
cril-1051	25	87	this	this	DET
cril-1051	25	88	workflow	workflow	NOUN
cril-1051	25	89	is	be	AUX
cril-1051	25	90	interlinear	interlinear	ADJ
cril-1051	25	91	glossed	glossed	ADJ
cril-1051	25	92	texts	text	NOUN
cril-1051	25	93	(	(	PUNCT
cril-1051	25	94	igt	igt	PROPN
cril-1051	25	95	)	)	PUNCT
cril-1051	25	96	,	,	PUNCT
cril-1051	25	97	a	a	DET
cril-1051	25	98	data	data	NOUN
cril-1051	25	99	format	format	NOUN
cril-1051	25	100	distinctive	distinctive	ADJ
cril-1051	25	101	to	to	ADP
cril-1051	25	102	linguistics	linguistic	NOUN
cril-1051	25	103	(	(	PUNCT
cril-1051	25	104	figure	figure	NOUN
cril-1051	25	105	1	1	NUM
cril-1051	25	106	)	)	PUNCT
cril-1051	25	107	.	.	PUNCT
cril-1051	26	1	interlinearization	interlinearization	NOUN
cril-1051	26	2	is	be	AUX
cril-1051	26	3	the	the	DET
cril-1051	26	4	primary	primary	ADJ
cril-1051	26	5	task	task	NOUN
cril-1051	26	6	after	after	ADP
cril-1051	26	7	transcription	transcription	NOUN
cril-1051	26	8	.	.	PUNCT
cril-1051	27	1	it	it	PRON
cril-1051	27	2	moves	move	VERB
cril-1051	27	3	the	the	DET
cril-1051	27	4	workflow	workflow	NOUN
cril-1051	27	5	beyond	beyond	ADP
cril-1051	27	6	simple	simple	ADJ
cril-1051	27	7	documentation	documentation	NOUN
cril-1051	27	8	but	but	CCONJ
cril-1051	27	9	still	still	ADV
cril-1051	27	10	serves	serve	VERB
cril-1051	27	11	as	as	ADP
cril-1051	27	12	a	a	DET
cril-1051	27	13	“	"	PUNCT
cril-1051	27	14	preprocessing	preprocessing	NOUN
cril-1051	27	15	step	step	NOUN
cril-1051	27	16	”	"	PUNCT
cril-1051	27	17	to	to	PART
cril-1051	27	18	language	language	NOUN
cril-1051	27	19	description	description	NOUN
cril-1051	27	20	(	(	PUNCT
cril-1051	27	21	strictly	strictly	ADV
cril-1051	27	22	defined	define	VERB
cril-1051	27	23	)	)	PUNCT
cril-1051	27	24	(	(	PUNCT
cril-1051	27	25	moon	moon	NOUN
cril-1051	27	26	,	,	PUNCT
cril-1051	27	27	erk	erk	PROPN
cril-1051	27	28	&	&	CCONJ
cril-1051	27	29	baldridge	baldridge	PROPN
cril-1051	27	30	2009	2009	NUM
cril-1051	27	31	)	)	PUNCT
cril-1051	27	32	.	.	PUNCT
cril-1051	28	1	it	it	PRON
cril-1051	28	2	comprises	comprise	VERB
cril-1051	28	3	annotation	annotation	NOUN
cril-1051	28	4	tasks	task	NOUN
cril-1051	28	5	that	that	PRON
cril-1051	28	6	enrich	enrich	VERB
cril-1051	28	7	the	the	DET
cril-1051	28	8	data	datum	NOUN
cril-1051	28	9	with	with	ADP
cril-1051	28	10	analytic	analytic	ADJ
cril-1051	28	11	information	information	NOUN
cril-1051	28	12	added	add	VERB
cril-1051	28	13	as	as	ADP
cril-1051	28	14	lines	line	NOUN
cril-1051	28	15	under	under	ADP
cril-1051	28	16	the	the	DET
cril-1051	28	17	transcribed	transcribe	VERB
cril-1051	28	18	text	text	NOUN
cril-1051	28	19	(	(	PUNCT
cril-1051	28	20	task	task	NOUN
cril-1051	28	21	1	1	NUM
cril-1051	28	22	in	in	ADP
cril-1051	28	23	above	above	ADP
cril-1051	28	24	workflow	workflow	NOUN
cril-1051	28	25	;	;	PUNCT
cril-1051	28	26	line	line	NOUN
cril-1051	28	27	1	1	NUM
cril-1051	28	28	in	in	ADP
cril-1051	28	29	figure	figure	NOUN
cril-1051	28	30	1	1	NUM
cril-1051	28	31	)	)	PUNCT
cril-1051	28	32	.	.	PUNCT
cril-1051	29	1	the	the	DET
cril-1051	29	2	most	most	ADV
cril-1051	29	3	common	common	ADJ
cril-1051	29	4	lines	line	NOUN
cril-1051	29	5	are	be	AUX
cril-1051	29	6	shown	show	VERB
cril-1051	29	7	in	in	ADP
cril-1051	29	8	figure	figure	NOUN
cril-1051	29	9	1	1	NUM
cril-1051	29	10	.	.	PUNCT
cril-1051	30	1	lines	line	NOUN
cril-1051	30	2	can	can	AUX
cril-1051	30	3	be	be	AUX
cril-1051	30	4	added	add	VERB
cril-1051	30	5	in	in	ADP
cril-1051	30	6	any	any	DET
cril-1051	30	7	order	order	NOUN
cril-1051	30	8	,	,	PUNCT
cril-1051	30	9	but	but	CCONJ
cril-1051	30	10	translations	translation	NOUN
cril-1051	30	11	(	(	PUNCT
cril-1051	30	12	task	task	NOUN
cril-1051	30	13	2a	2a	NUM
cril-1051	30	14	;	;	PUNCT
cril-1051	30	15	line	line	NOUN
cril-1051	30	16	7	7	NUM
cril-1051	30	17	)	)	PUNCT
cril-1051	30	18	morpheme	morpheme	NOUN
cril-1051	30	19	boundaries	boundary	NOUN
cril-1051	30	20	and	and	CCONJ
cril-1051	30	21	morpheme	morpheme	NOUN
cril-1051	30	22	glosses	gloss	NOUN
cril-1051	30	23	(	(	PUNCT
cril-1051	30	24	task	task	NOUN
cril-1051	30	25	3	3	NUM
cril-1051	30	26	;	;	PUNCT
cril-1051	30	27	lines	line	NOUN
cril-1051	30	28	2	2	NUM
cril-1051	30	29	and	and	CCONJ
cril-1051	30	30	3	3	NUM
cril-1051	30	31	,	,	PUNCT
cril-1051	30	32	respectively	respectively	ADV
cril-1051	30	33	)	)	PUNCT
cril-1051	30	34	are	be	AUX
cril-1051	30	35	usually	usually	ADV
cril-1051	30	36	added	add	VERB
cril-1051	30	37	first	first	ADV
cril-1051	30	38	.	.	PUNCT
cril-1051	31	1	doing	do	VERB
cril-1051	31	2	more	more	ADJ
cril-1051	31	3	annotation	annotation	NOUN
cril-1051	31	4	(	(	PUNCT
cril-1051	31	5	e.g.	e.g.	ADV
cril-1051	31	6	lines	line	NOUN
cril-1051	31	7	4	4	NUM
cril-1051	31	8	-	-	SYM
cril-1051	31	9	6	6	NUM
cril-1051	31	10	)	)	PUNCT
cril-1051	31	11	often	often	ADV
cril-1051	31	12	happens	happen	VERB
cril-1051	31	13	in	in	ADP
cril-1051	31	14	field	field	NOUN
cril-1051	31	15	projects	project	NOUN
cril-1051	31	16	,	,	PUNCT
cril-1051	31	17	but	but	CCONJ
cril-1051	31	18	translation	translation	NOUN
cril-1051	31	19	,	,	PUNCT
cril-1051	31	20	morpheme	morpheme	NOUN
cril-1051	31	21	segmentation	segmentation	NOUN
cril-1051	31	22	,	,	PUNCT
cril-1051	31	23	and	and	CCONJ
cril-1051	31	24	morpheme	morpheme	NOUN
cril-1051	31	25	glossing	glossing	NOUN
cril-1051	31	26	are	be	AUX
cril-1051	31	27	usually	usually	ADV
cril-1051	31	28	given	give	VERB
cril-1051	31	29	the	the	DET
cril-1051	31	30	highest	high	ADJ
cril-1051	31	31	priority	priority	NOUN
cril-1051	31	32	.	.	PUNCT
cril-1051	32	1	figure	figure	NOUN
cril-1051	32	2	1	1	NUM
cril-1051	32	3	.	.	PUNCT
cril-1051	33	1	interlinearization	interlinearization	NOUN
cril-1051	33	2	:	:	PUNCT
cril-1051	33	3	interlinear	interlinear	ADJ
cril-1051	33	4	glossed	gloss	VERB
cril-1051	33	5	texts	text	NOUN
cril-1051	33	6	add	add	VERB
cril-1051	33	7	lines	line	NOUN
cril-1051	33	8	of	of	ADP
cril-1051	33	9	annotation	annotation	NOUN
cril-1051	33	10	to	to	ADP
cril-1051	33	11	the	the	DET
cril-1051	33	12	original	original	ADJ
cril-1051	33	13	text	text	NOUN
cril-1051	33	14	.	.	PUNCT
cril-1051	34	1	interlinearizing	interlinearize	VERB
cril-1051	34	2	data	datum	NOUN
cril-1051	34	3	uncovers	uncover	NOUN
cril-1051	34	4	the	the	DET
cril-1051	34	5	rarer	rare	ADJ
cril-1051	34	6	and	and	CCONJ
cril-1051	34	7	unique	unique	ADJ
cril-1051	34	8	linguistic	linguistic	ADJ
cril-1051	34	9	phenomena	phenomenon	NOUN
cril-1051	34	10	.	.	PUNCT
cril-1051	35	1	interlinearization	interlinearization	NOUN
cril-1051	35	2	opens	open	VERB
cril-1051	35	3	the	the	DET
cril-1051	35	4	door	door	NOUN
cril-1051	35	5	for	for	ADP
cril-1051	35	6	deeper	deep	ADJ
cril-1051	35	7	linguistic	linguistic	ADJ
cril-1051	35	8	analysis	analysis	NOUN
cril-1051	35	9	and	and	CCONJ
cril-1051	35	10	lays	lay	VERB
cril-1051	35	11	the	the	DET
cril-1051	35	12	foundation	foundation	NOUN
cril-1051	35	13	for	for	ADP
cril-1051	35	14	reference	reference	NOUN
cril-1051	35	15	grammars	grammar	NOUN
cril-1051	35	16	,	,	PUNCT
cril-1051	35	17	dictionaries	dictionary	NOUN
cril-1051	35	18	,	,	PUNCT
cril-1051	35	19	and	and	CCONJ
cril-1051	35	20	language	language	NOUN
cril-1051	35	21	learning	learn	VERB
cril-1051	35	22	materials	material	NOUN
cril-1051	35	23	,	,	PUNCT
cril-1051	35	24	but	but	CCONJ
cril-1051	35	25	interlinearization	interlinearization	NOUN
cril-1051	35	26	is	be	AUX
cril-1051	35	27	not	not	PART
cril-1051	35	28	sufficient	sufficient	ADJ
cril-1051	35	29	to	to	PART
cril-1051	35	30	create	create	VERB
cril-1051	35	31	complete	complete	ADJ
cril-1051	35	32	grammars	grammar	NOUN
cril-1051	35	33	,	,	PUNCT
cril-1051	35	34	dictionaries	dictionary	NOUN
cril-1051	35	35	,	,	PUNCT
cril-1051	35	36	etc	etc	X
cril-1051	35	37	.	.	X
cril-1051	35	38	one	one	NUM
cril-1051	35	39	additional	additional	ADJ
cril-1051	35	40	descriptive	descriptive	ADJ
cril-1051	35	41	task	task	NOUN
cril-1051	35	42	is	be	AUX
cril-1051	35	43	often	often	ADV
cril-1051	35	44	included	include	VERB
cril-1051	35	45	:	:	PUNCT
cril-1051	35	46	the	the	DET
cril-1051	35	47	collection	collection	NOUN
cril-1051	35	48	of	of	ADP
cril-1051	35	49	morphological	morphological	ADJ
cril-1051	35	50	inflection	inflection	NOUN
cril-1051	35	51	patterns	pattern	NOUN
cril-1051	35	52	,	,	PUNCT
cril-1051	35	53	or	or	CCONJ
cril-1051	35	54	paradigms	paradigm	VERB
cril-1051	35	55	,	,	PUNCT
cril-1051	35	56	for	for	ADP
cril-1051	35	57	several	several	ADJ
cril-1051	35	58	lemmata	lemmata	ADJ
cril-1051	35	59	(	(	PUNCT
cril-1051	35	60	task	task	NOUN
cril-1051	35	61	4	4	NUM
cril-1051	35	62	)	)	PUNCT
cril-1051	35	63	.	.	PUNCT
cril-1051	36	1	inflectional	inflectional	ADJ
cril-1051	36	2	colorado	colorado	PROPN
cril-1051	36	3	research	research	NOUN
cril-1051	36	4	in	in	ADP
cril-1051	36	5	linguistics	linguistic	NOUN
cril-1051	36	6	,	,	PUNCT
cril-1051	36	7	volume	volume	NOUN
cril-1051	36	8	25	25	NUM
cril-1051	36	9	(	(	PUNCT
cril-1051	36	10	2021	2021	NUM
cril-1051	36	11	)	)	PUNCT
cril-1051	36	12	4	4	NUM
cril-1051	36	13	paradigms	paradigm	NOUN
cril-1051	36	14	are	be	AUX
cril-1051	36	15	elicited	elicit	VERB
cril-1051	36	16	because	because	SCONJ
cril-1051	36	17	complete	complete	ADJ
cril-1051	36	18	paradigms	paradigm	NOUN
cril-1051	36	19	are	be	AUX
cril-1051	36	20	rarely	rarely	ADV
cril-1051	36	21	found	find	VERB
cril-1051	36	22	in	in	ADP
cril-1051	36	23	natural	natural	ADJ
cril-1051	36	24	language	language	NOUN
cril-1051	36	25	.	.	PUNCT
cril-1051	37	1	complete	complete	ADJ
cril-1051	37	2	paradigms	paradigm	NOUN
cril-1051	37	3	are	be	AUX
cril-1051	37	4	needed	need	VERB
cril-1051	37	5	to	to	PART
cril-1051	37	6	infer	infer	VERB
cril-1051	37	7	general	general	ADJ
cril-1051	37	8	rules	rule	NOUN
cril-1051	37	9	of	of	ADP
cril-1051	37	10	inflection	inflection	NOUN
cril-1051	37	11	.	.	PUNCT
cril-1051	38	1	without	without	ADP
cril-1051	38	2	translations	translation	NOUN
cril-1051	38	3	,	,	PUNCT
cril-1051	38	4	morpheme	morpheme	ADJ
cril-1051	38	5	segmentation	segmentation	NOUN
cril-1051	38	6	,	,	PUNCT
cril-1051	38	7	and	and	CCONJ
cril-1051	38	8	glossing	gloss	VERB
cril-1051	38	9	,	,	PUNCT
cril-1051	38	10	the	the	DET
cril-1051	38	11	data	data	NOUN
cril-1051	38	12	is	be	AUX
cril-1051	38	13	understandable	understandable	ADJ
cril-1051	38	14	only	only	ADV
cril-1051	38	15	to	to	ADP
cril-1051	38	16	someone	someone	PRON
cril-1051	38	17	who	who	PRON
cril-1051	38	18	already	already	ADV
cril-1051	38	19	speaks	speak	VERB
cril-1051	38	20	the	the	DET
cril-1051	38	21	language	language	NOUN
cril-1051	38	22	.	.	PUNCT
cril-1051	39	1	if	if	SCONJ
cril-1051	39	2	no	no	DET
cril-1051	39	3	speakers	speaker	NOUN
cril-1051	39	4	are	be	AUX
cril-1051	39	5	left	leave	VERB
cril-1051	39	6	,	,	PUNCT
cril-1051	39	7	the	the	DET
cril-1051	39	8	data	data	NOUN
cril-1051	39	9	is	be	AUX
cril-1051	39	10	mostly	mostly	ADV
cril-1051	39	11	inaccessible	inaccessible	ADJ
cril-1051	39	12	,	,	PUNCT
cril-1051	39	13	much	much	ADV
cril-1051	39	14	like	like	ADP
cril-1051	39	15	egyptian	egyptian	ADJ
cril-1051	39	16	hieroglyphics	hieroglyphic	NOUN
cril-1051	39	17	before	before	SCONJ
cril-1051	39	18	the	the	DET
cril-1051	39	19	rosetta	rosetta	PROPN
cril-1051	39	20	stone	stone	NOUN
cril-1051	39	21	was	be	AUX
cril-1051	39	22	discovered	discover	VERB
cril-1051	39	23	.	.	PUNCT
cril-1051	40	1	a	a	DET
cril-1051	40	2	few	few	ADJ
cril-1051	40	3	specially	specially	ADV
cril-1051	40	4	designed	design	VERB
cril-1051	40	5	software	software	NOUN
cril-1051	40	6	tools	tool	NOUN
cril-1051	40	7	provide	provide	VERB
cril-1051	40	8	limited	limited	ADJ
cril-1051	40	9	automated	automate	VERB
cril-1051	40	10	assistance	assistance	NOUN
cril-1051	40	11	.	.	PUNCT
cril-1051	41	1	the	the	DET
cril-1051	41	2	two	two	NUM
cril-1051	41	3	most	most	ADV
cril-1051	41	4	popular	popular	ADJ
cril-1051	41	5	are	be	AUX
cril-1051	41	6	elan	elan	NOUN
cril-1051	41	7	(	(	PUNCT
cril-1051	41	8	auer	auer	NOUN
cril-1051	41	9	et	et	PROPN
cril-1051	41	10	al	al	PROPN
cril-1051	41	11	.	.	PROPN
cril-1051	41	12	2010	2010	NUM
cril-1051	41	13	)	)	PUNCT
cril-1051	41	14	and	and	CCONJ
cril-1051	41	15	flex	flex	ADJ
cril-1051	41	16	(	(	PUNCT
cril-1051	41	17	rogers	rogers	PROPN
cril-1051	41	18	2010	2010	NUM
cril-1051	41	19	)	)	PUNCT
cril-1051	41	20	.	.	PUNCT
cril-1051	42	1	examples	example	NOUN
cril-1051	42	2	of	of	ADP
cril-1051	42	3	their	their	PRON
cril-1051	42	4	interlinearization	interlinearization	NOUN
cril-1051	42	5	interfaces	interface	NOUN
cril-1051	42	6	are	be	AUX
cril-1051	42	7	shown	show	VERB
cril-1051	42	8	in	in	ADP
cril-1051	42	9	figures	figure	NOUN
cril-1051	42	10	2	2	NUM
cril-1051	42	11	and	and	CCONJ
cril-1051	42	12	3	3	NUM
cril-1051	42	13	.	.	PUNCT
cril-1051	43	1	these	these	DET
cril-1051	43	2	tools	tool	NOUN
cril-1051	43	3	implement	implement	VERB
cril-1051	43	4	hand	hand	NOUN
cril-1051	43	5	-	-	PUNCT
cril-1051	43	6	constructed	construct	VERB
cril-1051	43	7	,	,	PUNCT
cril-1051	43	8	rule	rule	NOUN
cril-1051	43	9	-	-	PUNCT
cril-1051	43	10	based	base	VERB
cril-1051	43	11	computational	computational	ADJ
cril-1051	43	12	morphological	morphological	ADJ
cril-1051	43	13	parsers	parser	NOUN
cril-1051	43	14	but	but	CCONJ
cril-1051	43	15	rule	rule	NOUN
cril-1051	43	16	-	-	PUNCT
cril-1051	43	17	based	base	VERB
cril-1051	43	18	parsers	parser	NOUN
cril-1051	43	19	do	do	AUX
cril-1051	43	20	not	not	PART
cril-1051	43	21	generalize	generalize	VERB
cril-1051	43	22	to	to	ADP
cril-1051	43	23	new	new	ADJ
cril-1051	43	24	data	datum	NOUN
cril-1051	43	25	.	.	PUNCT
cril-1051	44	1	flex	flex	ADJ
cril-1051	44	2	also	also	ADV
cril-1051	44	3	copies	copy	VERB
cril-1051	44	4	morpheme	morpheme	ADJ
cril-1051	44	5	boundaries	boundary	NOUN
cril-1051	44	6	and	and	CCONJ
cril-1051	44	7	glosses	gloss	NOUN
cril-1051	44	8	onto	onto	ADP
cril-1051	44	9	other	other	ADJ
cril-1051	44	10	words	word	NOUN
cril-1051	44	11	if	if	SCONJ
cril-1051	44	12	they	they	PRON
cril-1051	44	13	are	be	AUX
cril-1051	44	14	identical	identical	ADJ
cril-1051	44	15	to	to	ADP
cril-1051	44	16	words	word	NOUN
cril-1051	44	17	that	that	PRON
cril-1051	44	18	were	be	AUX
cril-1051	44	19	previously	previously	ADV
cril-1051	44	20	annotated	annotate	VERB
cril-1051	44	21	by	by	ADP
cril-1051	44	22	hand	hand	NOUN
cril-1051	44	23	.	.	PUNCT
cril-1051	45	1	neither	neither	DET
cril-1051	45	2	tool	tool	NOUN
cril-1051	45	3	incorporates	incorporate	VERB
cril-1051	45	4	machine	machine	NOUN
cril-1051	45	5	learning	learning	NOUN
cril-1051	45	6	.	.	PUNCT
cril-1051	46	1	figure	figure	NOUN
cril-1051	46	2	2	2	NUM
cril-1051	46	3	.	.	NOUN
cril-1051	46	4	user	user	NOUN
cril-1051	46	5	interface	interface	NOUN
cril-1051	46	6	for	for	ADP
cril-1051	46	7	interlinearization	interlinearization	NOUN
cril-1051	46	8	in	in	ADP
cril-1051	46	9	fieldworks	fieldwork	NOUN
cril-1051	46	10	language	language	NOUN
cril-1051	46	11	explorer	explorer	NOUN
cril-1051	46	12	(	(	PUNCT
cril-1051	46	13	flex	flex	ADJ
cril-1051	46	14	)	)	PUNCT
cril-1051	46	15	displaying	display	VERB
cril-1051	46	16	a	a	DET
cril-1051	46	17	manipuri	manipuri	PROPN
cril-1051	46	18	[	[	X
cril-1051	46	19	mni	mni	X
cril-1051	46	20	]	]	X
cril-1051	46	21	text	text	NOUN
cril-1051	46	22	computational	computational	ADJ
cril-1051	46	23	morphology	morphology	NOUN
cril-1051	46	24	for	for	ADP
cril-1051	46	25	language	language	NOUN
cril-1051	46	26	description	description	NOUN
cril-1051	46	27	and	and	CCONJ
cril-1051	46	28	documentation	documentation	NOUN
cril-1051	46	29	5	5	NUM
cril-1051	46	30	figure	figure	NOUN
cril-1051	46	31	3	3	NUM
cril-1051	46	32	.	.	PUNCT
cril-1051	46	33	user	user	NOUN
cril-1051	46	34	interface	interface	PROPN
cril-1051	46	35	for	for	ADP
cril-1051	46	36	interlinearization	interlinearization	NOUN
cril-1051	46	37	in	in	ADP
cril-1051	46	38	elan	elan	PROPN
cril-1051	46	39	showing	show	VERB
cril-1051	46	40	a	a	DET
cril-1051	46	41	practice	practice	NOUN
cril-1051	46	42	session	session	NOUN
cril-1051	46	43	in	in	ADP
cril-1051	46	44	english	english	PROPN
cril-1051	46	45	3	3	NUM
cril-1051	46	46	.	.	PUNCT
cril-1051	46	47	natural	natural	ADJ
cril-1051	46	48	language	language	NOUN
cril-1051	46	49	processing	processing	NOUN
cril-1051	46	50	(	(	PUNCT
cril-1051	46	51	nlp	nlp	NOUN
cril-1051	46	52	)	)	PUNCT
cril-1051	46	53	for	for	ADP
cril-1051	46	54	low	low	ADJ
cril-1051	46	55	-	-	PUNCT
cril-1051	46	56	resource	resource	NOUN
cril-1051	46	57	languages	language	NOUN
cril-1051	46	58	though	though	SCONJ
cril-1051	46	59	the	the	DET
cril-1051	46	60	line	line	NOUN
cril-1051	46	61	between	between	ADP
cril-1051	46	62	documentation	documentation	NOUN
cril-1051	46	63	and	and	CCONJ
cril-1051	46	64	description	description	NOUN
cril-1051	46	65	may	may	AUX
cril-1051	46	66	not	not	PART
cril-1051	46	67	be	be	AUX
cril-1051	46	68	clear	clear	ADJ
cril-1051	46	69	,	,	PUNCT
cril-1051	46	70	one	one	NUM
cril-1051	46	71	thing	thing	NOUN
cril-1051	46	72	is	be	AUX
cril-1051	46	73	clear	clear	ADJ
cril-1051	46	74	:	:	PUNCT
cril-1051	46	75	current	current	ADJ
cril-1051	46	76	methods	method	NOUN
cril-1051	46	77	can	can	AUX
cril-1051	46	78	not	not	PART
cril-1051	46	79	easily	easily	ADV
cril-1051	46	80	process	process	VERB
cril-1051	46	81	large	large	ADJ
cril-1051	46	82	amounts	amount	NOUN
cril-1051	46	83	of	of	ADP
cril-1051	46	84	data	datum	NOUN
cril-1051	46	85	.	.	PUNCT
cril-1051	47	1	most	most	ADV
cril-1051	47	2	archived	archive	VERB
cril-1051	47	3	corpora	corpora	NOUN
cril-1051	47	4	are	be	AUX
cril-1051	47	5	only	only	ADV
cril-1051	47	6	partly	partly	ADV
cril-1051	47	7	annotated	annotate	VERB
cril-1051	47	8	because	because	SCONJ
cril-1051	47	9	funding	funding	NOUN
cril-1051	47	10	and	and	CCONJ
cril-1051	47	11	time	time	NOUN
cril-1051	47	12	constraints	constraint	NOUN
cril-1051	47	13	do	do	AUX
cril-1051	47	14	not	not	PART
cril-1051	47	15	allow	allow	VERB
cril-1051	47	16	complete	complete	ADJ
cril-1051	47	17	interlinearization	interlinearization	NOUN
cril-1051	47	18	(	(	PUNCT
cril-1051	47	19	cox	cox	PROPN
cril-1051	47	20	,	,	PUNCT
cril-1051	47	21	bouliame	bouliame	PROPN
cril-1051	47	22	&	&	CCONJ
cril-1051	47	23	alam	alam	PROPN
cril-1051	47	24	2019	2019	NUM
cril-1051	47	25	)	)	PUNCT
cril-1051	47	26	.	.	PUNCT
cril-1051	48	1	methods	method	NOUN
cril-1051	48	2	currently	currently	ADV
cril-1051	48	3	that	that	PRON
cril-1051	48	4	are	be	AUX
cril-1051	48	5	today	today	NOUN
cril-1051	48	6	used	use	VERB
cril-1051	48	7	commonly	commonly	ADV
cril-1051	48	8	in	in	ADP
cril-1051	48	9	ldd	ldd	PROPN
cril-1051	48	10	rely	rely	VERB
cril-1051	48	11	primarily	primarily	ADV
cril-1051	48	12	on	on	ADP
cril-1051	48	13	hand	hand	NOUN
cril-1051	48	14	annotation	annotation	NOUN
cril-1051	48	15	which	which	PRON
cril-1051	48	16	is	be	AUX
cril-1051	48	17	extremely	extremely	ADV
cril-1051	48	18	inefficient	inefficient	ADJ
cril-1051	48	19	.	.	PUNCT
cril-1051	49	1	the	the	DET
cril-1051	49	2	typical	typical	ADJ
cril-1051	49	3	strategy	strategy	NOUN
cril-1051	49	4	of	of	ADP
cril-1051	49	5	annotating	annotate	VERB
cril-1051	49	6	texts	text	NOUN
cril-1051	49	7	from	from	ADP
cril-1051	49	8	top	top	NOUN
cril-1051	49	9	to	to	ADP
cril-1051	49	10	bottom	bottom	NOUN
cril-1051	49	11	is	be	AUX
cril-1051	49	12	non	non	ADJ
cril-1051	49	13	-	-	ADJ
cril-1051	49	14	optimal	optimal	ADJ
cril-1051	49	15	for	for	ADP
cril-1051	49	16	training	train	VERB
cril-1051	49	17	a	a	DET
cril-1051	49	18	supervised	supervised	ADJ
cril-1051	49	19	machine	machine	NOUN
cril-1051	49	20	learning	learning	NOUN
cril-1051	49	21	model	model	NOUN
cril-1051	49	22	(	(	PUNCT
cril-1051	49	23	baldridge	baldridge	NOUN
cril-1051	49	24	&	&	CCONJ
cril-1051	49	25	osborne	osborne	PROPN
cril-1051	49	26	2008	2008	NUM
cril-1051	49	27	;	;	PUNCT
cril-1051	49	28	baldridge	baldridge	NOUN
cril-1051	49	29	&	&	CCONJ
cril-1051	49	30	palmer	palmer	PROPN
cril-1051	49	31	2009	2009	NUM
cril-1051	49	32	;	;	PUNCT
cril-1051	49	33	palmer	palmer	PROPN
cril-1051	49	34	2009	2009	NUM
cril-1051	49	35	)	)	PUNCT
cril-1051	49	36	.	.	PUNCT
cril-1051	50	1	since	since	SCONJ
cril-1051	50	2	naturally	naturally	ADV
cril-1051	50	3	occurring	occur	VERB
cril-1051	50	4	speech	speech	NOUN
cril-1051	50	5	contains	contain	VERB
cril-1051	50	6	many	many	ADJ
cril-1051	50	7	repeated	repeat	VERB
cril-1051	50	8	linguistic	linguistic	ADJ
cril-1051	50	9	structures	structure	NOUN
cril-1051	50	10	,	,	PUNCT
cril-1051	50	11	manual	manual	ADJ
cril-1051	50	12	annotation	annotation	NOUN
cril-1051	50	13	has	have	AUX
cril-1051	50	14	been	be	AUX
cril-1051	50	15	described	describe	VERB
cril-1051	50	16	as	as	ADP
cril-1051	50	17	repetitive	repetitive	ADJ
cril-1051	50	18	,	,	PUNCT
cril-1051	50	19	monotonous	monotonous	ADJ
cril-1051	50	20	,	,	PUNCT
cril-1051	50	21	costly	costly	ADJ
cril-1051	50	22	,	,	PUNCT
cril-1051	50	23	and	and	CCONJ
cril-1051	50	24	time	time	NOUN
cril-1051	50	25	-	-	PUNCT
cril-1051	50	26	consuming	consume	VERB
cril-1051	50	27	(	(	PUNCT
cril-1051	50	28	duong	duong	PROPN
cril-1051	50	29	2017	2017	NUM
cril-1051	50	30	;	;	PUNCT
cril-1051	50	31	he	he	PRON
cril-1051	50	32	et	et	PROPN
cril-1051	50	33	al	al	PROPN
cril-1051	50	34	.	.	PROPN
cril-1051	50	35	2016	2016	NUM
cril-1051	50	36	)	)	PUNCT
cril-1051	50	37	.	.	PUNCT
cril-1051	51	1	it	it	PRON
cril-1051	51	2	can	can	AUX
cril-1051	51	3	take	take	VERB
cril-1051	51	4	anywhere	anywhere	ADV
cril-1051	51	5	from	from	ADP
cril-1051	51	6	20	20	NUM
cril-1051	51	7	to	to	PART
cril-1051	51	8	100	100	NUM
cril-1051	51	9	hours	hour	NOUN
cril-1051	51	10	to	to	PART
cril-1051	51	11	transcribe	transcribe	VERB
cril-1051	51	12	(	(	PUNCT
cril-1051	51	13	task	task	NOUN
cril-1051	51	14	2a	2a	NUM
cril-1051	51	15	)	)	PUNCT
cril-1051	51	16	a	a	DET
cril-1051	51	17	single	single	ADJ
cril-1051	51	18	hour	hour	NOUN
cril-1051	51	19	of	of	ADP
cril-1051	51	20	speech	speech	NOUN
cril-1051	51	21	(	(	PUNCT
cril-1051	51	22	seifart	seifart	NOUN
cril-1051	51	23	et	et	PROPN
cril-1051	51	24	al	al	PROPN
cril-1051	51	25	.	.	PROPN
cril-1051	51	26	2018	2018	NUM
cril-1051	51	27	)	)	PUNCT
cril-1051	51	28	and	and	CCONJ
cril-1051	51	29	it	it	PRON
cril-1051	51	30	is	be	AUX
cril-1051	51	31	reasonable	reasonable	ADJ
cril-1051	51	32	to	to	PART
cril-1051	51	33	assume	assume	VERB
cril-1051	51	34	that	that	SCONJ
cril-1051	51	35	interlinearization	interlinearization	NOUN
cril-1051	51	36	(	(	PUNCT
cril-1051	51	37	tasks	task	NOUN
cril-1051	51	38	2b	2b	NUM
cril-1051	51	39	and	and	CCONJ
cril-1051	51	40	3	3	NUM
cril-1051	51	41	)	)	PUNCT
cril-1051	51	42	and	and	CCONJ
cril-1051	51	43	eliciting	elicit	VERB
cril-1051	51	44	morphological	morphological	ADJ
cril-1051	51	45	paradigms	paradigm	NOUN
cril-1051	51	46	(	(	PUNCT
cril-1051	51	47	task	task	NOUN
cril-1051	51	48	4	4	NUM
cril-1051	51	49	)	)	PUNCT
cril-1051	51	50	require	require	VERB
cril-1051	51	51	significantly	significantly	ADV
cril-1051	51	52	more	more	ADJ
cril-1051	51	53	time	time	NOUN
cril-1051	51	54	.	.	PUNCT
cril-1051	52	1	a	a	DET
cril-1051	52	2	recent	recent	ADJ
cril-1051	52	3	growth	growth	NOUN
cril-1051	52	4	of	of	ADP
cril-1051	52	5	nlp	nlp	ADJ
cril-1051	52	6	interest	interest	NOUN
cril-1051	52	7	in	in	ADP
cril-1051	52	8	low	low	ADJ
cril-1051	52	9	-	-	PUNCT
cril-1051	52	10	resource	resource	NOUN
cril-1051	52	11	languages	language	NOUN
cril-1051	52	12	(	(	PUNCT
cril-1051	52	13	lrl	lrl	PROPN
cril-1051	52	14	)	)	PUNCT
cril-1051	52	15	has	have	AUX
cril-1051	52	16	brought	bring	VERB
cril-1051	52	17	machine	machine	NOUN
cril-1051	52	18	learning	learning	NOUN
cril-1051	52	19	models	model	NOUN
cril-1051	52	20	and	and	CCONJ
cril-1051	52	21	methods	method	NOUN
cril-1051	52	22	that	that	PRON
cril-1051	52	23	achieve	achieve	VERB
cril-1051	52	24	good	good	ADJ
cril-1051	52	25	results	result	NOUN
cril-1051	52	26	even	even	ADV
cril-1051	52	27	with	with	ADP
cril-1051	52	28	ldd	ldd	PROPN
cril-1051	52	29	field	field	NOUN
cril-1051	52	30	data	datum	NOUN
cril-1051	52	31	.	.	PUNCT
cril-1051	53	1	a	a	DET
cril-1051	53	2	notable	notable	ADJ
cril-1051	53	3	example	example	NOUN
cril-1051	53	4	is	be	AUX
cril-1051	53	5	elpis	elpi	VERB
cril-1051	53	6	(	(	PUNCT
cril-1051	53	7	foley	foley	PROPN
cril-1051	53	8	et	et	PROPN
cril-1051	53	9	al	al	PROPN
cril-1051	53	10	.	.	PROPN
cril-1051	53	11	2018	2018	NUM
cril-1051	53	12	)	)	PUNCT
cril-1051	53	13	,	,	PUNCT
cril-1051	53	14	an	an	DET
cril-1051	53	15	online	online	ADJ
cril-1051	53	16	tool	tool	NOUN
cril-1051	53	17	that	that	PRON
cril-1051	53	18	includes	include	VERB
cril-1051	53	19	a	a	DET
cril-1051	53	20	user	user	NOUN
cril-1051	53	21	interface	interface	NOUN
cril-1051	53	22	accessible	accessible	ADJ
cril-1051	53	23	to	to	ADP
cril-1051	53	24	those	those	PRON
cril-1051	53	25	with	with	ADP
cril-1051	53	26	no	no	DET
cril-1051	53	27	programming	programming	NOUN
cril-1051	53	28	background	background	NOUN
cril-1051	53	29	.	.	PUNCT
cril-1051	54	1	machine	machine	NOUN
cril-1051	54	2	translation	translation	NOUN
cril-1051	54	3	(	(	PUNCT
cril-1051	54	4	mt	mt	PROPN
cril-1051	54	5	)	)	PUNCT
cril-1051	54	6	has	have	AUX
cril-1051	54	7	also	also	ADV
cril-1051	54	8	been	be	AUX
cril-1051	54	9	applied	apply	VERB
cril-1051	54	10	to	to	ADP
cril-1051	54	11	documentary	documentary	NOUN
cril-1051	54	12	data	datum	NOUN
cril-1051	54	13	,	,	PUNCT
cril-1051	54	14	using	use	VERB
cril-1051	54	15	the	the	DET
cril-1051	54	16	output	output	NOUN
cril-1051	54	17	of	of	ADP
cril-1051	54	18	an	an	DET
cril-1051	54	19	automatic	automatic	ADJ
cril-1051	54	20	speech	speech	NOUN
cril-1051	54	21	recognition	recognition	NOUN
cril-1051	54	22	system	system	NOUN
cril-1051	54	23	as	as	ADP
cril-1051	54	24	input	input	NOUN
cril-1051	54	25	to	to	ADP
cril-1051	54	26	the	the	DET
cril-1051	54	27	mt	mt	PROPN
cril-1051	54	28	system	system	NOUN
cril-1051	54	29	(	(	PUNCT
cril-1051	54	30	anastasopoulos	anastasopoulos	PROPN
cril-1051	54	31	,	,	PUNCT
cril-1051	54	32	chiang	chiang	PROPN
cril-1051	54	33	&	&	CCONJ
cril-1051	54	34	duong	duong	PROPN
cril-1051	54	35	2016	2016	NUM
cril-1051	54	36	;	;	PUNCT
cril-1051	54	37	duong	duong	PROPN
cril-1051	54	38	et	et	PROPN
cril-1051	54	39	al	al	PROPN
cril-1051	54	40	.	.	PROPN
cril-1051	54	41	2016	2016	NUM
cril-1051	54	42	)	)	PUNCT
cril-1051	54	43	.	.	PUNCT
cril-1051	55	1	colorado	colorado	PROPN
cril-1051	55	2	research	research	PROPN
cril-1051	55	3	in	in	ADP
cril-1051	55	4	linguistics	linguistic	NOUN
cril-1051	55	5	,	,	PUNCT
cril-1051	55	6	volume	volume	NOUN
cril-1051	55	7	25	25	NUM
cril-1051	55	8	(	(	PUNCT
cril-1051	55	9	2021	2021	NUM
cril-1051	55	10	)	)	PUNCT
cril-1051	55	11	6	6	NUM
cril-1051	55	12	the	the	DET
cril-1051	55	13	potential	potential	NOUN
cril-1051	55	14	for	for	ADP
cril-1051	55	15	machine	machine	NOUN
cril-1051	55	16	learning	learn	VERB
cril-1051	55	17	to	to	PART
cril-1051	55	18	perform	perform	VERB
cril-1051	55	19	morphological	morphological	ADJ
cril-1051	55	20	analysis	analysis	NOUN
cril-1051	55	21	during	during	ADP
cril-1051	55	22	interlinearization	interlinearization	NOUN
cril-1051	55	23	has	have	AUX
cril-1051	55	24	been	be	AUX
cril-1051	55	25	clearly	clearly	ADV
cril-1051	55	26	demonstrated	demonstrate	VERB
cril-1051	55	27	(	(	PUNCT
cril-1051	55	28	baldridge	baldridge	NOUN
cril-1051	55	29	&	&	CCONJ
cril-1051	55	30	palmer	palmer	PROPN
cril-1051	55	31	2009	2009	NUM
cril-1051	55	32	;	;	PUNCT
cril-1051	55	33	palmer	palmer	PROPN
cril-1051	55	34	2009	2009	NUM
cril-1051	55	35	;	;	PUNCT
cril-1051	55	36	palmer	palmer	PROPN
cril-1051	55	37	et	et	PROPN
cril-1051	55	38	al	al	PROPN
cril-1051	55	39	.	.	PROPN
cril-1051	55	40	2010	2010	NUM
cril-1051	55	41	;	;	PUNCT
cril-1051	55	42	xia	xia	PROPN
cril-1051	55	43	et	et	PROPN
cril-1051	55	44	al	al	PROPN
cril-1051	55	45	.	.	PROPN
cril-1051	55	46	2016	2016	NUM
cril-1051	55	47	)	)	PUNCT
cril-1051	55	48	.	.	PUNCT
cril-1051	56	1	for	for	ADP
cril-1051	56	2	example	example	NOUN
cril-1051	56	3	,	,	PUNCT
cril-1051	56	4	felt	feel	VERB
cril-1051	56	5	(	(	PUNCT
cril-1051	56	6	2012	2012	NUM
cril-1051	56	7	)	)	PUNCT
cril-1051	56	8	found	find	VERB
cril-1051	56	9	that	that	SCONJ
cril-1051	56	10	when	when	SCONJ
cril-1051	56	11	a	a	DET
cril-1051	56	12	round	round	NOUN
cril-1051	56	13	of	of	ADP
cril-1051	56	14	annotation	annotation	NOUN
cril-1051	56	15	is	be	AUX
cril-1051	56	16	done	do	VERB
cril-1051	56	17	automatically	automatically	ADV
cril-1051	56	18	by	by	ADP
cril-1051	56	19	a	a	DET
cril-1051	56	20	machine	machine	NOUN
cril-1051	56	21	learning	learning	NOUN
cril-1051	56	22	model	model	NOUN
cril-1051	56	23	and	and	CCONJ
cril-1051	56	24	then	then	ADV
cril-1051	56	25	corrected	correct	VERB
cril-1051	56	26	by	by	ADP
cril-1051	56	27	the	the	DET
cril-1051	56	28	human	human	ADJ
cril-1051	56	29	annotators	annotator	NOUN
cril-1051	56	30	,	,	PUNCT
cril-1051	56	31	the	the	DET
cril-1051	56	32	annotators	annotator	NOUN
cril-1051	56	33	’	'	PUNCT
cril-1051	56	34	accuracy	accuracy	NOUN
cril-1051	56	35	is	be	AUX
cril-1051	56	36	improve	improve	VERB
cril-1051	56	37	if	if	SCONJ
cril-1051	56	38	the	the	DET
cril-1051	56	39	machine	machine	NOUN
cril-1051	56	40	learning	learn	VERB
cril-1051	56	41	model	model	NOUN
cril-1051	56	42	achieves	achieve	VERB
cril-1051	56	43	at	at	ADP
cril-1051	56	44	least	least	ADV
cril-1051	56	45	60	60	NUM
cril-1051	56	46	%	%	NOUN
cril-1051	56	47	accuracy	accuracy	NOUN
cril-1051	56	48	and	and	CCONJ
cril-1051	56	49	significantly	significantly	ADV
cril-1051	56	50	speeds	speed	VERB
cril-1051	56	51	manual	manual	ADJ
cril-1051	56	52	annotation	annotation	NOUN
cril-1051	56	53	if	if	SCONJ
cril-1051	56	54	it	it	PRON
cril-1051	56	55	achieves	achieve	VERB
cril-1051	56	56	an	an	DET
cril-1051	56	57	accuracy	accuracy	NOUN
cril-1051	56	58	of	of	ADP
cril-1051	56	59	80	80	NUM
cril-1051	56	60	%	%	NOUN
cril-1051	56	61	.	.	PUNCT
cril-1051	57	1	in	in	ADP
cril-1051	57	2	the	the	DET
cril-1051	57	3	area	area	NOUN
cril-1051	57	4	of	of	ADP
cril-1051	57	5	morphological	morphological	ADJ
cril-1051	57	6	paradigm	paradigm	NOUN
cril-1051	57	7	learning	learning	NOUN
cril-1051	57	8	,	,	PUNCT
cril-1051	57	9	the	the	DET
cril-1051	57	10	annual	annual	ADJ
cril-1051	57	11	sigmorphon	sigmorphon	NOUN
cril-1051	57	12	and	and	CCONJ
cril-1051	57	13	conll	conll	NOUN
cril-1051	57	14	-	-	PUNCT
cril-1051	57	15	sigmorphon	sigmorphon	NOUN
cril-1051	57	16	shared	share	VERB
cril-1051	57	17	tasks	task	NOUN
cril-1051	57	18	(	(	PUNCT
cril-1051	57	19	cotterell	cotterell	NOUN
cril-1051	57	20	et	et	PROPN
cril-1051	57	21	al	al	PROPN
cril-1051	57	22	.	.	PROPN
cril-1051	57	23	2016	2016	NUM
cril-1051	57	24	;	;	PUNCT
cril-1051	57	25	cotterell	cotterell	VERB
cril-1051	57	26	et	et	PROPN
cril-1051	57	27	al	al	PROPN
cril-1051	57	28	.	.	PROPN
cril-1051	57	29	2017	2017	NUM
cril-1051	57	30	;	;	PUNCT
cril-1051	57	31	cotterell	cotterell	VERB
cril-1051	57	32	et	et	PROPN
cril-1051	57	33	al	al	PROPN
cril-1051	57	34	.	.	PROPN
cril-1051	57	35	2018	2018	NUM
cril-1051	57	36	;	;	PUNCT
cril-1051	57	37	mccarthy	mccarthy	PROPN
cril-1051	57	38	et	et	PROPN
cril-1051	57	39	al	al	PROPN
cril-1051	57	40	.	.	PROPN
cril-1051	57	41	2019	2019	NUM
cril-1051	57	42	;	;	PUNCT
cril-1051	58	1	nicolai	nicolai	PROPN
cril-1051	58	2	,	,	PUNCT
cril-1051	58	3	gorman	gorman	PROPN
cril-1051	58	4	&	&	CCONJ
cril-1051	58	5	cotterell	cotterell	PROPN
cril-1051	58	6	2020	2020	NUM
cril-1051	58	7	)	)	PUNCT
cril-1051	58	8	have	have	AUX
cril-1051	58	9	developed	develop	VERB
cril-1051	58	10	successful	successful	ADJ
cril-1051	58	11	methods	method	NOUN
cril-1051	58	12	with	with	ADP
cril-1051	58	13	limited	limited	ADJ
cril-1051	58	14	training	training	NOUN
cril-1051	58	15	data	datum	NOUN
cril-1051	58	16	.	.	PUNCT
cril-1051	59	1	although	although	SCONJ
cril-1051	59	2	nlp	nlp	ADJ
cril-1051	59	3	interest	interest	NOUN
cril-1051	59	4	in	in	ADP
cril-1051	59	5	lrl	lrl	PROPN
cril-1051	59	6	has	have	AUX
cril-1051	59	7	grown	grow	VERB
cril-1051	59	8	noticeably	noticeably	ADV
cril-1051	59	9	in	in	ADP
cril-1051	59	10	the	the	DET
cril-1051	59	11	past	past	ADJ
cril-1051	59	12	few	few	ADJ
cril-1051	59	13	years	year	NOUN
cril-1051	59	14	,	,	PUNCT
cril-1051	59	15	it	it	PRON
cril-1051	59	16	is	be	AUX
cril-1051	59	17	not	not	PART
cril-1051	59	18	a	a	DET
cril-1051	59	19	new	new	ADJ
cril-1051	59	20	area	area	NOUN
cril-1051	59	21	of	of	ADP
cril-1051	59	22	research	research	NOUN
cril-1051	59	23	.	.	PUNCT
cril-1051	60	1	since	since	SCONJ
cril-1051	60	2	the	the	DET
cril-1051	60	3	late	late	ADJ
cril-1051	60	4	20th	20th	ADJ
cril-1051	60	5	century	century	NOUN
cril-1051	60	6	,	,	PUNCT
cril-1051	60	7	nlp	nlp	NOUN
cril-1051	60	8	has	have	AUX
cril-1051	60	9	taken	take	VERB
cril-1051	60	10	several	several	ADJ
cril-1051	60	11	approaches	approach	NOUN
cril-1051	60	12	to	to	ADP
cril-1051	60	13	low	low	ADJ
cril-1051	60	14	-	-	PUNCT
cril-1051	60	15	resource	resource	NOUN
cril-1051	60	16	languages	language	NOUN
cril-1051	60	17	that	that	PRON
cril-1051	60	18	can	can	AUX
cril-1051	60	19	be	be	AUX
cril-1051	60	20	classified	classify	VERB
cril-1051	60	21	as	as	ADP
cril-1051	60	22	either	either	CCONJ
cril-1051	60	23	rule	rule	NOUN
cril-1051	60	24	-	-	PUNCT
cril-1051	60	25	based	base	VERB
cril-1051	60	26	(	(	PUNCT
cril-1051	60	27	i.e.	i.e.	X
cril-1051	60	28	,	,	PUNCT
cril-1051	60	29	finite	finite	VERB
cril-1051	60	30	state	state	NOUN
cril-1051	60	31	transducers	transducer	NOUN
cril-1051	60	32	)	)	PUNCT
cril-1051	60	33	(	(	PUNCT
cril-1051	60	34	e.g.	e.g.	ADV
cril-1051	60	35	,	,	PUNCT
cril-1051	60	36	cotterell	cotterell	VERB
cril-1051	60	37	et	et	PROPN
cril-1051	60	38	al	al	PROPN
cril-1051	60	39	.	.	PROPN
cril-1051	60	40	2015	2015	NUM
cril-1051	60	41	;	;	PUNCT
cril-1051	60	42	forsberg	forsberg	PROPN
cril-1051	60	43	&	&	CCONJ
cril-1051	60	44	hulden	hulden	VERB
cril-1051	60	45	2016	2016	NUM
cril-1051	60	46	;	;	PUNCT
cril-1051	60	47	moeller	moeller	PROPN
cril-1051	60	48	et	et	PROPN
cril-1051	60	49	al	al	PROPN
cril-1051	60	50	.	.	PROPN
cril-1051	60	51	2018	2018	NUM
cril-1051	60	52	;	;	PUNCT
cril-1051	60	53	moeller	moeller	PROPN
cril-1051	60	54	et	et	PROPN
cril-1051	60	55	al	al	PROPN
cril-1051	60	56	.	.	PROPN
cril-1051	60	57	2019	2019	NUM
cril-1051	60	58	)	)	PUNCT
cril-1051	60	59	or	or	CCONJ
cril-1051	60	60	machine	machine	NOUN
cril-1051	60	61	learning	learning	NOUN
cril-1051	60	62	models	model	NOUN
cril-1051	60	63	that	that	SCONJ
cril-1051	60	64	“	"	PUNCT
cril-1051	60	65	learn	learn	VERB
cril-1051	60	66	”	"	PUNCT
cril-1051	60	67	rules	rule	NOUN
cril-1051	60	68	from	from	ADP
cril-1051	60	69	data	data	PROPN
cril-1051	60	70	.	.	PUNCT
cril-1051	61	1	machine	machine	NOUN
cril-1051	61	2	learning	learn	VERB
cril-1051	61	3	approaches	approach	NOUN
cril-1051	61	4	to	to	ADP
cril-1051	61	5	lrl	lrl	PROPN
cril-1051	61	6	can	can	AUX
cril-1051	61	7	be	be	AUX
cril-1051	61	8	classified	classify	VERB
cril-1051	61	9	according	accord	VERB
cril-1051	61	10	to	to	ADP
cril-1051	61	11	whether	whether	SCONJ
cril-1051	61	12	the	the	DET
cril-1051	61	13	training	training	NOUN
cril-1051	61	14	data	datum	NOUN
cril-1051	61	15	was	be	AUX
cril-1051	61	16	annotated	annotate	VERB
cril-1051	61	17	completely	completely	ADV
cril-1051	61	18	(	(	PUNCT
cril-1051	61	19	supervised	supervised	ADJ
cril-1051	61	20	)	)	PUNCT
cril-1051	61	21	(	(	PUNCT
cril-1051	61	22	e.g.	e.g.	ADV
cril-1051	61	23	,	,	PUNCT
cril-1051	61	24	bergmanis	bergmanis	NOUN
cril-1051	61	25	et	et	PROPN
cril-1051	61	26	al	al	PROPN
cril-1051	61	27	.	.	PROPN
cril-1051	61	28	2017	2017	NUM
cril-1051	61	29	;	;	PUNCT
cril-1051	61	30	sudhakar	sudhakar	PROPN
cril-1051	61	31	&	&	CCONJ
cril-1051	61	32	singh	singh	PROPN
cril-1051	61	33	2017	2017	NUM
cril-1051	61	34	;	;	PUNCT
cril-1051	61	35	makarov	makarov	NOUN
cril-1051	61	36	,	,	PUNCT
cril-1051	61	37	ruzsics	ruzsic	NOUN
cril-1051	61	38	&	&	CCONJ
cril-1051	61	39	clematide	clematide	ADJ
cril-1051	61	40	2017	2017	NUM
cril-1051	61	41	;	;	PUNCT
cril-1051	61	42	liu	liu	PROPN
cril-1051	61	43	et	et	PROPN
cril-1051	61	44	al	al	PROPN
cril-1051	61	45	.	.	PROPN
cril-1051	61	46	2018	2018	NUM
cril-1051	61	47	;	;	PUNCT
cril-1051	61	48	makarov	makarov	PROPN
cril-1051	61	49	&	&	CCONJ
cril-1051	61	50	clematide	clematide	ADJ
cril-1051	61	51	2018a	2018a	NUM
cril-1051	61	52	)	)	PUNCT
cril-1051	61	53	,	,	PUNCT
cril-1051	61	54	partially	partially	ADV
cril-1051	61	55	(	(	PUNCT
cril-1051	61	56	semi	semi	ADJ
cril-1051	61	57	-	-	ADJ
cril-1051	61	58	supervised	supervised	ADJ
cril-1051	61	59	)	)	PUNCT
cril-1051	61	60	(	(	PUNCT
cril-1051	61	61	e.g.	e.g.	ADV
cril-1051	61	62	,	,	PUNCT
cril-1051	61	63	ahlberg	ahlberg	PROPN
cril-1051	61	64	,	,	PUNCT
cril-1051	61	65	forsberg	forsberg	PROPN
cril-1051	61	66	&	&	CCONJ
cril-1051	61	67	hulden	hulden	VERB
cril-1051	61	68	2014	2014	NUM
cril-1051	61	69	)	)	PUNCT
cril-1051	61	70	,	,	PUNCT
cril-1051	61	71	or	or	CCONJ
cril-1051	61	72	not	not	PART
cril-1051	61	73	at	at	ADV
cril-1051	61	74	all	all	ADV
cril-1051	61	75	(	(	PUNCT
cril-1051	61	76	unsupervised	unsupervised	ADJ
cril-1051	61	77	)	)	PUNCT
cril-1051	61	78	(	(	PUNCT
cril-1051	61	79	e.g.	e.g.	ADV
cril-1051	61	80	,	,	PUNCT
cril-1051	61	81	moon	moon	PROPN
cril-1051	61	82	,	,	PUNCT
cril-1051	61	83	erk	erk	PROPN
cril-1051	61	84	&	&	CCONJ
cril-1051	61	85	baldridge	baldridge	PROPN
cril-1051	61	86	2009	2009	NUM
cril-1051	61	87	;	;	PUNCT
cril-1051	61	88	palmer	palmer	PROPN
cril-1051	61	89	et	et	PROPN
cril-1051	61	90	al	al	PROPN
cril-1051	61	91	.	.	PROPN
cril-1051	61	92	2010	2010	NUM
cril-1051	61	93	;	;	PUNCT
cril-1051	61	94	kirschenbaum	kirschenbaum	PROPN
cril-1051	61	95	,	,	PUNCT
cril-1051	61	96	wittenburg	wittenburg	PROPN
cril-1051	61	97	&	&	CCONJ
cril-1051	61	98	heyer	heyer	PROPN
cril-1051	61	99	2012	2012	NUM
cril-1051	61	100	;	;	PUNCT
cril-1051	61	101	soricut	soricut	VERB
cril-1051	61	102	&	&	CCONJ
cril-1051	61	103	och	och	NOUN
cril-1051	61	104	2015	2015	NUM
cril-1051	61	105	)	)	PUNCT
cril-1051	61	106	.	.	PUNCT
cril-1051	62	1	at	at	ADP
cril-1051	62	2	first	first	ADJ
cril-1051	62	3	glance	glance	NOUN
cril-1051	62	4	,	,	PUNCT
cril-1051	62	5	unsupervised	unsupervised	ADJ
cril-1051	62	6	and	and	CCONJ
cril-1051	62	7	semi	semi	ADJ
cril-1051	62	8	-	-	ADJ
cril-1051	62	9	supervised	supervised	ADJ
cril-1051	62	10	learning	learning	NOUN
cril-1051	62	11	seem	seem	VERB
cril-1051	62	12	most	most	ADV
cril-1051	62	13	promising	promising	ADJ
cril-1051	62	14	for	for	ADP
cril-1051	62	15	ldd	ldd	PROPN
cril-1051	62	16	because	because	SCONJ
cril-1051	62	17	they	they	PRON
cril-1051	62	18	do	do	AUX
cril-1051	62	19	not	not	PART
cril-1051	62	20	require	require	VERB
cril-1051	62	21	as	as	ADV
cril-1051	62	22	much	much	ADV
cril-1051	62	23	manually	manually	ADV
cril-1051	62	24	annotated	annotate	VERB
cril-1051	62	25	data	datum	NOUN
cril-1051	62	26	.	.	PUNCT
cril-1051	63	1	supervised	supervised	ADJ
cril-1051	63	2	learning	learning	NOUN
cril-1051	63	3	is	be	AUX
cril-1051	63	4	trained	train	VERB
cril-1051	63	5	on	on	ADP
cril-1051	63	6	“	"	PUNCT
cril-1051	63	7	gold	gold	ADJ
cril-1051	63	8	standard	standard	NOUN
cril-1051	63	9	”	"	PUNCT
cril-1051	63	10	annotated	annotate	VERB
cril-1051	63	11	data	datum	NOUN
cril-1051	63	12	.	.	PUNCT
cril-1051	64	1	however	however	ADV
cril-1051	64	2	,	,	PUNCT
cril-1051	64	3	even	even	ADV
cril-1051	64	4	though	though	SCONJ
cril-1051	64	5	supervised	supervised	ADJ
cril-1051	64	6	learning	learning	NOUN
cril-1051	64	7	requires	require	VERB
cril-1051	64	8	annotation	annotation	NOUN
cril-1051	64	9	,	,	PUNCT
cril-1051	64	10	it	it	PRON
cril-1051	64	11	needs	need	VERB
cril-1051	64	12	much	much	ADV
cril-1051	64	13	less	less	ADJ
cril-1051	64	14	data	datum	NOUN
cril-1051	64	15	than	than	ADP
cril-1051	64	16	unsupervised	unsupervised	ADJ
cril-1051	64	17	learning	learning	NOUN
cril-1051	64	18	and	and	CCONJ
cril-1051	64	19	almost	almost	ADV
cril-1051	64	20	always	always	ADV
cril-1051	64	21	yields	yield	VERB
cril-1051	64	22	better	well	ADJ
cril-1051	64	23	results	result	NOUN
cril-1051	64	24	(	(	PUNCT
cril-1051	64	25	ruokolainen	ruokolainen	NOUN
cril-1051	64	26	et	et	PROPN
cril-1051	64	27	al	al	PROPN
cril-1051	64	28	.	.	PROPN
cril-1051	64	29	2013	2013	NUM
cril-1051	64	30	;	;	PUNCT
cril-1051	64	31	cotterell	cotterell	VERB
cril-1051	64	32	et	et	PROPN
cril-1051	64	33	al	al	PROPN
cril-1051	64	34	.	.	PROPN
cril-1051	64	35	2015	2015	NUM
cril-1051	64	36	)	)	PUNCT
cril-1051	64	37	.	.	PUNCT
cril-1051	65	1	additionally	additionally	ADV
cril-1051	65	2	,	,	PUNCT
cril-1051	65	3	without	without	ADP
cril-1051	65	4	annotated	annotate	VERB
cril-1051	65	5	labels	label	NOUN
cril-1051	65	6	,	,	PUNCT
cril-1051	65	7	unsupervised	unsupervised	ADJ
cril-1051	65	8	learning	learning	NOUN
cril-1051	65	9	can	can	AUX
cril-1051	65	10	only	only	ADV
cril-1051	65	11	really	really	ADV
cril-1051	65	12	cluster	cluster	NOUN
cril-1051	65	13	data	datum	NOUN
cril-1051	65	14	by	by	ADP
cril-1051	65	15	the	the	DET
cril-1051	65	16	latent	latent	NOUN
cril-1051	65	17	patterns	pattern	NOUN
cril-1051	65	18	in	in	ADP
cril-1051	65	19	the	the	DET
cril-1051	65	20	data	datum	NOUN
cril-1051	65	21	.	.	PUNCT
cril-1051	66	1	discovering	discover	VERB
cril-1051	66	2	latent	latent	NOUN
cril-1051	66	3	patterns	pattern	NOUN
cril-1051	66	4	might	might	AUX
cril-1051	66	5	be	be	AUX
cril-1051	66	6	quite	quite	ADV
cril-1051	66	7	useful	useful	ADJ
cril-1051	66	8	for	for	ADP
cril-1051	66	9	linguists	linguist	NOUN
cril-1051	66	10	when	when	SCONJ
cril-1051	66	11	first	first	ADV
cril-1051	66	12	exploring	explore	VERB
cril-1051	66	13	the	the	DET
cril-1051	66	14	data	datum	NOUN
cril-1051	66	15	;	;	PUNCT
cril-1051	66	16	for	for	ADP
cril-1051	66	17	example	example	NOUN
cril-1051	66	18	,	,	PUNCT
cril-1051	66	19	frequent	frequent	ADJ
cril-1051	66	20	character	character	NOUN
cril-1051	66	21	patterns	pattern	NOUN
cril-1051	66	22	and	and	CCONJ
cril-1051	66	23	substrings	substring	NOUN
cril-1051	66	24	that	that	PRON
cril-1051	66	25	a	a	DET
cril-1051	66	26	model	model	NOUN
cril-1051	66	27	discovers	discover	VERB
cril-1051	66	28	could	could	AUX
cril-1051	66	29	provide	provide	VERB
cril-1051	66	30	an	an	DET
cril-1051	66	31	initial	initial	ADJ
cril-1051	66	32	hypothesis	hypothesis	NOUN
cril-1051	66	33	to	to	ADP
cril-1051	66	34	the	the	DET
cril-1051	66	35	linguist	linguist	NOUN
cril-1051	66	36	about	about	ADP
cril-1051	66	37	the	the	DET
cril-1051	66	38	language	language	NOUN
cril-1051	66	39	’s	’s	PART
cril-1051	66	40	morphological	morphological	ADJ
cril-1051	66	41	structure	structure	NOUN
cril-1051	66	42	.	.	PUNCT
cril-1051	67	1	however	however	ADV
cril-1051	67	2	,	,	PUNCT
cril-1051	67	3	no	no	ADV
cril-1051	67	4	matter	matter	ADV
cril-1051	67	5	how	how	SCONJ
cril-1051	67	6	accurate	accurate	ADJ
cril-1051	67	7	an	an	DET
cril-1051	67	8	unsupervised	unsupervised	ADJ
cril-1051	67	9	model	model	NOUN
cril-1051	67	10	may	may	AUX
cril-1051	67	11	be	be	AUX
cril-1051	67	12	,	,	PUNCT
cril-1051	67	13	it	it	PRON
cril-1051	67	14	can	can	AUX
cril-1051	67	15	not	not	PART
cril-1051	67	16	substitute	substitute	VERB
cril-1051	67	17	the	the	DET
cril-1051	67	18	valuable	valuable	ADJ
cril-1051	67	19	process	process	NOUN
cril-1051	67	20	of	of	ADP
cril-1051	67	21	manually	manually	ADV
cril-1051	67	22	analyzing	analyze	VERB
cril-1051	67	23	and	and	CCONJ
cril-1051	67	24	discovering	discover	VERB
cril-1051	67	25	patterns	pattern	NOUN
cril-1051	67	26	in	in	ADP
cril-1051	67	27	the	the	DET
cril-1051	67	28	data	datum	NOUN
cril-1051	67	29	.	.	PUNCT
cril-1051	68	1	detailed	detailed	ADJ
cril-1051	68	2	analysis	analysis	NOUN
cril-1051	68	3	of	of	ADP
cril-1051	68	4	new	new	ADJ
cril-1051	68	5	data	datum	NOUN
cril-1051	68	6	is	be	AUX
cril-1051	68	7	vital	vital	ADJ
cril-1051	68	8	for	for	ADP
cril-1051	68	9	linguists	linguist	NOUN
cril-1051	68	10	because	because	SCONJ
cril-1051	68	11	through	through	ADP
cril-1051	68	12	that	that	DET
cril-1051	68	13	process	process	NOUN
cril-1051	68	14	the	the	DET
cril-1051	68	15	linguist	linguist	NOUN
cril-1051	68	16	becomes	become	VERB
cril-1051	68	17	familiar	familiar	ADJ
cril-1051	68	18	with	with	ADP
cril-1051	68	19	the	the	DET
cril-1051	68	20	data	datum	NOUN
cril-1051	68	21	and	and	CCONJ
cril-1051	68	22	begins	begin	VERB
cril-1051	68	23	to	to	PART
cril-1051	68	24	absorb	absorb	VERB
cril-1051	68	25	an	an	DET
cril-1051	68	26	computational	computational	ADJ
cril-1051	68	27	morphology	morphology	NOUN
cril-1051	68	28	for	for	ADP
cril-1051	68	29	language	language	NOUN
cril-1051	68	30	description	description	NOUN
cril-1051	68	31	and	and	CCONJ
cril-1051	68	32	documentation	documentation	NOUN
cril-1051	68	33	7	7	NUM
cril-1051	68	34	intuitive	intuitive	ADJ
cril-1051	68	35	knowledge	knowledge	NOUN
cril-1051	68	36	of	of	ADP
cril-1051	68	37	the	the	DET
cril-1051	68	38	language	language	NOUN
cril-1051	68	39	.	.	PUNCT
cril-1051	69	1	nevertheless	nevertheless	ADV
cril-1051	69	2	,	,	PUNCT
cril-1051	69	3	the	the	DET
cril-1051	69	4	latent	latent	NOUN
cril-1051	69	5	patterns	pattern	NOUN
cril-1051	69	6	discovered	discover	VERB
cril-1051	69	7	by	by	ADP
cril-1051	69	8	unsupervised	unsupervised	ADJ
cril-1051	69	9	models	model	NOUN
cril-1051	69	10	can	can	AUX
cril-1051	69	11	have	have	VERB
cril-1051	69	12	many	many	ADJ
cril-1051	69	13	uses	use	NOUN
cril-1051	69	14	such	such	ADJ
cril-1051	69	15	as	as	ADP
cril-1051	69	16	being	be	AUX
cril-1051	69	17	leveraged	leverage	VERB
cril-1051	69	18	in	in	ADP
cril-1051	69	19	a	a	DET
cril-1051	69	20	semi	semi	ADJ
cril-1051	69	21	-	-	ADJ
cril-1051	69	22	supervised	supervised	ADJ
cril-1051	69	23	approach	approach	NOUN
cril-1051	69	24	.	.	PUNCT
cril-1051	70	1	semisupervised	semisupervise	VERB
cril-1051	70	2	learning	learning	NOUN
cril-1051	70	3	combines	combine	VERB
cril-1051	70	4	some	some	DET
cril-1051	70	5	supervised	supervised	ADJ
cril-1051	70	6	data	datum	NOUN
cril-1051	70	7	with	with	ADP
cril-1051	70	8	a	a	DET
cril-1051	70	9	larger	large	ADJ
cril-1051	70	10	set	set	NOUN
cril-1051	70	11	of	of	ADP
cril-1051	70	12	unsupervised	unsupervised	ADJ
cril-1051	70	13	data	datum	NOUN
cril-1051	70	14	(	(	PUNCT
cril-1051	70	15	kohonen	kohonen	PROPN
cril-1051	70	16	,	,	PUNCT
cril-1051	70	17	virpioja	virpioja	NOUN
cril-1051	70	18	&	&	CCONJ
cril-1051	70	19	lagus	lagus	PROPN
cril-1051	70	20	2010	2010	NUM
cril-1051	70	21	;	;	PUNCT
cril-1051	70	22	poon	poon	NOUN
cril-1051	70	23	,	,	PUNCT
cril-1051	70	24	cherry	cherry	NOUN
cril-1051	70	25	&	&	CCONJ
cril-1051	70	26	toutanova	toutanova	PROPN
cril-1051	70	27	2009	2009	NUM
cril-1051	70	28	)	)	PUNCT
cril-1051	70	29	.	.	PUNCT
cril-1051	71	1	this	this	DET
cril-1051	71	2	approach	approach	NOUN
cril-1051	71	3	is	be	AUX
cril-1051	71	4	suitable	suitable	ADJ
cril-1051	71	5	if	if	SCONJ
cril-1051	71	6	available	available	ADJ
cril-1051	71	7	annotated	annotated	ADJ
cril-1051	71	8	data	datum	NOUN
cril-1051	71	9	is	be	AUX
cril-1051	71	10	not	not	PART
cril-1051	71	11	adequate	adequate	ADJ
cril-1051	71	12	to	to	PART
cril-1051	71	13	effectively	effectively	ADV
cril-1051	71	14	train	train	VERB
cril-1051	71	15	a	a	DET
cril-1051	71	16	supervised	supervised	ADJ
cril-1051	71	17	model	model	NOUN
cril-1051	71	18	and	and	CCONJ
cril-1051	71	19	it	it	PRON
cril-1051	71	20	may	may	AUX
cril-1051	71	21	be	be	AUX
cril-1051	71	22	ideal	ideal	ADJ
cril-1051	71	23	for	for	ADP
cril-1051	71	24	ldd	ldd	PROPN
cril-1051	71	25	because	because	SCONJ
cril-1051	71	26	having	have	VERB
cril-1051	71	27	substantial	substantial	ADJ
cril-1051	71	28	amounts	amount	NOUN
cril-1051	71	29	of	of	ADP
cril-1051	71	30	unannotated	unannotated	ADJ
cril-1051	71	31	data	datum	NOUN
cril-1051	71	32	with	with	ADP
cril-1051	71	33	a	a	DET
cril-1051	71	34	small	small	ADJ
cril-1051	71	35	amount	amount	NOUN
cril-1051	71	36	of	of	ADP
cril-1051	71	37	annotated	annotate	VERB
cril-1051	71	38	data	data	NOUN
cril-1051	71	39	is	be	AUX
cril-1051	71	40	a	a	DET
cril-1051	71	41	common	common	ADJ
cril-1051	71	42	situation	situation	NOUN
cril-1051	71	43	.	.	PUNCT
cril-1051	72	1	unfortunately	unfortunately	ADV
cril-1051	72	2	,	,	PUNCT
cril-1051	72	3	real	real	ADJ
cril-1051	72	4	applications	application	NOUN
cril-1051	72	5	of	of	ADP
cril-1051	72	6	semi	semi	ADJ
cril-1051	72	7	-	-	ADJ
cril-1051	72	8	supervised	supervised	ADJ
cril-1051	72	9	learning	learning	NOUN
cril-1051	72	10	,	,	PUNCT
cril-1051	72	11	specifically	specifically	ADV
cril-1051	72	12	for	for	ADP
cril-1051	72	13	computational	computational	ADJ
cril-1051	72	14	morphology	morphology	NOUN
cril-1051	72	15	,	,	PUNCT
cril-1051	72	16	are	be	AUX
cril-1051	72	17	relatively	relatively	ADV
cril-1051	72	18	rare	rare	ADJ
cril-1051	72	19	,	,	PUNCT
cril-1051	72	20	particularly	particularly	ADV
cril-1051	72	21	with	with	ADP
cril-1051	72	22	neural	neural	ADJ
cril-1051	72	23	networks	network	NOUN
cril-1051	72	24	.	.	PUNCT
cril-1051	73	1	there	there	PRON
cril-1051	73	2	are	be	VERB
cril-1051	73	3	exceptions	exception	NOUN
cril-1051	73	4	,	,	PUNCT
cril-1051	73	5	such	such	ADJ
cril-1051	73	6	as	as	ADP
cril-1051	73	7	ahlberg	ahlberg	PROPN
cril-1051	73	8	et	et	PROPN
cril-1051	73	9	al	al	PROPN
cril-1051	73	10	.	.	PROPN
cril-1051	74	1	(	(	PUNCT
cril-1051	74	2	2014	2014	NUM
cril-1051	74	3	)	)	PUNCT
cril-1051	74	4	,	,	PUNCT
cril-1051	74	5	where	where	SCONJ
cril-1051	74	6	semi	semi	ADJ
cril-1051	74	7	-	-	ADJ
cril-1051	74	8	supervised	supervised	ADJ
cril-1051	74	9	learning	learning	NOUN
cril-1051	74	10	was	be	AUX
cril-1051	74	11	used	use	VERB
cril-1051	74	12	to	to	PART
cril-1051	74	13	induce	induce	VERB
cril-1051	74	14	morphological	morphological	ADJ
cril-1051	74	15	paradigms	paradigm	NOUN
cril-1051	74	16	in	in	ADP
cril-1051	74	17	low	low	ADJ
cril-1051	74	18	-	-	PUNCT
cril-1051	74	19	resource	resource	NOUN
cril-1051	74	20	settings	setting	NOUN
cril-1051	74	21	.	.	PUNCT
cril-1051	75	1	until	until	ADP
cril-1051	75	2	the	the	DET
cril-1051	75	3	2010s	2010s	NUM
cril-1051	75	4	,	,	PUNCT
cril-1051	75	5	most	most	ADJ
cril-1051	75	6	machine	machine	NOUN
cril-1051	75	7	learning	learning	NOUN
cril-1051	75	8	models	model	NOUN
cril-1051	75	9	were	be	AUX
cril-1051	75	10	feature	feature	NOUN
cril-1051	75	11	-	-	PUNCT
cril-1051	75	12	based	base	VERB
cril-1051	75	13	with	with	ADP
cril-1051	75	14	hand	hand	NOUN
cril-1051	75	15	-	-	PUNCT
cril-1051	75	16	designed	design	VERB
cril-1051	75	17	features	feature	NOUN
cril-1051	75	18	,	,	PUNCT
cril-1051	75	19	illustrated	illustrate	VERB
cril-1051	75	20	in	in	ADP
cril-1051	75	21	figure	figure	NOUN
cril-1051	75	22	4	4	NUM
cril-1051	75	23	.	.	PUNCT
cril-1051	76	1	the	the	DET
cril-1051	76	2	input	input	NOUN
cril-1051	76	3	would	would	AUX
cril-1051	76	4	be	be	AUX
cril-1051	76	5	a	a	DET
cril-1051	76	6	hand	hand	NOUN
cril-1051	76	7	-	-	PUNCT
cril-1051	76	8	designed	design	VERB
cril-1051	76	9	feature	feature	NOUN
cril-1051	76	10	function	function	NOUN
cril-1051	76	11	that	that	SCONJ
cril-1051	76	12	for	for	ADP
cril-1051	76	13	morphological	morphological	ADJ
cril-1051	76	14	analysis	analysis	NOUN
cril-1051	76	15	might	might	AUX
cril-1051	76	16	include	include	VERB
cril-1051	76	17	1	1	NUM
cril-1051	76	18	)	)	PUNCT
cril-1051	76	19	the	the	DET
cril-1051	76	20	whole	whole	ADJ
cril-1051	76	21	word	word	NOUN
cril-1051	76	22	,	,	PUNCT
cril-1051	76	23	2	2	X
cril-1051	76	24	)	)	PUNCT
cril-1051	76	25	the	the	DET
cril-1051	76	26	position	position	NOUN
cril-1051	76	27	of	of	ADP
cril-1051	76	28	the	the	DET
cril-1051	76	29	word	word	NOUN
cril-1051	76	30	in	in	ADP
cril-1051	76	31	the	the	DET
cril-1051	76	32	sentence	sentence	NOUN
cril-1051	76	33	,	,	PUNCT
cril-1051	76	34	3	3	X
cril-1051	76	35	)	)	PUNCT
cril-1051	76	36	surrounding	surround	VERB
cril-1051	76	37	words	word	NOUN
cril-1051	76	38	or	or	CCONJ
cril-1051	76	39	morphemes	morpheme	NOUN
cril-1051	76	40	,	,	PUNCT
cril-1051	76	41	4	4	NUM
cril-1051	76	42	)	)	PUNCT
cril-1051	76	43	the	the	DET
cril-1051	76	44	pos	pos	NOUN
cril-1051	76	45	tag	tag	NOUN
cril-1051	76	46	of	of	ADP
cril-1051	76	47	the	the	DET
cril-1051	76	48	previous	previous	ADJ
cril-1051	76	49	morpheme	morpheme	NOUN
cril-1051	76	50	/	/	SYM
cril-1051	76	51	word	word	NOUN
cril-1051	76	52	.	.	PUNCT
cril-1051	77	1	features	feature	NOUN
cril-1051	77	2	are	be	AUX
cril-1051	77	3	assigned	assign	VERB
cril-1051	77	4	weights	weight	NOUN
cril-1051	77	5	by	by	ADP
cril-1051	77	6	the	the	DET
cril-1051	77	7	model	model	NOUN
cril-1051	77	8	during	during	ADP
cril-1051	77	9	training	training	NOUN
cril-1051	77	10	to	to	PART
cril-1051	77	11	achieve	achieve	VERB
cril-1051	77	12	optimal	optimal	ADJ
cril-1051	77	13	performance	performance	NOUN
cril-1051	77	14	according	accord	VERB
cril-1051	77	15	to	to	ADP
cril-1051	77	16	some	some	DET
cril-1051	77	17	objective	objective	ADJ
cril-1051	77	18	function	function	NOUN
cril-1051	77	19	such	such	ADJ
cril-1051	77	20	as	as	ADP
cril-1051	77	21	classification	classification	NOUN
cril-1051	77	22	accuracy	accuracy	NOUN
cril-1051	77	23	.	.	PUNCT
cril-1051	78	1	for	for	ADP
cril-1051	78	2	example	example	NOUN
cril-1051	78	3	,	,	PUNCT
cril-1051	78	4	in	in	ADP
cril-1051	78	5	a	a	DET
cril-1051	78	6	morpheme	morpheme	ADJ
cril-1051	78	7	segmentation	segmentation	NOUN
cril-1051	78	8	task	task	NOUN
cril-1051	78	9	where	where	SCONJ
cril-1051	78	10	one	one	NUM
cril-1051	78	11	chosen	choose	VERB
cril-1051	78	12	feature	feature	NOUN
cril-1051	78	13	is	be	AUX
cril-1051	78	14	the	the	DET
cril-1051	78	15	previous	previous	ADJ
cril-1051	78	16	word	word	NOUN
cril-1051	78	17	and	and	CCONJ
cril-1051	78	18	the	the	DET
cril-1051	78	19	previous	previous	ADJ
cril-1051	78	20	word	word	NOUN
cril-1051	78	21	is	be	AUX
cril-1051	78	22	some	some	DET
cril-1051	78	23	form	form	NOUN
cril-1051	78	24	of	of	ADP
cril-1051	78	25	the	the	DET
cril-1051	78	26	english	english	PROPN
cril-1051	78	27	“	"	PUNCT
cril-1051	78	28	to	to	PART
cril-1051	78	29	be	be	AUX
cril-1051	78	30	”	"	PUNCT
cril-1051	78	31	verb	verb	NOUN
cril-1051	78	32	,	,	PUNCT
cril-1051	78	33	and	and	CCONJ
cril-1051	78	34	the	the	DET
cril-1051	78	35	target	target	NOUN
cril-1051	78	36	word	word	NOUN
cril-1051	78	37	ends	end	VERB
cril-1051	78	38	in	in	ADP
cril-1051	78	39	“	"	PUNCT
cril-1051	78	40	ing	ing	NOUN
cril-1051	78	41	”	"	PUNCT
cril-1051	78	42	,	,	PUNCT
cril-1051	78	43	then	then	ADV
cril-1051	78	44	the	the	DET
cril-1051	78	45	model	model	NOUN
cril-1051	78	46	might	might	AUX
cril-1051	78	47	give	give	VERB
cril-1051	78	48	a	a	DET
cril-1051	78	49	high	high	ADJ
cril-1051	78	50	weight	weight	NOUN
cril-1051	78	51	to	to	ADP
cril-1051	78	52	the	the	DET
cril-1051	78	53	previous	previous	ADJ
cril-1051	78	54	word	word	NOUN
cril-1051	78	55	so	so	SCONJ
cril-1051	78	56	that	that	SCONJ
cril-1051	78	57	the	the	DET
cril-1051	78	58	model	model	NOUN
cril-1051	78	59	pays	pay	VERB
cril-1051	78	60	attention	attention	NOUN
cril-1051	78	61	to	to	ADP
cril-1051	78	62	it	it	PRON
cril-1051	78	63	when	when	SCONJ
cril-1051	78	64	deciding	decide	VERB
cril-1051	78	65	how	how	SCONJ
cril-1051	78	66	to	to	PART
cril-1051	78	67	segment	segment	VERB
cril-1051	78	68	a	a	DET
cril-1051	78	69	word	word	NOUN
cril-1051	78	70	ending	end	VERB
cril-1051	78	71	in	in	ADP
cril-1051	78	72	“	"	PUNCT
cril-1051	78	73	ing	ing	NOUN
cril-1051	78	74	”	"	PUNCT
cril-1051	78	75	.	.	PUNCT
cril-1051	79	1	the	the	DET
cril-1051	79	2	performance	performance	NOUN
cril-1051	79	3	of	of	ADP
cril-1051	79	4	feature	feature	NOUN
cril-1051	79	5	-	-	PUNCT
cril-1051	79	6	based	base	VERB
cril-1051	79	7	models	model	NOUN
cril-1051	79	8	,	,	PUNCT
cril-1051	79	9	such	such	ADJ
cril-1051	79	10	as	as	ADP
cril-1051	79	11	conditional	conditional	ADJ
cril-1051	79	12	random	random	ADJ
cril-1051	79	13	fields	field	NOUN
cril-1051	79	14	(	(	PUNCT
cril-1051	79	15	crf	crf	NOUN
cril-1051	79	16	)	)	PUNCT
cril-1051	79	17	and	and	CCONJ
cril-1051	79	18	support	support	VERB
cril-1051	79	19	vector	vector	NOUN
cril-1051	79	20	machines	machine	NOUN
cril-1051	79	21	(	(	PUNCT
cril-1051	79	22	svm	svm	PROPN
cril-1051	79	23	)	)	PUNCT
cril-1051	79	24	,	,	PUNCT
cril-1051	79	25	relies	rely	VERB
cril-1051	79	26	heavily	heavily	ADV
cril-1051	79	27	on	on	ADP
cril-1051	79	28	the	the	DET
cril-1051	79	29	manual	manual	ADJ
cril-1051	79	30	choice	choice	NOUN
cril-1051	79	31	of	of	ADP
cril-1051	79	32	features	feature	NOUN
cril-1051	79	33	.	.	PUNCT
cril-1051	80	1	this	this	PRON
cril-1051	80	2	could	could	AUX
cril-1051	80	3	be	be	AUX
cril-1051	80	4	a	a	DET
cril-1051	80	5	drawback	drawback	NOUN
cril-1051	80	6	for	for	ADP
cril-1051	80	7	under	under	ADV
cril-1051	80	8	-	-	PUNCT
cril-1051	80	9	described	describe	VERB
cril-1051	80	10	languages	language	NOUN
cril-1051	80	11	,	,	PUNCT
cril-1051	80	12	because	because	SCONJ
cril-1051	80	13	if	if	SCONJ
cril-1051	80	14	little	little	ADJ
cril-1051	80	15	linguistic	linguistic	ADJ
cril-1051	80	16	description	description	NOUN
cril-1051	80	17	is	be	AUX
cril-1051	80	18	available	available	ADJ
cril-1051	80	19	,	,	PUNCT
cril-1051	80	20	how	how	SCONJ
cril-1051	80	21	does	do	AUX
cril-1051	80	22	one	one	PRON
cril-1051	80	23	know	know	VERB
cril-1051	80	24	which	which	PRON
cril-1051	80	25	features	feature	VERB
cril-1051	80	26	are	be	AUX
cril-1051	80	27	optimal	optimal	ADJ
cril-1051	80	28	for	for	ADP
cril-1051	80	29	that	that	DET
cril-1051	80	30	language	language	NOUN
cril-1051	80	31	?	?	PUNCT
cril-1051	81	1	fortunately	fortunately	ADV
cril-1051	81	2	,	,	PUNCT
cril-1051	81	3	some	some	DET
cril-1051	81	4	feature	feature	NOUN
cril-1051	81	5	-	-	PUNCT
cril-1051	81	6	based	base	VERB
cril-1051	81	7	models	model	NOUN
cril-1051	81	8	have	have	AUX
cril-1051	81	9	been	be	AUX
cril-1051	81	10	shown	show	VERB
cril-1051	81	11	to	to	PART
cril-1051	81	12	perform	perform	VERB
cril-1051	81	13	reasonably	reasonably	ADV
cril-1051	81	14	well	well	ADV
cril-1051	81	15	using	use	VERB
cril-1051	81	16	language	language	NOUN
cril-1051	81	17	-	-	PUNCT
cril-1051	81	18	independent	independent	ADJ
cril-1051	81	19	features	feature	NOUN
cril-1051	81	20	such	such	ADJ
cril-1051	81	21	as	as	ADP
cril-1051	81	22	length	length	NOUN
cril-1051	81	23	of	of	ADP
cril-1051	81	24	word	word	NOUN
cril-1051	81	25	or	or	CCONJ
cril-1051	81	26	placement	placement	NOUN
cril-1051	81	27	of	of	ADP
cril-1051	81	28	letter	letter	NOUN
cril-1051	81	29	in	in	ADP
cril-1051	81	30	word	word	NOUN
cril-1051	81	31	(	(	PUNCT
cril-1051	81	32	ruokolainen	ruokolainen	NOUN
cril-1051	81	33	et	et	PROPN
cril-1051	81	34	al	al	PROPN
cril-1051	81	35	.	.	PROPN
cril-1051	81	36	2016	2016	NUM
cril-1051	81	37	;	;	PUNCT
cril-1051	81	38	moeller	moeller	PROPN
cril-1051	81	39	&	&	CCONJ
cril-1051	81	40	hulden	hulden	VERB
cril-1051	81	41	2018	2018	NUM
cril-1051	81	42	)	)	PUNCT
cril-1051	81	43	.	.	PUNCT
cril-1051	82	1	colorado	colorado	PROPN
cril-1051	82	2	research	research	PROPN
cril-1051	82	3	in	in	ADP
cril-1051	82	4	linguistics	linguistic	NOUN
cril-1051	82	5	,	,	PUNCT
cril-1051	82	6	volume	volume	NOUN
cril-1051	82	7	25	25	NUM
cril-1051	82	8	(	(	PUNCT
cril-1051	82	9	2021	2021	NUM
cril-1051	82	10	)	)	PUNCT
cril-1051	82	11	8	8	NUM
cril-1051	82	12	figure	figure	NOUN
cril-1051	82	13	4	4	NUM
cril-1051	82	14	.	.	PUNCT
cril-1051	82	15	feature	feature	NOUN
cril-1051	82	16	-	-	PUNCT
cril-1051	82	17	based	base	VERB
cril-1051	82	18	machine	machine	NOUN
cril-1051	82	19	learning	learning	NOUN
cril-1051	82	20	requires	require	VERB
cril-1051	82	21	a	a	DET
cril-1051	82	22	human	human	NOUN
cril-1051	82	23	to	to	PART
cril-1051	82	24	identify	identify	VERB
cril-1051	82	25	and	and	CCONJ
cril-1051	82	26	extract	extract	VERB
cril-1051	82	27	features	feature	VERB
cril-1051	82	28	that	that	SCONJ
cril-1051	82	29	a	a	DET
cril-1051	82	30	feature	feature	NOUN
cril-1051	82	31	-	-	PUNCT
cril-1051	82	32	based	base	VERB
cril-1051	82	33	classification	classification	NOUN
cril-1051	82	34	model	model	NOUN
cril-1051	82	35	such	such	ADJ
cril-1051	82	36	as	as	ADP
cril-1051	82	37	a	a	DET
cril-1051	82	38	crf	crf	PROPN
cril-1051	82	39	uses	use	VERB
cril-1051	82	40	to	to	PART
cril-1051	82	41	provide	provide	VERB
cril-1051	82	42	the	the	DET
cril-1051	82	43	correct	correct	ADJ
cril-1051	82	44	output	output	NOUN
cril-1051	82	45	currently	currently	ADV
cril-1051	82	46	,	,	PUNCT
cril-1051	82	47	neural	neural	ADJ
cril-1051	82	48	networks	network	NOUN
cril-1051	82	49	models	model	NOUN
cril-1051	82	50	,	,	PUNCT
cril-1051	82	51	or	or	CCONJ
cril-1051	82	52	deep	deep	ADJ
cril-1051	82	53	learning	learning	NOUN
cril-1051	82	54	models	model	NOUN
cril-1051	82	55	,	,	PUNCT
cril-1051	82	56	are	be	AUX
cril-1051	82	57	dominating	dominate	VERB
cril-1051	82	58	nlp	nlp	NOUN
cril-1051	82	59	(	(	PUNCT
cril-1051	82	60	goldberg	goldberg	PROPN
cril-1051	82	61	2017	2017	NUM
cril-1051	82	62	)	)	PUNCT
cril-1051	82	63	.	.	PUNCT
cril-1051	83	1	even	even	ADV
cril-1051	83	2	though	though	SCONJ
cril-1051	83	3	they	they	PRON
cril-1051	83	4	outperform	outperform	VERB
cril-1051	83	5	older	old	ADJ
cril-1051	83	6	,	,	PUNCT
cril-1051	83	7	feature	feature	NOUN
cril-1051	83	8	-	-	PUNCT
cril-1051	83	9	based	base	VERB
cril-1051	83	10	models	model	NOUN
cril-1051	83	11	on	on	ADP
cril-1051	83	12	almost	almost	ADV
cril-1051	83	13	all	all	PRON
cril-1051	83	14	tasks	task	NOUN
cril-1051	83	15	,	,	PUNCT
cril-1051	83	16	they	they	PRON
cril-1051	83	17	did	do	AUX
cril-1051	83	18	not	not	PART
cril-1051	83	19	become	become	VERB
cril-1051	83	20	popular	popular	ADJ
cril-1051	83	21	until	until	ADP
cril-1051	83	22	the	the	DET
cril-1051	83	23	mid-2010s	mid-2010s	NOUN
cril-1051	83	24	because	because	SCONJ
cril-1051	83	25	they	they	PRON
cril-1051	83	26	require	require	VERB
cril-1051	83	27	greater	great	ADJ
cril-1051	83	28	computing	computing	NOUN
cril-1051	83	29	power	power	NOUN
cril-1051	83	30	and	and	CCONJ
cril-1051	83	31	,	,	PUNCT
cril-1051	83	32	for	for	ADP
cril-1051	83	33	some	some	DET
cril-1051	83	34	tasks	task	NOUN
cril-1051	83	35	,	,	PUNCT
cril-1051	83	36	train	train	VERB
cril-1051	83	37	more	more	ADV
cril-1051	83	38	slowly	slowly	ADV
cril-1051	83	39	(	(	PUNCT
cril-1051	83	40	cotterell	cotterell	NOUN
cril-1051	83	41	&	&	CCONJ
cril-1051	83	42	heigold	heigold	PROPN
cril-1051	83	43	2017	2017	NUM
cril-1051	83	44	)	)	PUNCT
cril-1051	83	45	.	.	PUNCT
cril-1051	84	1	neural	neural	ADJ
cril-1051	84	2	networks	network	NOUN
cril-1051	84	3	,	,	PUNCT
cril-1051	84	4	illustrated	illustrate	VERB
cril-1051	84	5	in	in	ADP
cril-1051	84	6	figure	figure	NOUN
cril-1051	84	7	5	5	NUM
cril-1051	84	8	,	,	PUNCT
cril-1051	84	9	refers	refer	VERB
cril-1051	84	10	to	to	ADP
cril-1051	84	11	a	a	DET
cril-1051	84	12	family	family	NOUN
cril-1051	84	13	of	of	ADP
cril-1051	84	14	supervised	supervised	ADJ
cril-1051	84	15	machine	machine	NOUN
cril-1051	84	16	learning	learning	NOUN
cril-1051	84	17	models	model	NOUN
cril-1051	84	18	that	that	PRON
cril-1051	84	19	are	be	AUX
cril-1051	84	20	composed	compose	VERB
cril-1051	84	21	of	of	ADP
cril-1051	84	22	layers	layer	NOUN
cril-1051	84	23	of	of	ADP
cril-1051	84	24	statistical	statistical	ADJ
cril-1051	84	25	units	unit	NOUN
cril-1051	84	26	.	.	PUNCT
cril-1051	85	1	the	the	DET
cril-1051	85	2	layers	layer	NOUN
cril-1051	85	3	essentially	essentially	ADV
cril-1051	85	4	substitute	substitute	VERB
cril-1051	85	5	the	the	DET
cril-1051	85	6	feature	feature	NOUN
cril-1051	85	7	engineering	engineering	NOUN
cril-1051	85	8	needed	need	VERB
cril-1051	85	9	in	in	ADP
cril-1051	85	10	non	non	ADJ
cril-1051	85	11	-	-	ADJ
cril-1051	85	12	neural	neural	ADJ
cril-1051	85	13	machine	machine	NOUN
cril-1051	85	14	learning	learning	NOUN
cril-1051	85	15	.	.	PUNCT
cril-1051	86	1	multiple	multiple	ADJ
cril-1051	86	2	embedded	embed	VERB
cril-1051	86	3	layers	layer	NOUN
cril-1051	86	4	allow	allow	VERB
cril-1051	86	5	the	the	DET
cril-1051	86	6	model	model	NOUN
cril-1051	86	7	to	to	PART
cril-1051	86	8	look	look	VERB
cril-1051	86	9	at	at	ADP
cril-1051	86	10	an	an	DET
cril-1051	86	11	exponential	exponential	ADJ
cril-1051	86	12	number	number	NOUN
cril-1051	86	13	of	of	ADP
cril-1051	86	14	“	"	PUNCT
cril-1051	86	15	semantically	semantically	ADV
cril-1051	86	16	”	"	PUNCT
cril-1051	86	17	neighboring	neighboring	NOUN
cril-1051	86	18	instances	instance	NOUN
cril-1051	86	19	of	of	ADP
cril-1051	86	20	each	each	DET
cril-1051	86	21	training	training	NOUN
cril-1051	86	22	instance	instance	NOUN
cril-1051	86	23	it	it	PRON
cril-1051	86	24	encounters	encounter	VERB
cril-1051	86	25	(	(	PUNCT
cril-1051	86	26	bengio	bengio	NOUN
cril-1051	86	27	et	et	PROPN
cril-1051	86	28	al	al	PROPN
cril-1051	86	29	.	.	PROPN
cril-1051	86	30	2003	2003	NUM
cril-1051	86	31	)	)	PUNCT
cril-1051	86	32	.	.	PUNCT
cril-1051	87	1	the	the	DET
cril-1051	87	2	layers	layer	NOUN
cril-1051	87	3	create	create	VERB
cril-1051	87	4	intermediate	intermediate	ADJ
cril-1051	87	5	representations	representation	NOUN
cril-1051	87	6	of	of	ADP
cril-1051	87	7	the	the	DET
cril-1051	87	8	data	datum	NOUN
cril-1051	87	9	that	that	PRON
cril-1051	87	10	allow	allow	VERB
cril-1051	87	11	the	the	DET
cril-1051	87	12	model	model	NOUN
cril-1051	87	13	to	to	PART
cril-1051	87	14	“	"	PUNCT
cril-1051	87	15	learn	learn	VERB
cril-1051	87	16	”	"	PUNCT
cril-1051	87	17	a	a	DET
cril-1051	87	18	distributed	distribute	VERB
cril-1051	87	19	representation	representation	NOUN
cril-1051	87	20	of	of	ADP
cril-1051	87	21	elements	element	NOUN
cril-1051	87	22	within	within	ADP
cril-1051	87	23	each	each	DET
cril-1051	87	24	instance	instance	NOUN
cril-1051	87	25	(	(	PUNCT
cril-1051	87	26	e.g.	e.g.	ADV
cril-1051	87	27	,	,	PUNCT
cril-1051	87	28	a	a	DET
cril-1051	87	29	distributed	distribute	VERB
cril-1051	87	30	representation	representation	NOUN
cril-1051	87	31	of	of	ADP
cril-1051	87	32	words	word	NOUN
cril-1051	87	33	within	within	ADP
cril-1051	87	34	a	a	DET
cril-1051	87	35	sentence	sentence	NOUN
cril-1051	87	36	)	)	PUNCT
cril-1051	87	37	.	.	PUNCT
cril-1051	88	1	this	this	DET
cril-1051	88	2	ability	ability	NOUN
cril-1051	88	3	of	of	ADP
cril-1051	88	4	the	the	DET
cril-1051	88	5	model	model	NOUN
cril-1051	88	6	to	to	PART
cril-1051	88	7	learn	learn	VERB
cril-1051	88	8	requires	require	VERB
cril-1051	88	9	no	no	DET
cril-1051	88	10	(	(	PUNCT
cril-1051	88	11	or	or	CCONJ
cril-1051	88	12	at	at	ADP
cril-1051	88	13	most	most	ADJ
cril-1051	88	14	,	,	PUNCT
cril-1051	88	15	quite	quite	ADV
cril-1051	88	16	simple	simple	ADJ
cril-1051	88	17	)	)	PUNCT
cril-1051	88	18	manual	manual	ADJ
cril-1051	88	19	feature	feature	NOUN
cril-1051	88	20	design	design	NOUN
cril-1051	88	21	.	.	PUNCT
cril-1051	89	1	each	each	DET
cril-1051	89	2	unit	unit	NOUN
cril-1051	89	3	in	in	ADP
cril-1051	89	4	each	each	DET
cril-1051	89	5	layer	layer	NOUN
cril-1051	89	6	is	be	AUX
cril-1051	89	7	connected	connect	VERB
cril-1051	89	8	to	to	ADP
cril-1051	89	9	each	each	DET
cril-1051	89	10	unit	unit	NOUN
cril-1051	89	11	in	in	ADP
cril-1051	89	12	the	the	DET
cril-1051	89	13	adjacent	adjacent	ADJ
cril-1051	89	14	layers	layer	NOUN
cril-1051	89	15	.	.	PUNCT
cril-1051	90	1	vector	vector	NOUN
cril-1051	90	2	representations	representation	NOUN
cril-1051	90	3	of	of	ADP
cril-1051	90	4	the	the	DET
cril-1051	90	5	data	datum	NOUN
cril-1051	90	6	are	be	AUX
cril-1051	90	7	received	receive	VERB
cril-1051	90	8	by	by	ADP
cril-1051	90	9	an	an	DET
cril-1051	90	10	input	input	NOUN
cril-1051	90	11	layer	layer	NOUN
cril-1051	90	12	and	and	CCONJ
cril-1051	90	13	transformed	transform	VERB
cril-1051	90	14	in	in	ADP
cril-1051	90	15	“	"	PUNCT
cril-1051	90	16	hidden	hidden	ADJ
cril-1051	90	17	”	"	PUNCT
cril-1051	90	18	layers	layer	NOUN
cril-1051	90	19	.	.	PUNCT
cril-1051	91	1	the	the	DET
cril-1051	91	2	hidden	hide	VERB
cril-1051	91	3	layers	layer	NOUN
cril-1051	91	4	feed	feed	VERB
cril-1051	91	5	into	into	ADP
cril-1051	91	6	a	a	DET
cril-1051	91	7	final	final	ADJ
cril-1051	91	8	logistic	logistic	ADJ
cril-1051	91	9	function	function	NOUN
cril-1051	91	10	layer	layer	NOUN
cril-1051	91	11	(	(	PUNCT
cril-1051	91	12	i.e.	i.e.	X
cril-1051	91	13	,	,	PUNCT
cril-1051	91	14	softmax	softmax	NOUN
cril-1051	91	15	)	)	PUNCT
cril-1051	91	16	that	that	PRON
cril-1051	91	17	outputs	output	VERB
cril-1051	91	18	a	a	DET
cril-1051	91	19	prediction	prediction	NOUN
cril-1051	91	20	of	of	ADP
cril-1051	91	21	each	each	DET
cril-1051	91	22	possible	possible	ADJ
cril-1051	91	23	class	class	NOUN
cril-1051	91	24	as	as	ADP
cril-1051	91	25	a	a	DET
cril-1051	91	26	probability	probability	NOUN
cril-1051	91	27	between	between	ADP
cril-1051	91	28	0	0	NUM
cril-1051	91	29	and	and	CCONJ
cril-1051	91	30	1	1	NUM
cril-1051	91	31	.	.	PUNCT
cril-1051	92	1	the	the	DET
cril-1051	92	2	connections	connection	NOUN
cril-1051	92	3	between	between	ADP
cril-1051	92	4	layers	layer	NOUN
cril-1051	92	5	are	be	AUX
cril-1051	92	6	represented	represent	VERB
cril-1051	92	7	by	by	ADP
cril-1051	92	8	learnable	learnable	ADJ
cril-1051	92	9	weights	weight	NOUN
cril-1051	92	10	;	;	PUNCT
cril-1051	92	11	the	the	PRON
cril-1051	92	12	higher	high	ADJ
cril-1051	92	13	the	the	DET
cril-1051	92	14	weight	weight	NOUN
cril-1051	92	15	the	the	DET
cril-1051	92	16	more	more	ADJ
cril-1051	92	17	influence	influence	NOUN
cril-1051	92	18	a	a	DET
cril-1051	92	19	unit	unit	NOUN
cril-1051	92	20	has	have	VERB
cril-1051	92	21	on	on	ADP
cril-1051	92	22	the	the	DET
cril-1051	92	23	result	result	NOUN
cril-1051	92	24	.	.	PUNCT
cril-1051	93	1	since	since	SCONJ
cril-1051	93	2	deep	deep	ADJ
cril-1051	93	3	learning	learning	NOUN
cril-1051	93	4	is	be	AUX
cril-1051	93	5	supervised	supervise	VERB
cril-1051	93	6	the	the	DET
cril-1051	93	7	weights	weight	NOUN
cril-1051	93	8	are	be	AUX
cril-1051	93	9	adjusted	adjust	VERB
cril-1051	93	10	with	with	ADP
cril-1051	93	11	feedback	feedback	NOUN
cril-1051	93	12	from	from	ADP
cril-1051	93	13	the	the	DET
cril-1051	93	14	gold	gold	NOUN
cril-1051	93	15	standard.2	standard.2	NOUN
cril-1051	93	16	this	this	PRON
cril-1051	93	17	is	be	AUX
cril-1051	93	18	done	do	VERB
cril-1051	93	19	via	via	ADP
cril-1051	93	20	stochastic	stochastic	ADJ
cril-1051	93	21	gradient	gradient	ADJ
cril-1051	93	22	descent	descent	NOUN
cril-1051	93	23	or	or	CCONJ
cril-1051	93	24	some	some	DET
cril-1051	93	25	similar	similar	ADJ
cril-1051	93	26	optimization	optimization	NOUN
cril-1051	93	27	algorithm	algorithm	NOUN
cril-1051	93	28	(	(	PUNCT
cril-1051	93	29	goldberg	goldberg	PROPN
cril-1051	93	30	2017	2017	NUM
cril-1051	93	31	)	)	PUNCT
cril-1051	93	32	and	and	CCONJ
cril-1051	93	33	backpropagation	backpropagation	NOUN
cril-1051	93	34	,	,	PUNCT
cril-1051	93	35	which	which	PRON
cril-1051	93	36	tells	tell	VERB
cril-1051	93	37	the	the	DET
cril-1051	93	38	model	model	NOUN
cril-1051	93	39	how	how	SCONJ
cril-1051	93	40	to	to	PART
cril-1051	93	41	change	change	VERB
cril-1051	93	42	the	the	DET
cril-1051	93	43	parameters	parameter	NOUN
cril-1051	93	44	which	which	PRON
cril-1051	93	45	build	build	VERB
cril-1051	93	46	the	the	DET
cril-1051	93	47	representation	representation	NOUN
cril-1051	93	48	of	of	ADP
cril-1051	93	49	each	each	DET
cril-1051	93	50	layer	layer	NOUN
cril-1051	93	51	from	from	ADP
cril-1051	93	52	the	the	DET
cril-1051	93	53	previous	previous	ADJ
cril-1051	93	54	layer	layer	NOUN
cril-1051	93	55	(	(	PUNCT
cril-1051	93	56	lecun	lecun	PROPN
cril-1051	93	57	,	,	PUNCT
cril-1051	93	58	bengio	bengio	PROPN
cril-1051	93	59	&	&	CCONJ
cril-1051	93	60	hinton	hinton	PROPN
cril-1051	93	61	2015	2015	NUM
cril-1051	93	62	)	)	PUNCT
cril-1051	93	63	.	.	PUNCT
cril-1051	94	1	computational	computational	ADJ
cril-1051	94	2	morphology	morphology	NOUN
cril-1051	94	3	for	for	ADP
cril-1051	94	4	language	language	NOUN
cril-1051	94	5	description	description	NOUN
cril-1051	94	6	and	and	CCONJ
cril-1051	94	7	documentation	documentation	NOUN
cril-1051	94	8	9	9	NUM
cril-1051	94	9	figure	figure	NOUN
cril-1051	94	10	5	5	NUM
cril-1051	94	11	.	.	PUNCT
cril-1051	94	12	neural	neural	ADJ
cril-1051	94	13	networks	network	NOUN
cril-1051	94	14	,	,	PUNCT
cril-1051	94	15	or	or	CCONJ
cril-1051	94	16	deep	deep	ADJ
cril-1051	94	17	learning	learning	NOUN
cril-1051	94	18	,	,	PUNCT
cril-1051	94	19	models	model	NOUN
cril-1051	94	20	learn	learn	VERB
cril-1051	94	21	what	what	PRON
cril-1051	94	22	features	feature	VERB
cril-1051	94	23	in	in	ADP
cril-1051	94	24	the	the	DET
cril-1051	94	25	data	datum	NOUN
cril-1051	94	26	are	be	AUX
cril-1051	94	27	important	important	ADJ
cril-1051	94	28	for	for	ADP
cril-1051	94	29	giving	give	VERB
cril-1051	94	30	the	the	DET
cril-1051	94	31	correct	correct	ADJ
cril-1051	94	32	output	output	NOUN
cril-1051	94	33	until	until	ADP
cril-1051	94	34	recently	recently	ADV
cril-1051	94	35	,	,	PUNCT
cril-1051	94	36	neural	neural	ADJ
cril-1051	94	37	networks	network	NOUN
cril-1051	94	38	had	have	VERB
cril-1051	94	39	the	the	DET
cril-1051	94	40	same	same	ADJ
cril-1051	94	41	great	great	ADJ
cril-1051	94	42	disadvantage	disadvantage	NOUN
cril-1051	94	43	that	that	SCONJ
cril-1051	94	44	unsupervised	unsupervised	ADJ
cril-1051	94	45	learning	learning	NOUN
cril-1051	94	46	has	have	VERB
cril-1051	94	47	–	–	PUNCT
cril-1051	94	48	superior	superior	ADJ
cril-1051	94	49	performance	performance	NOUN
cril-1051	94	50	required	require	VERB
cril-1051	94	51	a	a	DET
cril-1051	94	52	great	great	ADJ
cril-1051	94	53	deal	deal	NOUN
cril-1051	94	54	of	of	ADP
cril-1051	94	55	data	datum	NOUN
cril-1051	94	56	.	.	PUNCT
cril-1051	95	1	data	datum	NOUN
cril-1051	95	2	from	from	ADP
cril-1051	95	3	ldd	ldd	PROPN
cril-1051	95	4	would	would	AUX
cril-1051	95	5	have	have	AUX
cril-1051	95	6	been	be	AUX
cril-1051	95	7	considered	consider	VERB
cril-1051	95	8	inadequate	inadequate	ADJ
cril-1051	95	9	to	to	PART
cril-1051	95	10	train	train	VERB
cril-1051	95	11	neural	neural	ADJ
cril-1051	95	12	networks	network	NOUN
cril-1051	95	13	(	(	PUNCT
cril-1051	95	14	duong	duong	PROPN
cril-1051	95	15	2017	2017	NUM
cril-1051	95	16	)	)	PUNCT
cril-1051	95	17	.	.	PUNCT
cril-1051	96	1	even	even	ADV
cril-1051	96	2	now	now	ADV
cril-1051	96	3	,	,	PUNCT
cril-1051	96	4	a	a	DET
cril-1051	96	5	non	non	ADJ
cril-1051	96	6	-	-	ADJ
cril-1051	96	7	neural	neural	ADJ
cril-1051	96	8	model	model	NOUN
cril-1051	96	9	can	can	AUX
cril-1051	96	10	outperform	outperform	VERB
cril-1051	96	11	any	any	DET
cril-1051	96	12	given	give	VERB
cril-1051	96	13	neural	neural	ADJ
cril-1051	96	14	model	model	NOUN
cril-1051	96	15	that	that	PRON
cril-1051	96	16	is	be	AUX
cril-1051	96	17	not	not	PART
cril-1051	96	18	tuned	tune	VERB
cril-1051	96	19	to	to	ADP
cril-1051	96	20	low	low	ADJ
cril-1051	96	21	-	-	PUNCT
cril-1051	96	22	resource	resource	NOUN
cril-1051	96	23	settings	setting	NOUN
cril-1051	96	24	and	and	CCONJ
cril-1051	96	25	neural	neural	ADJ
cril-1051	96	26	models	model	NOUN
cril-1051	96	27	can	can	AUX
cril-1051	96	28	be	be	AUX
cril-1051	96	29	difficult	difficult	ADJ
cril-1051	96	30	to	to	PART
cril-1051	96	31	optimize	optimize	VERB
cril-1051	96	32	and	and	CCONJ
cril-1051	96	33	tune	tune	NOUN
cril-1051	96	34	for	for	ADP
cril-1051	96	35	low	low	ADJ
cril-1051	96	36	-	-	PUNCT
cril-1051	96	37	resource	resource	NOUN
cril-1051	96	38	settings	setting	NOUN
cril-1051	96	39	(	(	PUNCT
cril-1051	96	40	popel	popel	NOUN
cril-1051	96	41	&	&	CCONJ
cril-1051	96	42	bojar	bojar	PROPN
cril-1051	96	43	2018	2018	NUM
cril-1051	96	44	)	)	PUNCT
cril-1051	96	45	.	.	PUNCT
cril-1051	97	1	new	new	ADJ
cril-1051	97	2	methods	method	NOUN
cril-1051	97	3	are	be	AUX
cril-1051	97	4	being	be	AUX
cril-1051	97	5	explored	explore	VERB
cril-1051	97	6	to	to	PART
cril-1051	97	7	overcome	overcome	VERB
cril-1051	97	8	neural	neural	ADJ
cril-1051	97	9	models	model	NOUN
cril-1051	97	10	’	'	PUNCT
cril-1051	97	11	dependence	dependence	NOUN
cril-1051	97	12	on	on	ADP
cril-1051	97	13	large	large	ADJ
cril-1051	97	14	corpora	corpus	NOUN
cril-1051	97	15	.	.	PUNCT
cril-1051	98	1	examples	example	NOUN
cril-1051	98	2	include	include	VERB
cril-1051	98	3	fine	fine	ADV
cril-1051	98	4	-	-	PUNCT
cril-1051	98	5	tuning	tune	VERB
cril-1051	98	6	a	a	DET
cril-1051	98	7	model	model	NOUN
cril-1051	98	8	to	to	ADP
cril-1051	98	9	the	the	DET
cril-1051	98	10	specific	specific	ADJ
cril-1051	98	11	task	task	NOUN
cril-1051	98	12	and	and	CCONJ
cril-1051	98	13	input	input	NOUN
cril-1051	98	14	data	datum	NOUN
cril-1051	98	15	,	,	PUNCT
cril-1051	98	16	training	train	VERB
cril-1051	98	17	intermediate	intermediate	ADJ
cril-1051	98	18	steps	step	NOUN
cril-1051	98	19	,	,	PUNCT
cril-1051	98	20	or	or	CCONJ
cril-1051	98	21	augmenting	augment	VERB
cril-1051	98	22	the	the	DET
cril-1051	98	23	training	training	NOUN
cril-1051	98	24	data	datum	NOUN
cril-1051	98	25	.	.	PUNCT
cril-1051	99	1	van	van	PROPN
cril-1051	99	2	biljon	biljon	PROPN
cril-1051	99	3	et	et	PROPN
cril-1051	99	4	al	al	PROPN
cril-1051	99	5	.	.	PROPN
cril-1051	99	6	(	(	PUNCT
cril-1051	99	7	2020	2020	NUM
cril-1051	99	8	)	)	PUNCT
cril-1051	99	9	looked	look	VERB
cril-1051	99	10	at	at	ADP
cril-1051	99	11	fine	fine	ADV
cril-1051	99	12	-	-	PUNCT
cril-1051	99	13	tuning	tuning	VERB
cril-1051	99	14	a	a	DET
cril-1051	99	15	model	model	NOUN
cril-1051	99	16	and	and	CCONJ
cril-1051	99	17	determined	determine	VERB
cril-1051	99	18	that	that	SCONJ
cril-1051	99	19	shallowor	shallowor	NOUN
cril-1051	99	20	medium	medium	ADJ
cril-1051	99	21	-	-	PUNCT
cril-1051	99	22	depth	depth	NOUN
cril-1051	99	23	size	size	NOUN
cril-1051	99	24	transformer	transformer	NOUN
cril-1051	99	25	models	model	NOUN
cril-1051	99	26	,	,	PUNCT
cril-1051	99	27	for	for	ADP
cril-1051	99	28	example	example	NOUN
cril-1051	99	29	only	only	ADV
cril-1051	99	30	3	3	NUM
cril-1051	99	31	encoder	encoder	NOUN
cril-1051	99	32	and	and	CCONJ
cril-1051	99	33	3	3	NUM
cril-1051	99	34	decoder	decoder	NOUN
cril-1051	99	35	layers	layer	NOUN
cril-1051	99	36	,	,	PUNCT
cril-1051	99	37	give	give	VERB
cril-1051	99	38	better	well	ADJ
cril-1051	99	39	results	result	NOUN
cril-1051	99	40	with	with	ADP
cril-1051	99	41	limited	limited	ADJ
cril-1051	99	42	training	training	NOUN
cril-1051	99	43	data	datum	NOUN
cril-1051	99	44	.	.	PUNCT
cril-1051	100	1	an	an	DET
cril-1051	100	2	example	example	NOUN
cril-1051	100	3	of	of	ADP
cril-1051	100	4	an	an	DET
cril-1051	100	5	intermediate	intermediate	ADJ
cril-1051	100	6	training	training	NOUN
cril-1051	100	7	step	step	NOUN
cril-1051	100	8	would	would	AUX
cril-1051	100	9	be	be	AUX
cril-1051	100	10	first	first	ADV
cril-1051	100	11	training	train	VERB
cril-1051	100	12	a	a	DET
cril-1051	100	13	segmentation	segmentation	NOUN
cril-1051	100	14	model	model	NOUN
cril-1051	100	15	to	to	PART
cril-1051	100	16	produce	produce	VERB
cril-1051	100	17	surface	surface	NOUN
cril-1051	100	18	segments	segment	NOUN
cril-1051	100	19	(	(	PUNCT
cril-1051	100	20	morphs	morph	NOUN
cril-1051	100	21	)	)	PUNCT
cril-1051	100	22	and	and	CCONJ
cril-1051	100	23	from	from	ADP
cril-1051	100	24	them	they	PRON
cril-1051	100	25	to	to	PART
cril-1051	100	26	learn	learn	VERB
cril-1051	100	27	underlying	underlying	ADJ
cril-1051	100	28	forms	form	NOUN
cril-1051	100	29	of	of	ADP
cril-1051	100	30	morphemes	morpheme	NOUN
cril-1051	100	31	(	(	PUNCT
cril-1051	100	32	e.g.	e.g.	ADV
cril-1051	100	33	,	,	PUNCT
cril-1051	100	34	“	"	PUNCT
cril-1051	100	35	impossible	impossible	ADJ
cril-1051	100	36	”	"	PUNCT
cril-1051	100	37	à	à	X
cril-1051	100	38	“	"	PUNCT
cril-1051	100	39	in	in	ADP
cril-1051	100	40	-	-	PUNCT
cril-1051	100	41	possible	possible	ADJ
cril-1051	100	42	”	"	PUNCT
cril-1051	100	43	à	à	X
cril-1051	100	44	“	"	PUNCT
cril-1051	100	45	neg	neg	VERB
cril-1051	100	46	-	-	PUNCT
cril-1051	100	47	possible	possible	ADJ
cril-1051	100	48	”	"	PUNCT
cril-1051	100	49	)	)	PUNCT
cril-1051	100	50	(	(	PUNCT
cril-1051	100	51	cotterell	cotterell	NOUN
cril-1051	100	52	,	,	PUNCT
cril-1051	100	53	vieira	vieira	PROPN
cril-1051	100	54	&	&	CCONJ
cril-1051	100	55	schütze	schütze	PROPN
cril-1051	100	56	2016	2016	NUM
cril-1051	100	57	;	;	PUNCT
cril-1051	100	58	liu	liu	PROPN
cril-1051	100	59	et	et	PROPN
cril-1051	100	60	al	al	PROPN
cril-1051	100	61	.	.	PROPN
cril-1051	100	62	2018	2018	NUM
cril-1051	100	63	;	;	PUNCT
cril-1051	100	64	moeller	moeller	PROPN
cril-1051	100	65	et	et	PROPN
cril-1051	100	66	al	al	PROPN
cril-1051	100	67	.	.	PROPN
cril-1051	100	68	2019	2019	NUM
cril-1051	100	69	)	)	PUNCT
cril-1051	100	70	.	.	PUNCT
cril-1051	101	1	the	the	DET
cril-1051	101	2	third	third	ADJ
cril-1051	101	3	successful	successful	ADJ
cril-1051	101	4	method	method	NOUN
cril-1051	101	5	is	be	AUX
cril-1051	101	6	augmenting	augment	VERB
cril-1051	101	7	training	training	NOUN
cril-1051	101	8	data	datum	NOUN
cril-1051	101	9	.	.	PUNCT
cril-1051	102	1	augmentation	augmentation	NOUN
cril-1051	102	2	can	can	AUX
cril-1051	102	3	be	be	AUX
cril-1051	102	4	done	do	VERB
cril-1051	102	5	with	with	ADP
cril-1051	102	6	artificial	artificial	ADJ
cril-1051	102	7	word	word	NOUN
cril-1051	102	8	forms	form	NOUN
cril-1051	102	9	(	(	PUNCT
cril-1051	102	10	liu	liu	PROPN
cril-1051	102	11	et	et	PROPN
cril-1051	102	12	al	al	PROPN
cril-1051	102	13	.	.	PROPN
cril-1051	102	14	2018	2018	NUM
cril-1051	102	15	)	)	PUNCT
cril-1051	102	16	or	or	CCONJ
cril-1051	102	17	with	with	ADP
cril-1051	102	18	information	information	NOUN
cril-1051	102	19	extracted	extract	VERB
cril-1051	102	20	from	from	ADP
cril-1051	102	21	other	other	ADJ
cril-1051	102	22	resources	resource	NOUN
cril-1051	102	23	such	such	ADJ
cril-1051	102	24	as	as	ADP
cril-1051	102	25	grammars	grammar	NOUN
cril-1051	102	26	and	and	CCONJ
cril-1051	102	27	dictionaries	dictionary	NOUN
cril-1051	102	28	.	.	PUNCT
cril-1051	103	1	these	these	PRON
cril-1051	103	2	are	be	AUX
cril-1051	103	3	just	just	ADV
cril-1051	103	4	a	a	DET
cril-1051	103	5	few	few	ADJ
cril-1051	103	6	of	of	ADP
cril-1051	103	7	techniques	technique	NOUN
cril-1051	103	8	that	that	PRON
cril-1051	103	9	have	have	AUX
cril-1051	103	10	been	be	AUX
cril-1051	103	11	investigate	investigate	ADJ
cril-1051	103	12	;	;	PUNCT
cril-1051	103	13	there	there	PRON
cril-1051	103	14	are	be	VERB
cril-1051	103	15	probably	probably	ADV
cril-1051	103	16	many	many	ADJ
cril-1051	103	17	more	more	ADJ
cril-1051	103	18	that	that	SCONJ
cril-1051	103	19	we	we	PRON
cril-1051	103	20	have	have	AUX
cril-1051	103	21	not	not	PART
cril-1051	103	22	yet	yet	ADV
cril-1051	103	23	discovered	discover	VERB
cril-1051	103	24	.	.	PUNCT
cril-1051	104	1	although	although	SCONJ
cril-1051	104	2	nlp	nlp	ADJ
cril-1051	104	3	research	research	NOUN
cril-1051	104	4	in	in	ADP
cril-1051	104	5	lrl	lrl	PROPN
cril-1051	104	6	has	have	AUX
cril-1051	104	7	been	be	AUX
cril-1051	104	8	growing	grow	VERB
cril-1051	104	9	since	since	SCONJ
cril-1051	104	10	the	the	DET
cril-1051	104	11	mid-2010	mid-2010	NOUN
cril-1051	104	12	’s	’s	ADV
cril-1051	104	13	,	,	PUNCT
cril-1051	104	14	very	very	ADV
cril-1051	104	15	little	little	ADJ
cril-1051	104	16	of	of	ADP
cril-1051	104	17	it	it	PRON
cril-1051	104	18	has	have	AUX
cril-1051	104	19	been	be	AUX
cril-1051	104	20	applied	apply	VERB
cril-1051	104	21	to	to	ADP
cril-1051	104	22	linguistic	linguistic	ADJ
cril-1051	104	23	on	on	ADP
cril-1051	104	24	under	under	ADV
cril-1051	104	25	-	-	PUNCT
cril-1051	104	26	documented	document	VERB
cril-1051	104	27	languages	language	NOUN
cril-1051	104	28	.	.	PUNCT
cril-1051	105	1	one	one	NUM
cril-1051	105	2	exception	exception	NOUN
cril-1051	105	3	is	be	AUX
cril-1051	105	4	the	the	DET
cril-1051	105	5	aggregation	aggregation	NOUN
cril-1051	105	6	project	project	NOUN
cril-1051	105	7	(	(	PUNCT
cril-1051	105	8	bender	bender	NOUN
cril-1051	105	9	2014	2014	NUM
cril-1051	105	10	)	)	PUNCT
cril-1051	105	11	which	which	PRON
cril-1051	105	12	has	have	AUX
cril-1051	105	13	used	use	VERB
cril-1051	105	14	igt	igt	PROPN
cril-1051	105	15	to	to	PART
cril-1051	105	16	automatically	automatically	ADV
cril-1051	105	17	infer	infer	VERB
cril-1051	105	18	grammatical	grammatical	ADJ
cril-1051	105	19	structure	structure	NOUN
cril-1051	105	20	for	for	ADP
cril-1051	105	21	multiple	multiple	ADJ
cril-1051	105	22	languages	language	NOUN
cril-1051	105	23	(	(	PUNCT
cril-1051	105	24	lepp	lepp	PROPN
cril-1051	105	25	,	,	PUNCT
cril-1051	105	26	zamaraeva	zamaraeva	PROPN
cril-1051	105	27	&	&	CCONJ
cril-1051	105	28	bender	bender	PROPN
cril-1051	105	29	2019	2019	NUM
cril-1051	105	30	;	;	PUNCT
cril-1051	105	31	wax	wax	NOUN
cril-1051	105	32	2014	2014	NUM
cril-1051	105	33	)	)	PUNCT
cril-1051	105	34	.	.	PUNCT
cril-1051	106	1	much	much	ADJ
cril-1051	106	2	of	of	ADP
cril-1051	106	3	their	their	PRON
cril-1051	106	4	data	datum	NOUN
cril-1051	106	5	comes	come	VERB
cril-1051	106	6	from	from	ADP
cril-1051	106	7	colorado	colorado	PROPN
cril-1051	106	8	research	research	PROPN
cril-1051	106	9	in	in	ADP
cril-1051	106	10	linguistics	linguistic	NOUN
cril-1051	106	11	,	,	PUNCT
cril-1051	106	12	volume	volume	NOUN
cril-1051	106	13	25	25	NUM
cril-1051	106	14	(	(	PUNCT
cril-1051	106	15	2021	2021	NUM
cril-1051	106	16	)	)	PUNCT
cril-1051	106	17	10	10	NUM
cril-1051	106	18	the	the	DET
cril-1051	106	19	online	online	ADJ
cril-1051	106	20	database	database	NOUN
cril-1051	106	21	of	of	ADP
cril-1051	106	22	interlinear	interlinear	ADJ
cril-1051	106	23	text	text	NOUN
cril-1051	106	24	(	(	PUNCT
cril-1051	106	25	lewis	lewis	PROPN
cril-1051	106	26	&	&	CCONJ
cril-1051	106	27	xia	xia	PROPN
cril-1051	106	28	2010	2010	NUM
cril-1051	106	29	,	,	PUNCT
cril-1051	106	30	odin	odin	NOUN
cril-1051	106	31	)	)	PUNCT
cril-1051	106	32	which	which	PRON
cril-1051	106	33	is	be	AUX
cril-1051	106	34	a	a	DET
cril-1051	106	35	collection	collection	NOUN
cril-1051	106	36	extracted	extract	VERB
cril-1051	106	37	from	from	ADP
cril-1051	106	38	published	publish	VERB
cril-1051	106	39	linguistic	linguistic	ADJ
cril-1051	106	40	articles	article	NOUN
cril-1051	106	41	or	or	CCONJ
cril-1051	106	42	books	book	NOUN
cril-1051	106	43	.	.	PUNCT
cril-1051	107	1	these	these	DET
cril-1051	107	2	igt	igt	PROPN
cril-1051	107	3	excerpts	excerpt	NOUN
cril-1051	107	4	differ	differ	VERB
cril-1051	107	5	from	from	ADP
cril-1051	107	6	igts	igts	NOUN
cril-1051	107	7	produced	produce	VERB
cril-1051	107	8	by	by	ADP
cril-1051	107	9	field	field	NOUN
cril-1051	107	10	linguists	linguist	NOUN
cril-1051	107	11	in	in	ADP
cril-1051	107	12	at	at	ADV
cril-1051	107	13	least	least	ADV
cril-1051	107	14	one	one	NUM
cril-1051	107	15	important	important	ADJ
cril-1051	107	16	way	way	NOUN
cril-1051	107	17	.	.	PUNCT
cril-1051	108	1	noise	noise	NOUN
cril-1051	108	2	(	(	PUNCT
cril-1051	108	3	i.e.	i.e.	X
cril-1051	108	4	,	,	PUNCT
cril-1051	108	5	typos	typo	NOUN
cril-1051	108	6	,	,	PUNCT
cril-1051	108	7	inconsistencies	inconsistency	NOUN
cril-1051	108	8	,	,	PUNCT
cril-1051	108	9	etc	etc	X
cril-1051	108	10	.	.	X
cril-1051	108	11	)	)	PUNCT
cril-1051	108	12	is	be	AUX
cril-1051	108	13	generally	generally	ADV
cril-1051	108	14	removed	remove	VERB
cril-1051	108	15	before	before	ADP
cril-1051	108	16	publication	publication	NOUN
cril-1051	108	17	,	,	PUNCT
cril-1051	108	18	so	so	SCONJ
cril-1051	108	19	that	that	SCONJ
cril-1051	108	20	odin	odin	NOUN
cril-1051	108	21	does	do	AUX
cril-1051	108	22	not	not	PART
cril-1051	108	23	have	have	VERB
cril-1051	108	24	the	the	DET
cril-1051	108	25	level	level	NOUN
cril-1051	108	26	of	of	ADP
cril-1051	108	27	noise	noise	NOUN
cril-1051	108	28	that	that	SCONJ
cril-1051	108	29	field	field	NOUN
cril-1051	108	30	igt	igt	PROPN
cril-1051	108	31	does	do	AUX
cril-1051	108	32	which	which	PRON
cril-1051	108	33	simplifies	simplifie	NOUN
cril-1051	108	34	pre	pre	ADJ
cril-1051	108	35	-	-	ADJ
cril-1051	108	36	processing	process	VERB
cril-1051	108	37	and	and	CCONJ
cril-1051	108	38	does	do	AUX
cril-1051	108	39	not	not	PART
cril-1051	108	40	distract	distract	VERB
cril-1051	108	41	machine	machine	NOUN
cril-1051	108	42	learning	learning	NOUN
cril-1051	108	43	models	model	NOUN
cril-1051	108	44	with	with	ADP
cril-1051	108	45	spurious	spurious	ADJ
cril-1051	108	46	patterns	pattern	NOUN
cril-1051	108	47	.	.	PUNCT
cril-1051	109	1	4	4	X
cril-1051	109	2	.	.	X
cril-1051	109	3	morphological	morphological	ADJ
cril-1051	109	4	analysis	analysis	NOUN
cril-1051	109	5	morphological	morphological	ADJ
cril-1051	109	6	analysis	analysis	NOUN
cril-1051	109	7	is	be	AUX
cril-1051	109	8	a	a	DET
cril-1051	109	9	key	key	ADJ
cril-1051	109	10	activity	activity	NOUN
cril-1051	109	11	in	in	ADP
cril-1051	109	12	ldd	ldd	PROPN
cril-1051	109	13	.	.	PUNCT
cril-1051	110	1	morphological	morphological	ADJ
cril-1051	110	2	analysis	analysis	NOUN
cril-1051	110	3	is	be	AUX
cril-1051	110	4	particularly	particularly	ADV
cril-1051	110	5	important	important	ADJ
cril-1051	110	6	when	when	SCONJ
cril-1051	110	7	working	work	VERB
cril-1051	110	8	with	with	ADP
cril-1051	110	9	morphologically	morphologically	ADV
cril-1051	110	10	complex	complex	ADJ
cril-1051	110	11	languages	language	NOUN
cril-1051	110	12	.	.	PUNCT
cril-1051	111	1	languages	language	NOUN
cril-1051	111	2	that	that	PRON
cril-1051	111	3	build	build	VERB
cril-1051	111	4	words	word	NOUN
cril-1051	111	5	from	from	ADP
cril-1051	111	6	multiple	multiple	ADJ
cril-1051	111	7	morphemes	morpheme	NOUN
cril-1051	111	8	or	or	CCONJ
cril-1051	111	9	via	via	ADP
cril-1051	111	10	significant	significant	ADJ
cril-1051	111	11	morphophonological	morphophonological	ADJ
cril-1051	111	12	changes	change	NOUN
cril-1051	111	13	produce	produce	VERB
cril-1051	111	14	a	a	DET
cril-1051	111	15	high	high	ADJ
cril-1051	111	16	number	number	NOUN
cril-1051	111	17	of	of	ADP
cril-1051	111	18	inflected	inflected	ADJ
cril-1051	111	19	and	and	CCONJ
cril-1051	111	20	compound	compound	NOUN
cril-1051	111	21	words	word	NOUN
cril-1051	111	22	which	which	PRON
cril-1051	111	23	appear	appear	VERB
cril-1051	111	24	to	to	ADP
cril-1051	111	25	the	the	DET
cril-1051	111	26	machine	machine	NOUN
cril-1051	111	27	as	as	ADP
cril-1051	111	28	brand	brand	NOUN
cril-1051	111	29	new	new	ADJ
cril-1051	111	30	,	,	PUNCT
cril-1051	111	31	unrelated	unrelated	ADJ
cril-1051	111	32	words	word	NOUN
cril-1051	111	33	(	(	PUNCT
cril-1051	111	34	dreyer	dreyer	PROPN
cril-1051	111	35	&	&	CCONJ
cril-1051	111	36	eisner	eisner	PROPN
cril-1051	111	37	2011	2011	NUM
cril-1051	111	38	;	;	PUNCT
cril-1051	111	39	goldsmith	goldsmith	PROPN
cril-1051	111	40	,	,	PUNCT
cril-1051	111	41	lee	lee	PROPN
cril-1051	111	42	&	&	CCONJ
cril-1051	111	43	xanthos	xanthos	PROPN
cril-1051	111	44	2017	2017	NUM
cril-1051	111	45	;	;	PUNCT
cril-1051	111	46	hammarström	hammarström	NUM
cril-1051	111	47	&	&	CCONJ
cril-1051	111	48	borin	borin	ADJ
cril-1051	111	49	2011	2011	NUM
cril-1051	111	50	;	;	PUNCT
cril-1051	111	51	kann	kann	PROPN
cril-1051	111	52	,	,	PUNCT
cril-1051	111	53	cotterell	cotterell	PROPN
cril-1051	111	54	&	&	CCONJ
cril-1051	111	55	schütze	schütze	ADJ
cril-1051	111	56	2016	2016	NUM
cril-1051	111	57	;	;	PUNCT
cril-1051	111	58	ruokolainen	ruokolainen	PROPN
cril-1051	111	59	et	et	PROPN
cril-1051	111	60	al	al	PROPN
cril-1051	111	61	.	.	PROPN
cril-1051	111	62	2013	2013	NUM
cril-1051	111	63	)	)	PUNCT
cril-1051	111	64	.	.	PUNCT
cril-1051	112	1	nlp	nlp	PROPN
cril-1051	112	2	systems	system	NOUN
cril-1051	112	3	that	that	PRON
cril-1051	112	4	account	account	VERB
cril-1051	112	5	for	for	ADP
cril-1051	112	6	morphology	morphology	NOUN
cril-1051	112	7	can	can	AUX
cril-1051	112	8	reduce	reduce	VERB
cril-1051	112	9	data	data	NOUN
cril-1051	112	10	sparsity	sparsity	NOUN
cril-1051	112	11	caused	cause	VERB
cril-1051	112	12	by	by	ADP
cril-1051	112	13	an	an	DET
cril-1051	112	14	abundance	abundance	NOUN
cril-1051	112	15	of	of	ADP
cril-1051	112	16	individual	individual	ADJ
cril-1051	112	17	word	word	NOUN
cril-1051	112	18	forms	form	NOUN
cril-1051	112	19	(	(	PUNCT
cril-1051	112	20	mccarthy	mccarthy	PROPN
cril-1051	112	21	et	et	PROPN
cril-1051	112	22	al	al	PROPN
cril-1051	112	23	.	.	PROPN
cril-1051	112	24	2019	2019	NUM
cril-1051	112	25	;	;	PUNCT
cril-1051	112	26	vylomova	vylomova	PROPN
cril-1051	112	27	et	et	PROPN
cril-1051	112	28	al	al	PROPN
cril-1051	112	29	.	.	PROPN
cril-1051	112	30	2020	2020	NUM
cril-1051	112	31	)	)	PUNCT
cril-1051	112	32	and	and	CCONJ
cril-1051	112	33	help	help	VERB
cril-1051	112	34	mitigate	mitigate	VERB
cril-1051	112	35	bias	bias	NOUN
cril-1051	112	36	in	in	ADP
cril-1051	112	37	training	training	NOUN
cril-1051	112	38	data	datum	NOUN
cril-1051	112	39	(	(	PUNCT
cril-1051	112	40	zmigrod	zmigrod	NOUN
cril-1051	112	41	et	et	PROPN
cril-1051	112	42	al	al	PROPN
cril-1051	112	43	.	.	PROPN
cril-1051	112	44	2019	2019	NUM
cril-1051	112	45	)	)	PUNCT
cril-1051	112	46	.	.	PUNCT
cril-1051	113	1	computational	computational	ADJ
cril-1051	113	2	morphological	morphological	ADJ
cril-1051	113	3	systems	system	NOUN
cril-1051	113	4	have	have	AUX
cril-1051	113	5	often	often	ADV
cril-1051	113	6	been	be	AUX
cril-1051	113	7	limited	limit	VERB
cril-1051	113	8	to	to	ADP
cril-1051	113	9	languages	language	NOUN
cril-1051	113	10	with	with	ADP
cril-1051	113	11	publicly	publicly	ADV
cril-1051	113	12	available	available	ADJ
cril-1051	113	13	structured	structured	ADJ
cril-1051	113	14	data	datum	NOUN
cril-1051	113	15	,	,	PUNCT
cril-1051	113	16	for	for	ADP
cril-1051	113	17	example	example	NOUN
cril-1051	113	18	,	,	PUNCT
cril-1051	113	19	tables	table	NOUN
cril-1051	113	20	of	of	ADP
cril-1051	113	21	inflectional	inflectional	ADJ
cril-1051	113	22	patterns	pattern	NOUN
cril-1051	113	23	in	in	ADP
cril-1051	113	24	online	online	ADJ
cril-1051	113	25	dictionaries	dictionary	NOUN
cril-1051	113	26	like	like	ADP
cril-1051	113	27	wiktionary	wiktionary	ADJ
cril-1051	113	28	.	.	PUNCT
cril-1051	114	1	unfortunately	unfortunately	ADV
cril-1051	114	2	,	,	PUNCT
cril-1051	114	3	complete	complete	ADJ
cril-1051	114	4	inflectional	inflectional	ADJ
cril-1051	114	5	tables	table	NOUN
cril-1051	114	6	are	be	AUX
cril-1051	114	7	not	not	PART
cril-1051	114	8	easily	easily	ADV
cril-1051	114	9	available	available	ADJ
cril-1051	114	10	for	for	ADP
cril-1051	114	11	many	many	ADJ
cril-1051	114	12	of	of	ADP
cril-1051	114	13	the	the	DET
cril-1051	114	14	world	world	NOUN
cril-1051	114	15	’s	’s	PART
cril-1051	114	16	languages	language	NOUN
cril-1051	114	17	.	.	PUNCT
cril-1051	115	1	morphological	morphological	ADJ
cril-1051	115	2	analysis	analysis	NOUN
cril-1051	115	3	can	can	AUX
cril-1051	115	4	be	be	AUX
cril-1051	115	5	separated	separate	VERB
cril-1051	115	6	into	into	ADP
cril-1051	115	7	two	two	NUM
cril-1051	115	8	core	core	NOUN
cril-1051	115	9	tasks	task	NOUN
cril-1051	115	10	(	(	PUNCT
cril-1051	115	11	cotterell	cotterell	NOUN
cril-1051	115	12	et	et	PROPN
cril-1051	115	13	al	al	PROPN
cril-1051	115	14	.	.	PROPN
cril-1051	115	15	2015	2015	NUM
cril-1051	115	16	;	;	PUNCT
cril-1051	115	17	hammarström	hammarström	NOUN
cril-1051	115	18	&	&	CCONJ
cril-1051	115	19	borin	borin	NOUN
cril-1051	115	20	2011	2011	NUM
cril-1051	115	21	;	;	PUNCT
cril-1051	115	22	nicolai	nicolai	PROPN
cril-1051	115	23	&	&	CCONJ
cril-1051	115	24	kondrak	kondrak	PROPN
cril-1051	115	25	2017	2017	NUM
cril-1051	115	26	;	;	PUNCT
cril-1051	115	27	palmer	palmer	PROPN
cril-1051	115	28	2009	2009	NUM
cril-1051	115	29	)	)	PUNCT
cril-1051	115	30	.	.	PUNCT
cril-1051	116	1	the	the	DET
cril-1051	116	2	first	first	ADJ
cril-1051	116	3	task	task	NOUN
cril-1051	116	4	is	be	AUX
cril-1051	116	5	identifying	identify	VERB
cril-1051	116	6	morphemes	morpheme	NOUN
cril-1051	116	7	by	by	ADP
cril-1051	116	8	determining	determine	VERB
cril-1051	116	9	their	their	PRON
cril-1051	116	10	shapes	shape	NOUN
cril-1051	116	11	and	and	CCONJ
cril-1051	116	12	marking	mark	VERB
cril-1051	116	13	boundaries	boundary	NOUN
cril-1051	116	14	between	between	ADP
cril-1051	116	15	them	they	PRON
cril-1051	116	16	,	,	PUNCT
cril-1051	116	17	as	as	SCONJ
cril-1051	116	18	was	be	AUX
cril-1051	116	19	done	do	VERB
cril-1051	116	20	for	for	ADP
cril-1051	116	21	the	the	DET
cril-1051	116	22	lezgi	lezgi	ADJ
cril-1051	116	23	noun	noun	NOUN
cril-1051	116	24	in	in	ADP
cril-1051	116	25	example	example	NOUN
cril-1051	116	26	1b	1b	NUM
cril-1051	116	27	below	below	ADV
cril-1051	116	28	.	.	PUNCT
cril-1051	117	1	this	this	PRON
cril-1051	117	2	is	be	AUX
cril-1051	117	3	known	know	VERB
cril-1051	117	4	as	as	ADP
cril-1051	117	5	(	(	PUNCT
cril-1051	117	6	unlabeled	unlabeled	ADJ
cril-1051	117	7	)	)	PUNCT
cril-1051	117	8	morpheme	morpheme	NOUN
cril-1051	117	9	segmentation	segmentation	NOUN
cril-1051	117	10	(	(	PUNCT
cril-1051	117	11	creutz	creutz	ADJ
cril-1051	117	12	&	&	CCONJ
cril-1051	117	13	lagus	lagus	ADJ
cril-1051	117	14	2007	2007	NUM
cril-1051	117	15	;	;	PUNCT
cril-1051	117	16	snyder	snyder	PROPN
cril-1051	117	17	&	&	CCONJ
cril-1051	117	18	barzilay	barzilay	PROPN
cril-1051	117	19	2008	2008	NUM
cril-1051	117	20	)	)	PUNCT
cril-1051	117	21	.	.	PUNCT
cril-1051	118	1	the	the	DET
cril-1051	118	2	second	second	ADJ
cril-1051	118	3	task	task	NOUN
cril-1051	118	4	is	be	AUX
cril-1051	118	5	deducing	deduce	VERB
cril-1051	118	6	each	each	DET
cril-1051	118	7	morpheme	morpheme	NOUN
cril-1051	118	8	’s	’s	PART
cril-1051	118	9	meaning	meaning	NOUN
cril-1051	118	10	,	,	PUNCT
cril-1051	118	11	which	which	PRON
cril-1051	118	12	is	be	AUX
cril-1051	118	13	known	know	VERB
cril-1051	118	14	as	as	ADP
cril-1051	118	15	parsing	parse	VERB
cril-1051	118	16	,	,	PUNCT
cril-1051	118	17	or	or	CCONJ
cril-1051	118	18	sometimes	sometimes	ADV
cril-1051	118	19	called	call	VERB
cril-1051	118	20	morphological	morphological	ADJ
cril-1051	118	21	analysis	analysis	NOUN
cril-1051	118	22	by	by	ADP
cril-1051	118	23	itself.3	itself.3	PROPN
cril-1051	118	24	this	this	DET
cril-1051	118	25	single	single	ADJ
cril-1051	118	26	step	step	NOUN
cril-1051	118	27	is	be	AUX
cril-1051	118	28	known	know	VERB
cril-1051	118	29	in	in	ADP
cril-1051	118	30	linguistics	linguistic	NOUN
cril-1051	118	31	as	as	ADP
cril-1051	118	32	glossing	gloss	VERB
cril-1051	118	33	,	,	PUNCT
cril-1051	118	34	and	and	CCONJ
cril-1051	118	35	in	in	ADP
cril-1051	118	36	computational	computational	ADJ
cril-1051	118	37	linguists	linguist	NOUN
cril-1051	118	38	as	as	ADP
cril-1051	118	39	labeled	label	VERB
cril-1051	118	40	morpheme	morpheme	ADJ
cril-1051	118	41	segmentation	segmentation	NOUN
cril-1051	118	42	or	or	CCONJ
cril-1051	118	43	,	,	PUNCT
cril-1051	118	44	merely	merely	ADV
cril-1051	118	45	,	,	PUNCT
cril-1051	118	46	labeling	labeling	NOUN
cril-1051	118	47	,	,	PUNCT
cril-1051	118	48	or	or	CCONJ
cril-1051	118	49	tagging	tagging	NOUN
cril-1051	118	50	.	.	PUNCT
cril-1051	119	1	together	together	ADV
cril-1051	119	2	segmentation	segmentation	NOUN
cril-1051	119	3	and	and	CCONJ
cril-1051	119	4	glossing	gloss	VERB
cril-1051	119	5	make	make	VERB
cril-1051	119	6	up	up	ADP
cril-1051	119	7	a	a	DET
cril-1051	119	8	significant	significant	ADJ
cril-1051	119	9	part	part	NOUN
cril-1051	119	10	of	of	ADP
cril-1051	119	11	interlinearization	interlinearization	NOUN
cril-1051	119	12	in	in	ADP
cril-1051	119	13	documentary	documentary	NOUN
cril-1051	119	14	and	and	CCONJ
cril-1051	119	15	descriptive	descriptive	ADJ
cril-1051	119	16	linguistics	linguistic	NOUN
cril-1051	119	17	.	.	PUNCT
cril-1051	120	1	these	these	DET
cril-1051	120	2	two	two	NUM
cril-1051	120	3	tasks	task	NOUN
cril-1051	120	4	(	(	PUNCT
cril-1051	120	5	step	step	NOUN
cril-1051	120	6	3	3	NUM
cril-1051	120	7	of	of	ADP
cril-1051	120	8	bird	bird	NOUN
cril-1051	120	9	and	and	CCONJ
cril-1051	120	10	chiang	chiang	PROPN
cril-1051	120	11	’s	’s	PART
cril-1051	120	12	workflow	workflow	NOUN
cril-1051	120	13	on	on	ADP
cril-1051	120	14	)	)	PUNCT
cril-1051	120	15	are	be	AUX
cril-1051	120	16	often	often	ADV
cril-1051	120	17	the	the	DET
cril-1051	120	18	most	most	ADV
cril-1051	120	19	detailed	detailed	ADJ
cril-1051	120	20	analytical	analytical	ADJ
cril-1051	120	21	tasks	task	NOUN
cril-1051	120	22	undertaken	undertake	VERB
cril-1051	120	23	while	while	SCONJ
cril-1051	120	24	still	still	ADV
cril-1051	120	25	in	in	ADP
cril-1051	120	26	the	the	DET
cril-1051	120	27	field	field	NOUN
cril-1051	120	28	.	.	PUNCT
cril-1051	121	1	they	they	PRON
cril-1051	121	2	are	be	AUX
cril-1051	121	3	also	also	ADV
cril-1051	121	4	computational	computational	ADJ
cril-1051	121	5	morphology	morphology	NOUN
cril-1051	121	6	for	for	ADP
cril-1051	121	7	language	language	NOUN
cril-1051	121	8	description	description	NOUN
cril-1051	121	9	and	and	CCONJ
cril-1051	121	10	documentation	documentation	NOUN
cril-1051	121	11	11	11	NUM
cril-1051	121	12	perhaps	perhaps	ADV
cril-1051	121	13	the	the	DET
cril-1051	121	14	most	most	ADJ
cril-1051	121	15	time	time	NOUN
cril-1051	121	16	-	-	PUNCT
cril-1051	121	17	consuming	consume	VERB
cril-1051	121	18	tasks	task	NOUN
cril-1051	121	19	,	,	PUNCT
cril-1051	121	20	requiring	require	VERB
cril-1051	121	21	at	at	ADV
cril-1051	121	22	least	least	ADJ
cril-1051	121	23	as	as	ADV
cril-1051	121	24	much	much	ADJ
cril-1051	121	25	,	,	PUNCT
cril-1051	121	26	and	and	CCONJ
cril-1051	121	27	probably	probably	ADV
cril-1051	121	28	more	more	ADJ
cril-1051	121	29	,	,	PUNCT
cril-1051	121	30	time	time	NOUN
cril-1051	121	31	than	than	ADP
cril-1051	121	32	transcription	transcription	NOUN
cril-1051	121	33	which	which	PRON
cril-1051	121	34	can	can	AUX
cril-1051	121	35	take	take	VERB
cril-1051	121	36	up	up	ADP
cril-1051	121	37	to	to	PART
cril-1051	121	38	100	100	NUM
cril-1051	121	39	hours	hour	NOUN
cril-1051	121	40	for	for	ADP
cril-1051	121	41	each	each	DET
cril-1051	121	42	hour	hour	NOUN
cril-1051	121	43	of	of	ADP
cril-1051	121	44	recorded	record	VERB
cril-1051	121	45	speech	speech	NOUN
cril-1051	121	46	.	.	PUNCT
cril-1051	122	1	the	the	DET
cril-1051	122	2	linguistic	linguistic	ADJ
cril-1051	122	3	information	information	NOUN
cril-1051	122	4	provided	provide	VERB
cril-1051	122	5	by	by	ADP
cril-1051	122	6	morpheme	morpheme	ADJ
cril-1051	122	7	segments	segment	NOUN
cril-1051	122	8	and	and	CCONJ
cril-1051	122	9	glosses	gloss	NOUN
cril-1051	122	10	lays	lay	VERB
cril-1051	122	11	a	a	DET
cril-1051	122	12	vital	vital	ADJ
cril-1051	122	13	foundation	foundation	NOUN
cril-1051	122	14	for	for	ADP
cril-1051	122	15	subsequent	subsequent	ADJ
cril-1051	122	16	descriptive	descriptive	ADJ
cril-1051	122	17	work	work	NOUN
cril-1051	122	18	.	.	PUNCT
cril-1051	123	1	many	many	ADJ
cril-1051	123	2	nlp	nlp	NOUN
cril-1051	123	3	models	model	NOUN
cril-1051	123	4	have	have	AUX
cril-1051	123	5	been	be	AUX
cril-1051	123	6	applied	apply	VERB
cril-1051	123	7	to	to	ADP
cril-1051	123	8	morpheme	morpheme	ADJ
cril-1051	123	9	segmentation	segmentation	NOUN
cril-1051	123	10	and	and	CCONJ
cril-1051	123	11	glossing	gloss	VERB
cril-1051	123	12	.	.	PUNCT
cril-1051	124	1	automatic	automatic	ADJ
cril-1051	124	2	morpheme	morpheme	NOUN
cril-1051	124	3	segmentation	segmentation	NOUN
cril-1051	124	4	is	be	AUX
cril-1051	124	5	commonly	commonly	ADV
cril-1051	124	6	traced	trace	VERB
cril-1051	124	7	to	to	ADP
cril-1051	124	8	the	the	DET
cril-1051	124	9	early	early	ADJ
cril-1051	124	10	work	work	NOUN
cril-1051	124	11	of	of	ADP
cril-1051	124	12	harris	harris	PROPN
cril-1051	124	13	(	(	PUNCT
cril-1051	124	14	1955	1955	NUM
cril-1051	124	15	)	)	PUNCT
cril-1051	124	16	and	and	CCONJ
cril-1051	124	17	much	much	ADJ
cril-1051	124	18	segmentation	segmentation	NOUN
cril-1051	124	19	research	research	NOUN
cril-1051	124	20	since	since	SCONJ
cril-1051	124	21	then	then	ADV
cril-1051	124	22	has	have	AUX
cril-1051	124	23	implemented	implement	VERB
cril-1051	124	24	unsupervised	unsupervised	ADJ
cril-1051	124	25	learning	learning	NOUN
cril-1051	124	26	which	which	PRON
cril-1051	124	27	he	he	PRON
cril-1051	124	28	inspired	inspire	VERB
cril-1051	124	29	(	(	PUNCT
cril-1051	124	30	goldsmith	goldsmith	PROPN
cril-1051	124	31	2001	2001	NUM
cril-1051	124	32	;	;	PUNCT
cril-1051	124	33	creutz	creutz	ADJ
cril-1051	124	34	&	&	CCONJ
cril-1051	124	35	lagus	lagus	PROPN
cril-1051	124	36	2002	2002	NUM
cril-1051	124	37	;	;	PUNCT
cril-1051	124	38	poon	poon	NOUN
cril-1051	124	39	,	,	PUNCT
cril-1051	124	40	cherry	cherry	NOUN
cril-1051	124	41	&	&	CCONJ
cril-1051	124	42	toutanova	toutanova	PROPN
cril-1051	124	43	2009	2009	NUM
cril-1051	124	44	)	)	PUNCT
cril-1051	124	45	.	.	PUNCT
cril-1051	125	1	the	the	DET
cril-1051	125	2	preponderance	preponderance	NOUN
cril-1051	125	3	of	of	ADP
cril-1051	125	4	unsupervised	unsupervised	ADJ
cril-1051	125	5	models	model	NOUN
cril-1051	125	6	was	be	AUX
cril-1051	125	7	probably	probably	ADV
cril-1051	125	8	motivated	motivate	VERB
cril-1051	125	9	by	by	ADP
cril-1051	125	10	the	the	DET
cril-1051	125	11	difficulty	difficulty	NOUN
cril-1051	125	12	of	of	ADP
cril-1051	125	13	finding	find	VERB
cril-1051	125	14	the	the	DET
cril-1051	125	15	high	high	ADJ
cril-1051	125	16	quantity	quantity	NOUN
cril-1051	125	17	and	and	CCONJ
cril-1051	125	18	quality	quality	NOUN
cril-1051	125	19	manually	manually	ADV
cril-1051	125	20	segmented	segment	VERB
cril-1051	125	21	data	datum	NOUN
cril-1051	125	22	needed	need	VERB
cril-1051	125	23	to	to	PART
cril-1051	125	24	train	train	VERB
cril-1051	125	25	supervised	supervised	ADJ
cril-1051	125	26	models	model	NOUN
cril-1051	125	27	.	.	PUNCT
cril-1051	126	1	lack	lack	NOUN
cril-1051	126	2	of	of	ADP
cril-1051	126	3	sufficient	sufficient	ADJ
cril-1051	126	4	training	training	NOUN
cril-1051	126	5	data	datum	NOUN
cril-1051	126	6	is	be	AUX
cril-1051	126	7	illustrated	illustrate	VERB
cril-1051	126	8	by	by	ADP
cril-1051	126	9	a	a	DET
cril-1051	126	10	recent	recent	ADJ
cril-1051	126	11	supervised	supervised	ADJ
cril-1051	126	12	segmentation	segmentation	NOUN
cril-1051	126	13	experiment	experiment	NOUN
cril-1051	126	14	(	(	PUNCT
cril-1051	126	15	ansari	ansari	PROPN
cril-1051	126	16	et	et	PROPN
cril-1051	126	17	al	al	PROPN
cril-1051	126	18	.	.	PROPN
cril-1051	126	19	2019	2019	NUM
cril-1051	126	20	)	)	PUNCT
cril-1051	126	21	which	which	PRON
cril-1051	126	22	needed	need	VERB
cril-1051	126	23	to	to	PART
cril-1051	126	24	manually	manually	ADV
cril-1051	126	25	segment	segment	VERB
cril-1051	126	26	a	a	DET
cril-1051	126	27	corpus	corpus	NOUN
cril-1051	126	28	before	before	ADP
cril-1051	126	29	conducting	conduct	VERB
cril-1051	126	30	the	the	DET
cril-1051	126	31	experiment	experiment	NOUN
cril-1051	126	32	.	.	PUNCT
cril-1051	127	1	in	in	ADP
cril-1051	127	2	ldd	ldd	PROPN
cril-1051	127	3	,	,	PUNCT
cril-1051	127	4	segmentation	segmentation	NOUN
cril-1051	127	5	and	and	CCONJ
cril-1051	127	6	glossing	glossing	NOUN
cril-1051	127	7	are	be	AUX
cril-1051	127	8	typically	typically	ADV
cril-1051	127	9	tackled	tackle	VERB
cril-1051	127	10	simultaneously	simultaneously	ADV
cril-1051	127	11	.	.	PUNCT
cril-1051	128	1	segmentation	segmentation	NOUN
cril-1051	128	2	finds	find	VERB
cril-1051	128	3	breaks	break	NOUN
cril-1051	128	4	between	between	ADP
cril-1051	128	5	morphemes	morpheme	NOUN
cril-1051	128	6	as	as	ADP
cril-1051	128	7	for	for	ADP
cril-1051	128	8	the	the	DET
cril-1051	128	9	lezgi	lezgi	ADJ
cril-1051	128	10	noun	noun	NOUN
cril-1051	128	11	in	in	ADP
cril-1051	128	12	1b	1b	NUM
cril-1051	128	13	.	.	PUNCT
cril-1051	129	1	glossing	gloss	VERB
cril-1051	129	2	labels	label	NOUN
cril-1051	129	3	morphemes	morpheme	NOUN
cril-1051	129	4	with	with	ADP
cril-1051	129	5	their	their	PRON
cril-1051	129	6	meaning	meaning	NOUN
cril-1051	129	7	or	or	CCONJ
cril-1051	129	8	function	function	NOUN
cril-1051	129	9	,	,	PUNCT
cril-1051	129	10	as	as	ADP
cril-1051	129	11	in	in	ADP
cril-1051	129	12	1c	1c	NUM
cril-1051	129	13	.	.	PUNCT
cril-1051	130	1	glossing	gloss	VERB
cril-1051	130	2	does	do	AUX
cril-1051	130	3	not	not	PART
cril-1051	130	4	require	require	VERB
cril-1051	130	5	segmentation	segmentation	NOUN
cril-1051	130	6	,	,	PUNCT
cril-1051	130	7	and	and	CCONJ
cril-1051	130	8	if	if	SCONJ
cril-1051	130	9	done	do	VERB
cril-1051	130	10	independently	independently	ADV
cril-1051	130	11	is	be	AUX
cril-1051	130	12	sometimes	sometimes	ADV
cril-1051	130	13	referred	refer	VERB
cril-1051	130	14	to	to	ADP
cril-1051	130	15	as	as	ADP
cril-1051	130	16	parsing	parse	VERB
cril-1051	130	17	.	.	PUNCT
cril-1051	131	1	parsing	parse	VERB
cril-1051	131	2	by	by	ADP
cril-1051	131	3	itself	itself	PRON
cril-1051	131	4	would	would	AUX
cril-1051	131	5	only	only	ADV
cril-1051	131	6	provide	provide	VERB
cril-1051	131	7	the	the	DET
cril-1051	131	8	information	information	NOUN
cril-1051	131	9	in	in	ADP
cril-1051	131	10	1c	1c	NUM
cril-1051	131	11	without	without	ADP
cril-1051	131	12	indication	indication	NOUN
cril-1051	131	13	of	of	ADP
cril-1051	131	14	morpheme	morpheme	ADJ
cril-1051	131	15	boundaries	boundary	NOUN
cril-1051	131	16	.	.	PUNCT
cril-1051	132	1	(	(	PUNCT
cril-1051	132	2	1	1	X
cril-1051	132	3	)	)	PUNCT
cril-1051	132	4	a.	a.	NOUN
cril-1051	132	5	paçahdin	paçahdin	PROPN
cril-1051	132	6	b.	b.	PROPN
cril-1051	132	7	paçah	paçah	PROPN
cril-1051	132	8	-	-	PUNCT
cril-1051	132	9	di	di	NOUN
cril-1051	132	10	-	-	PUNCT
cril-1051	132	11	n	n	CCONJ
cril-1051	132	12	c.	c.	NOUN
cril-1051	132	13	king	king	PROPN
cril-1051	132	14	-	-	PUNCT
cril-1051	132	15	obl	obl	PROPN
cril-1051	132	16	-	-	PUNCT
cril-1051	132	17	gen	gen	PROPN
cril-1051	132	18	d.	d.	PROPN
cril-1051	132	19	‘	'	PUNCT
cril-1051	132	20	king	king	PROPN
cril-1051	132	21	’s	’s	PART
cril-1051	132	22	’	'	PUNCT
cril-1051	132	23	nlp	nlp	ADJ
cril-1051	132	24	experiments	experiment	NOUN
cril-1051	132	25	with	with	ADP
cril-1051	132	26	lrl	lrl	PROPN
cril-1051	132	27	often	often	ADV
cril-1051	132	28	treat	treat	VERB
cril-1051	132	29	segmentation	segmentation	NOUN
cril-1051	132	30	and	and	CCONJ
cril-1051	132	31	glossing	gloss	VERB
cril-1051	132	32	as	as	ADP
cril-1051	132	33	separate	separate	ADJ
cril-1051	132	34	tasks	task	NOUN
cril-1051	132	35	.	.	PUNCT
cril-1051	133	1	other	other	ADJ
cril-1051	133	2	nlp	nlp	NOUN
cril-1051	133	3	works	work	NOUN
cril-1051	133	4	have	have	AUX
cril-1051	133	5	taken	take	VERB
cril-1051	133	6	a	a	DET
cril-1051	133	7	tip	tip	NOUN
cril-1051	133	8	from	from	ADP
cril-1051	133	9	ldd	ldd	PROPN
cril-1051	133	10	and	and	CCONJ
cril-1051	133	11	joined	join	VERB
cril-1051	133	12	the	the	DET
cril-1051	133	13	two	two	NUM
cril-1051	133	14	tasks	task	NOUN
cril-1051	133	15	.	.	PUNCT
cril-1051	134	1	joint	joint	ADJ
cril-1051	134	2	learning	learning	NOUN
cril-1051	134	3	of	of	ADP
cril-1051	134	4	segmentation	segmentation	NOUN
cril-1051	134	5	and	and	CCONJ
cril-1051	134	6	glossing	gloss	VERB
cril-1051	134	7	,	,	PUNCT
cril-1051	134	8	or	or	CCONJ
cril-1051	134	9	labeled	label	VERB
cril-1051	134	10	segmentation	segmentation	NOUN
cril-1051	134	11	,	,	PUNCT
cril-1051	134	12	is	be	AUX
cril-1051	134	13	less	less	ADV
cril-1051	134	14	common	common	ADJ
cril-1051	134	15	but	but	CCONJ
cril-1051	134	16	has	have	AUX
cril-1051	134	17	been	be	AUX
cril-1051	134	18	successful	successful	ADJ
cril-1051	134	19	for	for	ADP
cril-1051	134	20	lrl	lrl	PROPN
cril-1051	134	21	(	(	PUNCT
cril-1051	134	22	cotterell	cotterell	VERB
cril-1051	134	23	et	et	PROPN
cril-1051	134	24	al	al	PROPN
cril-1051	134	25	.	.	PROPN
cril-1051	134	26	,	,	PUNCT
cril-1051	134	27	2015	2015	NUM
cril-1051	134	28	;	;	PUNCT
cril-1051	134	29	moeller	moeller	PROPN
cril-1051	134	30	&	&	CCONJ
cril-1051	134	31	hulden	hulden	PROPN
cril-1051	134	32	,	,	PUNCT
cril-1051	134	33	2018	2018	NUM
cril-1051	134	34	)	)	PUNCT
cril-1051	134	35	.	.	PUNCT
cril-1051	135	1	in	in	ADP
cril-1051	135	2	general	general	ADJ
cril-1051	135	3	,	,	PUNCT
cril-1051	135	4	joint	joint	ADJ
cril-1051	135	5	learning	learning	NOUN
cril-1051	135	6	is	be	AUX
cril-1051	135	7	characterized	characterize	VERB
cril-1051	135	8	by	by	ADP
cril-1051	135	9	training	training	NOUN
cril-1051	135	10	on	on	ADP
cril-1051	135	11	different	different	ADJ
cril-1051	135	12	types	type	NOUN
cril-1051	135	13	of	of	ADP
cril-1051	135	14	information	information	NOUN
cril-1051	135	15	and	and	CCONJ
cril-1051	135	16	is	be	AUX
cril-1051	135	17	based	base	VERB
cril-1051	135	18	on	on	ADP
cril-1051	135	19	the	the	DET
cril-1051	135	20	intuition	intuition	NOUN
cril-1051	135	21	that	that	SCONJ
cril-1051	135	22	one	one	NUM
cril-1051	135	23	type	type	NOUN
cril-1051	135	24	of	of	ADP
cril-1051	135	25	linguistic	linguistic	ADJ
cril-1051	135	26	knowledge	knowledge	NOUN
cril-1051	135	27	(	(	PUNCT
cril-1051	135	28	e.g.	e.g.	ADV
cril-1051	135	29	,	,	PUNCT
cril-1051	135	30	syntax	syntax	NOUN
cril-1051	135	31	)	)	PUNCT
cril-1051	135	32	can	can	AUX
cril-1051	135	33	improve	improve	VERB
cril-1051	135	34	results	result	NOUN
cril-1051	135	35	in	in	ADP
cril-1051	135	36	another	another	DET
cril-1051	135	37	domain	domain	NOUN
cril-1051	135	38	(	(	PUNCT
cril-1051	135	39	e.g.	e.g.	ADV
cril-1051	135	40	,	,	PUNCT
cril-1051	135	41	morphology	morphology	NOUN
cril-1051	135	42	)	)	PUNCT
cril-1051	136	1	(	(	PUNCT
cril-1051	136	2	goldsmith	goldsmith	NOUN
cril-1051	136	3	et	et	PROPN
cril-1051	136	4	al	al	PROPN
cril-1051	136	5	.	.	PROPN
cril-1051	136	6	,	,	PUNCT
cril-1051	136	7	2017	2017	NUM
cril-1051	136	8	)	)	PUNCT
cril-1051	136	9	.	.	PUNCT
cril-1051	137	1	much	much	ADJ
cril-1051	137	2	nlp	nlp	ADJ
cril-1051	137	3	work	work	NOUN
cril-1051	137	4	focuses	focus	VERB
cril-1051	137	5	on	on	ADP
cril-1051	137	6	glossing	gloss	VERB
cril-1051	137	7	only	only	ADV
cril-1051	137	8	,	,	PUNCT
cril-1051	137	9	which	which	PRON
cril-1051	137	10	assumes	assume	VERB
cril-1051	137	11	that	that	SCONJ
cril-1051	137	12	the	the	DET
cril-1051	137	13	data	data	NOUN
cril-1051	137	14	is	be	AUX
cril-1051	137	15	already	already	ADV
cril-1051	137	16	is	be	AUX
cril-1051	137	17	,	,	PUNCT
cril-1051	137	18	or	or	CCONJ
cril-1051	137	19	does	do	AUX
cril-1051	137	20	not	not	PART
cril-1051	137	21	need	need	VERB
cril-1051	137	22	to	to	PART
cril-1051	137	23	be	be	AUX
cril-1051	137	24	,	,	PUNCT
cril-1051	137	25	segmented	segment	VERB
cril-1051	137	26	into	into	ADP
cril-1051	137	27	morphemes	morpheme	NOUN
cril-1051	137	28	.	.	PUNCT
cril-1051	138	1	mcmillan	mcmillan	NOUN
cril-1051	138	2	-	-	PUNCT
cril-1051	138	3	major	major	ADJ
cril-1051	138	4	(	(	PUNCT
cril-1051	138	5	2020	2020	NUM
cril-1051	138	6	)	)	PUNCT
cril-1051	138	7	trained	train	VERB
cril-1051	138	8	systems	system	NOUN
cril-1051	138	9	to	to	PART
cril-1051	138	10	produce	produce	VERB
cril-1051	138	11	a	a	DET
cril-1051	138	12	gloss	gloss	NOUN
cril-1051	138	13	colorado	colorado	PROPN
cril-1051	138	14	research	research	PROPN
cril-1051	138	15	in	in	ADP
cril-1051	138	16	linguistics	linguistic	NOUN
cril-1051	138	17	,	,	PUNCT
cril-1051	138	18	volume	volume	NOUN
cril-1051	138	19	25	25	NUM
cril-1051	138	20	(	(	PUNCT
cril-1051	138	21	2021	2021	NUM
cril-1051	138	22	)	)	PUNCT
cril-1051	138	23	12	12	NUM
cril-1051	138	24	line	line	NOUN
cril-1051	138	25	by	by	ADP
cril-1051	138	26	incorporating	incorporate	VERB
cril-1051	138	27	predictions	prediction	NOUN
cril-1051	138	28	made	make	VERB
cril-1051	138	29	from	from	ADP
cril-1051	138	30	existing	exist	VERB
cril-1051	138	31	segmentations	segmentation	NOUN
cril-1051	138	32	and	and	CCONJ
cril-1051	138	33	from	from	ADP
cril-1051	138	34	the	the	DET
cril-1051	138	35	free	free	ADJ
cril-1051	138	36	translation	translation	NOUN
cril-1051	138	37	enriched	enrich	VERB
cril-1051	138	38	with	with	ADP
cril-1051	138	39	intent	intent	NOUN
cril-1051	138	40	(	(	PUNCT
cril-1051	138	41	georgi	georgi	PROPN
cril-1051	138	42	2016	2016	NUM
cril-1051	138	43	)	)	PUNCT
cril-1051	138	44	.	.	PUNCT
cril-1051	139	1	both	both	DET
cril-1051	139	2	samardzic	samardzic	PROPN
cril-1051	139	3	et	et	PROPN
cril-1051	139	4	al	al	PROPN
cril-1051	139	5	.	.	PUNCT
cril-1051	140	1	(	(	PUNCT
cril-1051	140	2	2015	2015	NUM
cril-1051	140	3	)	)	PUNCT
cril-1051	140	4	also	also	ADV
cril-1051	140	5	used	use	VERB
cril-1051	140	6	information	information	NOUN
cril-1051	140	7	from	from	ADP
cril-1051	140	8	other	other	ADJ
cril-1051	140	9	igt	igt	PROPN
cril-1051	140	10	lines	line	NOUN
cril-1051	140	11	such	such	ADJ
cril-1051	140	12	as	as	ADP
cril-1051	140	13	translation	translation	NOUN
cril-1051	140	14	and	and	CCONJ
cril-1051	140	15	part	part	NOUN
cril-1051	140	16	-	-	PUNCT
cril-1051	140	17	of	of	ADP
cril-1051	140	18	-	-	PUNCT
cril-1051	140	19	speech	speech	NOUN
cril-1051	140	20	tags	tag	NOUN
cril-1051	140	21	to	to	PART
cril-1051	140	22	train	train	VERB
cril-1051	140	23	a	a	DET
cril-1051	140	24	system	system	NOUN
cril-1051	140	25	to	to	PART
cril-1051	140	26	gloss	gloss	VERB
cril-1051	140	27	.	.	PUNCT
cril-1051	141	1	5	5	NUM
cril-1051	141	2	.	.	NOUN
cril-1051	141	3	inflectional	inflectional	ADJ
cril-1051	141	4	paradigm	paradigm	NOUN
cril-1051	141	5	learning	learn	VERB
cril-1051	141	6	morphology	morphology	NOUN
cril-1051	141	7	includes	include	VERB
cril-1051	141	8	the	the	DET
cril-1051	141	9	inference	inference	NOUN
cril-1051	141	10	of	of	ADP
cril-1051	141	11	rules	rule	NOUN
cril-1051	141	12	that	that	PRON
cril-1051	141	13	govern	govern	VERB
cril-1051	141	14	word	word	NOUN
cril-1051	141	15	building	building	NOUN
cril-1051	141	16	strategies	strategy	NOUN
cril-1051	141	17	and	and	CCONJ
cril-1051	141	18	the	the	DET
cril-1051	141	19	discovery	discovery	NOUN
cril-1051	141	20	of	of	ADP
cril-1051	141	21	how	how	SCONJ
cril-1051	141	22	word	word	NOUN
cril-1051	141	23	forms	form	NOUN
cril-1051	141	24	are	be	AUX
cril-1051	141	25	systematically	systematically	ADV
cril-1051	141	26	related	relate	VERB
cril-1051	141	27	(	(	PUNCT
cril-1051	141	28	roark	roark	NOUN
cril-1051	141	29	&	&	CCONJ
cril-1051	141	30	sproat	sproat	NOUN
cril-1051	141	31	2007	2007	NUM
cril-1051	141	32	)	)	PUNCT
cril-1051	141	33	.	.	PUNCT
cril-1051	142	1	therefore	therefore	ADV
cril-1051	142	2	,	,	PUNCT
cril-1051	142	3	virpioja	virpioja	NOUN
cril-1051	142	4	et	et	PROPN
cril-1051	142	5	al	al	PROPN
cril-1051	142	6	.	.	PROPN
cril-1051	143	1	(	(	PUNCT
cril-1051	143	2	2011	2011	NUM
cril-1051	143	3	)	)	PUNCT
cril-1051	143	4	add	add	VERB
cril-1051	143	5	a	a	DET
cril-1051	143	6	third	third	ADJ
cril-1051	143	7	task	task	NOUN
cril-1051	143	8	to	to	ADP
cril-1051	143	9	morphological	morphological	ADJ
cril-1051	143	10	analysis	analysis	NOUN
cril-1051	143	11	:	:	PUNCT
cril-1051	143	12	identification	identification	NOUN
cril-1051	143	13	of	of	ADP
cril-1051	143	14	morphologically	morphologically	ADV
cril-1051	143	15	related	relate	VERB
cril-1051	143	16	words	word	NOUN
cril-1051	143	17	through	through	ADP
cril-1051	143	18	patterns	pattern	NOUN
cril-1051	143	19	of	of	ADP
cril-1051	143	20	inflection	inflection	NOUN
cril-1051	143	21	.	.	PUNCT
cril-1051	144	1	durrett	durrett	PROPN
cril-1051	144	2	and	and	CCONJ
cril-1051	144	3	denero	denero	NOUN
cril-1051	144	4	(	(	PUNCT
cril-1051	144	5	2013	2013	NUM
cril-1051	144	6	)	)	PUNCT
cril-1051	144	7	claim	claim	VERB
cril-1051	144	8	that	that	SCONJ
cril-1051	144	9	the	the	DET
cril-1051	144	10	inference	inference	NOUN
cril-1051	144	11	of	of	ADP
cril-1051	144	12	inflectional	inflectional	ADJ
cril-1051	144	13	patterns	pattern	NOUN
cril-1051	144	14	must	must	AUX
cril-1051	144	15	be	be	AUX
cril-1051	144	16	based	base	VERB
cril-1051	144	17	on	on	ADP
cril-1051	144	18	three	three	NUM
cril-1051	144	19	assumptions	assumption	NOUN
cril-1051	144	20	.	.	PUNCT
cril-1051	145	1	first	first	ADV
cril-1051	145	2	,	,	PUNCT
cril-1051	145	3	each	each	DET
cril-1051	145	4	lexical	lexical	ADJ
cril-1051	145	5	category	category	NOUN
cril-1051	145	6	is	be	AUX
cril-1051	145	7	dictated	dictate	VERB
cril-1051	145	8	by	by	ADP
cril-1051	145	9	a	a	DET
cril-1051	145	10	subsystem	subsystem	NOUN
cril-1051	145	11	of	of	ADP
cril-1051	145	12	rules	rule	NOUN
cril-1051	145	13	.	.	PUNCT
cril-1051	146	1	russian	russian	ADJ
cril-1051	146	2	nouns	noun	NOUN
cril-1051	146	3	,	,	PUNCT
cril-1051	146	4	for	for	ADP
cril-1051	146	5	example	example	NOUN
cril-1051	146	6	,	,	PUNCT
cril-1051	146	7	can	can	AUX
cril-1051	146	8	be	be	AUX
cril-1051	146	9	generalized	generalize	VERB
cril-1051	146	10	into	into	ADP
cril-1051	146	11	three	three	NUM
cril-1051	146	12	simplified	simplified	ADJ
cril-1051	146	13	patterns	pattern	NOUN
cril-1051	146	14	of	of	ADP
cril-1051	146	15	inflection	inflection	NOUN
cril-1051	146	16	that	that	PRON
cril-1051	146	17	are	be	AUX
cril-1051	146	18	usually	usually	ADV
cril-1051	146	19	labeled	label	VERB
cril-1051	146	20	“	"	PUNCT
cril-1051	146	21	masculine	masculine	ADJ
cril-1051	146	22	”	"	PUNCT
cril-1051	146	23	,	,	PUNCT
cril-1051	146	24	“	"	PUNCT
cril-1051	146	25	feminine	feminine	NOUN
cril-1051	146	26	”	"	PUNCT
cril-1051	146	27	,	,	PUNCT
cril-1051	146	28	and	and	CCONJ
cril-1051	146	29	“	"	PUNCT
cril-1051	146	30	neuter	neuter	ADJ
cril-1051	146	31	”	"	PUNCT
cril-1051	146	32	.	.	PUNCT
cril-1051	147	1	lexemes	lexeme	NOUN
cril-1051	147	2	that	that	PRON
cril-1051	147	3	adhere	adhere	VERB
cril-1051	147	4	to	to	ADP
cril-1051	147	5	the	the	DET
cril-1051	147	6	same	same	ADJ
cril-1051	147	7	pattern	pattern	NOUN
cril-1051	147	8	are	be	AUX
cril-1051	147	9	grouped	group	VERB
cril-1051	147	10	into	into	ADP
cril-1051	147	11	inflectional	inflectional	ADJ
cril-1051	147	12	classes	class	NOUN
cril-1051	147	13	(	(	PUNCT
cril-1051	147	14	sometimes	sometimes	ADV
cril-1051	147	15	called	call	VERB
cril-1051	147	16	“	"	PUNCT
cril-1051	147	17	declensions	declension	NOUN
cril-1051	147	18	”	"	PUNCT
cril-1051	147	19	for	for	ADP
cril-1051	147	20	nouns	noun	NOUN
cril-1051	147	21	and	and	CCONJ
cril-1051	147	22	adjectives	adjective	NOUN
cril-1051	147	23	and	and	CCONJ
cril-1051	147	24	“	"	PUNCT
cril-1051	147	25	conjugations	conjugation	NOUN
cril-1051	147	26	”	"	PUNCT
cril-1051	147	27	for	for	ADP
cril-1051	147	28	verbs	verb	NOUN
cril-1051	147	29	)	)	PUNCT
cril-1051	147	30	.	.	PUNCT
cril-1051	148	1	the	the	DET
cril-1051	148	2	patterns	pattern	NOUN
cril-1051	148	3	themselves	themselves	PRON
cril-1051	148	4	are	be	AUX
cril-1051	148	5	known	know	VERB
cril-1051	148	6	as	as	ADP
cril-1051	148	7	inflectional	inflectional	ADJ
cril-1051	148	8	paradigms	paradigm	NOUN
cril-1051	148	9	.	.	PUNCT
cril-1051	149	1	second	second	ADJ
cril-1051	149	2	,	,	PUNCT
cril-1051	149	3	inflectional	inflectional	ADJ
cril-1051	149	4	changes	change	NOUN
cril-1051	149	5	are	be	AUX
cril-1051	149	6	triggered	trigger	VERB
cril-1051	149	7	by	by	ADP
cril-1051	149	8	context	context	NOUN
cril-1051	149	9	and	and	CCONJ
cril-1051	149	10	,	,	PUNCT
cril-1051	149	11	therefore	therefore	ADV
cril-1051	149	12	,	,	PUNCT
cril-1051	149	13	the	the	DET
cril-1051	149	14	patterns	pattern	NOUN
cril-1051	149	15	can	can	AUX
cril-1051	149	16	be	be	AUX
cril-1051	149	17	inferred	infer	VERB
cril-1051	149	18	from	from	ADP
cril-1051	149	19	context	context	NOUN
cril-1051	149	20	.	.	PUNCT
cril-1051	150	1	descriptive	descriptive	ADJ
cril-1051	150	2	studies	study	NOUN
cril-1051	150	3	look	look	VERB
cril-1051	150	4	to	to	ADP
cril-1051	150	5	phonology	phonology	NOUN
cril-1051	150	6	or	or	CCONJ
cril-1051	150	7	else	else	ADV
cril-1051	150	8	to	to	ADP
cril-1051	150	9	both	both	DET
cril-1051	150	10	phonological	phonological	ADJ
cril-1051	150	11	structure	structure	NOUN
cril-1051	150	12	and	and	CCONJ
cril-1051	150	13	the	the	DET
cril-1051	150	14	semantic	semantic	ADJ
cril-1051	150	15	content	content	NOUN
cril-1051	150	16	of	of	ADP
cril-1051	150	17	the	the	DET
cril-1051	150	18	lexeme	lexeme	NOUN
cril-1051	150	19	for	for	ADP
cril-1051	150	20	the	the	DET
cril-1051	150	21	triggering	trigger	VERB
cril-1051	150	22	context	context	NOUN
cril-1051	150	23	.	.	PUNCT
cril-1051	151	1	computational	computational	ADJ
cril-1051	151	2	models	model	NOUN
cril-1051	151	3	,	,	PUNCT
cril-1051	151	4	due	due	ADP
cril-1051	151	5	to	to	ADP
cril-1051	151	6	the	the	DET
cril-1051	151	7	nature	nature	NOUN
cril-1051	151	8	of	of	ADP
cril-1051	151	9	their	their	PRON
cril-1051	151	10	input	input	NOUN
cril-1051	151	11	,	,	PUNCT
cril-1051	151	12	look	look	VERB
cril-1051	151	13	to	to	ADP
cril-1051	151	14	orthographic	orthographic	ADJ
cril-1051	151	15	context	context	NOUN
cril-1051	151	16	.	.	PUNCT
cril-1051	152	1	the	the	DET
cril-1051	152	2	third	third	ADJ
cril-1051	152	3	assumption	assumption	NOUN
cril-1051	152	4	is	be	AUX
cril-1051	152	5	that	that	SCONJ
cril-1051	152	6	each	each	DET
cril-1051	152	7	stem	stem	NOUN
cril-1051	152	8	morpheme	morpheme	NOUN
cril-1051	152	9	is	be	AUX
cril-1051	152	10	inflected	inflect	VERB
cril-1051	152	11	consistently	consistently	ADV
cril-1051	152	12	according	accord	VERB
cril-1051	152	13	to	to	ADP
cril-1051	152	14	the	the	DET
cril-1051	152	15	inflectional	inflectional	ADJ
cril-1051	152	16	class	class	NOUN
cril-1051	152	17	it	it	PRON
cril-1051	152	18	belongs	belong	VERB
cril-1051	152	19	to	to	ADP
cril-1051	152	20	with	with	ADP
cril-1051	152	21	any	any	DET
cril-1051	152	22	idiosyncrasies	idiosyncrasy	NOUN
cril-1051	152	23	of	of	ADP
cril-1051	152	24	the	the	DET
cril-1051	152	25	stem	stem	NOUN
cril-1051	152	26	itself	itself	PRON
cril-1051	152	27	.	.	PUNCT
cril-1051	153	1	monson	monson	PROPN
cril-1051	153	2	et	et	PROPN
cril-1051	153	3	al	al	PROPN
cril-1051	153	4	.	.	PROPN
cril-1051	154	1	(	(	PUNCT
cril-1051	154	2	2007	2007	NUM
cril-1051	154	3	)	)	PUNCT
cril-1051	154	4	give	give	VERB
cril-1051	154	5	two	two	NUM
cril-1051	154	6	guiding	guide	VERB
cril-1051	154	7	principles	principle	NOUN
cril-1051	154	8	for	for	ADP
cril-1051	154	9	computational	computational	ADJ
cril-1051	154	10	paradigm	paradigm	NOUN
cril-1051	154	11	learning	learning	NOUN
cril-1051	154	12	.	.	PUNCT
cril-1051	155	1	one	one	NUM
cril-1051	155	2	is	be	AUX
cril-1051	155	3	that	that	SCONJ
cril-1051	155	4	inflected	inflect	VERB
cril-1051	155	5	forms	form	NOUN
cril-1051	155	6	of	of	ADP
cril-1051	155	7	a	a	DET
cril-1051	155	8	lemma	lemma	PROPN
cril-1051	155	9	look	look	VERB
cril-1051	155	10	similar	similar	ADJ
cril-1051	155	11	to	to	ADP
cril-1051	155	12	each	each	DET
cril-1051	155	13	other	other	ADJ
cril-1051	155	14	.	.	PUNCT
cril-1051	156	1	this	this	DET
cril-1051	156	2	principle	principle	NOUN
cril-1051	156	3	holds	hold	VERB
cril-1051	156	4	well	well	ADV
cril-1051	156	5	enough	enough	ADV
cril-1051	156	6	to	to	PART
cril-1051	156	7	serve	serve	VERB
cril-1051	156	8	as	as	ADP
cril-1051	156	9	a	a	DET
cril-1051	156	10	solid	solid	ADJ
cril-1051	156	11	working	working	NOUN
cril-1051	156	12	assumption	assumption	NOUN
cril-1051	156	13	,	,	PUNCT
cril-1051	156	14	although	although	SCONJ
cril-1051	156	15	languages	language	NOUN
cril-1051	156	16	abound	abound	VERB
cril-1051	156	17	with	with	ADP
cril-1051	156	18	exceptions	exception	NOUN
cril-1051	156	19	and	and	CCONJ
cril-1051	156	20	inflection	inflection	NOUN
cril-1051	156	21	can	can	AUX
cril-1051	156	22	even	even	ADV
cril-1051	156	23	be	be	AUX
cril-1051	156	24	suppletive	suppletive	ADJ
cril-1051	156	25	(	(	PUNCT
cril-1051	156	26	e.g.	e.g.	ADV
cril-1051	156	27	,	,	PUNCT
cril-1051	156	28	go	go	VERB
cril-1051	156	29	vs.	vs.	CCONJ
cril-1051	156	30	went	went	NOUN
cril-1051	156	31	,	,	PUNCT
cril-1051	156	32	etc	etc	X
cril-1051	156	33	.	.	X
cril-1051	156	34	)	)	PUNCT
cril-1051	156	35	.	.	PUNCT
cril-1051	157	1	the	the	DET
cril-1051	157	2	second	second	ADJ
cril-1051	157	3	principle	principle	NOUN
cril-1051	157	4	is	be	AUX
cril-1051	157	5	that	that	SCONJ
cril-1051	157	6	“	"	PUNCT
cril-1051	157	7	in	in	ADP
cril-1051	157	8	any	any	DET
cril-1051	157	9	given	give	VERB
cril-1051	157	10	corpus	corpus	NOUN
cril-1051	157	11	,	,	PUNCT
cril-1051	157	12	a	a	DET
cril-1051	157	13	particular	particular	ADJ
cril-1051	157	14	lexeme	lexeme	NOUN
cril-1051	157	15	will	will	AUX
cril-1051	157	16	likely	likely	ADV
cril-1051	157	17	not	not	PART
cril-1051	157	18	occur	occur	VERB
cril-1051	157	19	in	in	ADP
cril-1051	157	20	all	all	DET
cril-1051	157	21	possible	possible	ADJ
cril-1051	157	22	inflected	inflected	ADJ
cril-1051	157	23	forms	form	NOUN
cril-1051	157	24	”	"	PUNCT
cril-1051	157	25	.	.	PUNCT
cril-1051	158	1	inflectional	inflectional	ADJ
cril-1051	158	2	paradigms	paradigm	NOUN
cril-1051	158	3	can	can	AUX
cril-1051	158	4	be	be	AUX
cril-1051	158	5	quite	quite	ADV
cril-1051	158	6	large	large	ADJ
cril-1051	158	7	.	.	PUNCT
cril-1051	159	1	languages	language	NOUN
cril-1051	159	2	may	may	AUX
cril-1051	159	3	have	have	VERB
cril-1051	159	4	hundreds	hundred	NOUN
cril-1051	159	5	or	or	CCONJ
cril-1051	159	6	even	even	ADV
cril-1051	159	7	thousands	thousand	NOUN
cril-1051	159	8	of	of	ADP
cril-1051	159	9	forms	form	NOUN
cril-1051	159	10	per	per	ADP
cril-1051	159	11	lemma	lemma	PROPN
cril-1051	159	12	(	(	PUNCT
cril-1051	159	13	corbett	corbett	NOUN
cril-1051	159	14	2013	2013	NUM
cril-1051	159	15	)	)	PUNCT
cril-1051	159	16	.	.	PUNCT
cril-1051	160	1	even	even	ADV
cril-1051	160	2	with	with	ADP
cril-1051	160	3	a	a	DET
cril-1051	160	4	large	large	ADJ
cril-1051	160	5	corpus	corpus	NOUN
cril-1051	160	6	,	,	PUNCT
cril-1051	160	7	attempts	attempt	VERB
cril-1051	160	8	to	to	PART
cril-1051	160	9	learn	learn	VERB
cril-1051	160	10	paradigms	paradigm	NOUN
cril-1051	160	11	,	,	PUNCT
cril-1051	160	12	like	like	ADP
cril-1051	160	13	the	the	DET
cril-1051	160	14	one	one	NOUN
cril-1051	160	15	illustrated	illustrate	VERB
cril-1051	160	16	in	in	ADP
cril-1051	160	17	figure	figure	NOUN
cril-1051	160	18	6	6	NUM
cril-1051	160	19	,	,	PUNCT
cril-1051	160	20	by	by	ADP
cril-1051	160	21	only	only	ADV
cril-1051	160	22	using	use	VERB
cril-1051	160	23	the	the	DET
cril-1051	160	24	corpus	corpus	NOUN
cril-1051	160	25	will	will	AUX
cril-1051	160	26	leave	leave	VERB
cril-1051	160	27	empty	empty	ADJ
cril-1051	160	28	cells	cell	NOUN
cril-1051	160	29	in	in	ADP
cril-1051	160	30	the	the	DET
cril-1051	160	31	paradigm	paradigm	NOUN
cril-1051	160	32	’s	’s	PART
cril-1051	160	33	table	table	NOUN
cril-1051	160	34	.	.	PUNCT
cril-1051	161	1	it	it	PRON
cril-1051	161	2	is	be	AUX
cril-1051	161	3	possible	possible	ADJ
cril-1051	161	4	that	that	SCONJ
cril-1051	161	5	certain	certain	ADJ
cril-1051	161	6	forms	form	NOUN
cril-1051	161	7	may	may	AUX
cril-1051	161	8	never	never	ADV
cril-1051	161	9	occur	occur	VERB
cril-1051	161	10	in	in	ADP
cril-1051	161	11	natural	natural	ADJ
cril-1051	161	12	language	language	NOUN
cril-1051	161	13	even	even	ADV
cril-1051	161	14	though	though	SCONJ
cril-1051	161	15	they	they	PRON
cril-1051	161	16	are	be	AUX
cril-1051	161	17	grammatically	grammatically	ADV
cril-1051	161	18	possible	possible	ADJ
cril-1051	161	19	(	(	PUNCT
cril-1051	161	20	silfverberg	silfverberg	PROPN
cril-1051	161	21	&	&	CCONJ
cril-1051	161	22	hulden	hulden	ADJ
cril-1051	161	23	2018	2018	NUM
cril-1051	161	24	)	)	PUNCT
cril-1051	161	25	.	.	PUNCT
cril-1051	162	1	additionally	additionally	ADV
cril-1051	162	2	,	,	PUNCT
cril-1051	162	3	frequent	frequent	ADJ
cril-1051	162	4	words	word	NOUN
cril-1051	162	5	often	often	ADV
cril-1051	162	6	follow	follow	VERB
cril-1051	162	7	irregular	irregular	ADJ
cril-1051	162	8	patterns	pattern	NOUN
cril-1051	162	9	,	,	PUNCT
cril-1051	162	10	as	as	SCONJ
cril-1051	162	11	does	do	VERB
cril-1051	162	12	computational	computational	ADJ
cril-1051	162	13	morphology	morphology	NOUN
cril-1051	162	14	for	for	ADP
cril-1051	162	15	language	language	NOUN
cril-1051	162	16	description	description	NOUN
cril-1051	162	17	and	and	CCONJ
cril-1051	162	18	documentation	documentation	NOUN
cril-1051	162	19	13	13	NUM
cril-1051	162	20	the	the	DET
cril-1051	162	21	english	english	PROPN
cril-1051	162	22	verb	verb	NOUN
cril-1051	162	23	be	be	AUX
cril-1051	162	24	.	.	PUNCT
cril-1051	163	1	for	for	ADP
cril-1051	163	2	these	these	DET
cril-1051	163	3	reasons	reason	NOUN
cril-1051	163	4	,	,	PUNCT
cril-1051	163	5	the	the	DET
cril-1051	163	6	ldd	ldd	PROPN
cril-1051	163	7	workflow	workflow	NOUN
cril-1051	163	8	includes	include	VERB
cril-1051	163	9	elicitation	elicitation	NOUN
cril-1051	163	10	of	of	ADP
cril-1051	163	11	morphological	morphological	ADJ
cril-1051	163	12	paradigms	paradigm	NOUN
cril-1051	163	13	(	(	PUNCT
cril-1051	163	14	lupke	lupke	PROPN
cril-1051	163	15	2010	2010	NUM
cril-1051	163	16	;	;	PUNCT
cril-1051	163	17	boerger	boerger	NOUN
cril-1051	163	18	et	et	PROPN
cril-1051	163	19	al	al	PROPN
cril-1051	163	20	.	.	PROPN
cril-1051	163	21	2016	2016	NUM
cril-1051	163	22	)	)	PUNCT
cril-1051	163	23	.	.	PUNCT
cril-1051	164	1	figure	figure	VERB
cril-1051	164	2	6	6	NUM
cril-1051	164	3	.	.	PUNCT
cril-1051	164	4	inflectional	inflectional	ADJ
cril-1051	164	5	paradigm	paradigm	NOUN
cril-1051	164	6	of	of	ADP
cril-1051	164	7	the	the	DET
cril-1051	164	8	english	english	ADJ
cril-1051	164	9	verb	verb	NOUN
cril-1051	164	10	“	"	PUNCT
cril-1051	164	11	to	to	PART
cril-1051	164	12	be	be	AUX
cril-1051	164	13	”	"	PUNCT
cril-1051	164	14	computational	computational	ADJ
cril-1051	164	15	models	model	NOUN
cril-1051	164	16	can	can	AUX
cril-1051	164	17	learn	learn	VERB
cril-1051	164	18	frequent	frequent	ADJ
cril-1051	164	19	and	and	CCONJ
cril-1051	164	20	regular	regular	ADJ
cril-1051	164	21	paradigmatic	paradigmatic	ADJ
cril-1051	164	22	patterns	pattern	NOUN
cril-1051	164	23	with	with	ADP
cril-1051	164	24	over	over	ADP
cril-1051	164	25	90	90	NUM
cril-1051	164	26	%	%	NOUN
cril-1051	164	27	accuracy	accuracy	NOUN
cril-1051	164	28	even	even	ADV
cril-1051	164	29	in	in	ADP
cril-1051	164	30	low	low	ADJ
cril-1051	164	31	-	-	PUNCT
cril-1051	164	32	resource	resource	NOUN
cril-1051	164	33	settings	setting	NOUN
cril-1051	164	34	(	(	PUNCT
cril-1051	164	35	hammarström	hammarström	NOUN
cril-1051	164	36	&	&	CCONJ
cril-1051	164	37	borin	borin	ADJ
cril-1051	164	38	2011	2011	NUM
cril-1051	164	39	;	;	PUNCT
cril-1051	164	40	durrett	durrett	PROPN
cril-1051	164	41	&	&	CCONJ
cril-1051	164	42	denero	denero	PROPN
cril-1051	164	43	2013	2013	NUM
cril-1051	164	44	;	;	PUNCT
cril-1051	164	45	ahlberg	ahlberg	PROPN
cril-1051	164	46	,	,	PUNCT
cril-1051	164	47	forsberg	forsberg	PROPN
cril-1051	164	48	&	&	CCONJ
cril-1051	164	49	hulden	hulden	VERB
cril-1051	164	50	2014	2014	NUM
cril-1051	164	51	)	)	PUNCT
cril-1051	164	52	.	.	PUNCT
cril-1051	165	1	most	most	ADJ
cril-1051	165	2	early	early	ADJ
cril-1051	165	3	work	work	NOUN
cril-1051	165	4	on	on	ADP
cril-1051	165	5	paradigm	paradigm	NOUN
cril-1051	165	6	induction	induction	NOUN
cril-1051	165	7	applied	apply	VERB
cril-1051	165	8	unsupervised	unsupervised	ADJ
cril-1051	165	9	learning	learning	NOUN
cril-1051	165	10	to	to	ADP
cril-1051	165	11	concatenative	concatenative	ADJ
cril-1051	165	12	morphology	morphology	NOUN
cril-1051	165	13	(	(	PUNCT
cril-1051	165	14	goldsmith	goldsmith	PROPN
cril-1051	165	15	2001	2001	NUM
cril-1051	165	16	;	;	PUNCT
cril-1051	165	17	chan	chan	PROPN
cril-1051	165	18	2006	2006	NUM
cril-1051	165	19	;	;	PUNCT
cril-1051	165	20	monson	monson	PROPN
cril-1051	165	21	et	et	PROPN
cril-1051	165	22	al	al	PROPN
cril-1051	165	23	.	.	PROPN
cril-1051	165	24	2007	2007	NUM
cril-1051	165	25	)	)	PUNCT
cril-1051	165	26	.	.	PUNCT
cril-1051	166	1	the	the	DET
cril-1051	166	2	unsupervised	unsupervised	ADJ
cril-1051	166	3	version	version	NOUN
cril-1051	166	4	of	of	ADP
cril-1051	166	5	the	the	DET
cril-1051	166	6	paradigm	paradigm	NOUN
cril-1051	166	7	completion	completion	NOUN
cril-1051	166	8	task	task	NOUN
cril-1051	166	9	(	(	PUNCT
cril-1051	166	10	jin	jin	NOUN
cril-1051	166	11	et	et	PROPN
cril-1051	166	12	al	al	PROPN
cril-1051	166	13	.	.	PROPN
cril-1051	166	14	2020	2020	NUM
cril-1051	166	15	)	)	PUNCT
cril-1051	166	16	has	have	AUX
cril-1051	166	17	been	be	AUX
cril-1051	166	18	the	the	DET
cril-1051	166	19	subject	subject	NOUN
cril-1051	166	20	of	of	ADP
cril-1051	166	21	a	a	DET
cril-1051	166	22	recent	recent	ADJ
cril-1051	166	23	shared	shared	ADJ
cril-1051	166	24	task	task	NOUN
cril-1051	166	25	(	(	PUNCT
cril-1051	166	26	kann	kann	PROPN
cril-1051	166	27	et	et	PROPN
cril-1051	166	28	al	al	PROPN
cril-1051	166	29	.	.	PROPN
cril-1051	166	30	2020	2020	NUM
cril-1051	166	31	)	)	PUNCT
cril-1051	166	32	,	,	PUNCT
cril-1051	166	33	with	with	ADP
cril-1051	166	34	the	the	DET
cril-1051	166	35	conclusion	conclusion	NOUN
cril-1051	166	36	that	that	SCONJ
cril-1051	166	37	it	it	PRON
cril-1051	166	38	is	be	AUX
cril-1051	166	39	extremely	extremely	ADV
cril-1051	166	40	challenging	challenging	ADJ
cril-1051	166	41	for	for	ADP
cril-1051	166	42	current	current	ADJ
cril-1051	166	43	state	state	NOUN
cril-1051	166	44	-	-	PUNCT
cril-1051	166	45	of	of	ADP
cril-1051	166	46	-	-	PUNCT
cril-1051	166	47	the	the	DET
cril-1051	166	48	-	-	PUNCT
cril-1051	166	49	art	art	NOUN
cril-1051	166	50	systems	system	NOUN
cril-1051	166	51	.	.	PUNCT
cril-1051	167	1	semi	semi	ADJ
cril-1051	167	2	-	-	ADJ
cril-1051	167	3	supervised	supervised	ADJ
cril-1051	167	4	models	model	NOUN
cril-1051	167	5	have	have	AUX
cril-1051	167	6	been	be	AUX
cril-1051	167	7	more	more	ADV
cril-1051	167	8	recently	recently	ADV
cril-1051	167	9	applied	apply	VERB
cril-1051	167	10	on	on	ADP
cril-1051	167	11	concatenative	concatenative	ADJ
cril-1051	167	12	and	and	CCONJ
cril-1051	167	13	non	non	ADJ
cril-1051	167	14	-	-	ADJ
cril-1051	167	15	concatenative	concatenative	ADJ
cril-1051	167	16	languages	language	NOUN
cril-1051	167	17	(	(	PUNCT
cril-1051	167	18	dreyer	dreyer	PROPN
cril-1051	167	19	&	&	CCONJ
cril-1051	167	20	eisner	eisner	PROPN
cril-1051	167	21	2011	2011	NUM
cril-1051	167	22	;	;	PUNCT
cril-1051	167	23	durrett	durrett	PROPN
cril-1051	167	24	&	&	CCONJ
cril-1051	167	25	denero	denero	NOUN
cril-1051	167	26	2013	2013	NUM
cril-1051	167	27	)	)	PUNCT
cril-1051	167	28	.	.	PUNCT
cril-1051	168	1	supervised	supervised	ADJ
cril-1051	168	2	learning	learning	NOUN
cril-1051	168	3	has	have	AUX
cril-1051	168	4	also	also	ADV
cril-1051	168	5	been	be	AUX
cril-1051	168	6	applied	apply	VERB
cril-1051	168	7	to	to	ADP
cril-1051	168	8	inflectional	inflectional	ADJ
cril-1051	168	9	morphology	morphology	NOUN
cril-1051	168	10	.	.	PUNCT
cril-1051	169	1	some	some	DET
cril-1051	169	2	work	work	NOUN
cril-1051	169	3	focuses	focus	VERB
cril-1051	169	4	on	on	ADP
cril-1051	169	5	generating	generate	VERB
cril-1051	169	6	inflected	inflected	ADJ
cril-1051	169	7	forms	form	NOUN
cril-1051	169	8	,	,	PUNCT
cril-1051	169	9	including	include	VERB
cril-1051	169	10	work	work	NOUN
cril-1051	169	11	motivated	motivate	VERB
cril-1051	169	12	by	by	ADP
cril-1051	169	13	the	the	DET
cril-1051	169	14	paradigm	paradigm	NOUN
cril-1051	169	15	cell	cell	NOUN
cril-1051	169	16	filling	filling	NOUN
cril-1051	169	17	problem	problem	NOUN
cril-1051	169	18	(	(	PUNCT
cril-1051	169	19	pcfp	pcfp	NOUN
cril-1051	169	20	)	)	PUNCT
cril-1051	169	21	,	,	PUNCT
cril-1051	169	22	illustrated	illustrate	VERB
cril-1051	169	23	in	in	ADP
cril-1051	169	24	figure	figure	NOUN
cril-1051	169	25	7	7	NUM
cril-1051	169	26	(	(	PUNCT
cril-1051	169	27	ackerman	ackerman	PROPN
cril-1051	169	28	,	,	PUNCT
cril-1051	169	29	blevins	blevins	PROPN
cril-1051	169	30	&	&	CCONJ
cril-1051	169	31	malouf	malouf	PROPN
cril-1051	169	32	2009	2009	NUM
cril-1051	169	33	)	)	PUNCT
cril-1051	169	34	.	.	PUNCT
cril-1051	170	1	the	the	DET
cril-1051	170	2	pcfp	pcfp	NOUN
cril-1051	170	3	is	be	AUX
cril-1051	170	4	framed	frame	VERB
cril-1051	170	5	as	as	ADP
cril-1051	170	6	an	an	DET
cril-1051	170	7	attempt	attempt	NOUN
cril-1051	170	8	to	to	PART
cril-1051	170	9	model	model	VERB
cril-1051	170	10	how	how	SCONJ
cril-1051	170	11	new	new	ADJ
cril-1051	170	12	speakers	speaker	NOUN
cril-1051	170	13	(	(	PUNCT
cril-1051	170	14	e.g.	e.g.	ADV
cril-1051	170	15	,	,	PUNCT
cril-1051	170	16	young	young	ADJ
cril-1051	170	17	children	child	NOUN
cril-1051	170	18	)	)	PUNCT
cril-1051	170	19	infer	infer	VERB
cril-1051	170	20	the	the	DET
cril-1051	170	21	inflected	inflected	ADJ
cril-1051	170	22	forms	form	NOUN
cril-1051	170	23	they	they	PRON
cril-1051	170	24	have	have	AUX
cril-1051	170	25	not	not	PART
cril-1051	170	26	yet	yet	ADV
cril-1051	170	27	encountered	encounter	VERB
cril-1051	170	28	(	(	PUNCT
cril-1051	170	29	dreyer	dreyer	PROPN
cril-1051	170	30	&	&	CCONJ
cril-1051	170	31	eisner	eisner	PROPN
cril-1051	170	32	2011	2011	NUM
cril-1051	170	33	;	;	PUNCT
cril-1051	170	34	ahlberg	ahlberg	PROPN
cril-1051	170	35	,	,	PUNCT
cril-1051	170	36	forsberg	forsberg	PROPN
cril-1051	170	37	&	&	CCONJ
cril-1051	170	38	hulden	hulden	VERB
cril-1051	170	39	2015	2015	NUM
cril-1051	170	40	;	;	PUNCT
cril-1051	170	41	malouf	malouf	PROPN
cril-1051	170	42	2016	2016	NUM
cril-1051	170	43	;	;	PUNCT
cril-1051	170	44	silfverberg	silfverberg	PROPN
cril-1051	170	45	&	&	CCONJ
cril-1051	170	46	hulden	hulden	ADJ
cril-1051	170	47	2018	2018	NUM
cril-1051	170	48	)	)	PUNCT
cril-1051	170	49	.	.	PUNCT
cril-1051	171	1	colorado	colorado	PROPN
cril-1051	171	2	research	research	PROPN
cril-1051	171	3	in	in	ADP
cril-1051	171	4	linguistics	linguistic	NOUN
cril-1051	171	5	,	,	PUNCT
cril-1051	171	6	volume	volume	NOUN
cril-1051	171	7	25	25	NUM
cril-1051	171	8	(	(	PUNCT
cril-1051	171	9	2021	2021	NUM
cril-1051	171	10	)	)	PUNCT
cril-1051	171	11	14	14	NUM
cril-1051	171	12	figure	figure	NOUN
cril-1051	171	13	7	7	NUM
cril-1051	171	14	.	.	PUNCT
cril-1051	171	15	illustration	illustration	NOUN
cril-1051	171	16	of	of	ADP
cril-1051	171	17	the	the	DET
cril-1051	171	18	paradigm	paradigm	NOUN
cril-1051	171	19	cell	cell	NOUN
cril-1051	171	20	filling	filling	NOUN
cril-1051	171	21	problem	problem	NOUN
cril-1051	171	22	(	(	PUNCT
cril-1051	171	23	silfverberg	silfverberg	NOUN
cril-1051	171	24	and	and	CCONJ
cril-1051	171	25	hulden	hulden	NOUN
cril-1051	171	26	,	,	PUNCT
cril-1051	171	27	2018	2018	NUM
cril-1051	171	28	)	)	PUNCT
cril-1051	171	29	with	with	ADP
cril-1051	171	30	spanish	spanish	ADJ
cril-1051	171	31	verb	verb	NOUN
cril-1051	171	32	paradigms	paradigm	VERB
cril-1051	171	33	other	other	ADJ
cril-1051	171	34	work	work	NOUN
cril-1051	171	35	with	with	ADP
cril-1051	171	36	supervised	supervised	ADJ
cril-1051	171	37	learning	learning	NOUN
cril-1051	171	38	has	have	AUX
cril-1051	171	39	attempted	attempt	VERB
cril-1051	171	40	to	to	PART
cril-1051	171	41	induce	induce	VERB
cril-1051	171	42	inflectional	inflectional	ADJ
cril-1051	171	43	paradigms	paradigm	NOUN
cril-1051	171	44	from	from	ADP
cril-1051	171	45	text	text	NOUN
cril-1051	171	46	.	.	PUNCT
cril-1051	172	1	with	with	ADP
cril-1051	172	2	this	this	DET
cril-1051	172	3	method	method	NOUN
cril-1051	172	4	,	,	PUNCT
cril-1051	172	5	paradigms	paradigm	NOUN
cril-1051	172	6	are	be	AUX
cril-1051	172	7	completed	complete	VERB
cril-1051	172	8	by	by	ADP
cril-1051	172	9	finding	find	VERB
cril-1051	172	10	overlapping	overlap	VERB
cril-1051	172	11	patterns	pattern	NOUN
cril-1051	172	12	from	from	ADP
cril-1051	172	13	several	several	ADJ
cril-1051	172	14	incomplete	incomplete	ADJ
cril-1051	172	15	paradigms	paradigm	NOUN
cril-1051	172	16	in	in	ADP
cril-1051	172	17	text	text	NOUN
cril-1051	172	18	.	.	PUNCT
cril-1051	173	1	one	one	NUM
cril-1051	173	2	method	method	NOUN
cril-1051	173	3	does	do	VERB
cril-1051	173	4	this	this	PRON
cril-1051	173	5	by	by	ADP
cril-1051	173	6	abstracting	abstract	VERB
cril-1051	173	7	the	the	DET
cril-1051	173	8	longest	long	ADJ
cril-1051	173	9	common	common	ADJ
cril-1051	173	10	subsequence	subsequence	NOUN
cril-1051	173	11	of	of	ADP
cril-1051	173	12	characters	character	NOUN
cril-1051	173	13	in	in	ADP
cril-1051	173	14	inflected	inflected	ADJ
cril-1051	173	15	forms	form	NOUN
cril-1051	173	16	of	of	ADP
cril-1051	173	17	the	the	DET
cril-1051	173	18	same	same	ADJ
cril-1051	173	19	lexeme	lexeme	NOUN
cril-1051	173	20	and	and	CCONJ
cril-1051	173	21	then	then	ADV
cril-1051	173	22	clustering	cluster	VERB
cril-1051	173	23	words	word	NOUN
cril-1051	173	24	with	with	ADP
cril-1051	173	25	same	same	ADJ
cril-1051	173	26	or	or	CCONJ
cril-1051	173	27	similar	similar	ADJ
cril-1051	173	28	patterns	pattern	NOUN
cril-1051	173	29	(	(	PUNCT
cril-1051	173	30	ahlberg	ahlberg	PROPN
cril-1051	173	31	,	,	PUNCT
cril-1051	173	32	forsberg	forsberg	PROPN
cril-1051	173	33	&	&	CCONJ
cril-1051	173	34	hulden	hulden	VERB
cril-1051	173	35	2014	2014	NUM
cril-1051	173	36	;	;	PUNCT
cril-1051	173	37	ahlberg	ahlberg	PROPN
cril-1051	173	38	,	,	PUNCT
cril-1051	173	39	forsberg	forsberg	PROPN
cril-1051	173	40	&	&	CCONJ
cril-1051	173	41	hulden	hulden	VERB
cril-1051	173	42	2015	2015	NUM
cril-1051	173	43	)	)	PUNCT
cril-1051	173	44	.	.	PUNCT
cril-1051	174	1	this	this	PRON
cril-1051	174	2	is	be	AUX
cril-1051	174	3	illustrated	illustrate	VERB
cril-1051	174	4	in	in	ADP
cril-1051	174	5	figure	figure	NOUN
cril-1051	174	6	8	8	NUM
cril-1051	174	7	.	.	PUNCT
cril-1051	175	1	exceptions	exception	NOUN
cril-1051	175	2	or	or	CCONJ
cril-1051	175	3	irregularities	irregularity	NOUN
cril-1051	175	4	in	in	ADP
cril-1051	175	5	the	the	DET
cril-1051	175	6	paradigms	paradigms	NOUN
cril-1051	175	7	can	can	AUX
cril-1051	175	8	be	be	AUX
cril-1051	175	9	accounted	account	VERB
cril-1051	175	10	for	for	ADP
cril-1051	175	11	by	by	ADP
cril-1051	175	12	collapsing	collapse	VERB
cril-1051	175	13	the	the	DET
cril-1051	175	14	similar	similar	ADJ
cril-1051	175	15	patterns	pattern	NOUN
cril-1051	175	16	.	.	PUNCT
cril-1051	176	1	the	the	DET
cril-1051	176	2	experiment	experiment	NOUN
cril-1051	176	3	has	have	AUX
cril-1051	176	4	been	be	AUX
cril-1051	176	5	quite	quite	ADV
cril-1051	176	6	successful	successful	ADJ
cril-1051	176	7	for	for	ADP
cril-1051	176	8	a	a	DET
cril-1051	176	9	few	few	ADJ
cril-1051	176	10	indoeuropean	indoeuropean	NOUN
cril-1051	176	11	languages	language	NOUN
cril-1051	176	12	(	(	PUNCT
cril-1051	176	13	german	german	ADJ
cril-1051	176	14	,	,	PUNCT
cril-1051	176	15	spanish	spanish	ADJ
cril-1051	176	16	,	,	PUNCT
cril-1051	176	17	catalan	catalan	NOUN
cril-1051	176	18	,	,	PUNCT
cril-1051	176	19	french	french	ADJ
cril-1051	176	20	,	,	PUNCT
cril-1051	176	21	galician	galician	ADJ
cril-1051	176	22	,	,	PUNCT
cril-1051	176	23	italian	italian	ADJ
cril-1051	176	24	,	,	PUNCT
cril-1051	176	25	portuguese	portuguese	ADJ
cril-1051	176	26	,	,	PUNCT
cril-1051	176	27	russian	russian	PROPN
cril-1051	176	28	)	)	PUNCT
cril-1051	176	29	,	,	PUNCT
cril-1051	176	30	as	as	ADV
cril-1051	176	31	well	well	ADV
cril-1051	176	32	as	as	ADP
cril-1051	176	33	maltese	maltese	ADJ
cril-1051	176	34	and	and	CCONJ
cril-1051	176	35	finnish	finnish	ADJ
cril-1051	176	36	.	.	PUNCT
cril-1051	177	1	kann	kann	PROPN
cril-1051	177	2	et	et	PROPN
cril-1051	177	3	al	al	PROPN
cril-1051	177	4	.	.	PROPN
cril-1051	177	5	(	(	PUNCT
cril-1051	177	6	2017a	2017a	NUM
cril-1051	177	7	)	)	PUNCT
cril-1051	177	8	differed	differ	VERB
cril-1051	177	9	from	from	ADP
cril-1051	177	10	other	other	ADJ
cril-1051	177	11	approaches	approach	NOUN
cril-1051	177	12	in	in	ADP
cril-1051	177	13	that	that	SCONJ
cril-1051	177	14	they	they	PRON
cril-1051	177	15	encoded	encode	VERB
cril-1051	177	16	multiple	multiple	ADJ
cril-1051	177	17	inflected	inflected	ADJ
cril-1051	177	18	forms	form	NOUN
cril-1051	177	19	of	of	ADP
cril-1051	177	20	a	a	DET
cril-1051	177	21	lemma	lemma	PROPN
cril-1051	177	22	to	to	PART
cril-1051	177	23	provide	provide	VERB
cril-1051	177	24	complementary	complementary	ADJ
cril-1051	177	25	information	information	NOUN
cril-1051	177	26	in	in	ADP
cril-1051	177	27	order	order	NOUN
cril-1051	177	28	to	to	PART
cril-1051	177	29	generate	generate	VERB
cril-1051	177	30	unknown	unknown	ADJ
cril-1051	177	31	forms	form	NOUN
cril-1051	177	32	.	.	PUNCT
cril-1051	178	1	cotterell	cotterell	VERB
cril-1051	178	2	et	et	PROPN
cril-1051	178	3	al	al	PROPN
cril-1051	178	4	.	.	PROPN
cril-1051	179	1	(	(	PUNCT
cril-1051	179	2	2017	2017	NUM
cril-1051	179	3	)	)	PUNCT
cril-1051	179	4	introduced	introduce	VERB
cril-1051	179	5	neural	neural	ADJ
cril-1051	179	6	graphical	graphical	ADJ
cril-1051	179	7	models	model	NOUN
cril-1051	179	8	which	which	PRON
cril-1051	179	9	completed	complete	VERB
cril-1051	179	10	paradigms	paradigm	NOUN
cril-1051	179	11	based	base	VERB
cril-1051	179	12	on	on	ADP
cril-1051	179	13	principal	principal	ADJ
cril-1051	179	14	parts	part	NOUN
cril-1051	179	15	.	.	PUNCT
cril-1051	180	1	computational	computational	ADJ
cril-1051	180	2	morphology	morphology	NOUN
cril-1051	180	3	for	for	ADP
cril-1051	180	4	language	language	NOUN
cril-1051	180	5	description	description	NOUN
cril-1051	180	6	and	and	CCONJ
cril-1051	180	7	documentation	documentation	NOUN
cril-1051	180	8	15	15	NUM
cril-1051	180	9	figure	figure	NOUN
cril-1051	180	10	8	8	NUM
cril-1051	180	11	.	.	PUNCT
cril-1051	181	1	ahlberg	ahlberg	PROPN
cril-1051	181	2	et	et	PROPN
cril-1051	181	3	al	al	PROPN
cril-1051	181	4	.	.	PROPN
cril-1051	182	1	(	(	PUNCT
cril-1051	182	2	2015	2015	NUM
cril-1051	182	3	):	):	PUNCT
cril-1051	182	4	inducing	induce	VERB
cril-1051	182	5	paradigms	paradigm	NOUN
cril-1051	182	6	the	the	DET
cril-1051	182	7	longest	long	ADJ
cril-1051	182	8	common	common	ADJ
cril-1051	182	9	subsequences	subsequence	NOUN
cril-1051	182	10	(	(	PUNCT
cril-1051	182	11	lcs	lcs	NOUN
cril-1051	182	12	)	)	PUNCT
cril-1051	182	13	rng	rng	NOUN
cril-1051	182	14	or	or	CCONJ
cril-1051	182	15	swm	swm	NOUN
cril-1051	182	16	are	be	AUX
cril-1051	182	17	extracted	extract	VERB
cril-1051	182	18	(	(	PUNCT
cril-1051	182	19	step	step	NOUN
cril-1051	182	20	1	1	NUM
cril-1051	182	21	)	)	PUNCT
cril-1051	182	22	and	and	CCONJ
cril-1051	182	23	represented	represent	VERB
cril-1051	182	24	as	as	ADP
cril-1051	182	25	x1	x1	PROPN
cril-1051	182	26	and	and	CCONJ
cril-1051	182	27	x2	x2	PROPN
cril-1051	182	28	which	which	PRON
cril-1051	182	29	replace	replace	VERB
cril-1051	182	30	the	the	DET
cril-1051	182	31	lcs	lcs	NOUN
cril-1051	182	32	(	(	PUNCT
cril-1051	182	33	step	step	NOUN
cril-1051	182	34	2	2	NUM
cril-1051	182	35	)	)	PUNCT
cril-1051	182	36	.	.	PUNCT
cril-1051	183	1	words	word	NOUN
cril-1051	183	2	with	with	ADP
cril-1051	183	3	the	the	DET
cril-1051	183	4	same	same	ADJ
cril-1051	183	5	inflectional	inflectional	ADJ
cril-1051	183	6	patterns	pattern	NOUN
cril-1051	183	7	will	will	AUX
cril-1051	183	8	be	be	AUX
cril-1051	183	9	identical	identical	ADJ
cril-1051	183	10	(	(	PUNCT
cril-1051	183	11	step	step	NOUN
cril-1051	183	12	3	3	NUM
cril-1051	183	13	)	)	PUNCT
cril-1051	183	14	and	and	CCONJ
cril-1051	183	15	can	can	AUX
cril-1051	183	16	be	be	AUX
cril-1051	183	17	generalized	generalize	VERB
cril-1051	183	18	into	into	ADP
cril-1051	183	19	paradigms	paradigms	NOUN
cril-1051	183	20	(	(	PUNCT
cril-1051	183	21	step	step	NOUN
cril-1051	183	22	4	4	NUM
cril-1051	183	23	)	)	PUNCT
cril-1051	183	24	.	.	PUNCT
cril-1051	184	1	the	the	DET
cril-1051	184	2	remaining	remain	VERB
cril-1051	184	3	characters	character	NOUN
cril-1051	184	4	i	i	PRON
cril-1051	184	5	,	,	PUNCT
cril-1051	184	6	a	a	X
cril-1051	184	7	,	,	PUNCT
cril-1051	184	8	u	u	NOUN
cril-1051	184	9	are	be	AUX
cril-1051	184	10	assumed	assume	VERB
cril-1051	184	11	to	to	PART
cril-1051	184	12	be	be	AUX
cril-1051	184	13	inflectional	inflectional	ADJ
cril-1051	184	14	affixes	affix	NOUN
cril-1051	184	15	.	.	PUNCT
cril-1051	185	1	most	most	ADJ
cril-1051	185	2	recent	recent	ADJ
cril-1051	185	3	work	work	NOUN
cril-1051	185	4	in	in	ADP
cril-1051	185	5	paradigm	paradigm	NOUN
cril-1051	185	6	induction	induction	NOUN
cril-1051	185	7	has	have	AUX
cril-1051	185	8	been	be	AUX
cril-1051	185	9	concerned	concern	VERB
cril-1051	185	10	with	with	ADP
cril-1051	185	11	generation	generation	NOUN
cril-1051	185	12	(	(	PUNCT
cril-1051	185	13	as	as	SCONJ
cril-1051	185	14	opposed	oppose	VERB
cril-1051	185	15	to	to	ADP
cril-1051	185	16	analysis	analysis	NOUN
cril-1051	185	17	)	)	PUNCT
cril-1051	185	18	of	of	ADP
cril-1051	185	19	inflected	inflect	VERB
cril-1051	185	20	words	word	NOUN
cril-1051	185	21	and	and	CCONJ
cril-1051	185	22	has	have	AUX
cril-1051	185	23	focused	focus	VERB
cril-1051	185	24	on	on	ADP
cril-1051	185	25	morphological	morphological	ADJ
cril-1051	185	26	inflection	inflection	NOUN
cril-1051	185	27	or	or	CCONJ
cril-1051	185	28	reinflection	reinflection	NOUN
cril-1051	185	29	(	(	PUNCT
cril-1051	185	30	durrett	durrett	PROPN
cril-1051	185	31	&	&	CCONJ
cril-1051	185	32	denero	denero	PROPN
cril-1051	185	33	2013	2013	NUM
cril-1051	185	34	;	;	PUNCT
cril-1051	185	35	nicolai	nicolai	PROPN
cril-1051	185	36	,	,	PUNCT
cril-1051	185	37	cherry	cherry	PROPN
cril-1051	185	38	&	&	CCONJ
cril-1051	185	39	kondrak	kondrak	PROPN
cril-1051	185	40	2015	2015	NUM
cril-1051	185	41	;	;	PUNCT
cril-1051	185	42	faruqui	faruqui	NOUN
cril-1051	185	43	et	et	PROPN
cril-1051	185	44	al	al	PROPN
cril-1051	185	45	.	.	PROPN
cril-1051	185	46	2016	2016	NUM
cril-1051	185	47	;	;	PUNCT
cril-1051	185	48	kann	kann	PROPN
cril-1051	185	49	&	&	CCONJ
cril-1051	185	50	schütze	schütze	PROPN
cril-1051	185	51	2016	2016	NUM
cril-1051	185	52	;	;	PUNCT
cril-1051	185	53	aharoni	aharoni	PROPN
cril-1051	185	54	&	&	CCONJ
cril-1051	185	55	goldberg	goldberg	PROPN
cril-1051	185	56	2017	2017	NUM
cril-1051	185	57	)	)	PUNCT
cril-1051	185	58	.	.	PUNCT
cril-1051	186	1	partially	partially	ADV
cril-1051	186	2	building	build	VERB
cril-1051	186	3	on	on	ADP
cril-1051	186	4	these	these	PRON
cril-1051	186	5	,	,	PUNCT
cril-1051	186	6	other	other	ADJ
cril-1051	186	7	research	research	NOUN
cril-1051	186	8	has	have	AUX
cril-1051	186	9	developed	develop	VERB
cril-1051	186	10	machine	machine	NOUN
cril-1051	186	11	learning	learning	NOUN
cril-1051	186	12	models	model	NOUN
cril-1051	186	13	which	which	PRON
cril-1051	186	14	are	be	AUX
cril-1051	186	15	more	more	ADV
cril-1051	186	16	suitable	suitable	ADJ
cril-1051	186	17	for	for	ADP
cril-1051	186	18	lrl	lrl	PROPN
cril-1051	186	19	and	and	CCONJ
cril-1051	186	20	perform	perform	VERB
cril-1051	186	21	well	well	ADV
cril-1051	186	22	with	with	ADP
cril-1051	186	23	limited	limited	ADJ
cril-1051	186	24	data	datum	NOUN
cril-1051	186	25	(	(	PUNCT
cril-1051	186	26	kann	kann	PROPN
cril-1051	186	27	,	,	PUNCT
cril-1051	186	28	cotterell	cotterell	NOUN
cril-1051	186	29	&	&	CCONJ
cril-1051	186	30	schütze	schütze	PROPN
cril-1051	186	31	2017b	2017b	NUM
cril-1051	186	32	;	;	PUNCT
cril-1051	186	33	sharma	sharma	PROPN
cril-1051	186	34	,	,	PUNCT
cril-1051	186	35	katrapati	katrapati	PROPN
cril-1051	186	36	&	&	CCONJ
cril-1051	186	37	sharma	sharma	PROPN
cril-1051	186	38	2018	2018	NUM
cril-1051	186	39	;	;	PUNCT
cril-1051	186	40	makarov	makarov	PROPN
cril-1051	186	41	&	&	CCONJ
cril-1051	186	42	clematide	clematide	ADJ
cril-1051	186	43	2018b	2018b	NUM
cril-1051	186	44	;	;	PUNCT
cril-1051	186	45	wu	wu	PROPN
cril-1051	186	46	&	&	CCONJ
cril-1051	186	47	cotterell	cotterell	PROPN
cril-1051	186	48	2019	2019	NUM
cril-1051	186	49	;	;	PUNCT
cril-1051	186	50	kann	kann	PROPN
cril-1051	186	51	,	,	PUNCT
cril-1051	186	52	bowman	bowman	PROPN
cril-1051	186	53	&	&	CCONJ
cril-1051	186	54	cho	cho	PROPN
cril-1051	186	55	2020	2020	NUM
cril-1051	186	56	;	;	PUNCT
cril-1051	186	57	wu	wu	PROPN
cril-1051	186	58	,	,	PUNCT
cril-1051	186	59	cotterell	cotterell	VERB
cril-1051	186	60	&	&	CCONJ
cril-1051	186	61	hulden	hulden	ADJ
cril-1051	186	62	2021	2021	NUM
cril-1051	186	63	)	)	PUNCT
cril-1051	186	64	.	.	PUNCT
cril-1051	187	1	6	6	X
cril-1051	187	2	.	.	X
cril-1051	187	3	conclusion	conclusion	NOUN
cril-1051	187	4	morphology	morphology	NOUN
cril-1051	187	5	comprises	comprise	VERB
cril-1051	187	6	word	word	NOUN
cril-1051	187	7	-	-	PUNCT
cril-1051	187	8	building	building	NOUN
cril-1051	187	9	properties	property	NOUN
cril-1051	187	10	in	in	ADP
cril-1051	187	11	human	human	ADJ
cril-1051	187	12	languages	language	NOUN
cril-1051	187	13	and	and	CCONJ
cril-1051	187	14	their	their	PRON
cril-1051	187	15	accompanying	accompany	VERB
cril-1051	187	16	(	(	PUNCT
cril-1051	187	17	morpho)syntactic	morpho)syntactic	ADJ
cril-1051	187	18	phenomena	phenomenon	NOUN
cril-1051	187	19	.	.	PUNCT
cril-1051	188	1	historically	historically	ADV
cril-1051	188	2	,	,	PUNCT
cril-1051	188	3	nlp	nlp	ADJ
cril-1051	188	4	and	and	CCONJ
cril-1051	188	5	“	"	PUNCT
cril-1051	188	6	paper	paper	NOUN
cril-1051	188	7	-	-	PUNCT
cril-1051	188	8	and	and	CCONJ
cril-1051	188	9	-	-	PUNCT
cril-1051	188	10	pencil	pencil	NOUN
cril-1051	188	11	linguists	linguist	NOUN
cril-1051	188	12	”	"	PUNCT
cril-1051	188	13	have	have	AUX
cril-1051	188	14	taken	take	VERB
cril-1051	188	15	different	different	ADJ
cril-1051	188	16	and	and	CCONJ
cril-1051	188	17	sometimes	sometimes	ADV
cril-1051	188	18	seemingly	seemingly	ADV
cril-1051	188	19	incompatible	incompatible	ADJ
cril-1051	188	20	approaches	approach	NOUN
cril-1051	188	21	to	to	ADP
cril-1051	188	22	morphology	morphology	NOUN
cril-1051	188	23	(	(	PUNCT
cril-1051	188	24	karttunen	karttunen	PROPN
cril-1051	188	25	&	&	CCONJ
cril-1051	188	26	beesley	beesley	PROPN
cril-1051	188	27	2005	2005	NUM
cril-1051	188	28	)	)	PUNCT
cril-1051	188	29	.	.	PUNCT
cril-1051	189	1	despite	despite	SCONJ
cril-1051	189	2	their	their	PRON
cril-1051	189	3	out	out	ADP
cril-1051	189	4	-	-	PUNCT
cril-1051	189	5	of	of	ADP
cril-1051	189	6	-	-	PUNCT
cril-1051	189	7	sync	sync	NOUN
cril-1051	189	8	approaches	approach	NOUN
cril-1051	189	9	,	,	PUNCT
cril-1051	189	10	both	both	PRON
cril-1051	189	11	benefit	benefit	VERB
cril-1051	189	12	from	from	ADP
cril-1051	189	13	morphological	morphological	ADJ
cril-1051	189	14	analysis	analysis	NOUN
cril-1051	189	15	(	(	PUNCT
cril-1051	189	16	cotterell	cotterell	NOUN
cril-1051	189	17	et	et	PROPN
cril-1051	189	18	al	al	PROPN
cril-1051	189	19	.	.	PROPN
cril-1051	189	20	2015	2015	NUM
cril-1051	189	21	)	)	PUNCT
cril-1051	189	22	.	.	PUNCT
cril-1051	190	1	morphological	morphological	ADJ
cril-1051	190	2	analysis	analysis	NOUN
cril-1051	190	3	includes	include	VERB
cril-1051	190	4	morpheme	morpheme	ADJ
cril-1051	190	5	segmentation	segmentation	NOUN
cril-1051	190	6	,	,	PUNCT
cril-1051	190	7	glossing	gloss	VERB
cril-1051	190	8	,	,	PUNCT
cril-1051	190	9	and	and	CCONJ
cril-1051	190	10	learning	learn	VERB
cril-1051	190	11	inflectional	inflectional	ADJ
cril-1051	190	12	paradigmatic	paradigmatic	ADJ
cril-1051	190	13	patterns	pattern	NOUN
cril-1051	190	14	.	.	PUNCT
cril-1051	191	1	this	this	DET
cril-1051	191	2	paper	paper	NOUN
cril-1051	191	3	presented	present	VERB
cril-1051	191	4	the	the	DET
cril-1051	191	5	history	history	NOUN
cril-1051	191	6	of	of	ADP
cril-1051	191	7	computational	computational	ADJ
cril-1051	191	8	models	model	NOUN
cril-1051	191	9	for	for	ADP
cril-1051	191	10	morphological	morphological	ADJ
cril-1051	191	11	analysis	analysis	NOUN
cril-1051	191	12	and	and	CCONJ
cril-1051	191	13	looked	look	VERB
cril-1051	191	14	specifically	specifically	ADV
cril-1051	191	15	at	at	ADP
cril-1051	191	16	their	their	PRON
cril-1051	191	17	application	application	NOUN
cril-1051	191	18	and	and	CCONJ
cril-1051	191	19	success	success	NOUN
cril-1051	191	20	when	when	SCONJ
cril-1051	191	21	limited	limited	ADJ
cril-1051	191	22	training	training	NOUN
cril-1051	191	23	data	datum	NOUN
cril-1051	191	24	is	be	AUX
cril-1051	191	25	available	available	ADJ
cril-1051	191	26	.	.	PUNCT
cril-1051	192	1	the	the	DET
cril-1051	192	2	work	work	NOUN
cril-1051	192	3	discussed	discuss	VERB
cril-1051	192	4	in	in	ADP
cril-1051	192	5	this	this	DET
cril-1051	192	6	paper	paper	NOUN
cril-1051	192	7	demonstrate	demonstrate	NOUN
cril-1051	192	8	that	that	SCONJ
cril-1051	192	9	computational	computational	ADJ
cril-1051	192	10	models	model	NOUN
cril-1051	192	11	colorado	colorado	PROPN
cril-1051	192	12	research	research	PROPN
cril-1051	192	13	in	in	ADP
cril-1051	192	14	linguistics	linguistic	NOUN
cril-1051	192	15	,	,	PUNCT
cril-1051	192	16	volume	volume	NOUN
cril-1051	192	17	25	25	NUM
cril-1051	192	18	(	(	PUNCT
cril-1051	192	19	2021	2021	NUM
cril-1051	192	20	)	)	PUNCT
cril-1051	192	21	16	16	NUM
cril-1051	192	22	and	and	CCONJ
cril-1051	192	23	methods	method	NOUN
cril-1051	192	24	can	can	AUX
cril-1051	192	25	both	both	PRON
cril-1051	192	26	successfully	successfully	ADV
cril-1051	192	27	perform	perform	VERB
cril-1051	192	28	morphological	morphological	ADJ
cril-1051	192	29	analysis	analysis	NOUN
cril-1051	192	30	.	.	PUNCT
cril-1051	193	1	this	this	DET
cril-1051	193	2	success	success	NOUN
cril-1051	193	3	may	may	AUX
cril-1051	193	4	be	be	AUX
cril-1051	193	5	of	of	ADP
cril-1051	193	6	great	great	ADJ
cril-1051	193	7	benefit	benefit	NOUN
cril-1051	193	8	for	for	ADP
cril-1051	193	9	the	the	DET
cril-1051	193	10	documentation	documentation	NOUN
cril-1051	193	11	and	and	CCONJ
cril-1051	193	12	description	description	NOUN
cril-1051	193	13	of	of	ADP
cril-1051	193	14	endangered	endangered	ADJ
cril-1051	193	15	or	or	CCONJ
cril-1051	193	16	under	under	ADV
cril-1051	193	17	-	-	PUNCT
cril-1051	193	18	documented	document	VERB
cril-1051	193	19	languages	language	NOUN
cril-1051	193	20	.	.	PUNCT
cril-1051	194	1	references	reference	NOUN
cril-1051	194	2	ackerman	ackerman	PROPN
cril-1051	194	3	,	,	PUNCT
cril-1051	194	4	farrell	farrell	PROPN
cril-1051	194	5	,	,	PUNCT
cril-1051	194	6	james	james	PROPN
cril-1051	194	7	p.	p.	PROPN
cril-1051	194	8	blevins	blevins	PROPN
cril-1051	194	9	&	&	CCONJ
cril-1051	194	10	robert	robert	PROPN
cril-1051	194	11	malouf	malouf	PROPN
cril-1051	194	12	.	.	PUNCT
cril-1051	195	1	2009	2009	NUM
cril-1051	195	2	.	.	PUNCT
cril-1051	196	1	parts	part	NOUN
cril-1051	196	2	and	and	CCONJ
cril-1051	196	3	wholes	whole	NOUN
cril-1051	196	4	:	:	PUNCT
cril-1051	196	5	implicative	implicative	ADJ
cril-1051	196	6	patterns	pattern	NOUN
cril-1051	196	7	in	in	ADP
cril-1051	196	8	inflectional	inflectional	ADJ
cril-1051	196	9	paradigms	paradigm	NOUN
cril-1051	196	10	.	.	PUNCT
cril-1051	197	1	in	in	ADP
cril-1051	197	2	analogy	analogy	NOUN
cril-1051	197	3	in	in	ADP
cril-1051	197	4	grammar	grammar	NOUN
cril-1051	197	5	.	.	PUNCT
cril-1051	198	1	oxford	oxford	NOUN
cril-1051	198	2	:	:	PUNCT
cril-1051	198	3	oxford	oxford	PROPN
cril-1051	198	4	university	university	PROPN
cril-1051	198	5	press	press	NOUN
cril-1051	198	6	.	.	PUNCT
cril-1051	199	1	aharoni	aharoni	NOUN
cril-1051	199	2	,	,	PUNCT
cril-1051	199	3	roee	roee	PROPN
cril-1051	199	4	&	&	CCONJ
cril-1051	199	5	yoav	yoav	PROPN
cril-1051	199	6	goldberg	goldberg	PROPN
cril-1051	199	7	.	.	PROPN
cril-1051	199	8	2017	2017	NUM
cril-1051	199	9	.	.	PUNCT
cril-1051	200	1	morphological	morphological	ADJ
cril-1051	200	2	inflection	inflection	NOUN
cril-1051	200	3	generation	generation	NOUN
cril-1051	200	4	with	with	ADP
cril-1051	200	5	hard	hard	ADJ
cril-1051	200	6	monotonic	monotonic	ADJ
cril-1051	200	7	attention	attention	NOUN
cril-1051	200	8	.	.	PUNCT
cril-1051	201	1	proceedings	proceeding	NOUN
cril-1051	201	2	of	of	ADP
cril-1051	201	3	the	the	DET
cril-1051	201	4	55th	55th	ADJ
cril-1051	201	5	annual	annual	ADJ
cril-1051	201	6	meeting	meeting	NOUN
cril-1051	201	7	of	of	ADP
cril-1051	201	8	the	the	DET
cril-1051	201	9	association	association	NOUN
cril-1051	201	10	for	for	ADP
cril-1051	201	11	computational	computational	ADJ
cril-1051	201	12	linguistics	linguistic	NOUN
cril-1051	201	13	(	(	PUNCT
cril-1051	201	14	volume	volume	NOUN
cril-1051	201	15	1	1	NUM
cril-1051	201	16	:	:	PUNCT
cril-1051	201	17	long	long	ADJ
cril-1051	201	18	papers	paper	NOUN
cril-1051	201	19	)	)	PUNCT
cril-1051	201	20	1	1	NUM
cril-1051	201	21	.	.	X
cril-1051	201	22	2004–2015	2004–2015	NUM
cril-1051	201	23	.	.	PUNCT
cril-1051	202	1	https://aclanthology.coli.uni-saarland.de/papers/p17-1183/p17-1183	https://aclanthology.coli.uni-saarland.de/papers/p17-1183/p17-1183	PROPN
cril-1051	202	2	(	(	PUNCT
cril-1051	202	3	16	16	NUM
cril-1051	202	4	january	january	PROPN
cril-1051	202	5	,	,	PUNCT
cril-1051	202	6	2018	2018	NUM
cril-1051	202	7	)	)	PUNCT
cril-1051	202	8	.	.	PUNCT
cril-1051	203	1	ahlberg	ahlberg	PROPN
cril-1051	203	2	,	,	PUNCT
cril-1051	203	3	malin	malin	PROPN
cril-1051	203	4	,	,	PUNCT
cril-1051	203	5	markus	markus	PROPN
cril-1051	203	6	forsberg	forsberg	PROPN
cril-1051	203	7	&	&	CCONJ
cril-1051	203	8	mans	mans	PROPN
cril-1051	203	9	hulden	hulden	VERB
cril-1051	203	10	.	.	PUNCT
cril-1051	204	1	2014	2014	NUM
cril-1051	204	2	.	.	PUNCT
cril-1051	205	1	semi	semi	ADJ
cril-1051	205	2	-	-	ADJ
cril-1051	205	3	supervised	supervised	ADJ
cril-1051	205	4	learning	learning	NOUN
cril-1051	205	5	of	of	ADP
cril-1051	205	6	morphological	morphological	ADJ
cril-1051	205	7	paradigms	paradigm	NOUN
cril-1051	205	8	and	and	CCONJ
cril-1051	205	9	lexicons	lexicon	NOUN
cril-1051	205	10	.	.	PUNCT
cril-1051	206	1	in	in	ADP
cril-1051	206	2	proceedings	proceeding	NOUN
cril-1051	206	3	of	of	ADP
cril-1051	206	4	14th	14th	ADJ
cril-1051	206	5	conference	conference	NOUN
cril-1051	206	6	of	of	ADP
cril-1051	206	7	the	the	DET
cril-1051	206	8	european	european	ADJ
cril-1051	206	9	chapter	chapter	NOUN
cril-1051	206	10	of	of	ADP
cril-1051	206	11	the	the	DET
cril-1051	206	12	association	association	NOUN
cril-1051	206	13	for	for	ADP
cril-1051	206	14	computational	computational	ADJ
cril-1051	206	15	linguistics	linguistic	NOUN
cril-1051	206	16	,	,	PUNCT
cril-1051	206	17	569–578	569–578	NUM
cril-1051	206	18	.	.	PUNCT
cril-1051	206	19	gothenburg	gothenburg	PROPN
cril-1051	206	20	,	,	PUNCT
cril-1051	206	21	sweden	sweden	PROPN
cril-1051	206	22	:	:	PUNCT
cril-1051	206	23	association	association	NOUN
cril-1051	206	24	for	for	ADP
cril-1051	206	25	computational	computational	ADJ
cril-1051	206	26	linguistics	linguistic	NOUN
cril-1051	206	27	.	.	PUNCT
cril-1051	207	1	http://www.aclweb.org/anthology/e/e14/e14-1.pdf#page=569	http://www.aclweb.org/anthology/e/e14/e14-1.pdf#page=569	PROPN
cril-1051	207	2	(	(	PUNCT
cril-1051	207	3	1	1	NUM
cril-1051	207	4	november	november	PROPN
cril-1051	207	5	,	,	PUNCT
cril-1051	207	6	2016	2016	NUM
cril-1051	207	7	)	)	PUNCT
cril-1051	207	8	.	.	PUNCT
cril-1051	208	1	ahlberg	ahlberg	PROPN
cril-1051	208	2	,	,	PUNCT
cril-1051	208	3	malin	malin	PROPN
cril-1051	208	4	,	,	PUNCT
cril-1051	208	5	markus	markus	PROPN
cril-1051	208	6	forsberg	forsberg	PROPN
cril-1051	208	7	&	&	CCONJ
cril-1051	208	8	mans	mans	PROPN
cril-1051	208	9	hulden	hulden	VERB
cril-1051	208	10	.	.	PUNCT
cril-1051	209	1	2015	2015	NUM
cril-1051	209	2	.	.	PUNCT
cril-1051	210	1	paradigm	paradigm	NOUN
cril-1051	210	2	classification	classification	NOUN
cril-1051	210	3	in	in	ADP
cril-1051	210	4	supervised	supervised	ADJ
cril-1051	210	5	learning	learning	NOUN
cril-1051	210	6	of	of	ADP
cril-1051	210	7	morphology	morphology	NOUN
cril-1051	210	8	.	.	PUNCT
cril-1051	211	1	in	in	ADP
cril-1051	211	2	proceedings	proceeding	NOUN
cril-1051	211	3	of	of	ADP
cril-1051	211	4	the	the	DET
cril-1051	211	5	2015	2015	NUM
cril-1051	211	6	conference	conference	NOUN
cril-1051	211	7	of	of	ADP
cril-1051	211	8	the	the	DET
cril-1051	211	9	north	north	ADJ
cril-1051	211	10	american	american	ADJ
cril-1051	211	11	chapter	chapter	NOUN
cril-1051	211	12	of	of	ADP
cril-1051	211	13	the	the	DET
cril-1051	211	14	association	association	NOUN
cril-1051	211	15	for	for	ADP
cril-1051	211	16	computational	computational	ADJ
cril-1051	211	17	linguistics	linguistic	NOUN
cril-1051	211	18	:	:	PUNCT
cril-1051	211	19	human	human	ADJ
cril-1051	211	20	language	language	NOUN
cril-1051	211	21	technologies	technology	NOUN
cril-1051	211	22	,	,	PUNCT
cril-1051	211	23	1024–1029	1024–1029	NUM
cril-1051	211	24	.	.	PUNCT
cril-1051	212	1	denver	denver	PROPN
cril-1051	212	2	,	,	PUNCT
cril-1051	212	3	colorado	colorado	PROPN
cril-1051	212	4	:	:	PUNCT
cril-1051	212	5	association	association	NOUN
cril-1051	212	6	for	for	ADP
cril-1051	212	7	computational	computational	ADJ
cril-1051	212	8	linguistics	linguistic	NOUN
cril-1051	212	9	.	.	PUNCT
cril-1051	213	1	http://www.aclweb.org/anthology/n15-1107	http://www.aclweb.org/anthology/n15-1107	NOUN
cril-1051	213	2	(	(	PUNCT
cril-1051	213	3	18	18	NUM
cril-1051	213	4	january	january	PROPN
cril-1051	213	5	,	,	PUNCT
cril-1051	213	6	2019	2019	NUM
cril-1051	213	7	)	)	PUNCT
cril-1051	213	8	.	.	PUNCT
cril-1051	214	1	anastasopoulos	anastasopoulos	PROPN
cril-1051	214	2	,	,	PUNCT
cril-1051	214	3	antonios	antonios	PROPN
cril-1051	214	4	,	,	PUNCT
cril-1051	214	5	david	david	PROPN
cril-1051	214	6	chiang	chiang	PROPN
cril-1051	214	7	&	&	CCONJ
cril-1051	214	8	long	long	PROPN
cril-1051	214	9	duong	duong	PROPN
cril-1051	214	10	.	.	PUNCT
cril-1051	215	1	2016	2016	NUM
cril-1051	215	2	.	.	PUNCT
cril-1051	216	1	an	an	DET
cril-1051	216	2	unsupervised	unsupervised	ADJ
cril-1051	216	3	probability	probability	NOUN
cril-1051	216	4	model	model	NOUN
cril-1051	216	5	for	for	ADP
cril-1051	216	6	speech	speech	NOUN
cril-1051	216	7	-	-	PUNCT
cril-1051	216	8	to	to	ADP
cril-1051	216	9	-	-	PUNCT
cril-1051	216	10	translation	translation	NOUN
cril-1051	216	11	alignment	alignment	NOUN
cril-1051	216	12	of	of	ADP
cril-1051	216	13	low	low	ADJ
cril-1051	216	14	-	-	PUNCT
cril-1051	216	15	resource	resource	NOUN
cril-1051	216	16	languages	language	NOUN
cril-1051	216	17	.	.	PUNCT
cril-1051	217	1	in	in	ADP
cril-1051	217	2	proceedings	proceeding	NOUN
cril-1051	217	3	of	of	ADP
cril-1051	217	4	the	the	DET
cril-1051	217	5	2016	2016	NUM
cril-1051	217	6	conference	conference	NOUN
cril-1051	217	7	on	on	ADP
cril-1051	217	8	empirical	empirical	ADJ
cril-1051	217	9	methods	method	NOUN
cril-1051	217	10	in	in	ADP
cril-1051	217	11	natural	natural	ADJ
cril-1051	217	12	language	language	NOUN
cril-1051	217	13	processing	processing	NOUN
cril-1051	217	14	,	,	PUNCT
cril-1051	217	15	1255–1263	1255–1263	NUM
cril-1051	217	16	.	.	PUNCT
cril-1051	218	1	austin	austin	PROPN
cril-1051	218	2	,	,	PUNCT
cril-1051	218	3	texas	texas	PROPN
cril-1051	218	4	:	:	PUNCT
cril-1051	218	5	association	association	NOUN
cril-1051	218	6	for	for	ADP
cril-1051	218	7	computational	computational	ADJ
cril-1051	218	8	linguistics	linguistic	NOUN
cril-1051	218	9	.	.	PUNCT
cril-1051	219	1	https://aclweb.org/anthology/d16-1133	https://aclweb.org/anthology/d16-1133	PROPN
cril-1051	219	2	(	(	PUNCT
cril-1051	219	3	2	2	NUM
cril-1051	219	4	july	july	PROPN
cril-1051	219	5	,	,	PUNCT
cril-1051	219	6	2018	2018	NUM
cril-1051	219	7	)	)	PUNCT
cril-1051	219	8	.	.	PUNCT
cril-1051	220	1	ansari	ansari	PROPN
cril-1051	220	2	,	,	PUNCT
cril-1051	220	3	ebrahim	ebrahim	PROPN
cril-1051	220	4	,	,	PUNCT
cril-1051	220	5	zdeněk	zdeněk	PROPN
cril-1051	220	6	žabokrtský	žabokrtský	PROPN
cril-1051	220	7	,	,	PUNCT
cril-1051	220	8	mohammad	mohammad	PROPN
cril-1051	220	9	mahmoudi	mahmoudi	PROPN
cril-1051	220	10	,	,	PUNCT
cril-1051	220	11	hamid	hamid	PROPN
cril-1051	220	12	haghdoost	haghdoost	PROPN
cril-1051	220	13	&	&	CCONJ
cril-1051	220	14	jonáš	jonáš	ADJ
cril-1051	220	15	vidra	vidra	NOUN
cril-1051	220	16	.	.	PUNCT
cril-1051	221	1	2019	2019	NUM
cril-1051	221	2	.	.	PUNCT
cril-1051	222	1	supervised	supervise	VERB
cril-1051	222	2	morphological	morphological	ADJ
cril-1051	222	3	segmentation	segmentation	NOUN
cril-1051	222	4	using	use	VERB
cril-1051	222	5	rich	rich	ADJ
cril-1051	222	6	annotated	annotate	VERB
cril-1051	222	7	lexicon	lexicon	NOUN
cril-1051	222	8	.	.	PUNCT
cril-1051	223	1	in	in	ADP
cril-1051	223	2	proceedings	proceeding	NOUN
cril-1051	223	3	of	of	ADP
cril-1051	223	4	the	the	DET
cril-1051	223	5	international	international	ADJ
cril-1051	223	6	conference	conference	NOUN
cril-1051	223	7	on	on	ADP
cril-1051	223	8	recent	recent	ADJ
cril-1051	223	9	advances	advance	NOUN
cril-1051	223	10	in	in	ADP
cril-1051	223	11	natural	natural	ADJ
cril-1051	223	12	language	language	NOUN
cril-1051	223	13	processing	processing	NOUN
cril-1051	223	14	(	(	PUNCT
cril-1051	223	15	ranlp	ranlp	PROPN
cril-1051	223	16	2019	2019	NUM
cril-1051	223	17	)	)	PUNCT
cril-1051	223	18	,	,	PUNCT
cril-1051	223	19	52–61	52–61	NUM
cril-1051	223	20	.	.	PUNCT
cril-1051	223	21	varna	varna	PROPN
cril-1051	223	22	,	,	PUNCT
cril-1051	223	23	bulgaria	bulgaria	PROPN
cril-1051	223	24	:	:	PUNCT
cril-1051	223	25	incoma	incoma	PROPN
cril-1051	223	26	ltd	ltd	PROPN
cril-1051	223	27	.	.	PROPN
cril-1051	224	1	https://www.aclweb.org/anthology/r19-1007	https://www.aclweb.org/anthology/r19-1007	PUNCT
cril-1051	224	2	(	(	PUNCT
cril-1051	224	3	25	25	NUM
cril-1051	224	4	january	january	PROPN
cril-1051	224	5	,	,	PUNCT
cril-1051	224	6	2020	2020	NUM
cril-1051	224	7	)	)	PUNCT
cril-1051	224	8	.	.	PUNCT
cril-1051	225	1	computational	computational	ADJ
cril-1051	225	2	morphology	morphology	NOUN
cril-1051	225	3	for	for	ADP
cril-1051	225	4	language	language	NOUN
cril-1051	225	5	description	description	NOUN
cril-1051	225	6	and	and	CCONJ
cril-1051	225	7	documentation	documentation	NOUN
cril-1051	225	8	17	17	NUM
cril-1051	225	9	auer	auer	NOUN
cril-1051	225	10	,	,	PUNCT
cril-1051	225	11	eric	eric	PROPN
cril-1051	225	12	,	,	PUNCT
cril-1051	225	13	albert	albert	PROPN
cril-1051	225	14	russel	russel	PROPN
cril-1051	225	15	,	,	PUNCT
cril-1051	225	16	han	han	PROPN
cril-1051	225	17	sloetjes	sloetjes	PROPN
cril-1051	225	18	,	,	PUNCT
cril-1051	225	19	peter	peter	PROPN
cril-1051	225	20	wittenburg	wittenburg	PROPN
cril-1051	225	21	,	,	PUNCT
cril-1051	225	22	oliver	oliver	PROPN
cril-1051	225	23	schreer	schreer	PROPN
cril-1051	225	24	,	,	PUNCT
cril-1051	225	25	s.	s.	PROPN
cril-1051	225	26	masnieri	masnieri	PROPN
cril-1051	225	27	,	,	PUNCT
cril-1051	225	28	daniel	daniel	PROPN
cril-1051	225	29	schneider	schneider	PROPN
cril-1051	225	30	&	&	CCONJ
cril-1051	225	31	sebastian	sebastian	PROPN
cril-1051	225	32	tschöpel	tschöpel	PROPN
cril-1051	225	33	.	.	PUNCT
cril-1051	226	1	2010	2010	NUM
cril-1051	226	2	.	.	PUNCT
cril-1051	227	1	elan	elan	PROPN
cril-1051	227	2	as	as	ADP
cril-1051	227	3	flexible	flexible	ADJ
cril-1051	227	4	annotation	annotation	NOUN
cril-1051	227	5	framework	framework	NOUN
cril-1051	227	6	for	for	ADP
cril-1051	227	7	sound	sound	NOUN
cril-1051	227	8	and	and	CCONJ
cril-1051	227	9	image	image	NOUN
cril-1051	227	10	processing	processing	NOUN
cril-1051	227	11	detectors	detector	NOUN
cril-1051	227	12	.	.	PUNCT
cril-1051	228	1	in	in	ADP
cril-1051	228	2	nicoletta	nicoletta	PROPN
cril-1051	228	3	calzolari	calzolari	PROPN
cril-1051	228	4	,	,	PUNCT
cril-1051	228	5	khalid	khalid	PROPN
cril-1051	228	6	choukri	choukri	PROPN
cril-1051	228	7	,	,	PUNCT
cril-1051	228	8	bente	bente	PROPN
cril-1051	228	9	maegaard	maegaard	PROPN
cril-1051	228	10	,	,	PUNCT
cril-1051	228	11	joseph	joseph	PROPN
cril-1051	228	12	mariani	mariani	PROPN
cril-1051	228	13	,	,	PUNCT
cril-1051	228	14	jan	jan	PROPN
cril-1051	228	15	odijk	odijk	PROPN
cril-1051	228	16	,	,	PUNCT
cril-1051	228	17	stelios	stelios	PROPN
cril-1051	228	18	piperidis	piperidis	PROPN
cril-1051	228	19	,	,	PUNCT
cril-1051	228	20	mike	mike	PROPN
cril-1051	228	21	rosner	rosner	PROPN
cril-1051	228	22	&	&	CCONJ
cril-1051	228	23	daniel	daniel	PROPN
cril-1051	228	24	tapias	tapias	PROPN
cril-1051	228	25	(	(	PUNCT
cril-1051	228	26	eds	ed	NOUN
cril-1051	228	27	.	.	PUNCT
cril-1051	228	28	)	)	PUNCT
cril-1051	228	29	,	,	PUNCT
cril-1051	228	30	european	european	ADJ
cril-1051	228	31	language	language	NOUN
cril-1051	228	32	resources	resource	NOUN
cril-1051	228	33	association	association	NOUN
cril-1051	228	34	lrec	lrec	NOUN
cril-1051	228	35	2010	2010	NUM
cril-1051	228	36	:	:	PUNCT
cril-1051	228	37	proceedings	proceeding	NOUN
cril-1051	228	38	of	of	ADP
cril-1051	228	39	the	the	DET
cril-1051	228	40	7th	7th	ADJ
cril-1051	228	41	international	international	ADJ
cril-1051	228	42	language	language	NOUN
cril-1051	228	43	resources	resource	NOUN
cril-1051	228	44	and	and	CCONJ
cril-1051	228	45	evaluation	evaluation	NOUN
cril-1051	228	46	,	,	PUNCT
cril-1051	228	47	890–893	890–893	NUM
cril-1051	228	48	.	.	PUNCT
cril-1051	229	1	paris	paris	PROPN
cril-1051	229	2	:	:	PUNCT
cril-1051	229	3	elra	elra	ADV
cril-1051	229	4	:	:	PUNCT
cril-1051	229	5	european	european	ADJ
cril-1051	229	6	language	language	PROPN
cril-1051	229	7	resources	resource	NOUN
cril-1051	229	8	association	association	NOUN
cril-1051	229	9	.	.	PUNCT
cril-1051	230	1	http://dblp.unitrier.de/db/conf/lrec/lrec2010.html#auerrswsmst10	http://dblp.unitrier.de/db/conf/lrec/lrec2010.html#auerrswsmst10	PROPN
cril-1051	230	2	(	(	PUNCT
cril-1051	230	3	30	30	NUM
cril-1051	230	4	january	january	PROPN
cril-1051	230	5	,	,	PUNCT
cril-1051	230	6	2014	2014	NUM
cril-1051	230	7	)	)	PUNCT
cril-1051	230	8	.	.	PUNCT
cril-1051	231	1	baldridge	baldridge	PROPN
cril-1051	231	2	,	,	PUNCT
cril-1051	231	3	jason	jason	PROPN
cril-1051	231	4	&	&	CCONJ
cril-1051	231	5	miles	miles	PROPN
cril-1051	231	6	osborne	osborne	PROPN
cril-1051	231	7	.	.	PUNCT
cril-1051	232	1	2008	2008	NUM
cril-1051	232	2	.	.	PUNCT
cril-1051	233	1	active	active	ADJ
cril-1051	233	2	learning	learning	NOUN
cril-1051	233	3	and	and	CCONJ
cril-1051	233	4	logarithmic	logarithmic	ADJ
cril-1051	233	5	opinion	opinion	NOUN
cril-1051	233	6	pools	pool	NOUN
cril-1051	233	7	for	for	ADP
cril-1051	233	8	hpsg	hpsg	ADJ
cril-1051	233	9	parse	parse	NOUN
cril-1051	233	10	selection	selection	NOUN
cril-1051	233	11	.	.	PUNCT
cril-1051	234	1	natural	natural	ADJ
cril-1051	234	2	language	language	NOUN
cril-1051	234	3	engineering	engineering	NOUN
cril-1051	234	4	14(2	14(2	NUM
cril-1051	234	5	)	)	PUNCT
cril-1051	234	6	.	.	PUNCT
cril-1051	235	1	191–222	191–222	NUM
cril-1051	235	2	.	.	PUNCT
cril-1051	235	3	baldridge	baldridge	PROPN
cril-1051	235	4	,	,	PUNCT
cril-1051	235	5	jason	jason	PROPN
cril-1051	235	6	&	&	CCONJ
cril-1051	235	7	alexis	alexis	PROPN
cril-1051	235	8	palmer	palmer	PROPN
cril-1051	235	9	.	.	PUNCT
cril-1051	235	10	2009	2009	NUM
cril-1051	235	11	.	.	PUNCT
cril-1051	236	1	how	how	SCONJ
cril-1051	236	2	well	well	ADV
cril-1051	236	3	does	do	AUX
cril-1051	236	4	active	active	ADJ
cril-1051	236	5	learning	learning	NOUN
cril-1051	236	6	actually	actually	ADV
cril-1051	236	7	work	work	VERB
cril-1051	236	8	?	?	PUNCT
cril-1051	237	1	timebased	timebase	VERB
cril-1051	237	2	evaluation	evaluation	NOUN
cril-1051	237	3	of	of	ADP
cril-1051	237	4	cost	cost	NOUN
cril-1051	237	5	-	-	PUNCT
cril-1051	237	6	reduction	reduction	NOUN
cril-1051	237	7	strategies	strategy	NOUN
cril-1051	237	8	for	for	ADP
cril-1051	237	9	language	language	NOUN
cril-1051	237	10	documentation	documentation	NOUN
cril-1051	237	11	.	.	PUNCT
cril-1051	238	1	in	in	ADP
cril-1051	238	2	proceedings	proceeding	NOUN
cril-1051	238	3	of	of	ADP
cril-1051	238	4	the	the	DET
cril-1051	238	5	2009	2009	NUM
cril-1051	238	6	conference	conference	NOUN
cril-1051	238	7	on	on	ADP
cril-1051	238	8	empirical	empirical	ADJ
cril-1051	238	9	methods	method	NOUN
cril-1051	238	10	in	in	ADP
cril-1051	238	11	natural	natural	ADJ
cril-1051	238	12	language	language	NOUN
cril-1051	238	13	processing	processing	NOUN
cril-1051	238	14	,	,	PUNCT
cril-1051	238	15	296	296	NUM
cril-1051	238	16	–	–	PUNCT
cril-1051	238	17	305	305	NUM
cril-1051	238	18	.	.	PUNCT
cril-1051	238	19	singapore	singapore	PROPN
cril-1051	238	20	.	.	PUNCT
cril-1051	239	1	http://www.aclweb.org/anthology/d/d09/d09-1031.pdf	http://www.aclweb.org/anthology/d/d09/d09-1031.pdf	PROPN
cril-1051	239	2	.	.	PUNCT
cril-1051	240	1	bender	bender	PROPN
cril-1051	240	2	,	,	PUNCT
cril-1051	240	3	emily	emily	PROPN
cril-1051	240	4	m.	m.	PROPN
cril-1051	240	5	2014	2014	NUM
cril-1051	240	6	.	.	PUNCT
cril-1051	241	1	language	language	NOUN
cril-1051	241	2	collage	collage	NOUN
cril-1051	241	3	:	:	PUNCT
cril-1051	241	4	grammatical	grammatical	ADJ
cril-1051	241	5	description	description	NOUN
cril-1051	241	6	with	with	ADP
cril-1051	241	7	the	the	DET
cril-1051	241	8	lingo	lingo	NOUN
cril-1051	241	9	grammar	grammar	NOUN
cril-1051	241	10	matrix	matrix	NOUN
cril-1051	241	11	.	.	PUNCT
cril-1051	242	1	in	in	ADP
cril-1051	242	2	proceedings	proceeding	NOUN
cril-1051	242	3	of	of	ADP
cril-1051	242	4	the	the	DET
cril-1051	242	5	ninth	ninth	ADJ
cril-1051	242	6	international	international	ADJ
cril-1051	242	7	conference	conference	NOUN
cril-1051	242	8	of	of	ADP
cril-1051	242	9	language	language	NOUN
cril-1051	242	10	resources	resource	NOUN
cril-1051	242	11	and	and	CCONJ
cril-1051	242	12	evaluation	evaluation	NOUN
cril-1051	242	13	(	(	PUNCT
cril-1051	242	14	lrec-2014	lrec-2014	NOUN
cril-1051	242	15	)	)	PUNCT
cril-1051	242	16	,	,	PUNCT
cril-1051	242	17	2447–2451	2447–2451	NUM
cril-1051	242	18	.	.	PUNCT
cril-1051	243	1	http://www.lrecconf.org/proceedings/lrec2014/pdf/639_paper.pdf	http://www.lrecconf.org/proceedings/lrec2014/pdf/639_paper.pdf	NOUN
cril-1051	243	2	.	.	PUNCT
cril-1051	244	1	bengio	bengio	PROPN
cril-1051	244	2	,	,	PUNCT
cril-1051	244	3	yoshua	yoshua	PROPN
cril-1051	244	4	,	,	PUNCT
cril-1051	244	5	réjean	réjean	ADJ
cril-1051	244	6	ducharme	ducharme	NOUN
cril-1051	244	7	,	,	PUNCT
cril-1051	244	8	pascal	pascal	PROPN
cril-1051	244	9	vincent	vincent	PROPN
cril-1051	244	10	&	&	CCONJ
cril-1051	244	11	christian	christian	PROPN
cril-1051	244	12	jauvin	jauvin	PROPN
cril-1051	244	13	.	.	PROPN
cril-1051	244	14	2003	2003	NUM
cril-1051	244	15	.	.	PUNCT
cril-1051	245	1	a	a	DET
cril-1051	245	2	neural	neural	ADJ
cril-1051	245	3	probabilistic	probabilistic	ADJ
cril-1051	245	4	language	language	NOUN
cril-1051	245	5	model	model	NOUN
cril-1051	245	6	.	.	PUNCT
cril-1051	246	1	journal	journal	PROPN
cril-1051	246	2	of	of	ADP
cril-1051	246	3	machine	machine	NOUN
cril-1051	246	4	learning	learn	VERB
cril-1051	246	5	research	research	NOUN
cril-1051	246	6	3	3	NUM
cril-1051	246	7	.	.	NOUN
cril-1051	246	8	1137–1155	1137–1155	NUM
cril-1051	246	9	.	.	PUNCT
cril-1051	246	10	bergmanis	bergmanis	NOUN
cril-1051	246	11	,	,	PUNCT
cril-1051	246	12	toms	tom	NOUN
cril-1051	246	13	,	,	PUNCT
cril-1051	246	14	katharina	katharina	PROPN
cril-1051	246	15	kann	kann	PROPN
cril-1051	246	16	,	,	PUNCT
cril-1051	246	17	hinrich	hinrich	PROPN
cril-1051	246	18	schütze	schütze	PROPN
cril-1051	246	19	&	&	CCONJ
cril-1051	246	20	sharon	sharon	PROPN
cril-1051	246	21	goldwater	goldwater	PROPN
cril-1051	246	22	.	.	PUNCT
cril-1051	246	23	2017	2017	NUM
cril-1051	246	24	.	.	PUNCT
cril-1051	247	1	training	train	VERB
cril-1051	247	2	data	datum	NOUN
cril-1051	247	3	augmentation	augmentation	NOUN
cril-1051	247	4	for	for	ADP
cril-1051	247	5	low	low	ADJ
cril-1051	247	6	-	-	PUNCT
cril-1051	247	7	resource	resource	NOUN
cril-1051	247	8	morphological	morphological	ADJ
cril-1051	247	9	inflection	inflection	NOUN
cril-1051	247	10	.	.	PUNCT
cril-1051	248	1	proceedings	proceeding	NOUN
cril-1051	248	2	of	of	ADP
cril-1051	248	3	the	the	DET
cril-1051	248	4	conll	conll	PROPN
cril-1051	248	5	sigmorphon	sigmorphon	PROPN
cril-1051	248	6	2017	2017	NUM
cril-1051	248	7	shared	share	VERB
cril-1051	248	8	task	task	NOUN
cril-1051	248	9	:	:	PUNCT
cril-1051	248	10	universal	universal	ADJ
cril-1051	248	11	morphological	morphological	ADJ
cril-1051	248	12	reinflection	reinflection	NOUN
cril-1051	248	13	31–39	31–39	NUM
cril-1051	248	14	.	.	PUNCT
cril-1051	249	1	https://aclanthology.coli.uni-saarland.de/papers/k17-2002/k17-2002	https://aclanthology.coli.uni-saarland.de/papers/k17-2002/k17-2002	NOUN
cril-1051	249	2	(	(	PUNCT
cril-1051	249	3	16	16	NUM
cril-1051	249	4	january	january	PROPN
cril-1051	249	5	,	,	PUNCT
cril-1051	249	6	2018	2018	NUM
cril-1051	249	7	)	)	PUNCT
cril-1051	249	8	.	.	PUNCT
cril-1051	250	1	biljon	biljon	NOUN
cril-1051	250	2	,	,	PUNCT
cril-1051	250	3	elan	elan	PROPN
cril-1051	250	4	van	van	PROPN
cril-1051	250	5	,	,	PUNCT
cril-1051	250	6	arnu	arnu	PROPN
cril-1051	250	7	pretorius	pretorius	PROPN
cril-1051	250	8	&	&	CCONJ
cril-1051	250	9	julia	julia	PROPN
cril-1051	250	10	kreutzer	kreutzer	PROPN
cril-1051	250	11	.	.	PUNCT
cril-1051	251	1	2020	2020	NUM
cril-1051	251	2	.	.	PUNCT
cril-1051	252	1	on	on	ADP
cril-1051	252	2	optimal	optimal	ADJ
cril-1051	252	3	transformer	transformer	ADJ
cril-1051	252	4	depth	depth	NOUN
cril-1051	252	5	for	for	ADP
cril-1051	252	6	low	low	ADJ
cril-1051	252	7	-	-	PUNCT
cril-1051	252	8	resource	resource	NOUN
cril-1051	252	9	language	language	NOUN
cril-1051	252	10	translation	translation	NOUN
cril-1051	252	11	.	.	PUNCT
cril-1051	253	1	arxiv:2004.04418	arxiv:2004.04418	NOUN
cril-1051	254	1	[	[	X
cril-1051	254	2	cs	cs	X
cril-1051	254	3	]	]	X
cril-1051	254	4	.	.	PUNCT
cril-1051	255	1	http://arxiv.org/abs/2004.04418	http://arxiv.org/abs/2004.04418	PROPN
cril-1051	255	2	(	(	PUNCT
cril-1051	255	3	14	14	NUM
cril-1051	255	4	may	may	AUX
cril-1051	255	5	,	,	PUNCT
cril-1051	255	6	2020	2020	NUM
cril-1051	255	7	)	)	PUNCT
cril-1051	255	8	.	.	PUNCT
cril-1051	256	1	bird	bird	NOUN
cril-1051	256	2	,	,	PUNCT
cril-1051	256	3	steven	steven	PROPN
cril-1051	256	4	.	.	PROPN
cril-1051	256	5	2009	2009	NUM
cril-1051	256	6	.	.	PUNCT
cril-1051	256	7	natural	natural	ADJ
cril-1051	256	8	language	language	NOUN
cril-1051	256	9	processing	processing	NOUN
cril-1051	256	10	and	and	CCONJ
cril-1051	256	11	linguistic	linguistic	ADJ
cril-1051	256	12	fieldwork	fieldwork	NOUN
cril-1051	256	13	.	.	PUNCT
cril-1051	257	1	computational	computational	ADJ
cril-1051	257	2	linguistics	linguistic	NOUN
cril-1051	257	3	35(3	35(3	NUM
cril-1051	257	4	)	)	PUNCT
cril-1051	257	5	.	.	PUNCT
cril-1051	258	1	469–474	469–474	NUM
cril-1051	258	2	.	.	PUNCT
cril-1051	259	1	http://dx.doi.org/10.1162/coli.35.3.469	http://dx.doi.org/10.1162/coli.35.3.469	NOUN
cril-1051	259	2	(	(	PUNCT
cril-1051	259	3	16	16	NUM
cril-1051	259	4	march	march	NOUN
cril-1051	259	5	,	,	PUNCT
cril-1051	259	6	2015	2015	NUM
cril-1051	259	7	)	)	PUNCT
cril-1051	259	8	.	.	PUNCT
cril-1051	260	1	bird	bird	NOUN
cril-1051	260	2	,	,	PUNCT
cril-1051	260	3	steven	steven	PROPN
cril-1051	260	4	&	&	CCONJ
cril-1051	260	5	david	david	PROPN
cril-1051	260	6	chiang	chiang	PROPN
cril-1051	260	7	.	.	PUNCT
cril-1051	260	8	2012	2012	NUM
cril-1051	260	9	.	.	PUNCT
cril-1051	261	1	machine	machine	NOUN
cril-1051	261	2	translation	translation	NOUN
cril-1051	261	3	for	for	ADP
cril-1051	261	4	language	language	NOUN
cril-1051	261	5	preservation	preservation	NOUN
cril-1051	261	6	.	.	PUNCT
cril-1051	262	1	in	in	ADP
cril-1051	262	2	proceedings	proceeding	NOUN
cril-1051	262	3	of	of	ADP
cril-1051	262	4	coling	cole	VERB
cril-1051	262	5	2012	2012	NUM
cril-1051	262	6	,	,	PUNCT
cril-1051	262	7	125–134	125–134	NUM
cril-1051	262	8	.	.	PUNCT
cril-1051	262	9	mumbai	mumbai	PROPN
cril-1051	262	10	.	.	PUNCT
cril-1051	263	1	colorado	colorado	PROPN
cril-1051	263	2	research	research	PROPN
cril-1051	263	3	in	in	ADP
cril-1051	263	4	linguistics	linguistic	NOUN
cril-1051	263	5	,	,	PUNCT
cril-1051	263	6	volume	volume	NOUN
cril-1051	263	7	25	25	NUM
cril-1051	263	8	(	(	PUNCT
cril-1051	263	9	2021	2021	NUM
cril-1051	263	10	)	)	PUNCT
cril-1051	263	11	18	18	NUM
cril-1051	263	12	boerger	boerger	NOUN
cril-1051	263	13	,	,	PUNCT
cril-1051	263	14	brenda	brenda	PROPN
cril-1051	263	15	h.	h.	PROPN
cril-1051	263	16	,	,	PUNCT
cril-1051	263	17	sarah	sarah	PROPN
cril-1051	263	18	ruth	ruth	PROPN
cril-1051	263	19	moeller	moeller	PROPN
cril-1051	263	20	,	,	PUNCT
cril-1051	263	21	will	will	AUX
cril-1051	263	22	reiman	reiman	VERB
cril-1051	263	23	&	&	CCONJ
cril-1051	263	24	stephen	stephen	PROPN
cril-1051	263	25	self	self	PROPN
cril-1051	263	26	.	.	PUNCT
cril-1051	264	1	2016	2016	NUM
cril-1051	264	2	.	.	PUNCT
cril-1051	265	1	language	language	NOUN
cril-1051	265	2	and	and	CCONJ
cril-1051	265	3	culture	culture	NOUN
cril-1051	265	4	documentation	documentation	NOUN
cril-1051	265	5	manual	manual	NOUN
cril-1051	265	6	.	.	PUNCT
cril-1051	266	1	leanpub	leanpub	VERB
cril-1051	266	2	.	.	PUNCT
cril-1051	267	1	https://leanpub.com/languageandculturedocumentationmanual	https://leanpub.com/languageandculturedocumentationmanual	PUNCT
cril-1051	268	1	(	(	PUNCT
cril-1051	268	2	22	22	NUM
cril-1051	268	3	may	may	AUX
cril-1051	268	4	,	,	PUNCT
cril-1051	268	5	2020	2020	NUM
cril-1051	268	6	)	)	PUNCT
cril-1051	268	7	.	.	PUNCT
cril-1051	269	1	bowern	bowern	PROPN
cril-1051	269	2	,	,	PUNCT
cril-1051	269	3	claire	claire	NOUN
cril-1051	269	4	.	.	PUNCT
cril-1051	270	1	2008	2008	NUM
cril-1051	270	2	.	.	PUNCT
cril-1051	271	1	linguistic	linguistic	ADJ
cril-1051	271	2	fieldwork	fieldwork	PROPN
cril-1051	271	3	:	:	PUNCT
cril-1051	271	4	a	a	DET
cril-1051	271	5	practical	practical	ADJ
cril-1051	271	6	guide	guide	NOUN
cril-1051	271	7	.	.	PUNCT
cril-1051	272	1	houndmills	houndmill	NOUN
cril-1051	272	2	,	,	PUNCT
cril-1051	272	3	basingstoke	basingstoke	PROPN
cril-1051	272	4	,	,	PUNCT
cril-1051	272	5	hampshire	hampshire	PROPN
cril-1051	273	1	[	[	X
cril-1051	273	2	england	england	X
cril-1051	273	3	]	]	X
cril-1051	273	4	;	;	PUNCT
cril-1051	273	5	new	new	PROPN
cril-1051	273	6	york	york	PROPN
cril-1051	273	7	:	:	PUNCT
cril-1051	273	8	palgrave	palgrave	PROPN
cril-1051	273	9	macmillan	macmillan	PROPN
cril-1051	273	10	.	.	PUNCT
cril-1051	274	1	chan	chan	PROPN
cril-1051	274	2	,	,	PUNCT
cril-1051	274	3	erwin	erwin	PROPN
cril-1051	274	4	.	.	PUNCT
cril-1051	274	5	2006	2006	NUM
cril-1051	274	6	.	.	PUNCT
cril-1051	275	1	learning	learn	VERB
cril-1051	275	2	probabilistic	probabilistic	ADJ
cril-1051	275	3	paradigms	paradigm	NOUN
cril-1051	275	4	for	for	ADP
cril-1051	275	5	morphology	morphology	NOUN
cril-1051	275	6	in	in	ADP
cril-1051	275	7	a	a	DET
cril-1051	275	8	latent	latent	NOUN
cril-1051	275	9	class	class	NOUN
cril-1051	275	10	model	model	NOUN
cril-1051	275	11	.	.	PUNCT
cril-1051	276	1	in	in	ADP
cril-1051	276	2	proceedings	proceeding	NOUN
cril-1051	276	3	of	of	ADP
cril-1051	276	4	the	the	DET
cril-1051	276	5	eighth	eighth	ADJ
cril-1051	276	6	meeting	meeting	NOUN
cril-1051	276	7	of	of	ADP
cril-1051	276	8	the	the	DET
cril-1051	276	9	acl	acl	PROPN
cril-1051	276	10	special	special	ADJ
cril-1051	276	11	interest	interest	NOUN
cril-1051	276	12	group	group	NOUN
cril-1051	276	13	on	on	ADP
cril-1051	276	14	computational	computational	ADJ
cril-1051	276	15	phonology	phonology	NOUN
cril-1051	276	16	and	and	CCONJ
cril-1051	276	17	morphology	morphology	NOUN
cril-1051	276	18	(	(	PUNCT
cril-1051	276	19	sigphon	sigphon	ADJ
cril-1051	276	20	’	'	PUNCT
cril-1051	276	21	06	06	NUM
cril-1051	276	22	)	)	PUNCT
cril-1051	276	23	,	,	PUNCT
cril-1051	276	24	69–78	69–78	PROPN
cril-1051	276	25	.	.	PUNCT
cril-1051	276	26	stroudsburg	stroudsburg	PROPN
cril-1051	276	27	,	,	PUNCT
cril-1051	276	28	pa	pa	PROPN
cril-1051	276	29	,	,	PUNCT
cril-1051	276	30	usa	usa	PROPN
cril-1051	276	31	:	:	PUNCT
cril-1051	276	32	association	association	NOUN
cril-1051	276	33	for	for	ADP
cril-1051	276	34	computational	computational	ADJ
cril-1051	276	35	linguistics	linguistic	NOUN
cril-1051	276	36	.	.	PUNCT
cril-1051	277	1	http://dl.acm.org/citation.cfm?id=1622165.1622174	http://dl.acm.org/citation.cfm?id=1622165.1622174	PROPN
cril-1051	277	2	.	.	PUNCT
cril-1051	277	3	corbett	corbett	PROPN
cril-1051	277	4	,	,	PUNCT
cril-1051	277	5	greville	greville	NOUN
cril-1051	277	6	g.	g.	PROPN
cril-1051	277	7	2013	2013	NUM
cril-1051	277	8	.	.	PUNCT
cril-1051	278	1	the	the	DET
cril-1051	278	2	unique	unique	ADJ
cril-1051	278	3	challenge	challenge	NOUN
cril-1051	278	4	of	of	ADP
cril-1051	278	5	the	the	DET
cril-1051	278	6	archi	archi	NOUN
cril-1051	278	7	paradigm	paradigm	NOUN
cril-1051	278	8	.	.	PUNCT
cril-1051	279	1	in	in	ADP
cril-1051	279	2	chundra	chundra	PROPN
cril-1051	279	3	cathcart	cathcart	PROPN
cril-1051	279	4	,	,	PUNCT
cril-1051	279	5	shinae	shinae	PROPN
cril-1051	279	6	kang	kang	PROPN
cril-1051	279	7	&	&	CCONJ
cril-1051	279	8	clare	clare	PROPN
cril-1051	279	9	s.	s.	PROPN
cril-1051	279	10	sandy	sandy	PROPN
cril-1051	279	11	(	(	PUNCT
cril-1051	279	12	eds	ed	NOUN
cril-1051	279	13	.	.	PUNCT
cril-1051	279	14	)	)	PUNCT
cril-1051	279	15	,	,	PUNCT
cril-1051	279	16	proceedings	proceeding	NOUN
cril-1051	279	17	of	of	ADP
cril-1051	279	18	the	the	DET
cril-1051	279	19	37th	37th	ADJ
cril-1051	279	20	annual	annual	ADJ
cril-1051	279	21	meeting	meeting	NOUN
cril-1051	279	22	of	of	ADP
cril-1051	279	23	the	the	DET
cril-1051	279	24	berkeley	berkeley	PROPN
cril-1051	279	25	linguistics	linguistics	PROPN
cril-1051	279	26	society	society	NOUN
cril-1051	279	27	:	:	PUNCT
cril-1051	279	28	special	special	ADJ
cril-1051	279	29	session	session	NOUN
cril-1051	279	30	on	on	ADP
cril-1051	279	31	languages	language	NOUN
cril-1051	279	32	of	of	ADP
cril-1051	279	33	the	the	DET
cril-1051	279	34	caucasus	caucasus	NOUN
cril-1051	279	35	,	,	PUNCT
cril-1051	279	36	52–67	52–67	NUM
cril-1051	279	37	.	.	PUNCT
cril-1051	279	38	berkeley	berkeley	PROPN
cril-1051	279	39	,	,	PUNCT
cril-1051	279	40	ca	ca	NOUN
cril-1051	279	41	.	.	PUNCT
cril-1051	280	1	https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.657.9753&rep=rep1&type=pd	https://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.657.9753&rep=rep1&type=pd	PROPN
cril-1051	280	2	f	f	PROPN
cril-1051	280	3	(	(	PUNCT
cril-1051	280	4	12	12	NUM
cril-1051	280	5	january	january	PROPN
cril-1051	280	6	,	,	PUNCT
cril-1051	280	7	2021	2021	NUM
cril-1051	280	8	)	)	PUNCT
cril-1051	280	9	.	.	PUNCT
cril-1051	281	1	cotterell	cotterell	PROPN
cril-1051	281	2	,	,	PUNCT
cril-1051	281	3	ryan	ryan	PROPN
cril-1051	281	4	&	&	CCONJ
cril-1051	281	5	georg	georg	PROPN
cril-1051	281	6	heigold	heigold	PROPN
cril-1051	281	7	.	.	PUNCT
cril-1051	282	1	2017	2017	NUM
cril-1051	282	2	.	.	PUNCT
cril-1051	283	1	cross	cross	ADJ
cril-1051	283	2	-	-	ADJ
cril-1051	283	3	lingual	lingual	ADJ
cril-1051	283	4	character	character	NOUN
cril-1051	283	5	-	-	PUNCT
cril-1051	283	6	level	level	NOUN
cril-1051	283	7	neural	neural	ADJ
cril-1051	283	8	morphological	morphological	ADJ
cril-1051	283	9	tagging	tagging	NOUN
cril-1051	283	10	.	.	PUNCT
cril-1051	284	1	proceedings	proceeding	NOUN
cril-1051	284	2	of	of	ADP
cril-1051	284	3	the	the	DET
cril-1051	284	4	2017	2017	NUM
cril-1051	284	5	conference	conference	NOUN
cril-1051	284	6	on	on	ADP
cril-1051	284	7	empirical	empirical	ADJ
cril-1051	284	8	methods	method	NOUN
cril-1051	284	9	in	in	ADP
cril-1051	284	10	natural	natural	ADJ
cril-1051	284	11	language	language	NOUN
cril-1051	284	12	processing	process	VERB
cril-1051	284	13	748–759	748–759	NUM
cril-1051	284	14	.	.	PUNCT
cril-1051	285	1	https://aclanthology.coli.uni-saarland.de/papers/d171078/d17-1078	https://aclanthology.coli.uni-saarland.de/papers/d171078/d17-1078	NOUN
cril-1051	285	2	(	(	PUNCT
cril-1051	285	3	16	16	NUM
cril-1051	285	4	january	january	PROPN
cril-1051	285	5	,	,	PUNCT
cril-1051	285	6	2018	2018	NUM
cril-1051	285	7	)	)	PUNCT
cril-1051	285	8	.	.	PUNCT
cril-1051	286	1	cotterell	cotterell	NOUN
cril-1051	286	2	,	,	PUNCT
cril-1051	286	3	ryan	ryan	PROPN
cril-1051	286	4	,	,	PUNCT
cril-1051	286	5	christo	christo	PROPN
cril-1051	286	6	kirov	kirov	PROPN
cril-1051	286	7	,	,	PUNCT
cril-1051	286	8	john	john	PROPN
cril-1051	286	9	sylak	sylak	PROPN
cril-1051	286	10	-	-	PUNCT
cril-1051	286	11	glassman	glassman	PROPN
cril-1051	286	12	,	,	PUNCT
cril-1051	286	13	géraldine	géraldine	ADJ
cril-1051	286	14	walther	walther	NOUN
cril-1051	286	15	,	,	PUNCT
cril-1051	286	16	ekaterina	ekaterina	PROPN
cril-1051	286	17	vylomova	vylomova	PROPN
cril-1051	286	18	,	,	PUNCT
cril-1051	286	19	arya	arya	PROPN
cril-1051	286	20	d.	d.	PROPN
cril-1051	286	21	mccarthy	mccarthy	PROPN
cril-1051	286	22	,	,	PUNCT
cril-1051	286	23	katharina	katharina	PROPN
cril-1051	286	24	kann	kann	PROPN
cril-1051	286	25	,	,	PUNCT
cril-1051	286	26	et	et	PROPN
cril-1051	286	27	al	al	PROPN
cril-1051	286	28	.	.	PROPN
cril-1051	286	29	2018	2018	NUM
cril-1051	286	30	.	.	PUNCT
cril-1051	287	1	the	the	DET
cril-1051	287	2	conll	conll	PROPN
cril-1051	287	3	–	–	PUNCT
cril-1051	287	4	sigmorphon	sigmorphon	NOUN
cril-1051	287	5	2018	2018	NUM
cril-1051	287	6	shared	share	VERB
cril-1051	287	7	task	task	NOUN
cril-1051	287	8	:	:	PUNCT
cril-1051	287	9	universal	universal	ADJ
cril-1051	287	10	morphological	morphological	ADJ
cril-1051	287	11	reinflection	reinflection	NOUN
cril-1051	287	12	.	.	PUNCT
cril-1051	288	1	in	in	ADP
cril-1051	288	2	proceedings	proceeding	NOUN
cril-1051	288	3	of	of	ADP
cril-1051	288	4	the	the	DET
cril-1051	288	5	conll	conll	PROPN
cril-1051	288	6	sigmorphon	sigmorphon	PROPN
cril-1051	288	7	2018	2018	NUM
cril-1051	288	8	shared	share	VERB
cril-1051	288	9	task	task	NOUN
cril-1051	288	10	:	:	PUNCT
cril-1051	288	11	universal	universal	ADJ
cril-1051	288	12	morphological	morphological	ADJ
cril-1051	288	13	reinflection	reinflection	NOUN
cril-1051	288	14	,	,	PUNCT
cril-1051	288	15	1–27	1–27	PROPN
cril-1051	288	16	.	.	PUNCT
cril-1051	288	17	brussels	brussels	PROPN
cril-1051	288	18	:	:	PUNCT
cril-1051	288	19	association	association	NOUN
cril-1051	288	20	for	for	ADP
cril-1051	288	21	computational	computational	ADJ
cril-1051	288	22	linguistics	linguistic	NOUN
cril-1051	288	23	.	.	PUNCT
cril-1051	289	1	http://www.aclweb.org/anthology/k18-3001	http://www.aclweb.org/anthology/k18-3001	PUNCT
cril-1051	289	2	(	(	PUNCT
cril-1051	289	3	2	2	NUM
cril-1051	289	4	november	november	PROPN
cril-1051	289	5	,	,	PUNCT
cril-1051	289	6	2018	2018	NUM
cril-1051	289	7	)	)	PUNCT
cril-1051	289	8	.	.	PUNCT
cril-1051	290	1	cotterell	cotterell	NOUN
cril-1051	290	2	,	,	PUNCT
cril-1051	290	3	ryan	ryan	PROPN
cril-1051	290	4	,	,	PUNCT
cril-1051	290	5	christo	christo	PROPN
cril-1051	290	6	kirov	kirov	PROPN
cril-1051	290	7	,	,	PUNCT
cril-1051	290	8	john	john	PROPN
cril-1051	290	9	sylak	sylak	PROPN
cril-1051	290	10	-	-	PUNCT
cril-1051	290	11	glassman	glassman	PROPN
cril-1051	290	12	,	,	PUNCT
cril-1051	290	13	géraldine	géraldine	ADJ
cril-1051	290	14	walther	walther	NOUN
cril-1051	290	15	,	,	PUNCT
cril-1051	290	16	ekaterina	ekaterina	PROPN
cril-1051	290	17	vylomova	vylomova	PROPN
cril-1051	290	18	,	,	PUNCT
cril-1051	290	19	patrick	patrick	PROPN
cril-1051	290	20	xia	xia	PROPN
cril-1051	290	21	,	,	PUNCT
cril-1051	290	22	manaal	manaal	NOUN
cril-1051	290	23	faruqui	faruqui	NOUN
cril-1051	290	24	,	,	PUNCT
cril-1051	290	25	et	et	PROPN
cril-1051	290	26	al	al	PROPN
cril-1051	290	27	.	.	PROPN
cril-1051	290	28	2017	2017	NUM
cril-1051	290	29	.	.	PUNCT
cril-1051	291	1	conll	conll	PROPN
cril-1051	291	2	-	-	PUNCT
cril-1051	291	3	sigmorphon	sigmorphon	NOUN
cril-1051	291	4	2017	2017	NUM
cril-1051	291	5	shared	share	VERB
cril-1051	291	6	task	task	NOUN
cril-1051	291	7	:	:	PUNCT
cril-1051	291	8	universal	universal	ADJ
cril-1051	291	9	morphological	morphological	ADJ
cril-1051	291	10	reinflection	reinflection	NOUN
cril-1051	291	11	in	in	ADP
cril-1051	291	12	52	52	NUM
cril-1051	291	13	languages	language	NOUN
cril-1051	291	14	.	.	PUNCT
cril-1051	292	1	in	in	ADP
cril-1051	292	2	proceedings	proceeding	NOUN
cril-1051	292	3	of	of	ADP
cril-1051	292	4	the	the	DET
cril-1051	292	5	conll	conll	PROPN
cril-1051	292	6	sigmorphon	sigmorphon	PROPN
cril-1051	292	7	2017	2017	NUM
cril-1051	292	8	shared	share	VERB
cril-1051	292	9	task	task	NOUN
cril-1051	292	10	:	:	PUNCT
cril-1051	292	11	universal	universal	ADJ
cril-1051	292	12	morphological	morphological	ADJ
cril-1051	292	13	reinflection	reinflection	NOUN
cril-1051	292	14	,	,	PUNCT
cril-1051	292	15	1–30	1–30	PROPN
cril-1051	292	16	.	.	PUNCT
cril-1051	292	17	computational	computational	ADJ
cril-1051	292	18	morphology	morphology	NOUN
cril-1051	292	19	for	for	ADP
cril-1051	292	20	language	language	NOUN
cril-1051	292	21	description	description	NOUN
cril-1051	292	22	and	and	CCONJ
cril-1051	292	23	documentation	documentation	NOUN
cril-1051	292	24	19	19	NUM
cril-1051	292	25	vancouver	vancouver	NOUN
cril-1051	292	26	:	:	PUNCT
cril-1051	292	27	association	association	NOUN
cril-1051	292	28	for	for	ADP
cril-1051	292	29	computational	computational	ADJ
cril-1051	292	30	linguistics	linguistic	NOUN
cril-1051	292	31	.	.	PUNCT
cril-1051	293	1	http://www.aclweb.org/anthology/k17-2001	http://www.aclweb.org/anthology/k17-2001	PROPN
cril-1051	293	2	.	.	PUNCT
cril-1051	294	1	cotterell	cotterell	NOUN
cril-1051	294	2	,	,	PUNCT
cril-1051	294	3	ryan	ryan	PROPN
cril-1051	294	4	,	,	PUNCT
cril-1051	294	5	christo	christo	PROPN
cril-1051	294	6	kirov	kirov	PROPN
cril-1051	294	7	,	,	PUNCT
cril-1051	294	8	john	john	PROPN
cril-1051	294	9	sylak	sylak	PROPN
cril-1051	294	10	-	-	PUNCT
cril-1051	294	11	glassman	glassman	PROPN
cril-1051	294	12	,	,	PUNCT
cril-1051	294	13	david	david	PROPN
cril-1051	294	14	yarowsky	yarowsky	PROPN
cril-1051	294	15	,	,	PUNCT
cril-1051	294	16	jason	jason	PROPN
cril-1051	294	17	eisner	eisner	PROPN
cril-1051	294	18	&	&	CCONJ
cril-1051	294	19	mans	mans	PROPN
cril-1051	294	20	hulden	hulden	VERB
cril-1051	294	21	.	.	PUNCT
cril-1051	295	1	2016	2016	NUM
cril-1051	295	2	.	.	PUNCT
cril-1051	296	1	the	the	DET
cril-1051	296	2	sigmorphon	sigmorphon	NOUN
cril-1051	296	3	2016	2016	NUM
cril-1051	296	4	shared	share	VERB
cril-1051	296	5	task	task	NOUN
cril-1051	296	6	—	—	PUNCT
cril-1051	296	7	morphological	morphological	ADJ
cril-1051	296	8	reinflection	reinflection	NOUN
cril-1051	296	9	.	.	PUNCT
cril-1051	297	1	in	in	ADP
cril-1051	297	2	proceedings	proceeding	NOUN
cril-1051	297	3	of	of	ADP
cril-1051	297	4	the	the	DET
cril-1051	297	5	14th	14th	ADJ
cril-1051	297	6	sigmorphon	sigmorphon	NOUN
cril-1051	297	7	workshop	workshop	NOUN
cril-1051	297	8	on	on	ADP
cril-1051	297	9	computational	computational	ADJ
cril-1051	297	10	research	research	NOUN
cril-1051	297	11	in	in	ADP
cril-1051	297	12	phonetics	phonetic	NOUN
cril-1051	297	13	,	,	PUNCT
cril-1051	297	14	phonology	phonology	NOUN
cril-1051	297	15	,	,	PUNCT
cril-1051	297	16	and	and	CCONJ
cril-1051	297	17	morphology	morphology	NOUN
cril-1051	297	18	,	,	PUNCT
cril-1051	297	19	10–22	10–22	NUM
cril-1051	297	20	.	.	PUNCT
cril-1051	297	21	cotterell	cotterell	PROPN
cril-1051	297	22	,	,	PUNCT
cril-1051	297	23	ryan	ryan	PROPN
cril-1051	297	24	,	,	PUNCT
cril-1051	297	25	thomas	thomas	PROPN
cril-1051	297	26	müller	müller	PROPN
cril-1051	297	27	,	,	PUNCT
cril-1051	297	28	alexander	alexander	PROPN
cril-1051	297	29	m.	m.	PROPN
cril-1051	297	30	fraser	fraser	PROPN
cril-1051	297	31	&	&	CCONJ
cril-1051	297	32	hinrich	hinrich	PROPN
cril-1051	297	33	schütze	schütze	PROPN
cril-1051	297	34	.	.	PROPN
cril-1051	297	35	2015	2015	NUM
cril-1051	297	36	.	.	PUNCT
cril-1051	298	1	labeled	label	VERB
cril-1051	298	2	morphological	morphological	ADJ
cril-1051	298	3	segmentation	segmentation	NOUN
cril-1051	298	4	with	with	ADP
cril-1051	298	5	semi	semi	ADJ
cril-1051	298	6	-	-	ADJ
cril-1051	298	7	markov	markov	ADJ
cril-1051	298	8	models	model	NOUN
cril-1051	298	9	.	.	PUNCT
cril-1051	299	1	in	in	ADP
cril-1051	299	2	proceedings	proceeding	NOUN
cril-1051	299	3	of	of	ADP
cril-1051	299	4	the	the	DET
cril-1051	299	5	nineteenth	nineteenth	ADJ
cril-1051	299	6	conference	conference	NOUN
cril-1051	299	7	on	on	ADP
cril-1051	299	8	computational	computational	ADJ
cril-1051	299	9	natural	natural	ADJ
cril-1051	299	10	language	language	NOUN
cril-1051	299	11	learning	learning	NOUN
cril-1051	299	12	,	,	PUNCT
cril-1051	299	13	164–174	164–174	NUM
cril-1051	299	14	.	.	PUNCT
cril-1051	300	1	beijing	beijing	PROPN
cril-1051	300	2	,	,	PUNCT
cril-1051	300	3	china	china	PROPN
cril-1051	300	4	:	:	PUNCT
cril-1051	300	5	association	association	NOUN
cril-1051	300	6	for	for	ADP
cril-1051	300	7	computational	computational	ADJ
cril-1051	300	8	linguistics	linguistic	NOUN
cril-1051	300	9	.	.	PUNCT
cril-1051	301	1	cotterell	cotterell	NOUN
cril-1051	301	2	,	,	PUNCT
cril-1051	301	3	ryan	ryan	PROPN
cril-1051	301	4	,	,	PUNCT
cril-1051	301	5	john	john	PROPN
cril-1051	301	6	sylak	sylak	PROPN
cril-1051	301	7	-	-	PUNCT
cril-1051	301	8	glassman	glassman	PROPN
cril-1051	301	9	&	&	CCONJ
cril-1051	301	10	christo	christo	PROPN
cril-1051	301	11	kirov	kirov	PROPN
cril-1051	301	12	.	.	PROPN
cril-1051	301	13	2017	2017	NUM
cril-1051	301	14	.	.	PUNCT
cril-1051	302	1	neural	neural	ADJ
cril-1051	302	2	graphical	graphical	ADJ
cril-1051	302	3	models	model	NOUN
cril-1051	302	4	over	over	ADP
cril-1051	302	5	strings	string	NOUN
cril-1051	302	6	for	for	ADP
cril-1051	302	7	principal	principal	ADJ
cril-1051	302	8	parts	part	NOUN
cril-1051	302	9	morphological	morphological	ADJ
cril-1051	302	10	paradigm	paradigm	NOUN
cril-1051	302	11	completion	completion	NOUN
cril-1051	302	12	.	.	PUNCT
cril-1051	303	1	proceedings	proceeding	NOUN
cril-1051	303	2	of	of	ADP
cril-1051	303	3	the	the	DET
cril-1051	303	4	15th	15th	ADJ
cril-1051	303	5	conference	conference	NOUN
cril-1051	303	6	of	of	ADP
cril-1051	303	7	the	the	DET
cril-1051	303	8	european	european	ADJ
cril-1051	303	9	chapter	chapter	NOUN
cril-1051	303	10	of	of	ADP
cril-1051	303	11	the	the	DET
cril-1051	303	12	association	association	NOUN
cril-1051	303	13	for	for	ADP
cril-1051	303	14	computational	computational	ADJ
cril-1051	303	15	linguistics	linguistic	NOUN
cril-1051	303	16	:	:	PUNCT
cril-1051	303	17	volume	volume	NOUN
cril-1051	303	18	2	2	NUM
cril-1051	303	19	,	,	PUNCT
cril-1051	303	20	short	short	ADJ
cril-1051	303	21	papers	paper	NOUN
cril-1051	303	22	2	2	NUM
cril-1051	303	23	.	.	PUNCT
cril-1051	304	1	759–765	759–765	NUM
cril-1051	304	2	.	.	PUNCT
cril-1051	305	1	https://aclanthology.coli.uni-saarland.de/papers/e172120/e17-2120	https://aclanthology.coli.uni-saarland.de/papers/e172120/e17-2120	PROPN
cril-1051	305	2	(	(	PUNCT
cril-1051	305	3	16	16	NUM
cril-1051	305	4	january	january	PROPN
cril-1051	305	5	,	,	PUNCT
cril-1051	305	6	2018	2018	NUM
cril-1051	305	7	)	)	PUNCT
cril-1051	305	8	.	.	PUNCT
cril-1051	306	1	cotterell	cotterell	NOUN
cril-1051	306	2	,	,	PUNCT
cril-1051	306	3	ryan	ryan	PROPN
cril-1051	306	4	,	,	PUNCT
cril-1051	306	5	tim	tim	PROPN
cril-1051	306	6	vieira	vieira	PROPN
cril-1051	306	7	&	&	CCONJ
cril-1051	306	8	hinrich	hinrich	PROPN
cril-1051	306	9	schütze	schütze	PROPN
cril-1051	306	10	.	.	PROPN
cril-1051	306	11	2016	2016	NUM
cril-1051	306	12	.	.	PUNCT
cril-1051	307	1	a	a	DET
cril-1051	307	2	joint	joint	ADJ
cril-1051	307	3	model	model	NOUN
cril-1051	307	4	of	of	ADP
cril-1051	307	5	orthography	orthography	NOUN
cril-1051	307	6	and	and	CCONJ
cril-1051	307	7	morphological	morphological	ADJ
cril-1051	307	8	segmentation	segmentation	NOUN
cril-1051	307	9	.	.	PUNCT
cril-1051	308	1	proceedings	proceeding	NOUN
cril-1051	308	2	of	of	ADP
cril-1051	308	3	the	the	DET
cril-1051	308	4	2016	2016	NUM
cril-1051	308	5	conference	conference	NOUN
cril-1051	308	6	of	of	ADP
cril-1051	308	7	the	the	DET
cril-1051	308	8	north	north	ADJ
cril-1051	308	9	american	american	ADJ
cril-1051	308	10	chapter	chapter	NOUN
cril-1051	308	11	of	of	ADP
cril-1051	308	12	the	the	DET
cril-1051	308	13	association	association	NOUN
cril-1051	308	14	for	for	ADP
cril-1051	308	15	computational	computational	ADJ
cril-1051	308	16	linguistics	linguistic	NOUN
cril-1051	308	17	:	:	PUNCT
cril-1051	308	18	human	human	ADJ
cril-1051	308	19	language	language	NOUN
cril-1051	308	20	technologies	technology	NOUN
cril-1051	308	21	664–669	664–669	NUM
cril-1051	308	22	.	.	PUNCT
cril-1051	309	1	https://aclanthology.coli.uni-saarland.de/papers/n16-1080/n161080	https://aclanthology.coli.uni-saarland.de/papers/n16-1080/n161080	NOUN
cril-1051	309	2	(	(	PUNCT
cril-1051	309	3	16	16	NUM
cril-1051	309	4	january	january	PROPN
cril-1051	309	5	,	,	PUNCT
cril-1051	309	6	2018	2018	NUM
cril-1051	309	7	)	)	PUNCT
cril-1051	309	8	.	.	PUNCT
cril-1051	310	1	cox	cox	PROPN
cril-1051	310	2	,	,	PUNCT
cril-1051	310	3	christopher	christopher	PROPN
cril-1051	310	4	,	,	PUNCT
cril-1051	310	5	gilles	gilles	PROPN
cril-1051	310	6	bouliame	bouliame	PROPN
cril-1051	310	7	&	&	CCONJ
cril-1051	310	8	jahangir	jahangir	PROPN
cril-1051	310	9	alam	alam	PROPN
cril-1051	310	10	.	.	PUNCT
cril-1051	310	11	2019	2019	NUM
cril-1051	310	12	.	.	PUNCT
cril-1051	311	1	taking	take	VERB
cril-1051	311	2	aim	aim	NOUN
cril-1051	311	3	at	at	ADP
cril-1051	311	4	the	the	DET
cril-1051	311	5	transcription	transcription	NOUN
cril-1051	311	6	bottleneck	bottleneck	NOUN
cril-1051	311	7	:	:	PUNCT
cril-1051	311	8	integrating	integrate	VERB
cril-1051	311	9	speech	speech	NOUN
cril-1051	311	10	technology	technology	NOUN
cril-1051	311	11	into	into	ADP
cril-1051	311	12	language	language	NOUN
cril-1051	311	13	documentation	documentation	NOUN
cril-1051	311	14	and	and	CCONJ
cril-1051	311	15	conservation	conservation	NOUN
cril-1051	311	16	.	.	PUNCT
cril-1051	312	1	slideshow	slideshow	PROPN
cril-1051	312	2	presented	present	VERB
cril-1051	312	3	at	at	ADP
cril-1051	312	4	the	the	DET
cril-1051	312	5	6th	6th	ADJ
cril-1051	312	6	international	international	ADJ
cril-1051	312	7	conference	conference	NOUN
cril-1051	312	8	on	on	ADP
cril-1051	312	9	language	language	NOUN
cril-1051	312	10	documentation	documentation	NOUN
cril-1051	312	11	and	and	CCONJ
cril-1051	312	12	conservation	conservation	NOUN
cril-1051	312	13	(	(	PUNCT
cril-1051	312	14	icdlc	icdlc	NOUN
cril-1051	312	15	)	)	PUNCT
cril-1051	312	16	,	,	PUNCT
cril-1051	312	17	honolulu	honolulu	PROPN
cril-1051	312	18	,	,	PUNCT
cril-1051	312	19	hi	hi	INTJ
cril-1051	312	20	.	.	PUNCT
cril-1051	312	21	https://scholarspace.manoa.hawaii.edu/handle/10125/44841	https://scholarspace.manoa.hawaii.edu/handle/10125/44841	PROPN
cril-1051	312	22	.	.	PUNCT
cril-1051	313	1	creutz	creutz	ADJ
cril-1051	313	2	,	,	PUNCT
cril-1051	313	3	mathias	mathias	PROPN
cril-1051	313	4	&	&	CCONJ
cril-1051	313	5	krista	krista	PROPN
cril-1051	313	6	lagus	lagus	PROPN
cril-1051	313	7	.	.	PUNCT
cril-1051	314	1	2002	2002	NUM
cril-1051	314	2	.	.	PUNCT
cril-1051	315	1	unsupervised	unsupervised	ADJ
cril-1051	315	2	discovery	discovery	NOUN
cril-1051	315	3	of	of	ADP
cril-1051	315	4	morphemes	morpheme	NOUN
cril-1051	315	5	.	.	PUNCT
cril-1051	316	1	in	in	ADP
cril-1051	316	2	proceedings	proceeding	NOUN
cril-1051	316	3	of	of	ADP
cril-1051	316	4	the	the	DET
cril-1051	316	5	acl-02	acl-02	ADJ
cril-1051	316	6	workshop	workshop	NOUN
cril-1051	316	7	on	on	ADP
cril-1051	316	8	morphological	morphological	ADJ
cril-1051	316	9	and	and	CCONJ
cril-1051	316	10	phonological	phonological	ADJ
cril-1051	316	11	learning	learning	NOUN
cril-1051	316	12	-	-	PUNCT
cril-1051	316	13	volume	volume	NOUN
cril-1051	316	14	6	6	NUM
cril-1051	316	15	,	,	PUNCT
cril-1051	316	16	21–30	21–30	NUM
cril-1051	316	17	.	.	PUNCT
cril-1051	316	18	philadelphia	philadelphia	PROPN
cril-1051	316	19	,	,	PUNCT
cril-1051	316	20	pa	pa	PROPN
cril-1051	316	21	:	:	PUNCT
cril-1051	316	22	association	association	NOUN
cril-1051	316	23	for	for	ADP
cril-1051	316	24	computational	computational	ADJ
cril-1051	316	25	linguistics	linguistic	NOUN
cril-1051	316	26	.	.	PUNCT
cril-1051	317	1	creutz	creutz	ADJ
cril-1051	317	2	,	,	PUNCT
cril-1051	317	3	mathias	mathias	PROPN
cril-1051	317	4	&	&	CCONJ
cril-1051	317	5	krista	krista	PROPN
cril-1051	317	6	lagus	lagus	PROPN
cril-1051	317	7	.	.	PUNCT
cril-1051	318	1	2007	2007	NUM
cril-1051	318	2	.	.	PUNCT
cril-1051	319	1	unsupervised	unsupervised	ADJ
cril-1051	319	2	models	model	NOUN
cril-1051	319	3	for	for	ADP
cril-1051	319	4	morpheme	morpheme	ADJ
cril-1051	319	5	segmentation	segmentation	NOUN
cril-1051	319	6	and	and	CCONJ
cril-1051	319	7	morphology	morphology	NOUN
cril-1051	319	8	learning	learn	VERB
cril-1051	319	9	.	.	PUNCT
cril-1051	320	1	acm	acm	PROPN
cril-1051	320	2	trans	trans	PROPN
cril-1051	320	3	.	.	PROPN
cril-1051	320	4	speech	speech	PROPN
cril-1051	320	5	lang	lang	PROPN
cril-1051	320	6	.	.	PUNCT
cril-1051	320	7	process	process	NOUN
cril-1051	320	8	.	.	PUNCT
cril-1051	321	1	4(1	4(1	NOUN
cril-1051	321	2	)	)	PUNCT
cril-1051	321	3	.	.	PUNCT
cril-1051	322	1	3:1	3:1	X
cril-1051	322	2	-	-	PUNCT
cril-1051	322	3	3:34	3:34	NUM
cril-1051	322	4	.	.	PUNCT
cril-1051	323	1	http://users.ics.aalto.fi/krista/papers/creutz07acmtslp.pdf	http://users.ics.aalto.fi/krista/papers/creutz07acmtslp.pdf	PROPN
cril-1051	323	2	.	.	PROPN
cril-1051	323	3	colorado	colorado	PROPN
cril-1051	323	4	research	research	PROPN
cril-1051	323	5	in	in	ADP
cril-1051	323	6	linguistics	linguistic	NOUN
cril-1051	323	7	,	,	PUNCT
cril-1051	323	8	volume	volume	NOUN
cril-1051	323	9	25	25	NUM
cril-1051	323	10	(	(	PUNCT
cril-1051	323	11	2021	2021	NUM
cril-1051	323	12	)	)	PUNCT
cril-1051	323	13	20	20	NUM
cril-1051	323	14	czaykowska	czaykowska	NOUN
cril-1051	323	15	-	-	PUNCT
cril-1051	323	16	higgins	higgins	PROPN
cril-1051	323	17	,	,	PUNCT
cril-1051	323	18	ewa	ewa	PROPN
cril-1051	323	19	.	.	PROPN
cril-1051	323	20	2009	2009	NUM
cril-1051	323	21	.	.	PUNCT
cril-1051	324	1	research	research	NOUN
cril-1051	324	2	models	model	NOUN
cril-1051	324	3	,	,	PUNCT
cril-1051	324	4	community	community	NOUN
cril-1051	324	5	engagement	engagement	NOUN
cril-1051	324	6	,	,	PUNCT
cril-1051	324	7	and	and	CCONJ
cril-1051	324	8	linguistic	linguistic	ADJ
cril-1051	324	9	fieldwork	fieldwork	NOUN
cril-1051	324	10	:	:	PUNCT
cril-1051	324	11	reflections	reflection	NOUN
cril-1051	324	12	on	on	ADP
cril-1051	324	13	working	work	VERB
cril-1051	324	14	within	within	ADP
cril-1051	324	15	canadian	canadian	ADJ
cril-1051	324	16	indigenous	indigenous	ADJ
cril-1051	324	17	communities	community	NOUN
cril-1051	324	18	.	.	PUNCT
cril-1051	325	1	language	language	NOUN
cril-1051	325	2	documentation	documentation	NOUN
cril-1051	325	3	&	&	CCONJ
cril-1051	325	4	conservation	conservation	NOUN
cril-1051	325	5	3(1	3(1	NUM
cril-1051	325	6	)	)	PUNCT
cril-1051	325	7	.	.	PUNCT
cril-1051	326	1	15–50	15–50	NUM
cril-1051	326	2	.	.	PUNCT
cril-1051	327	1	http://scholarspace.manoa.hawaii.edu/bitstream/handle/10125/4423/czaykowskahiggins.p	http://scholarspace.manoa.hawaii.edu/bitstream/handle/10125/4423/czaykowskahiggins.p	X
cril-1051	327	2	df?sequence=1	df?sequence=1	PROPN
cril-1051	327	3	.	.	PUNCT
cril-1051	328	1	dreyer	dreyer	PROPN
cril-1051	328	2	,	,	PUNCT
cril-1051	328	3	markus	markus	PROPN
cril-1051	328	4	&	&	CCONJ
cril-1051	328	5	jason	jason	PROPN
cril-1051	328	6	eisner	eisner	PROPN
cril-1051	328	7	.	.	PUNCT
cril-1051	329	1	2011	2011	NUM
cril-1051	329	2	.	.	PUNCT
cril-1051	330	1	discovering	discover	VERB
cril-1051	330	2	morphological	morphological	ADJ
cril-1051	330	3	paradigms	paradigm	NOUN
cril-1051	330	4	from	from	ADP
cril-1051	330	5	plain	plain	ADJ
cril-1051	330	6	text	text	NOUN
cril-1051	330	7	using	use	VERB
cril-1051	330	8	a	a	DET
cril-1051	330	9	dirichlet	dirichlet	NOUN
cril-1051	330	10	process	process	NOUN
cril-1051	330	11	mixture	mixture	NOUN
cril-1051	330	12	model	model	NOUN
cril-1051	330	13	.	.	PUNCT
cril-1051	331	1	in	in	ADP
cril-1051	331	2	proceedings	proceeding	NOUN
cril-1051	331	3	of	of	ADP
cril-1051	331	4	the	the	DET
cril-1051	331	5	conference	conference	NOUN
cril-1051	331	6	on	on	ADP
cril-1051	331	7	empirical	empirical	ADJ
cril-1051	331	8	methods	method	NOUN
cril-1051	331	9	in	in	ADP
cril-1051	331	10	natural	natural	ADJ
cril-1051	331	11	language	language	NOUN
cril-1051	331	12	processing	processing	NOUN
cril-1051	331	13	,	,	PUNCT
cril-1051	331	14	616–627	616–627	NUM
cril-1051	331	15	.	.	PUNCT
cril-1051	332	1	association	association	NOUN
cril-1051	332	2	for	for	ADP
cril-1051	332	3	computational	computational	ADJ
cril-1051	332	4	linguistics	linguistic	NOUN
cril-1051	332	5	.	.	PUNCT
cril-1051	333	1	duong	duong	PROPN
cril-1051	333	2	,	,	PUNCT
cril-1051	333	3	long	long	ADV
cril-1051	333	4	.	.	PUNCT
cril-1051	334	1	2017	2017	NUM
cril-1051	334	2	.	.	PUNCT
cril-1051	334	3	natural	natural	ADJ
cril-1051	334	4	language	language	NOUN
cril-1051	334	5	processing	processing	NOUN
cril-1051	334	6	for	for	ADP
cril-1051	334	7	resource	resource	NOUN
cril-1051	334	8	-	-	PUNCT
cril-1051	334	9	poor	poor	ADJ
cril-1051	334	10	languages	language	NOUN
cril-1051	334	11	.	.	PUNCT
cril-1051	335	1	melbourne	melbourne	PROPN
cril-1051	335	2	,	,	PUNCT
cril-1051	335	3	australia	australia	PROPN
cril-1051	335	4	:	:	PUNCT
cril-1051	335	5	university	university	NOUN
cril-1051	335	6	of	of	ADP
cril-1051	335	7	melbourne	melbourne	PROPN
cril-1051	335	8	phd	phd	NOUN
cril-1051	335	9	thesis	thesis	NOUN
cril-1051	335	10	.	.	PUNCT
cril-1051	336	1	http://minervaaccess.unimelb.edu.au/handle/11343/192938	http://minervaaccess.unimelb.edu.au/handle/11343/192938	PROPN
cril-1051	336	2	(	(	PUNCT
cril-1051	336	3	29	29	NUM
cril-1051	336	4	june	june	PROPN
cril-1051	336	5	,	,	PUNCT
cril-1051	336	6	2018	2018	NUM
cril-1051	336	7	)	)	PUNCT
cril-1051	336	8	.	.	PUNCT
cril-1051	337	1	duong	duong	PROPN
cril-1051	337	2	,	,	PUNCT
cril-1051	337	3	long	long	ADV
cril-1051	337	4	,	,	PUNCT
cril-1051	337	5	antonios	antonios	PROPN
cril-1051	337	6	anastasopoulos	anastasopoulos	PROPN
cril-1051	337	7	,	,	PUNCT
cril-1051	337	8	david	david	PROPN
cril-1051	337	9	chiang	chiang	PROPN
cril-1051	337	10	,	,	PUNCT
cril-1051	337	11	steven	steven	PROPN
cril-1051	337	12	bird	bird	PROPN
cril-1051	337	13	&	&	CCONJ
cril-1051	337	14	trevor	trevor	PROPN
cril-1051	337	15	cohn	cohn	PROPN
cril-1051	337	16	.	.	PROPN
cril-1051	337	17	2016	2016	NUM
cril-1051	337	18	.	.	PUNCT
cril-1051	338	1	an	an	DET
cril-1051	338	2	attentional	attentional	ADJ
cril-1051	338	3	model	model	NOUN
cril-1051	338	4	for	for	ADP
cril-1051	338	5	speech	speech	NOUN
cril-1051	338	6	translation	translation	NOUN
cril-1051	338	7	without	without	ADP
cril-1051	338	8	transcription	transcription	NOUN
cril-1051	338	9	.	.	PUNCT
cril-1051	339	1	in	in	ADP
cril-1051	339	2	proceedings	proceeding	NOUN
cril-1051	339	3	of	of	ADP
cril-1051	339	4	the	the	DET
cril-1051	339	5	2016	2016	NUM
cril-1051	339	6	conference	conference	NOUN
cril-1051	339	7	of	of	ADP
cril-1051	339	8	the	the	DET
cril-1051	339	9	north	north	ADJ
cril-1051	339	10	american	american	ADJ
cril-1051	339	11	chapter	chapter	NOUN
cril-1051	339	12	of	of	ADP
cril-1051	339	13	the	the	DET
cril-1051	339	14	association	association	NOUN
cril-1051	339	15	for	for	ADP
cril-1051	339	16	computational	computational	ADJ
cril-1051	339	17	linguistics	linguistic	NOUN
cril-1051	339	18	:	:	PUNCT
cril-1051	339	19	human	human	ADJ
cril-1051	339	20	language	language	NOUN
cril-1051	339	21	technologies	technology	NOUN
cril-1051	339	22	,	,	PUNCT
cril-1051	339	23	949–959	949–959	NUM
cril-1051	339	24	.	.	PUNCT
cril-1051	340	1	san	san	PROPN
cril-1051	340	2	diego	diego	PROPN
cril-1051	340	3	,	,	PUNCT
cril-1051	340	4	california	california	PROPN
cril-1051	340	5	:	:	PUNCT
cril-1051	340	6	association	association	NOUN
cril-1051	340	7	for	for	ADP
cril-1051	340	8	computational	computational	ADJ
cril-1051	340	9	linguistics	linguistic	NOUN
cril-1051	340	10	.	.	PUNCT
cril-1051	341	1	http://www.aclweb.org/anthology/n16-1109	http://www.aclweb.org/anthology/n16-1109	NUM
cril-1051	341	2	(	(	PUNCT
cril-1051	341	3	6	6	NUM
cril-1051	341	4	july	july	PROPN
cril-1051	341	5	,	,	PUNCT
cril-1051	341	6	2018	2018	NUM
cril-1051	341	7	)	)	PUNCT
cril-1051	341	8	.	.	PUNCT
cril-1051	342	1	durrett	durrett	PROPN
cril-1051	342	2	,	,	PUNCT
cril-1051	342	3	greg	greg	PROPN
cril-1051	342	4	&	&	CCONJ
cril-1051	342	5	john	john	PROPN
cril-1051	342	6	denero	denero	PROPN
cril-1051	342	7	.	.	PROPN
cril-1051	343	1	2013	2013	NUM
cril-1051	343	2	.	.	PUNCT
cril-1051	344	1	supervised	supervised	ADJ
cril-1051	344	2	learning	learning	NOUN
cril-1051	344	3	of	of	ADP
cril-1051	344	4	complete	complete	ADJ
cril-1051	344	5	morphological	morphological	ADJ
cril-1051	344	6	paradigms	paradigm	NOUN
cril-1051	344	7	.	.	PUNCT
cril-1051	345	1	in	in	ADP
cril-1051	345	2	proceedings	proceeding	NOUN
cril-1051	345	3	of	of	ADP
cril-1051	345	4	the	the	DET
cril-1051	345	5	2013	2013	NUM
cril-1051	345	6	conference	conference	NOUN
cril-1051	345	7	of	of	ADP
cril-1051	345	8	the	the	DET
cril-1051	345	9	north	north	ADJ
cril-1051	345	10	american	american	ADJ
cril-1051	345	11	chapter	chapter	NOUN
cril-1051	345	12	of	of	ADP
cril-1051	345	13	the	the	DET
cril-1051	345	14	association	association	NOUN
cril-1051	345	15	for	for	ADP
cril-1051	345	16	computational	computational	ADJ
cril-1051	345	17	linguistics	linguistic	NOUN
cril-1051	345	18	:	:	PUNCT
cril-1051	345	19	human	human	ADJ
cril-1051	345	20	language	language	NOUN
cril-1051	345	21	technologies	technology	NOUN
cril-1051	345	22	,	,	PUNCT
cril-1051	345	23	1185–1195	1185–1195	NUM
cril-1051	345	24	.	.	PUNCT
cril-1051	345	25	atlanta	atlanta	PROPN
cril-1051	345	26	,	,	PUNCT
cril-1051	345	27	georgia	georgia	PROPN
cril-1051	345	28	:	:	PUNCT
cril-1051	345	29	association	association	NOUN
cril-1051	345	30	for	for	ADP
cril-1051	345	31	computational	computational	ADJ
cril-1051	345	32	linguistics	linguistic	NOUN
cril-1051	345	33	.	.	PUNCT
cril-1051	346	1	https://research.google.com/pubs/pub41850.html	https://research.google.com/pubs/pub41850.html	PROPN
cril-1051	346	2	(	(	PUNCT
cril-1051	346	3	2	2	NUM
cril-1051	346	4	january	january	PROPN
cril-1051	346	5	,	,	PUNCT
cril-1051	346	6	2018	2018	NUM
cril-1051	346	7	)	)	PUNCT
cril-1051	346	8	.	.	PUNCT
cril-1051	347	1	faruqui	faruqui	NOUN
cril-1051	347	2	,	,	PUNCT
cril-1051	347	3	manaal	manaal	PROPN
cril-1051	347	4	,	,	PUNCT
cril-1051	347	5	yulia	yulia	PROPN
cril-1051	347	6	tsvetkov	tsvetkov	PROPN
cril-1051	347	7	,	,	PUNCT
cril-1051	347	8	graham	graham	PROPN
cril-1051	347	9	neubig	neubig	PROPN
cril-1051	347	10	&	&	CCONJ
cril-1051	347	11	chris	chris	PROPN
cril-1051	347	12	dyer	dyer	PROPN
cril-1051	347	13	.	.	PUNCT
cril-1051	348	1	2016	2016	NUM
cril-1051	348	2	.	.	PUNCT
cril-1051	349	1	morphological	morphological	ADJ
cril-1051	349	2	inflection	inflection	NOUN
cril-1051	349	3	generation	generation	NOUN
cril-1051	349	4	using	use	VERB
cril-1051	349	5	character	character	NOUN
cril-1051	349	6	sequence	sequence	NOUN
cril-1051	349	7	to	to	ADP
cril-1051	349	8	sequence	sequence	NOUN
cril-1051	349	9	learning	learning	NOUN
cril-1051	349	10	.	.	PUNCT
cril-1051	350	1	in	in	ADP
cril-1051	350	2	proceedings	proceeding	NOUN
cril-1051	350	3	of	of	ADP
cril-1051	350	4	the	the	DET
cril-1051	350	5	2016	2016	NUM
cril-1051	350	6	conference	conference	NOUN
cril-1051	350	7	of	of	ADP
cril-1051	350	8	the	the	DET
cril-1051	350	9	north	north	ADJ
cril-1051	350	10	american	american	ADJ
cril-1051	350	11	chapter	chapter	NOUN
cril-1051	350	12	of	of	ADP
cril-1051	350	13	the	the	DET
cril-1051	350	14	association	association	NOUN
cril-1051	350	15	for	for	ADP
cril-1051	350	16	computational	computational	ADJ
cril-1051	350	17	linguistics	linguistic	NOUN
cril-1051	350	18	:	:	PUNCT
cril-1051	350	19	human	human	ADJ
cril-1051	350	20	language	language	NOUN
cril-1051	350	21	technologies	technology	NOUN
cril-1051	350	22	,	,	PUNCT
cril-1051	350	23	634–643	634–643	NUM
cril-1051	350	24	.	.	PUNCT
cril-1051	351	1	san	san	PROPN
cril-1051	351	2	diego	diego	PROPN
cril-1051	351	3	,	,	PUNCT
cril-1051	351	4	california	california	PROPN
cril-1051	351	5	:	:	PUNCT
cril-1051	351	6	association	association	NOUN
cril-1051	351	7	for	for	ADP
cril-1051	351	8	computational	computational	ADJ
cril-1051	351	9	linguistics	linguistic	NOUN
cril-1051	351	10	.	.	PUNCT
cril-1051	352	1	https://www.aclweb.org/anthology/n16-1077	https://www.aclweb.org/anthology/n16-1077	X
cril-1051	352	2	(	(	PUNCT
cril-1051	352	3	28	28	NUM
cril-1051	352	4	april	april	PROPN
cril-1051	352	5	,	,	PUNCT
cril-1051	352	6	2021	2021	NUM
cril-1051	352	7	)	)	PUNCT
cril-1051	352	8	.	.	PUNCT
cril-1051	353	1	felt	feel	VERB
cril-1051	353	2	,	,	PUNCT
cril-1051	353	3	paul	paul	PROPN
cril-1051	353	4	.	.	PROPN
cril-1051	353	5	2012	2012	NUM
cril-1051	353	6	.	.	PUNCT
cril-1051	354	1	improving	improve	VERB
cril-1051	354	2	the	the	DET
cril-1051	354	3	effectiveness	effectiveness	NOUN
cril-1051	354	4	of	of	ADP
cril-1051	354	5	machine	machine	NOUN
cril-1051	354	6	-	-	PUNCT
cril-1051	354	7	assisted	assist	VERB
cril-1051	354	8	annotation	annotation	NOUN
cril-1051	354	9	.	.	PUNCT
cril-1051	355	1	brigham	brigham	PROPN
cril-1051	355	2	young	young	PROPN
cril-1051	355	3	university	university	PROPN
cril-1051	355	4	ma	ma	PROPN
cril-1051	355	5	thesis	thesis	PROPN
cril-1051	355	6	.	.	PUNCT
cril-1051	356	1	https://scholarsarchive.byu.edu/etd/3214	https://scholarsarchive.byu.edu/etd/3214	PROPN
cril-1051	356	2	.	.	PUNCT
cril-1051	357	1	computational	computational	ADJ
cril-1051	357	2	morphology	morphology	NOUN
cril-1051	357	3	for	for	ADP
cril-1051	357	4	language	language	NOUN
cril-1051	357	5	description	description	NOUN
cril-1051	357	6	and	and	CCONJ
cril-1051	357	7	documentation	documentation	NOUN
cril-1051	357	8	21	21	NUM
cril-1051	357	9	foley	foley	PROPN
cril-1051	357	10	,	,	PUNCT
cril-1051	357	11	ben	ben	PROPN
cril-1051	357	12	,	,	PUNCT
cril-1051	357	13	josh	josh	PROPN
cril-1051	357	14	arnold	arnold	PROPN
cril-1051	357	15	,	,	PUNCT
cril-1051	357	16	rolando	rolando	PROPN
cril-1051	357	17	coto	coto	PROPN
cril-1051	357	18	-	-	PUNCT
cril-1051	357	19	solano	solano	PROPN
cril-1051	357	20	,	,	PUNCT
cril-1051	357	21	gautier	gauti	ADJ
cril-1051	357	22	durantin	durantin	PROPN
cril-1051	357	23	,	,	PUNCT
cril-1051	357	24	t.	t.	PROPN
cril-1051	357	25	mark	mark	PROPN
cril-1051	357	26	ellison	ellison	PROPN
cril-1051	357	27	,	,	PUNCT
cril-1051	357	28	daan	daan	PROPN
cril-1051	357	29	van	van	PROPN
cril-1051	357	30	esch	esch	PROPN
cril-1051	357	31	,	,	PUNCT
cril-1051	357	32	scott	scott	PROPN
cril-1051	357	33	heath	heath	PROPN
cril-1051	357	34	,	,	PUNCT
cril-1051	357	35	et	et	PROPN
cril-1051	357	36	al	al	PROPN
cril-1051	357	37	.	.	PROPN
cril-1051	357	38	2018	2018	NUM
cril-1051	357	39	.	.	PUNCT
cril-1051	358	1	building	build	VERB
cril-1051	358	2	speech	speech	NOUN
cril-1051	358	3	recognition	recognition	NOUN
cril-1051	358	4	systems	system	NOUN
cril-1051	358	5	for	for	ADP
cril-1051	358	6	language	language	NOUN
cril-1051	358	7	documentation	documentation	NOUN
cril-1051	358	8	:	:	PUNCT
cril-1051	358	9	the	the	DET
cril-1051	358	10	coedl	coedl	PROPN
cril-1051	358	11	endangered	endanger	VERB
cril-1051	358	12	language	language	NOUN
cril-1051	358	13	pipeline	pipeline	NOUN
cril-1051	358	14	and	and	CCONJ
cril-1051	358	15	inference	inference	NOUN
cril-1051	358	16	system	system	NOUN
cril-1051	358	17	.	.	PUNCT
cril-1051	359	1	in	in	ADP
cril-1051	359	2	proceedings	proceeding	NOUN
cril-1051	359	3	of	of	ADP
cril-1051	359	4	the	the	DET
cril-1051	359	5	6th	6th	ADJ
cril-1051	359	6	international	international	ADJ
cril-1051	359	7	workshop	workshop	NOUN
cril-1051	359	8	on	on	ADP
cril-1051	359	9	spoken	spoken	ADJ
cril-1051	359	10	language	language	NOUN
cril-1051	359	11	technologies	technology	NOUN
cril-1051	359	12	for	for	ADP
cril-1051	359	13	under	under	ADV
cril-1051	359	14	-	-	PUNCT
cril-1051	359	15	resourced	resource	VERB
cril-1051	359	16	languages	language	NOUN
cril-1051	359	17	(	(	PUNCT
cril-1051	359	18	sltu	sltu	VERB
cril-1051	359	19	2018	2018	NUM
cril-1051	359	20	)	)	PUNCT
cril-1051	359	21	.	.	PUNCT
cril-1051	360	1	https://www.iscaspeech.org/archive/sltu_2018/pdfs/ben.pdf	https://www.iscaspeech.org/archive/sltu_2018/pdfs/ben.pdf	NOUN
cril-1051	360	2	(	(	PUNCT
cril-1051	360	3	7	7	NUM
cril-1051	360	4	march	march	NOUN
cril-1051	360	5	,	,	PUNCT
cril-1051	360	6	2019	2019	NUM
cril-1051	360	7	)	)	PUNCT
cril-1051	360	8	.	.	PUNCT
cril-1051	361	1	forsberg	forsberg	PROPN
cril-1051	361	2	,	,	PUNCT
cril-1051	361	3	markus	markus	PROPN
cril-1051	361	4	&	&	CCONJ
cril-1051	361	5	mans	mans	PROPN
cril-1051	361	6	hulden	hulden	VERB
cril-1051	361	7	.	.	PUNCT
cril-1051	362	1	2016	2016	NUM
cril-1051	362	2	.	.	PUNCT
cril-1051	363	1	learning	learn	VERB
cril-1051	363	2	transducer	transducer	NOUN
cril-1051	363	3	models	model	NOUN
cril-1051	363	4	for	for	ADP
cril-1051	363	5	morphological	morphological	ADJ
cril-1051	363	6	analysis	analysis	NOUN
cril-1051	363	7	from	from	ADP
cril-1051	363	8	example	example	NOUN
cril-1051	363	9	inflections	inflection	NOUN
cril-1051	363	10	.	.	PUNCT
cril-1051	364	1	in	in	ADP
cril-1051	364	2	proceedings	proceeding	NOUN
cril-1051	364	3	of	of	ADP
cril-1051	364	4	the	the	DET
cril-1051	364	5	sigfsm	sigfsm	ADJ
cril-1051	364	6	workshop	workshop	NOUN
cril-1051	364	7	on	on	ADP
cril-1051	364	8	statistical	statistical	ADJ
cril-1051	364	9	nlp	nlp	NOUN
cril-1051	364	10	and	and	CCONJ
cril-1051	364	11	weighted	weighted	ADJ
cril-1051	364	12	automata	automata	NOUN
cril-1051	364	13	,	,	PUNCT
cril-1051	364	14	42–50	42–50	NOUN
cril-1051	364	15	.	.	PUNCT
cril-1051	365	1	berlin	berlin	PROPN
cril-1051	365	2	,	,	PUNCT
cril-1051	365	3	germany	germany	PROPN
cril-1051	365	4	:	:	PUNCT
cril-1051	365	5	association	association	NOUN
cril-1051	365	6	for	for	ADP
cril-1051	365	7	computational	computational	ADJ
cril-1051	365	8	linguistics	linguistic	NOUN
cril-1051	365	9	.	.	PUNCT
cril-1051	366	1	http://anthology.aclweb.org/w16-2405	http://anthology.aclweb.org/w16-2405	PROPN
cril-1051	366	2	(	(	PUNCT
cril-1051	366	3	18	18	NUM
cril-1051	366	4	january	january	PROPN
cril-1051	366	5	,	,	PUNCT
cril-1051	366	6	2019	2019	NUM
cril-1051	366	7	)	)	PUNCT
cril-1051	366	8	.	.	PUNCT
cril-1051	367	1	georgi	georgi	PROPN
cril-1051	367	2	,	,	PUNCT
cril-1051	367	3	ryan	ryan	PROPN
cril-1051	367	4	alden	alden	PROPN
cril-1051	367	5	.	.	PUNCT
cril-1051	367	6	2016	2016	NUM
cril-1051	367	7	.	.	PUNCT
cril-1051	368	1	from	from	ADP
cril-1051	368	2	aari	aari	ADJ
cril-1051	368	3	to	to	ADP
cril-1051	368	4	zulu	zulu	NUM
cril-1051	368	5	:	:	PUNCT
cril-1051	368	6	massively	massively	ADV
cril-1051	368	7	multilingual	multilingual	ADJ
cril-1051	368	8	creation	creation	NOUN
cril-1051	368	9	of	of	ADP
cril-1051	368	10	language	language	NOUN
cril-1051	368	11	tools	tool	NOUN
cril-1051	368	12	using	use	VERB
cril-1051	368	13	interlinear	interlinear	ADJ
cril-1051	368	14	glossed	glossed	ADJ
cril-1051	368	15	text	text	NOUN
cril-1051	368	16	.	.	PUNCT
cril-1051	369	1	phd	phd	NOUN
cril-1051	369	2	thesis	thesis	PROPN
cril-1051	369	3	.	.	PUNCT
cril-1051	370	1	https://digital.lib.washington.edu:443/researchworks/handle/1773/37168	https://digital.lib.washington.edu:443/researchworks/handle/1773/37168	NOUN
cril-1051	370	2	(	(	PUNCT
cril-1051	370	3	28	28	NUM
cril-1051	370	4	september	september	PROPN
cril-1051	370	5	,	,	PUNCT
cril-1051	370	6	2020	2020	NUM
cril-1051	370	7	)	)	PUNCT
cril-1051	370	8	.	.	PUNCT
cril-1051	371	1	goldberg	goldberg	PROPN
cril-1051	371	2	,	,	PUNCT
cril-1051	371	3	yoav	yoav	PROPN
cril-1051	371	4	.	.	PROPN
cril-1051	371	5	2017	2017	NUM
cril-1051	371	6	.	.	PUNCT
cril-1051	372	1	neural	neural	ADJ
cril-1051	372	2	network	network	NOUN
cril-1051	372	3	methods	method	NOUN
cril-1051	372	4	for	for	ADP
cril-1051	372	5	natural	natural	ADJ
cril-1051	372	6	language	language	NOUN
cril-1051	372	7	processing	processing	NOUN
cril-1051	372	8	(	(	PUNCT
cril-1051	372	9	synthesis	synthesis	NOUN
cril-1051	372	10	lectures	lecture	NOUN
cril-1051	372	11	on	on	ADP
cril-1051	372	12	human	human	ADJ
cril-1051	372	13	language	language	NOUN
cril-1051	372	14	technologies	technology	NOUN
cril-1051	372	15	37	37	NUM
cril-1051	372	16	)	)	PUNCT
cril-1051	372	17	.	.	PUNCT
cril-1051	373	1	morgan	morgan	PROPN
cril-1051	373	2	&	&	CCONJ
cril-1051	373	3	claypool	claypool	PROPN
cril-1051	373	4	.	.	PUNCT
cril-1051	374	1	http://www.morganclaypool.com/doi/10.2200/s00762ed1v01y201703hlt037	http://www.morganclaypool.com/doi/10.2200/s00762ed1v01y201703hlt037	PROPN
cril-1051	374	2	(	(	PUNCT
cril-1051	374	3	12	12	NUM
cril-1051	374	4	january	january	PROPN
cril-1051	374	5	,	,	PUNCT
cril-1051	374	6	2019	2019	NUM
cril-1051	374	7	)	)	PUNCT
cril-1051	374	8	.	.	PUNCT
cril-1051	375	1	goldsmith	goldsmith	PROPN
cril-1051	375	2	,	,	PUNCT
cril-1051	375	3	john	john	PROPN
cril-1051	375	4	.	.	PROPN
cril-1051	375	5	2001	2001	NUM
cril-1051	375	6	.	.	PUNCT
cril-1051	376	1	unsupervised	unsupervised	ADJ
cril-1051	376	2	learning	learning	NOUN
cril-1051	376	3	of	of	ADP
cril-1051	376	4	the	the	DET
cril-1051	376	5	morphology	morphology	NOUN
cril-1051	376	6	of	of	ADP
cril-1051	376	7	a	a	DET
cril-1051	376	8	natural	natural	ADJ
cril-1051	376	9	language	language	NOUN
cril-1051	376	10	.	.	PUNCT
cril-1051	377	1	computational	computational	ADJ
cril-1051	377	2	linguistics	linguistic	NOUN
cril-1051	377	3	27(2	27(2	NUM
cril-1051	377	4	)	)	PUNCT
cril-1051	377	5	.	.	PUNCT
cril-1051	378	1	153–198	153–198	NUM
cril-1051	378	2	.	.	PUNCT
cril-1051	378	3	goldsmith	goldsmith	PROPN
cril-1051	378	4	,	,	PUNCT
cril-1051	378	5	john	john	PROPN
cril-1051	378	6	,	,	PUNCT
cril-1051	378	7	jackson	jackson	PROPN
cril-1051	378	8	lee	lee	PROPN
cril-1051	378	9	&	&	CCONJ
cril-1051	378	10	aris	aris	PROPN
cril-1051	378	11	xanthos	xanthos	PROPN
cril-1051	378	12	.	.	PUNCT
cril-1051	379	1	2017	2017	NUM
cril-1051	379	2	.	.	PUNCT
cril-1051	380	1	computational	computational	ADJ
cril-1051	380	2	learning	learning	NOUN
cril-1051	380	3	of	of	ADP
cril-1051	380	4	morphology	morphology	NOUN
cril-1051	380	5	.	.	PUNCT
cril-1051	381	1	annual	annual	ADJ
cril-1051	381	2	review	review	NOUN
cril-1051	381	3	of	of	ADP
cril-1051	381	4	linguistics	linguistic	NOUN
cril-1051	381	5	3	3	NUM
cril-1051	381	6	.	.	PUNCT
cril-1051	382	1	85–106	85–106	NUM
cril-1051	382	2	.	.	PUNCT
cril-1051	383	1	hammarström	hammarström	PROPN
cril-1051	383	2	,	,	PUNCT
cril-1051	383	3	harald	harald	PROPN
cril-1051	383	4	&	&	CCONJ
cril-1051	383	5	lars	lars	PROPN
cril-1051	383	6	borin	borin	PROPN
cril-1051	383	7	.	.	PUNCT
cril-1051	384	1	2011	2011	NUM
cril-1051	384	2	.	.	PUNCT
cril-1051	385	1	unsupervised	unsupervised	ADJ
cril-1051	385	2	learning	learning	NOUN
cril-1051	385	3	of	of	ADP
cril-1051	385	4	morphology	morphology	NOUN
cril-1051	385	5	.	.	PUNCT
cril-1051	386	1	computational	computational	ADJ
cril-1051	386	2	linguistics	linguistic	NOUN
cril-1051	386	3	37(2	37(2	NUM
cril-1051	386	4	)	)	PUNCT
cril-1051	386	5	.	.	PUNCT
cril-1051	387	1	309–350	309–350	NUM
cril-1051	387	2	.	.	PUNCT
cril-1051	388	1	http://www.mitpressjournals.org/doi/abs/10.1162/coli_a_00050	http://www.mitpressjournals.org/doi/abs/10.1162/coli_a_00050	VERB
cril-1051	388	2	(	(	PUNCT
cril-1051	388	3	29	29	NUM
cril-1051	388	4	august	august	PROPN
cril-1051	388	5	,	,	PUNCT
cril-1051	388	6	2017	2017	NUM
cril-1051	388	7	)	)	PUNCT
cril-1051	388	8	.	.	PUNCT
cril-1051	389	1	harris	harris	PROPN
cril-1051	389	2	,	,	PUNCT
cril-1051	389	3	zellig	zellig	PROPN
cril-1051	389	4	.	.	PUNCT
cril-1051	389	5	1955	1955	NUM
cril-1051	389	6	.	.	PUNCT
cril-1051	390	1	from	from	ADP
cril-1051	390	2	phoneme	phoneme	NOUN
cril-1051	390	3	to	to	ADP
cril-1051	390	4	morpheme	morpheme	NOUN
cril-1051	390	5	.	.	PUNCT
cril-1051	391	1	language	language	NOUN
cril-1051	391	2	31(2	31(2	NUM
cril-1051	391	3	)	)	PUNCT
cril-1051	391	4	.	.	PUNCT
cril-1051	392	1	190–222	190–222	NUM
cril-1051	392	2	.	.	PUNCT
cril-1051	393	1	he	he	PRON
cril-1051	393	2	,	,	PUNCT
cril-1051	393	3	luheng	luheng	PROPN
cril-1051	393	4	,	,	PUNCT
cril-1051	393	5	julian	julian	PROPN
cril-1051	393	6	michael	michael	PROPN
cril-1051	393	7	,	,	PUNCT
cril-1051	393	8	mike	mike	PROPN
cril-1051	393	9	lewis	lewis	PROPN
cril-1051	393	10	&	&	CCONJ
cril-1051	393	11	luke	luke	PROPN
cril-1051	393	12	zettlemoyer	zettlemoyer	PROPN
cril-1051	393	13	.	.	PUNCT
cril-1051	394	1	2016	2016	NUM
cril-1051	394	2	.	.	PUNCT
cril-1051	395	1	human	human	NOUN
cril-1051	395	2	-	-	PUNCT
cril-1051	395	3	in	in	ADP
cril-1051	395	4	-	-	PUNCT
cril-1051	395	5	the	the	DET
cril-1051	395	6	-	-	PUNCT
cril-1051	395	7	loop	loop	NOUN
cril-1051	395	8	parsing	parsing	NOUN
cril-1051	395	9	.	.	PUNCT
cril-1051	396	1	in	in	ADP
cril-1051	396	2	proceedings	proceeding	NOUN
cril-1051	396	3	of	of	ADP
cril-1051	396	4	the	the	DET
cril-1051	396	5	2016	2016	NUM
cril-1051	396	6	conference	conference	NOUN
cril-1051	396	7	on	on	ADP
cril-1051	396	8	empirical	empirical	ADJ
cril-1051	396	9	methods	method	NOUN
cril-1051	396	10	in	in	ADP
cril-1051	396	11	natural	natural	ADJ
cril-1051	396	12	language	language	NOUN
cril-1051	396	13	processing	processing	NOUN
cril-1051	396	14	,	,	PUNCT
cril-1051	396	15	2337–2342	2337–2342	NUM
cril-1051	396	16	.	.	PUNCT
cril-1051	397	1	austin	austin	PROPN
cril-1051	397	2	,	,	PUNCT
cril-1051	397	3	texas	texas	PROPN
cril-1051	397	4	:	:	PUNCT
cril-1051	397	5	association	association	NOUN
cril-1051	397	6	for	for	ADP
cril-1051	397	7	computational	computational	ADJ
cril-1051	397	8	linguistic	linguistic	NOUN
cril-1051	397	9	.	.	PUNCT
cril-1051	398	1	colorado	colorado	PROPN
cril-1051	398	2	research	research	PROPN
cril-1051	398	3	in	in	ADP
cril-1051	398	4	linguistics	linguistic	NOUN
cril-1051	398	5	,	,	PUNCT
cril-1051	398	6	volume	volume	NOUN
cril-1051	398	7	25	25	NUM
cril-1051	398	8	(	(	PUNCT
cril-1051	398	9	2021	2021	NUM
cril-1051	398	10	)	)	PUNCT
cril-1051	398	11	22	22	NUM
cril-1051	398	12	himmelmann	himmelmann	NOUN
cril-1051	398	13	,	,	PUNCT
cril-1051	398	14	nikolaus	nikolaus	PROPN
cril-1051	398	15	p.	p.	NOUN
cril-1051	398	16	1998	1998	NUM
cril-1051	398	17	.	.	PUNCT
cril-1051	399	1	documentary	documentary	NOUN
cril-1051	399	2	and	and	CCONJ
cril-1051	399	3	descriptive	descriptive	ADJ
cril-1051	399	4	linguistics	linguistic	NOUN
cril-1051	399	5	.	.	PUNCT
cril-1051	400	1	linguistics	linguistic	NOUN
cril-1051	400	2	(	(	PUNCT
cril-1051	400	3	36	36	NUM
cril-1051	400	4	)	)	PUNCT
cril-1051	400	5	.	.	PUNCT
cril-1051	401	1	161–195	161–195	NUM
cril-1051	401	2	.	.	PUNCT
cril-1051	402	1	jin	jin	NOUN
cril-1051	402	2	,	,	PUNCT
cril-1051	402	3	huiming	huiming	NOUN
cril-1051	402	4	,	,	PUNCT
cril-1051	402	5	liwei	liwei	ADJ
cril-1051	402	6	cai	cai	PROPN
cril-1051	402	7	,	,	PUNCT
cril-1051	402	8	yihui	yihui	PROPN
cril-1051	402	9	peng	peng	PROPN
cril-1051	402	10	,	,	PUNCT
cril-1051	402	11	chen	chen	PROPN
cril-1051	402	12	xia	xia	PROPN
cril-1051	402	13	,	,	PUNCT
cril-1051	402	14	arya	arya	PROPN
cril-1051	402	15	mccarthy	mccarthy	PROPN
cril-1051	402	16	&	&	CCONJ
cril-1051	402	17	katharina	katharina	PROPN
cril-1051	402	18	kann	kann	PROPN
cril-1051	402	19	.	.	PROPN
cril-1051	402	20	2020	2020	NUM
cril-1051	402	21	.	.	PUNCT
cril-1051	403	1	unsupervised	unsupervised	ADJ
cril-1051	403	2	morphological	morphological	ADJ
cril-1051	403	3	paradigm	paradigm	NOUN
cril-1051	403	4	completion	completion	NOUN
cril-1051	403	5	.	.	PUNCT
cril-1051	404	1	in	in	ADP
cril-1051	404	2	proceedings	proceeding	NOUN
cril-1051	404	3	of	of	ADP
cril-1051	404	4	the	the	DET
cril-1051	404	5	58th	58th	ADJ
cril-1051	404	6	annual	annual	ADJ
cril-1051	404	7	meeting	meeting	NOUN
cril-1051	404	8	of	of	ADP
cril-1051	404	9	the	the	DET
cril-1051	404	10	association	association	NOUN
cril-1051	404	11	for	for	ADP
cril-1051	404	12	computational	computational	ADJ
cril-1051	404	13	linguistics	linguistic	NOUN
cril-1051	404	14	,	,	PUNCT
cril-1051	404	15	6696–6707	6696–6707	NUM
cril-1051	404	16	.	.	PUNCT
cril-1051	405	1	online	online	ADJ
cril-1051	405	2	:	:	PUNCT
cril-1051	405	3	association	association	NOUN
cril-1051	405	4	for	for	ADP
cril-1051	405	5	computational	computational	ADJ
cril-1051	405	6	linguistics	linguistic	NOUN
cril-1051	405	7	.	.	PUNCT
cril-1051	406	1	https://doi.org/10.18653/v1/2020.aclmain.598	https://doi.org/10.18653/v1/2020.aclmain.598	NOUN
cril-1051	406	2	.	.	PUNCT
cril-1051	407	1	https://www.aclweb.org/anthology/2020.acl-main.598	https://www.aclweb.org/anthology/2020.acl-main.598	ADJ
cril-1051	407	2	(	(	PUNCT
cril-1051	407	3	8	8	NUM
cril-1051	407	4	april	april	PROPN
cril-1051	407	5	,	,	PUNCT
cril-1051	407	6	2021	2021	NUM
cril-1051	407	7	)	)	PUNCT
cril-1051	407	8	.	.	PUNCT
cril-1051	408	1	kann	kann	PROPN
cril-1051	408	2	,	,	PUNCT
cril-1051	408	3	katharina	katharina	PROPN
cril-1051	408	4	,	,	PUNCT
cril-1051	408	5	samuel	samuel	PROPN
cril-1051	408	6	r.	r.	PROPN
cril-1051	408	7	bowman	bowman	PROPN
cril-1051	408	8	&	&	CCONJ
cril-1051	408	9	kyunghyun	kyunghyun	PROPN
cril-1051	408	10	cho	cho	PROPN
cril-1051	408	11	.	.	PUNCT
cril-1051	409	1	2020	2020	NUM
cril-1051	409	2	.	.	PUNCT
cril-1051	410	1	learning	learn	VERB
cril-1051	410	2	to	to	PART
cril-1051	410	3	learn	learn	VERB
cril-1051	410	4	morphological	morphological	ADJ
cril-1051	410	5	inflection	inflection	NOUN
cril-1051	410	6	for	for	ADP
cril-1051	410	7	resource	resource	NOUN
cril-1051	410	8	-	-	PUNCT
cril-1051	410	9	poor	poor	ADJ
cril-1051	410	10	languages	language	NOUN
cril-1051	410	11	.	.	PUNCT
cril-1051	411	1	proceedings	proceeding	NOUN
cril-1051	411	2	of	of	ADP
cril-1051	411	3	the	the	DET
cril-1051	411	4	aaai	aaai	PROPN
cril-1051	411	5	conference	conference	NOUN
cril-1051	411	6	on	on	ADP
cril-1051	411	7	artificial	artificial	ADJ
cril-1051	411	8	intelligence	intelligence	NOUN
cril-1051	411	9	34(05	34(05	PROPN
cril-1051	411	10	)	)	PUNCT
cril-1051	411	11	.	.	PUNCT
cril-1051	412	1	8058–8065	8058–8065	NUM
cril-1051	412	2	.	.	PUNCT
cril-1051	413	1	https://ojs.aaai.org/index.php/aaai/article/view/6316	https://ojs.aaai.org/index.php/aaai/article/view/6316	PROPN
cril-1051	413	2	(	(	PUNCT
cril-1051	413	3	26	26	NUM
cril-1051	413	4	january	january	PROPN
cril-1051	413	5	,	,	PUNCT
cril-1051	413	6	2021	2021	NUM
cril-1051	413	7	)	)	PUNCT
cril-1051	413	8	.	.	PUNCT
cril-1051	414	1	kann	kann	PROPN
cril-1051	414	2	,	,	PUNCT
cril-1051	414	3	katharina	katharina	PROPN
cril-1051	414	4	,	,	PUNCT
cril-1051	414	5	ryan	ryan	PROPN
cril-1051	414	6	cotterell	cotterell	PROPN
cril-1051	414	7	&	&	CCONJ
cril-1051	414	8	hinrich	hinrich	PROPN
cril-1051	414	9	schütze	schütze	PROPN
cril-1051	414	10	.	.	PROPN
cril-1051	414	11	2016	2016	NUM
cril-1051	414	12	.	.	PUNCT
cril-1051	415	1	neural	neural	ADJ
cril-1051	415	2	morphological	morphological	ADJ
cril-1051	415	3	analysis	analysis	NOUN
cril-1051	415	4	:	:	PUNCT
cril-1051	415	5	encoding	encoding	NOUN
cril-1051	415	6	-	-	PUNCT
cril-1051	415	7	decoding	decode	VERB
cril-1051	415	8	canonical	canonical	ADJ
cril-1051	415	9	segments	segment	NOUN
cril-1051	415	10	.	.	PUNCT
cril-1051	416	1	in	in	ADP
cril-1051	416	2	proceedings	proceeding	NOUN
cril-1051	416	3	of	of	ADP
cril-1051	416	4	the	the	DET
cril-1051	416	5	2016	2016	NUM
cril-1051	416	6	conference	conference	NOUN
cril-1051	416	7	on	on	ADP
cril-1051	416	8	empirical	empirical	ADJ
cril-1051	416	9	methods	method	NOUN
cril-1051	416	10	in	in	ADP
cril-1051	416	11	natural	natural	ADJ
cril-1051	416	12	language	language	NOUN
cril-1051	416	13	processing	processing	NOUN
cril-1051	416	14	,	,	PUNCT
cril-1051	416	15	961–967	961–967	NUM
cril-1051	416	16	.	.	PUNCT
cril-1051	417	1	austin	austin	PROPN
cril-1051	417	2	,	,	PUNCT
cril-1051	417	3	texas	texas	PROPN
cril-1051	417	4	:	:	PUNCT
cril-1051	417	5	association	association	NOUN
cril-1051	417	6	for	for	ADP
cril-1051	417	7	computational	computational	ADJ
cril-1051	417	8	linguistics	linguistic	NOUN
cril-1051	417	9	.	.	PUNCT
cril-1051	418	1	http://aclweb.org/anthology/d16-1097	http://aclweb.org/anthology/d16-1097	X
cril-1051	418	2	(	(	PUNCT
cril-1051	418	3	6	6	NUM
cril-1051	418	4	november	november	PROPN
cril-1051	418	5	,	,	PUNCT
cril-1051	418	6	2020	2020	NUM
cril-1051	418	7	)	)	PUNCT
cril-1051	418	8	.	.	PUNCT
cril-1051	419	1	kann	kann	PROPN
cril-1051	419	2	,	,	PUNCT
cril-1051	419	3	katharina	katharina	PROPN
cril-1051	419	4	,	,	PUNCT
cril-1051	419	5	ryan	ryan	PROPN
cril-1051	419	6	cotterell	cotterell	PROPN
cril-1051	419	7	&	&	CCONJ
cril-1051	419	8	hinrich	hinrich	PROPN
cril-1051	419	9	schütze	schütze	PROPN
cril-1051	419	10	.	.	PROPN
cril-1051	419	11	2017a	2017a	NUM
cril-1051	419	12	.	.	PUNCT
cril-1051	420	1	neural	neural	ADJ
cril-1051	420	2	multi	multi	ADJ
cril-1051	420	3	-	-	ADJ
cril-1051	420	4	source	source	ADJ
cril-1051	420	5	morphological	morphological	ADJ
cril-1051	420	6	reinflection	reinflection	NOUN
cril-1051	420	7	.	.	PUNCT
cril-1051	421	1	proceedings	proceeding	NOUN
cril-1051	421	2	of	of	ADP
cril-1051	421	3	the	the	DET
cril-1051	421	4	15th	15th	ADJ
cril-1051	421	5	conference	conference	NOUN
cril-1051	421	6	of	of	ADP
cril-1051	421	7	the	the	DET
cril-1051	421	8	european	european	ADJ
cril-1051	421	9	chapter	chapter	NOUN
cril-1051	421	10	of	of	ADP
cril-1051	421	11	the	the	DET
cril-1051	421	12	association	association	NOUN
cril-1051	421	13	for	for	ADP
cril-1051	421	14	computational	computational	ADJ
cril-1051	421	15	linguistics	linguistic	NOUN
cril-1051	421	16	:	:	PUNCT
cril-1051	421	17	volume	volume	NOUN
cril-1051	421	18	1	1	NUM
cril-1051	421	19	,	,	PUNCT
cril-1051	421	20	long	long	ADJ
cril-1051	421	21	papers	paper	NOUN
cril-1051	421	22	514–524	514–524	NUM
cril-1051	421	23	.	.	PUNCT
cril-1051	422	1	kann	kann	PROPN
cril-1051	422	2	,	,	PUNCT
cril-1051	422	3	katharina	katharina	PROPN
cril-1051	422	4	,	,	PUNCT
cril-1051	422	5	ryan	ryan	PROPN
cril-1051	422	6	cotterell	cotterell	PROPN
cril-1051	422	7	&	&	CCONJ
cril-1051	422	8	hinrich	hinrich	PROPN
cril-1051	422	9	schütze	schütze	PROPN
cril-1051	422	10	.	.	PROPN
cril-1051	422	11	2017b	2017b	NUM
cril-1051	422	12	.	.	PUNCT
cril-1051	423	1	one	one	NUM
cril-1051	423	2	-	-	PUNCT
cril-1051	423	3	shot	shot	NOUN
cril-1051	423	4	neural	neural	ADJ
cril-1051	423	5	cross	cross	ADJ
cril-1051	423	6	-	-	ADJ
cril-1051	423	7	lingual	lingual	ADJ
cril-1051	423	8	transfer	transfer	NOUN
cril-1051	423	9	for	for	ADP
cril-1051	423	10	paradigm	paradigm	NOUN
cril-1051	423	11	completion	completion	NOUN
cril-1051	423	12	.	.	PUNCT
cril-1051	424	1	in	in	ADP
cril-1051	424	2	proceedings	proceeding	NOUN
cril-1051	424	3	of	of	ADP
cril-1051	424	4	the	the	DET
cril-1051	424	5	55th	55th	ADJ
cril-1051	424	6	annual	annual	ADJ
cril-1051	424	7	meeting	meeting	NOUN
cril-1051	424	8	of	of	ADP
cril-1051	424	9	the	the	DET
cril-1051	424	10	association	association	NOUN
cril-1051	424	11	for	for	ADP
cril-1051	424	12	computational	computational	ADJ
cril-1051	424	13	linguistics	linguistic	NOUN
cril-1051	424	14	(	(	PUNCT
cril-1051	424	15	volume	volume	NOUN
cril-1051	424	16	1	1	NUM
cril-1051	424	17	:	:	PUNCT
cril-1051	424	18	long	long	ADJ
cril-1051	424	19	papers	paper	NOUN
cril-1051	424	20	)	)	PUNCT
cril-1051	424	21	,	,	PUNCT
cril-1051	424	22	1993–2003	1993–2003	NUM
cril-1051	424	23	.	.	PUNCT
cril-1051	425	1	vancouver	vancouver	PROPN
cril-1051	425	2	,	,	PUNCT
cril-1051	425	3	canada	canada	PROPN
cril-1051	425	4	:	:	PUNCT
cril-1051	425	5	association	association	NOUN
cril-1051	425	6	for	for	ADP
cril-1051	425	7	computational	computational	ADJ
cril-1051	425	8	linguistics	linguistic	NOUN
cril-1051	425	9	.	.	PUNCT
cril-1051	426	1	https://www.aclweb.org/anthology/p17-1182	https://www.aclweb.org/anthology/p17-1182	X
cril-1051	426	2	(	(	PUNCT
cril-1051	426	3	26	26	NUM
cril-1051	426	4	january	january	PROPN
cril-1051	426	5	,	,	PUNCT
cril-1051	426	6	2021	2021	NUM
cril-1051	426	7	)	)	PUNCT
cril-1051	426	8	.	.	PUNCT
cril-1051	427	1	kann	kann	PROPN
cril-1051	427	2	,	,	PUNCT
cril-1051	427	3	katharina	katharina	PROPN
cril-1051	427	4	,	,	PUNCT
cril-1051	427	5	arya	arya	PROPN
cril-1051	427	6	d.	d.	PROPN
cril-1051	427	7	mccarthy	mccarthy	PROPN
cril-1051	427	8	,	,	PUNCT
cril-1051	427	9	garrett	garrett	PROPN
cril-1051	427	10	nicolai	nicolai	PROPN
cril-1051	427	11	&	&	CCONJ
cril-1051	427	12	mans	mans	PROPN
cril-1051	427	13	hulden	hulden	VERB
cril-1051	427	14	.	.	PUNCT
cril-1051	428	1	2020	2020	NUM
cril-1051	428	2	.	.	PUNCT
cril-1051	429	1	the	the	DET
cril-1051	429	2	sigmorphon	sigmorphon	NOUN
cril-1051	429	3	2020	2020	NUM
cril-1051	429	4	shared	share	VERB
cril-1051	429	5	task	task	NOUN
cril-1051	429	6	on	on	ADP
cril-1051	429	7	unsupervised	unsupervised	ADJ
cril-1051	429	8	morphological	morphological	ADJ
cril-1051	429	9	paradigm	paradigm	NOUN
cril-1051	429	10	completion	completion	NOUN
cril-1051	429	11	.	.	PUNCT
cril-1051	430	1	in	in	ADP
cril-1051	430	2	proceedings	proceeding	NOUN
cril-1051	430	3	of	of	ADP
cril-1051	430	4	the	the	DET
cril-1051	430	5	17th	17th	ADJ
cril-1051	430	6	sigmorphon	sigmorphon	NOUN
cril-1051	430	7	workshop	workshop	NOUN
cril-1051	430	8	on	on	ADP
cril-1051	430	9	computational	computational	ADJ
cril-1051	430	10	research	research	NOUN
cril-1051	430	11	in	in	ADP
cril-1051	430	12	phonetics	phonetic	NOUN
cril-1051	430	13	,	,	PUNCT
cril-1051	430	14	phonology	phonology	NOUN
cril-1051	430	15	,	,	PUNCT
cril-1051	430	16	and	and	CCONJ
cril-1051	430	17	morphology	morphology	NOUN
cril-1051	430	18	,	,	PUNCT
cril-1051	430	19	51–62	51–62	NUM
cril-1051	430	20	.	.	PUNCT
cril-1051	430	21	online	online	ADJ
cril-1051	430	22	:	:	PUNCT
cril-1051	430	23	association	association	NOUN
cril-1051	430	24	for	for	ADP
cril-1051	430	25	computational	computational	ADJ
cril-1051	430	26	linguistics	linguistic	NOUN
cril-1051	430	27	.	.	PUNCT
cril-1051	431	1	https://www.aclweb.org/anthology/2020.sigmorphon-1.3	https://www.aclweb.org/anthology/2020.sigmorphon-1.3	PROPN
cril-1051	431	2	(	(	PUNCT
cril-1051	431	3	8	8	NUM
cril-1051	431	4	april	april	PROPN
cril-1051	431	5	,	,	PUNCT
cril-1051	431	6	2021	2021	NUM
cril-1051	431	7	)	)	PUNCT
cril-1051	431	8	.	.	PUNCT
cril-1051	432	1	computational	computational	ADJ
cril-1051	432	2	morphology	morphology	NOUN
cril-1051	432	3	for	for	ADP
cril-1051	432	4	language	language	NOUN
cril-1051	432	5	description	description	NOUN
cril-1051	432	6	and	and	CCONJ
cril-1051	432	7	documentation	documentation	NOUN
cril-1051	432	8	23	23	NUM
cril-1051	432	9	kann	kann	PROPN
cril-1051	432	10	,	,	PUNCT
cril-1051	432	11	katharina	katharina	PROPN
cril-1051	432	12	&	&	CCONJ
cril-1051	432	13	hinrich	hinrich	PROPN
cril-1051	432	14	schütze	schütze	PROPN
cril-1051	432	15	.	.	PROPN
cril-1051	432	16	2016	2016	NUM
cril-1051	432	17	.	.	PUNCT
cril-1051	433	1	single	single	ADJ
cril-1051	433	2	-	-	PUNCT
cril-1051	433	3	model	model	NOUN
cril-1051	433	4	encoder	encoder	NOUN
cril-1051	433	5	-	-	PUNCT
cril-1051	433	6	decoder	decoder	NOUN
cril-1051	433	7	with	with	ADP
cril-1051	433	8	explicit	explicit	ADJ
cril-1051	433	9	morphological	morphological	ADJ
cril-1051	433	10	representation	representation	NOUN
cril-1051	433	11	for	for	ADP
cril-1051	433	12	reinflection	reinflection	NOUN
cril-1051	433	13	.	.	PUNCT
cril-1051	434	1	in	in	ADP
cril-1051	434	2	proceedings	proceeding	NOUN
cril-1051	434	3	of	of	ADP
cril-1051	434	4	the	the	DET
cril-1051	434	5	54th	54th	ADJ
cril-1051	434	6	annual	annual	ADJ
cril-1051	434	7	meeting	meeting	NOUN
cril-1051	434	8	of	of	ADP
cril-1051	434	9	the	the	DET
cril-1051	434	10	association	association	NOUN
cril-1051	434	11	for	for	ADP
cril-1051	434	12	computational	computational	ADJ
cril-1051	434	13	linguistics	linguistic	NOUN
cril-1051	434	14	(	(	PUNCT
cril-1051	434	15	volume	volume	NOUN
cril-1051	434	16	2	2	NUM
cril-1051	434	17	:	:	PUNCT
cril-1051	434	18	short	short	ADJ
cril-1051	434	19	papers	paper	NOUN
cril-1051	434	20	)	)	PUNCT
cril-1051	434	21	,	,	PUNCT
cril-1051	434	22	555–560	555–560	NUM
cril-1051	434	23	.	.	PUNCT
cril-1051	435	1	berlin	berlin	PROPN
cril-1051	435	2	,	,	PUNCT
cril-1051	435	3	germany	germany	PROPN
cril-1051	435	4	:	:	PUNCT
cril-1051	435	5	association	association	NOUN
cril-1051	435	6	for	for	ADP
cril-1051	435	7	computational	computational	ADJ
cril-1051	435	8	linguistics	linguistic	NOUN
cril-1051	435	9	.	.	PUNCT
cril-1051	436	1	http://anthology.aclweb.org/p16-2090	http://anthology.aclweb.org/p16-2090	PROPN
cril-1051	436	2	(	(	PUNCT
cril-1051	436	3	5	5	NUM
cril-1051	436	4	november	november	PROPN
cril-1051	436	5	,	,	PUNCT
cril-1051	436	6	2018	2018	NUM
cril-1051	436	7	)	)	PUNCT
cril-1051	436	8	.	.	PUNCT
cril-1051	437	1	karttunen	karttunen	PROPN
cril-1051	437	2	,	,	PUNCT
cril-1051	437	3	lauri	lauri	PROPN
cril-1051	437	4	&	&	CCONJ
cril-1051	437	5	kenneth	kenneth	PROPN
cril-1051	437	6	r.	r.	PROPN
cril-1051	437	7	beesley	beesley	PROPN
cril-1051	437	8	.	.	PUNCT
cril-1051	438	1	2005	2005	NUM
cril-1051	438	2	.	.	PUNCT
cril-1051	439	1	twenty	twenty	NUM
cril-1051	439	2	-	-	PUNCT
cril-1051	439	3	five	five	NUM
cril-1051	439	4	years	year	NOUN
cril-1051	439	5	of	of	ADP
cril-1051	439	6	finite	finite	ADJ
cril-1051	439	7	-	-	ADJ
cril-1051	439	8	state	state	ADJ
cril-1051	439	9	morphology	morphology	NOUN
cril-1051	439	10	.	.	PUNCT
cril-1051	440	1	in	in	ADP
cril-1051	440	2	inquiries	inquiry	NOUN
cril-1051	440	3	into	into	ADP
cril-1051	440	4	words	word	NOUN
cril-1051	440	5	,	,	PUNCT
cril-1051	440	6	a	a	DET
cril-1051	440	7	festschrift	festschrift	NOUN
cril-1051	440	8	for	for	ADP
cril-1051	440	9	kimmo	kimmo	PROPN
cril-1051	440	10	koskenniemi	koskenniemi	PROPN
cril-1051	440	11	on	on	ADP
cril-1051	440	12	his	his	PRON
cril-1051	440	13	60th	60th	ADJ
cril-1051	440	14	birthday	birthday	NOUN
cril-1051	440	15	,	,	PUNCT
cril-1051	440	16	71–83	71–83	NUM
cril-1051	440	17	.	.	PUNCT
cril-1051	441	1	csli	csli	PROPN
cril-1051	441	2	publications	publication	NOUN
cril-1051	441	3	.	.	PUNCT
cril-1051	442	1	kirschenbaum	kirschenbaum	PROPN
cril-1051	442	2	,	,	PUNCT
cril-1051	442	3	amit	amit	PROPN
cril-1051	442	4	,	,	PUNCT
cril-1051	442	5	peter	peter	PROPN
cril-1051	442	6	wittenburg	wittenburg	PROPN
cril-1051	442	7	&	&	CCONJ
cril-1051	442	8	gerhard	gerhard	PROPN
cril-1051	442	9	heyer	heyer	PROPN
cril-1051	442	10	.	.	PUNCT
cril-1051	442	11	2012	2012	NUM
cril-1051	442	12	.	.	PUNCT
cril-1051	443	1	unsupervised	unsupervised	ADJ
cril-1051	443	2	morphological	morphological	ADJ
cril-1051	443	3	analysis	analysis	NOUN
cril-1051	443	4	of	of	ADP
cril-1051	443	5	small	small	ADJ
cril-1051	443	6	corpora	corpora	NOUN
cril-1051	443	7	:	:	PUNCT
cril-1051	443	8	first	first	ADJ
cril-1051	443	9	experiments	experiment	NOUN
cril-1051	443	10	with	with	ADP
cril-1051	443	11	kilivila	kilivila	PROPN
cril-1051	443	12	.	.	PUNCT
cril-1051	444	1	language	language	NOUN
cril-1051	444	2	documentation	documentation	NOUN
cril-1051	444	3	&	&	CCONJ
cril-1051	444	4	conservation	conservation	NOUN
cril-1051	444	5	special	special	ADJ
cril-1051	444	6	publication	publication	NOUN
cril-1051	444	7	(	(	PUNCT
cril-1051	444	8	potentials	potential	NOUN
cril-1051	444	9	of	of	ADP
cril-1051	444	10	language	language	NOUN
cril-1051	444	11	documentation	documentation	NOUN
cril-1051	444	12	:	:	PUNCT
cril-1051	444	13	methods	method	NOUN
cril-1051	444	14	,	,	PUNCT
cril-1051	444	15	analyses	analysis	NOUN
cril-1051	444	16	,	,	PUNCT
cril-1051	444	17	and	and	CCONJ
cril-1051	444	18	utilization	utilization	NOUN
cril-1051	444	19	)	)	PUNCT
cril-1051	444	20	3	3	NUM
cril-1051	444	21	.	.	X
cril-1051	444	22	25–31	25–31	NUM
cril-1051	444	23	.	.	PUNCT
cril-1051	444	24	http://scholarspace.manoa.hawaii.edu/bitstream/handle/10125/4513/04kirschenbaumetal	http://scholarspace.manoa.hawaii.edu/bitstream/handle/10125/4513/04kirschenbaumetal	PROPN
cril-1051	444	25	.	.	PUNCT
cril-1051	445	1	pdf	pdf	PROPN
cril-1051	445	2	.	.	PUNCT
cril-1051	446	1	kohonen	kohonen	PROPN
cril-1051	446	2	,	,	PUNCT
cril-1051	446	3	oskar	oskar	PROPN
cril-1051	446	4	,	,	PUNCT
cril-1051	446	5	sami	sami	PROPN
cril-1051	446	6	virpioja	virpioja	PROPN
cril-1051	446	7	&	&	CCONJ
cril-1051	446	8	krista	krista	PROPN
cril-1051	446	9	lagus	lagus	PROPN
cril-1051	446	10	.	.	PUNCT
cril-1051	447	1	2010	2010	NUM
cril-1051	447	2	.	.	PUNCT
cril-1051	448	1	semi	semi	ADJ
cril-1051	448	2	-	-	ADJ
cril-1051	448	3	supervised	supervised	ADJ
cril-1051	448	4	learning	learning	NOUN
cril-1051	448	5	of	of	ADP
cril-1051	448	6	concatenative	concatenative	ADJ
cril-1051	448	7	morphology	morphology	NOUN
cril-1051	448	8	.	.	PUNCT
cril-1051	449	1	in	in	ADP
cril-1051	449	2	proceedings	proceeding	NOUN
cril-1051	449	3	of	of	ADP
cril-1051	449	4	the	the	DET
cril-1051	449	5	11th	11th	ADJ
cril-1051	449	6	meeting	meeting	NOUN
cril-1051	449	7	of	of	ADP
cril-1051	449	8	the	the	DET
cril-1051	449	9	acl	acl	PROPN
cril-1051	449	10	special	special	ADJ
cril-1051	449	11	interest	interest	NOUN
cril-1051	449	12	group	group	NOUN
cril-1051	449	13	on	on	ADP
cril-1051	449	14	computational	computational	ADJ
cril-1051	449	15	morphology	morphology	NOUN
cril-1051	449	16	and	and	CCONJ
cril-1051	449	17	phonology	phonology	NOUN
cril-1051	449	18	,	,	PUNCT
cril-1051	449	19	78–86	78–86	NUM
cril-1051	449	20	.	.	PUNCT
cril-1051	449	21	association	association	NOUN
cril-1051	449	22	for	for	ADP
cril-1051	449	23	computational	computational	ADJ
cril-1051	449	24	linguistics	linguistic	NOUN
cril-1051	449	25	.	.	PUNCT
cril-1051	450	1	lecun	lecun	PROPN
cril-1051	450	2	,	,	PUNCT
cril-1051	450	3	yann	yann	PROPN
cril-1051	450	4	,	,	PUNCT
cril-1051	450	5	yoshua	yoshua	PROPN
cril-1051	450	6	bengio	bengio	PROPN
cril-1051	450	7	&	&	CCONJ
cril-1051	450	8	geoffrey	geoffrey	PROPN
cril-1051	450	9	hinton	hinton	PROPN
cril-1051	450	10	.	.	PUNCT
cril-1051	451	1	2015	2015	NUM
cril-1051	451	2	.	.	PUNCT
cril-1051	452	1	deep	deep	ADJ
cril-1051	452	2	learning	learning	NOUN
cril-1051	452	3	.	.	PUNCT
cril-1051	453	1	nature	nature	NOUN
cril-1051	453	2	521(7553	521(7553	NUM
cril-1051	453	3	)	)	PUNCT
cril-1051	453	4	.	.	PUNCT
cril-1051	454	1	436	436	NUM
cril-1051	454	2	–	–	PUNCT
cril-1051	454	3	444	444	NUM
cril-1051	454	4	.	.	PUNCT
cril-1051	454	5	https://doi.org/10.1038/nature14539	https://doi.org/10.1038/nature14539	NOUN
cril-1051	454	6	.	.	PUNCT
cril-1051	455	1	https://doi.org/10.1038/nature14539	https://doi.org/10.1038/nature14539	PROPN
cril-1051	455	2	.	.	PUNCT
cril-1051	456	1	lepp	lepp	PROPN
cril-1051	456	2	,	,	PUNCT
cril-1051	456	3	haley	haley	NOUN
cril-1051	456	4	,	,	PUNCT
cril-1051	456	5	olga	olga	PROPN
cril-1051	456	6	zamaraeva	zamaraeva	PROPN
cril-1051	456	7	&	&	CCONJ
cril-1051	456	8	emily	emily	PROPN
cril-1051	456	9	m.	m.	PROPN
cril-1051	456	10	bender	bender	PROPN
cril-1051	456	11	.	.	PUNCT
cril-1051	457	1	2019	2019	NUM
cril-1051	457	2	.	.	PUNCT
cril-1051	458	1	visualizing	visualize	VERB
cril-1051	458	2	inferred	infer	VERB
cril-1051	458	3	morphotactic	morphotactic	ADJ
cril-1051	458	4	systems	system	NOUN
cril-1051	458	5	.	.	PUNCT
cril-1051	459	1	in	in	ADP
cril-1051	459	2	proceedings	proceeding	NOUN
cril-1051	459	3	of	of	ADP
cril-1051	459	4	the	the	DET
cril-1051	459	5	2019	2019	NUM
cril-1051	459	6	conference	conference	NOUN
cril-1051	459	7	of	of	ADP
cril-1051	459	8	the	the	DET
cril-1051	459	9	north	north	ADJ
cril-1051	459	10	american	american	ADJ
cril-1051	459	11	chapter	chapter	NOUN
cril-1051	459	12	of	of	ADP
cril-1051	459	13	the	the	DET
cril-1051	459	14	association	association	NOUN
cril-1051	459	15	for	for	ADP
cril-1051	459	16	computational	computational	ADJ
cril-1051	459	17	linguistics	linguistic	NOUN
cril-1051	459	18	(	(	PUNCT
cril-1051	459	19	demonstrations	demonstration	NOUN
cril-1051	459	20	)	)	PUNCT
cril-1051	459	21	,	,	PUNCT
cril-1051	459	22	127–131	127–131	NUM
cril-1051	459	23	.	.	PUNCT
cril-1051	459	24	minneapolis	minneapolis	PROPN
cril-1051	459	25	,	,	PUNCT
cril-1051	459	26	minnesota	minnesota	PROPN
cril-1051	459	27	:	:	PUNCT
cril-1051	459	28	association	association	NOUN
cril-1051	459	29	for	for	ADP
cril-1051	459	30	computational	computational	ADJ
cril-1051	459	31	linguistics	linguistic	NOUN
cril-1051	459	32	.	.	PUNCT
cril-1051	459	33	https://www.aclweb.org/anthology/n19-4022	https://www.aclweb.org/anthology/n19-4022	X
cril-1051	459	34	(	(	PUNCT
cril-1051	459	35	6	6	NUM
cril-1051	459	36	april	april	PROPN
cril-1051	459	37	,	,	PUNCT
cril-1051	459	38	2020	2020	NUM
cril-1051	459	39	)	)	PUNCT
cril-1051	459	40	.	.	PUNCT
cril-1051	460	1	lewis	lewis	PROPN
cril-1051	460	2	,	,	PUNCT
cril-1051	460	3	william	william	PROPN
cril-1051	460	4	d.	d.	PROPN
cril-1051	460	5	&	&	CCONJ
cril-1051	460	6	fei	fei	PROPN
cril-1051	460	7	xia	xia	PROPN
cril-1051	460	8	.	.	PROPN
cril-1051	460	9	2010	2010	NUM
cril-1051	460	10	.	.	PUNCT
cril-1051	461	1	developing	develop	VERB
cril-1051	461	2	odin	odin	NOUN
cril-1051	461	3	:	:	PUNCT
cril-1051	461	4	a	a	DET
cril-1051	461	5	multilingual	multilingual	ADJ
cril-1051	461	6	repository	repository	NOUN
cril-1051	461	7	of	of	ADP
cril-1051	461	8	annotated	annotate	VERB
cril-1051	461	9	language	language	NOUN
cril-1051	461	10	data	datum	NOUN
cril-1051	461	11	for	for	ADP
cril-1051	461	12	hundreds	hundred	NOUN
cril-1051	461	13	of	of	ADP
cril-1051	461	14	the	the	DET
cril-1051	461	15	world	world	NOUN
cril-1051	461	16	’s	’s	PART
cril-1051	461	17	languages	language	NOUN
cril-1051	461	18	.	.	PUNCT
cril-1051	462	1	literary	literary	ADJ
cril-1051	462	2	and	and	CCONJ
cril-1051	462	3	linguistic	linguistic	ADJ
cril-1051	462	4	computing	computing	NOUN
cril-1051	462	5	25(3	25(3	NUM
cril-1051	462	6	)	)	PUNCT
cril-1051	462	7	.	.	PUNCT
cril-1051	463	1	303–319	303–319	NUM
cril-1051	463	2	.	.	PUNCT
cril-1051	464	1	http://llc.oxfordjournals.org/content/25/3/303	http://llc.oxfordjournals.org/content/25/3/303	NOUN
cril-1051	464	2	(	(	PUNCT
cril-1051	464	3	29	29	NUM
cril-1051	464	4	january	january	PROPN
cril-1051	464	5	,	,	PUNCT
cril-1051	464	6	2014	2014	NUM
cril-1051	464	7	)	)	PUNCT
cril-1051	464	8	.	.	PUNCT
cril-1051	465	1	liu	liu	PROPN
cril-1051	465	2	,	,	PUNCT
cril-1051	465	3	ling	ling	PROPN
cril-1051	465	4	,	,	PUNCT
cril-1051	465	5	ilamvazhuthy	ilamvazhuthy	ADJ
cril-1051	465	6	subbiah	subbiah	PROPN
cril-1051	465	7	,	,	PUNCT
cril-1051	465	8	adam	adam	PROPN
cril-1051	465	9	wiemerslage	wiemerslage	PROPN
cril-1051	465	10	,	,	PUNCT
cril-1051	465	11	jonathan	jonathan	PROPN
cril-1051	465	12	lilley	lilley	PROPN
cril-1051	465	13	&	&	CCONJ
cril-1051	465	14	sarah	sarah	PROPN
cril-1051	465	15	moeller	moeller	PROPN
cril-1051	465	16	.	.	PUNCT
cril-1051	465	17	2018	2018	NUM
cril-1051	465	18	.	.	PUNCT
cril-1051	466	1	morphological	morphological	ADJ
cril-1051	466	2	reinflection	reinflection	NOUN
cril-1051	466	3	in	in	ADP
cril-1051	466	4	context	context	NOUN
cril-1051	466	5	:	:	PUNCT
cril-1051	466	6	cu	cu	PROPN
cril-1051	466	7	boulder	boulder	PROPN
cril-1051	466	8	’s	’s	PART
cril-1051	466	9	submission	submission	NOUN
cril-1051	466	10	to	to	PART
cril-1051	466	11	conllcolorado	conllcolorado	VERB
cril-1051	466	12	research	research	NOUN
cril-1051	466	13	in	in	ADP
cril-1051	466	14	linguistics	linguistic	NOUN
cril-1051	466	15	,	,	PUNCT
cril-1051	466	16	volume	volume	NOUN
cril-1051	466	17	25	25	NUM
cril-1051	466	18	(	(	PUNCT
cril-1051	466	19	2021	2021	NUM
cril-1051	466	20	)	)	PUNCT
cril-1051	466	21	24	24	NUM
cril-1051	466	22	sigmorphon	sigmorphon	NOUN
cril-1051	466	23	2018	2018	NUM
cril-1051	466	24	shared	share	VERB
cril-1051	466	25	task	task	NOUN
cril-1051	466	26	.	.	PUNCT
cril-1051	467	1	in	in	ADP
cril-1051	467	2	proceedings	proceeding	NOUN
cril-1051	467	3	of	of	ADP
cril-1051	467	4	the	the	DET
cril-1051	467	5	conll	conll	PROPN
cril-1051	467	6	sigmorphon	sigmorphon	PROPN
cril-1051	467	7	2018	2018	NUM
cril-1051	467	8	shared	share	VERB
cril-1051	467	9	task	task	NOUN
cril-1051	467	10	:	:	PUNCT
cril-1051	467	11	universal	universal	ADJ
cril-1051	467	12	morphological	morphological	ADJ
cril-1051	467	13	reinflection	reinflection	NOUN
cril-1051	467	14	,	,	PUNCT
cril-1051	467	15	86–92	86–92	NUM
cril-1051	467	16	.	.	PUNCT
cril-1051	468	1	brussels	brussels	PROPN
cril-1051	468	2	:	:	PUNCT
cril-1051	468	3	association	association	NOUN
cril-1051	468	4	for	for	ADP
cril-1051	468	5	computational	computational	ADJ
cril-1051	468	6	linguistics	linguistic	NOUN
cril-1051	468	7	.	.	PUNCT
cril-1051	469	1	http://www.aclweb.org/anthology/k18-3010	http://www.aclweb.org/anthology/k18-3010	X
cril-1051	469	2	(	(	PUNCT
cril-1051	469	3	2	2	NUM
cril-1051	469	4	november	november	PROPN
cril-1051	469	5	,	,	PUNCT
cril-1051	469	6	2018	2018	NUM
cril-1051	469	7	)	)	PUNCT
cril-1051	469	8	.	.	PUNCT
cril-1051	470	1	lupke	lupke	PROPN
cril-1051	470	2	,	,	PUNCT
cril-1051	470	3	friederike	friederike	INTJ
cril-1051	470	4	.	.	PUNCT
cril-1051	471	1	2010	2010	NUM
cril-1051	471	2	.	.	PUNCT
cril-1051	472	1	data	datum	NOUN
cril-1051	472	2	collection	collection	NOUN
cril-1051	472	3	methods	method	NOUN
cril-1051	472	4	for	for	ADP
cril-1051	472	5	field	field	NOUN
cril-1051	472	6	-	-	PUNCT
cril-1051	472	7	based	base	VERB
cril-1051	472	8	language	language	NOUN
cril-1051	472	9	documentation	documentation	NOUN
cril-1051	472	10	.	.	PUNCT
cril-1051	473	1	language	language	NOUN
cril-1051	473	2	documentation	documentation	NOUN
cril-1051	473	3	and	and	CCONJ
cril-1051	473	4	description	description	NOUN
cril-1051	473	5	7	7	NUM
cril-1051	473	6	.	.	PUNCT
cril-1051	474	1	55–104	55–104	PROPN
cril-1051	474	2	.	.	PUNCT
cril-1051	474	3	makarov	makarov	PROPN
cril-1051	474	4	,	,	PUNCT
cril-1051	474	5	peter	peter	PROPN
cril-1051	474	6	&	&	CCONJ
cril-1051	474	7	simon	simon	PROPN
cril-1051	474	8	clematide	clematide	PROPN
cril-1051	474	9	.	.	PUNCT
cril-1051	474	10	2018a	2018a	NUM
cril-1051	474	11	.	.	PUNCT
cril-1051	475	1	uzh	uzh	PROPN
cril-1051	475	2	at	at	ADP
cril-1051	475	3	conll	conll	PROPN
cril-1051	475	4	–	–	PUNCT
cril-1051	475	5	sigmorphon	sigmorphon	NOUN
cril-1051	475	6	2018	2018	NUM
cril-1051	475	7	shared	share	VERB
cril-1051	475	8	task	task	NOUN
cril-1051	475	9	on	on	ADP
cril-1051	475	10	universal	universal	ADJ
cril-1051	475	11	morphological	morphological	ADJ
cril-1051	475	12	reinflection	reinflection	NOUN
cril-1051	475	13	.	.	PUNCT
cril-1051	476	1	in	in	ADP
cril-1051	476	2	proceedings	proceeding	NOUN
cril-1051	476	3	of	of	ADP
cril-1051	476	4	the	the	DET
cril-1051	476	5	conll	conll	PROPN
cril-1051	476	6	–	–	PUNCT
cril-1051	476	7	sigmorphon	sigmorphon	NOUN
cril-1051	476	8	2018	2018	NUM
cril-1051	476	9	shared	share	VERB
cril-1051	476	10	task	task	NOUN
cril-1051	476	11	:	:	PUNCT
cril-1051	476	12	universal	universal	ADJ
cril-1051	476	13	morphological	morphological	ADJ
cril-1051	476	14	reinflection	reinflection	NOUN
cril-1051	476	15	,	,	PUNCT
cril-1051	476	16	69–75	69–75	NUM
cril-1051	476	17	.	.	PUNCT
cril-1051	477	1	brussels	brussels	PROPN
cril-1051	477	2	:	:	PUNCT
cril-1051	477	3	association	association	NOUN
cril-1051	477	4	for	for	ADP
cril-1051	477	5	computational	computational	ADJ
cril-1051	477	6	linguistics	linguistic	NOUN
cril-1051	477	7	.	.	PUNCT
cril-1051	478	1	https://www.aclweb.org/anthology/k18-3008	https://www.aclweb.org/anthology/k18-3008	NOUN
cril-1051	478	2	(	(	PUNCT
cril-1051	478	3	16	16	NUM
cril-1051	478	4	may	may	NOUN
cril-1051	478	5	,	,	PUNCT
cril-1051	478	6	2019	2019	NUM
cril-1051	478	7	)	)	PUNCT
cril-1051	478	8	.	.	PUNCT
cril-1051	479	1	makarov	makarov	PROPN
cril-1051	479	2	,	,	PUNCT
cril-1051	479	3	peter	peter	PROPN
cril-1051	479	4	&	&	CCONJ
cril-1051	479	5	simon	simon	PROPN
cril-1051	479	6	clematide	clematide	PROPN
cril-1051	479	7	.	.	PUNCT
cril-1051	479	8	2018b	2018b	NUM
cril-1051	479	9	.	.	PUNCT
cril-1051	480	1	imitation	imitation	NOUN
cril-1051	480	2	learning	learn	VERB
cril-1051	480	3	for	for	ADP
cril-1051	480	4	neural	neural	ADJ
cril-1051	480	5	morphological	morphological	ADJ
cril-1051	480	6	string	string	NOUN
cril-1051	480	7	transduction	transduction	NOUN
cril-1051	480	8	.	.	PUNCT
cril-1051	481	1	in	in	ADP
cril-1051	481	2	proceedings	proceeding	NOUN
cril-1051	481	3	of	of	ADP
cril-1051	481	4	the	the	DET
cril-1051	481	5	2018	2018	NUM
cril-1051	481	6	conference	conference	NOUN
cril-1051	481	7	on	on	ADP
cril-1051	481	8	empirical	empirical	ADJ
cril-1051	481	9	methods	method	NOUN
cril-1051	481	10	in	in	ADP
cril-1051	481	11	natural	natural	ADJ
cril-1051	481	12	language	language	NOUN
cril-1051	481	13	processing	processing	NOUN
cril-1051	481	14	,	,	PUNCT
cril-1051	481	15	2877–2882	2877–2882	NUM
cril-1051	481	16	.	.	PUNCT
cril-1051	481	17	brussels	brussels	PROPN
cril-1051	481	18	,	,	PUNCT
cril-1051	481	19	belgium	belgium	PROPN
cril-1051	481	20	:	:	PUNCT
cril-1051	481	21	association	association	NOUN
cril-1051	481	22	for	for	ADP
cril-1051	481	23	computational	computational	ADJ
cril-1051	481	24	linguistics	linguistic	NOUN
cril-1051	481	25	.	.	PUNCT
cril-1051	482	1	https://www.aclweb.org/anthology/d18-1314	https://www.aclweb.org/anthology/d18-1314	PUNCT
cril-1051	482	2	(	(	PUNCT
cril-1051	482	3	28	28	NUM
cril-1051	482	4	april	april	PROPN
cril-1051	482	5	,	,	PUNCT
cril-1051	482	6	2021	2021	NUM
cril-1051	482	7	)	)	PUNCT
cril-1051	482	8	.	.	PUNCT
cril-1051	483	1	makarov	makarov	PROPN
cril-1051	483	2	,	,	PUNCT
cril-1051	483	3	peter	peter	PROPN
cril-1051	483	4	,	,	PUNCT
cril-1051	483	5	tatiana	tatiana	PROPN
cril-1051	483	6	ruzsics	ruzsics	PROPN
cril-1051	483	7	&	&	CCONJ
cril-1051	483	8	simon	simon	PROPN
cril-1051	483	9	clematide	clematide	PROPN
cril-1051	483	10	.	.	PUNCT
cril-1051	484	1	2017	2017	NUM
cril-1051	484	2	.	.	PUNCT
cril-1051	485	1	align	align	VERB
cril-1051	485	2	and	and	CCONJ
cril-1051	485	3	copy	copy	VERB
cril-1051	485	4	:	:	PUNCT
cril-1051	485	5	uzh	uzh	PROPN
cril-1051	485	6	at	at	ADP
cril-1051	485	7	sigmorphon	sigmorphon	NOUN
cril-1051	485	8	2017	2017	NUM
cril-1051	485	9	shared	share	VERB
cril-1051	485	10	task	task	NOUN
cril-1051	485	11	for	for	ADP
cril-1051	485	12	morphological	morphological	ADJ
cril-1051	485	13	reinflection	reinflection	NOUN
cril-1051	485	14	.	.	PUNCT
cril-1051	486	1	in	in	ADP
cril-1051	486	2	proceedings	proceeding	NOUN
cril-1051	486	3	of	of	ADP
cril-1051	486	4	the	the	DET
cril-1051	486	5	conll	conll	PROPN
cril-1051	486	6	sigmorphon	sigmorphon	PROPN
cril-1051	486	7	2017	2017	NUM
cril-1051	486	8	shared	share	VERB
cril-1051	486	9	task	task	NOUN
cril-1051	486	10	:	:	PUNCT
cril-1051	486	11	universal	universal	ADJ
cril-1051	486	12	morphological	morphological	ADJ
cril-1051	486	13	reinflection	reinflection	NOUN
cril-1051	486	14	,	,	PUNCT
cril-1051	486	15	49	49	NUM
cril-1051	486	16	–	–	SYM
cril-1051	486	17	57	57	NUM
cril-1051	486	18	.	.	PUNCT
cril-1051	487	1	vancouver	vancouver	PROPN
cril-1051	487	2	:	:	PUNCT
cril-1051	487	3	association	association	NOUN
cril-1051	487	4	for	for	ADP
cril-1051	487	5	computational	computational	ADJ
cril-1051	487	6	linguistics	linguistic	NOUN
cril-1051	487	7	.	.	PUNCT
cril-1051	488	1	http://www.aclweb.org/anthology/k17-2004	http://www.aclweb.org/anthology/k17-2004	PUNCT
cril-1051	488	2	(	(	PUNCT
cril-1051	488	3	3	3	NUM
cril-1051	488	4	november	november	PROPN
cril-1051	488	5	,	,	PUNCT
cril-1051	488	6	2018	2018	NUM
cril-1051	488	7	)	)	PUNCT
cril-1051	488	8	.	.	PUNCT
cril-1051	489	1	malouf	malouf	PROPN
cril-1051	489	2	,	,	PUNCT
cril-1051	489	3	robert	robert	PROPN
cril-1051	489	4	.	.	PROPN
cril-1051	489	5	2016	2016	NUM
cril-1051	489	6	.	.	PUNCT
cril-1051	490	1	generating	generate	VERB
cril-1051	490	2	morphological	morphological	ADJ
cril-1051	490	3	paradigms	paradigm	NOUN
cril-1051	490	4	with	with	ADP
cril-1051	490	5	a	a	DET
cril-1051	490	6	recurrent	recurrent	ADJ
cril-1051	490	7	neural	neural	ADJ
cril-1051	490	8	network	network	NOUN
cril-1051	490	9	.	.	PUNCT
cril-1051	491	1	san	san	PROPN
cril-1051	491	2	diego	diego	PROPN
cril-1051	491	3	linguistic	linguistic	ADJ
cril-1051	491	4	papers	paper	NOUN
cril-1051	491	5	6	6	NUM
cril-1051	491	6	.	.	PUNCT
cril-1051	492	1	122–129	122–129	NUM
cril-1051	492	2	.	.	PUNCT
cril-1051	493	1	mccarthy	mccarthy	PROPN
cril-1051	493	2	,	,	PUNCT
cril-1051	493	3	arya	arya	PROPN
cril-1051	493	4	d	d	PROPN
cril-1051	493	5	,	,	PUNCT
cril-1051	493	6	ekaterina	ekaterina	PROPN
cril-1051	493	7	vylomova	vylomova	PROPN
cril-1051	493	8	,	,	PUNCT
cril-1051	493	9	shijie	shijie	PROPN
cril-1051	493	10	wu	wu	PROPN
cril-1051	493	11	,	,	PUNCT
cril-1051	493	12	chaitanya	chaitanya	PROPN
cril-1051	493	13	malaviya	malaviya	PROPN
cril-1051	493	14	,	,	PUNCT
cril-1051	493	15	lawrence	lawrence	PROPN
cril-1051	493	16	wolfsonkin	wolfsonkin	PROPN
cril-1051	493	17	,	,	PUNCT
cril-1051	493	18	garrett	garrett	PROPN
cril-1051	493	19	nicolai	nicolai	PROPN
cril-1051	493	20	,	,	PUNCT
cril-1051	493	21	christo	christo	PROPN
cril-1051	493	22	kirov	kirov	PROPN
cril-1051	493	23	,	,	PUNCT
cril-1051	493	24	et	et	PROPN
cril-1051	493	25	al	al	PROPN
cril-1051	493	26	.	.	PROPN
cril-1051	493	27	2019	2019	NUM
cril-1051	493	28	.	.	PUNCT
cril-1051	494	1	the	the	DET
cril-1051	494	2	sigmorphon	sigmorphon	ADJ
cril-1051	494	3	2019	2019	NUM
cril-1051	494	4	shared	share	VERB
cril-1051	494	5	task	task	NOUN
cril-1051	494	6	:	:	PUNCT
cril-1051	494	7	crosslinguality	crosslinguality	NOUN
cril-1051	494	8	and	and	CCONJ
cril-1051	494	9	context	context	NOUN
cril-1051	494	10	in	in	ADP
cril-1051	494	11	morphology	morphology	NOUN
cril-1051	494	12	.	.	PUNCT
cril-1051	495	1	in	in	ADP
cril-1051	495	2	proceedings	proceeding	NOUN
cril-1051	495	3	of	of	ADP
cril-1051	495	4	the	the	DET
cril-1051	495	5	16th	16th	ADJ
cril-1051	495	6	sigmorphon	sigmorphon	NOUN
cril-1051	495	7	workshop	workshop	NOUN
cril-1051	495	8	on	on	ADP
cril-1051	495	9	computational	computational	ADJ
cril-1051	495	10	research	research	NOUN
cril-1051	495	11	in	in	ADP
cril-1051	495	12	phonetics	phonetic	NOUN
cril-1051	495	13	,	,	PUNCT
cril-1051	495	14	phonology	phonology	NOUN
cril-1051	495	15	,	,	PUNCT
cril-1051	495	16	and	and	CCONJ
cril-1051	495	17	morphology	morphology	NOUN
cril-1051	495	18	.	.	PUNCT
cril-1051	495	19	florence	florence	PROPN
cril-1051	495	20	,	,	PUNCT
cril-1051	495	21	italy	italy	PROPN
cril-1051	495	22	:	:	PUNCT
cril-1051	495	23	association	association	NOUN
cril-1051	495	24	for	for	ADP
cril-1051	495	25	computational	computational	ADJ
cril-1051	495	26	linguistics	linguistic	NOUN
cril-1051	495	27	.	.	PUNCT
cril-1051	496	1	mcmillan	mcmillan	NOUN
cril-1051	496	2	-	-	PUNCT
cril-1051	496	3	major	major	PROPN
cril-1051	496	4	,	,	PUNCT
cril-1051	496	5	angelina	angelina	PROPN
cril-1051	496	6	.	.	PROPN
cril-1051	496	7	2020	2020	NUM
cril-1051	496	8	.	.	PUNCT
cril-1051	497	1	automating	automate	VERB
cril-1051	497	2	gloss	gloss	NOUN
cril-1051	497	3	generation	generation	NOUN
cril-1051	497	4	in	in	ADP
cril-1051	497	5	interlinear	interlinear	ADJ
cril-1051	497	6	glossed	glossed	ADJ
cril-1051	497	7	text	text	NOUN
cril-1051	497	8	.	.	PUNCT
cril-1051	498	1	in	in	ADP
cril-1051	498	2	proceedings	proceeding	NOUN
cril-1051	498	3	of	of	ADP
cril-1051	498	4	the	the	DET
cril-1051	498	5	society	society	NOUN
cril-1051	498	6	for	for	ADP
cril-1051	498	7	computation	computation	NOUN
cril-1051	498	8	in	in	ADP
cril-1051	498	9	linguistics	linguistic	NOUN
cril-1051	498	10	,	,	PUNCT
cril-1051	498	11	vol	vol	NOUN
cril-1051	498	12	.	.	PROPN
cril-1051	498	13	3	3	NUM
cril-1051	498	14	,	,	PUNCT
cril-1051	498	15	338–349	338–349	NUM
cril-1051	498	16	.	.	PUNCT
cril-1051	499	1	https://doi.org/10.7275/tsmk-sa32	https://doi.org/10.7275/tsmk-sa32	PROPN
cril-1051	499	2	.	.	PUNCT
cril-1051	499	3	https://scholarworks.umass.edu/scil/vol3/iss1/33	https://scholarworks.umass.edu/scil/vol3/iss1/33	PROPN
cril-1051	499	4	.	.	PROPN
cril-1051	499	5	computational	computational	ADJ
cril-1051	499	6	morphology	morphology	NOUN
cril-1051	499	7	for	for	ADP
cril-1051	499	8	language	language	NOUN
cril-1051	499	9	description	description	NOUN
cril-1051	499	10	and	and	CCONJ
cril-1051	499	11	documentation	documentation	NOUN
cril-1051	499	12	25	25	NUM
cril-1051	499	13	moeller	moeller	PROPN
cril-1051	499	14	,	,	PUNCT
cril-1051	499	15	sarah	sarah	PROPN
cril-1051	499	16	&	&	CCONJ
cril-1051	499	17	mans	mans	PROPN
cril-1051	499	18	hulden	hulden	VERB
cril-1051	499	19	.	.	PUNCT
cril-1051	500	1	2018	2018	NUM
cril-1051	500	2	.	.	PUNCT
cril-1051	501	1	automatic	automatic	ADJ
cril-1051	501	2	glossing	glossing	NOUN
cril-1051	501	3	in	in	ADP
cril-1051	501	4	a	a	DET
cril-1051	501	5	low	low	ADJ
cril-1051	501	6	-	-	PUNCT
cril-1051	501	7	resource	resource	NOUN
cril-1051	501	8	setting	setting	NOUN
cril-1051	501	9	for	for	ADP
cril-1051	501	10	language	language	NOUN
cril-1051	501	11	documentation	documentation	NOUN
cril-1051	501	12	.	.	PUNCT
cril-1051	502	1	in	in	ADP
cril-1051	502	2	proceedings	proceeding	NOUN
cril-1051	502	3	of	of	ADP
cril-1051	502	4	the	the	DET
cril-1051	502	5	workshop	workshop	NOUN
cril-1051	502	6	on	on	ADP
cril-1051	502	7	computational	computational	ADJ
cril-1051	502	8	modeling	modeling	NOUN
cril-1051	502	9	of	of	ADP
cril-1051	502	10	polysynthetic	polysynthetic	ADJ
cril-1051	502	11	languages	language	NOUN
cril-1051	502	12	,	,	PUNCT
cril-1051	502	13	84–93	84–93	NUM
cril-1051	502	14	.	.	PUNCT
cril-1051	502	15	santa	santa	PROPN
cril-1051	502	16	fe	fe	PROPN
cril-1051	502	17	,	,	PUNCT
cril-1051	502	18	new	new	PROPN
cril-1051	502	19	mexico	mexico	PROPN
cril-1051	502	20	,	,	PUNCT
cril-1051	502	21	usa	usa	PROPN
cril-1051	502	22	:	:	PUNCT
cril-1051	502	23	association	association	NOUN
cril-1051	502	24	for	for	ADP
cril-1051	502	25	computational	computational	ADJ
cril-1051	502	26	linguistics	linguistic	NOUN
cril-1051	502	27	.	.	PUNCT
cril-1051	503	1	http://www.aclweb.org/anthology/w18-4809	http://www.aclweb.org/anthology/w18-4809	PUNCT
cril-1051	503	2	(	(	PUNCT
cril-1051	503	3	22	22	NUM
cril-1051	503	4	august	august	PROPN
cril-1051	503	5	,	,	PUNCT
cril-1051	503	6	2018	2018	NUM
cril-1051	503	7	)	)	PUNCT
cril-1051	503	8	.	.	PUNCT
cril-1051	504	1	moeller	moeller	PROPN
cril-1051	504	2	,	,	PUNCT
cril-1051	504	3	sarah	sarah	PROPN
cril-1051	504	4	,	,	PUNCT
cril-1051	504	5	ghazaleh	ghazaleh	PROPN
cril-1051	504	6	kazeminejad	kazeminejad	PROPN
cril-1051	504	7	,	,	PUNCT
cril-1051	504	8	andrew	andrew	PROPN
cril-1051	504	9	cowell	cowell	PROPN
cril-1051	504	10	&	&	CCONJ
cril-1051	504	11	mans	mans	PROPN
cril-1051	504	12	hulden	hulden	VERB
cril-1051	504	13	.	.	PUNCT
cril-1051	505	1	2018	2018	NUM
cril-1051	505	2	.	.	PUNCT
cril-1051	506	1	a	a	DET
cril-1051	506	2	neural	neural	ADJ
cril-1051	506	3	morphological	morphological	ADJ
cril-1051	506	4	analyzer	analyzer	NOUN
cril-1051	506	5	for	for	ADP
cril-1051	506	6	arapaho	arapaho	PROPN
cril-1051	506	7	verbs	verbs	PROPN
cril-1051	506	8	learned	learn	VERB
cril-1051	506	9	from	from	ADP
cril-1051	506	10	a	a	DET
cril-1051	506	11	finite	finite	ADJ
cril-1051	506	12	state	state	NOUN
cril-1051	506	13	transducer	transducer	NOUN
cril-1051	506	14	.	.	PUNCT
cril-1051	507	1	in	in	ADP
cril-1051	507	2	proceedings	proceeding	NOUN
cril-1051	507	3	of	of	ADP
cril-1051	507	4	the	the	DET
cril-1051	507	5	workshop	workshop	NOUN
cril-1051	507	6	on	on	ADP
cril-1051	507	7	computational	computational	ADJ
cril-1051	507	8	modeling	modeling	NOUN
cril-1051	507	9	of	of	ADP
cril-1051	507	10	polysynthetic	polysynthetic	ADJ
cril-1051	507	11	languages	language	NOUN
cril-1051	507	12	,	,	PUNCT
cril-1051	507	13	12–20	12–20	NUM
cril-1051	507	14	.	.	PUNCT
cril-1051	507	15	santa	santa	PROPN
cril-1051	507	16	fe	fe	PROPN
cril-1051	507	17	,	,	PUNCT
cril-1051	507	18	new	new	PROPN
cril-1051	507	19	mexico	mexico	PROPN
cril-1051	507	20	,	,	PUNCT
cril-1051	507	21	usa	usa	PROPN
cril-1051	507	22	:	:	PUNCT
cril-1051	507	23	association	association	NOUN
cril-1051	507	24	for	for	ADP
cril-1051	507	25	computational	computational	ADJ
cril-1051	507	26	linguistics	linguistic	NOUN
cril-1051	507	27	.	.	PUNCT
cril-1051	508	1	http://www.aclweb.org/anthology/w18-4802	http://www.aclweb.org/anthology/w18-4802	X
cril-1051	508	2	(	(	PUNCT
cril-1051	508	3	22	22	NUM
cril-1051	508	4	august	august	PROPN
cril-1051	508	5	,	,	PUNCT
cril-1051	508	6	2018	2018	NUM
cril-1051	508	7	)	)	PUNCT
cril-1051	508	8	.	.	PUNCT
cril-1051	509	1	moeller	moeller	PROPN
cril-1051	509	2	,	,	PUNCT
cril-1051	509	3	sarah	sarah	PROPN
cril-1051	509	4	,	,	PUNCT
cril-1051	509	5	ghazaleh	ghazaleh	PROPN
cril-1051	509	6	kazeminejad	kazeminejad	PROPN
cril-1051	509	7	,	,	PUNCT
cril-1051	509	8	andrew	andrew	PROPN
cril-1051	509	9	cowell	cowell	PROPN
cril-1051	509	10	&	&	CCONJ
cril-1051	509	11	mans	mans	PROPN
cril-1051	509	12	hulden	hulden	VERB
cril-1051	509	13	.	.	PUNCT
cril-1051	510	1	2019	2019	NUM
cril-1051	510	2	.	.	PUNCT
cril-1051	511	1	improving	improve	VERB
cril-1051	511	2	lowresource	lowresource	PROPN
cril-1051	511	3	morphological	morphological	ADJ
cril-1051	511	4	learning	learn	VERB
cril-1051	511	5	with	with	ADP
cril-1051	511	6	intermediate	intermediate	ADJ
cril-1051	511	7	forms	form	NOUN
cril-1051	511	8	from	from	ADP
cril-1051	511	9	finite	finite	ADJ
cril-1051	511	10	state	state	NOUN
cril-1051	511	11	transducers	transducer	NOUN
cril-1051	511	12	.	.	PUNCT
cril-1051	512	1	in	in	ADP
cril-1051	512	2	proceedings	proceeding	NOUN
cril-1051	512	3	of	of	ADP
cril-1051	512	4	the	the	DET
cril-1051	512	5	workshop	workshop	NOUN
cril-1051	512	6	on	on	ADP
cril-1051	512	7	computational	computational	ADJ
cril-1051	512	8	methods	method	NOUN
cril-1051	512	9	for	for	ADP
cril-1051	512	10	endangered	endanger	VERB
cril-1051	512	11	languages	language	NOUN
cril-1051	512	12	,	,	PUNCT
cril-1051	512	13	vol	vol	NOUN
cril-1051	512	14	.	.	PROPN
cril-1051	513	1	1	1	NUM
cril-1051	513	2	.	.	X
cril-1051	514	1	honolulu	honolulu	PROPN
cril-1051	514	2	,	,	PUNCT
cril-1051	514	3	hi	hi	INTJ
cril-1051	514	4	.	.	PUNCT
cril-1051	514	5	https://www.aclweb.org/anthology/w196011/.	https://www.aclweb.org/anthology/w196011/.	PROPN
cril-1051	515	1	monson	monson	PROPN
cril-1051	515	2	,	,	PUNCT
cril-1051	515	3	christian	christian	PROPN
cril-1051	515	4	,	,	PUNCT
cril-1051	515	5	jaime	jaime	NOUN
cril-1051	515	6	carbonell	carbonell	PROPN
cril-1051	515	7	,	,	PUNCT
cril-1051	515	8	alon	alon	PROPN
cril-1051	515	9	lavie	lavie	PROPN
cril-1051	515	10	&	&	CCONJ
cril-1051	515	11	lori	lori	PROPN
cril-1051	515	12	levin	levin	PROPN
cril-1051	515	13	.	.	PROPN
cril-1051	515	14	2007	2007	NUM
cril-1051	515	15	.	.	PUNCT
cril-1051	516	1	paramor	paramor	NOUN
cril-1051	516	2	:	:	PUNCT
cril-1051	516	3	finding	find	VERB
cril-1051	516	4	paradigms	paradigm	NOUN
cril-1051	516	5	across	across	ADP
cril-1051	516	6	morphology	morphology	NOUN
cril-1051	516	7	.	.	PUNCT
cril-1051	517	1	in	in	ADP
cril-1051	517	2	advances	advance	NOUN
cril-1051	517	3	in	in	ADP
cril-1051	517	4	multilingual	multilingual	ADJ
cril-1051	517	5	and	and	CCONJ
cril-1051	517	6	multimodal	multimodal	NOUN
cril-1051	517	7	information	information	NOUN
cril-1051	517	8	retrieval	retrieval	NOUN
cril-1051	517	9	(	(	PUNCT
cril-1051	517	10	lecture	lecture	NOUN
cril-1051	517	11	notes	note	NOUN
cril-1051	517	12	in	in	ADP
cril-1051	517	13	computer	computer	NOUN
cril-1051	517	14	science	science	NOUN
cril-1051	517	15	)	)	PUNCT
cril-1051	517	16	,	,	PUNCT
cril-1051	517	17	900–907	900–907	NUM
cril-1051	517	18	.	.	PUNCT
cril-1051	517	19	springer	springer	NOUN
cril-1051	517	20	,	,	PUNCT
cril-1051	517	21	berlin	berlin	PROPN
cril-1051	517	22	,	,	PUNCT
cril-1051	517	23	heidelberg	heidelberg	PROPN
cril-1051	517	24	.	.	PUNCT
cril-1051	518	1	https://link.springer.com/chapter/10.1007/978-3-540-85760-0_115	https://link.springer.com/chapter/10.1007/978-3-540-85760-0_115	PROPN
cril-1051	518	2	(	(	PUNCT
cril-1051	518	3	27	27	NUM
cril-1051	518	4	february	february	NOUN
cril-1051	518	5	,	,	PUNCT
cril-1051	518	6	2018	2018	NUM
cril-1051	518	7	)	)	PUNCT
cril-1051	518	8	.	.	PUNCT
cril-1051	519	1	moon	moon	PROPN
cril-1051	519	2	,	,	PUNCT
cril-1051	519	3	taesun	taesun	NOUN
cril-1051	519	4	,	,	PUNCT
cril-1051	519	5	katrin	katrin	PROPN
cril-1051	519	6	erk	erk	PROPN
cril-1051	519	7	&	&	CCONJ
cril-1051	519	8	jason	jason	PROPN
cril-1051	519	9	baldridge	baldridge	PROPN
cril-1051	519	10	.	.	PUNCT
cril-1051	520	1	2009	2009	NUM
cril-1051	520	2	.	.	PUNCT
cril-1051	521	1	unsupervised	unsupervised	ADJ
cril-1051	521	2	morphological	morphological	ADJ
cril-1051	521	3	segmentation	segmentation	NOUN
cril-1051	521	4	and	and	CCONJ
cril-1051	521	5	clustering	cluster	VERB
cril-1051	521	6	with	with	ADP
cril-1051	521	7	document	document	NOUN
cril-1051	521	8	boundaries	boundary	NOUN
cril-1051	521	9	.	.	PUNCT
cril-1051	522	1	in	in	ADP
cril-1051	522	2	proceedings	proceeding	NOUN
cril-1051	522	3	of	of	ADP
cril-1051	522	4	the	the	DET
cril-1051	522	5	2009	2009	NUM
cril-1051	522	6	conference	conference	NOUN
cril-1051	522	7	on	on	ADP
cril-1051	522	8	empirical	empirical	ADJ
cril-1051	522	9	methods	method	NOUN
cril-1051	522	10	in	in	ADP
cril-1051	522	11	natural	natural	ADJ
cril-1051	522	12	language	language	NOUN
cril-1051	522	13	processing	processing	NOUN
cril-1051	522	14	:	:	PUNCT
cril-1051	522	15	volume	volume	NOUN
cril-1051	522	16	2	2	NUM
cril-1051	522	17	,	,	PUNCT
cril-1051	522	18	668–677	668–677	NUM
cril-1051	522	19	.	.	PUNCT
cril-1051	523	1	association	association	NOUN
cril-1051	523	2	for	for	ADP
cril-1051	523	3	computational	computational	ADJ
cril-1051	523	4	linguistics	linguistic	NOUN
cril-1051	523	5	.	.	PUNCT
cril-1051	524	1	nicolai	nicolai	PROPN
cril-1051	524	2	,	,	PUNCT
cril-1051	524	3	garrett	garrett	PROPN
cril-1051	524	4	,	,	PUNCT
cril-1051	524	5	colin	colin	PROPN
cril-1051	524	6	cherry	cherry	PROPN
cril-1051	524	7	&	&	CCONJ
cril-1051	524	8	grzegorz	grzegorz	PROPN
cril-1051	524	9	kondrak	kondrak	PROPN
cril-1051	524	10	.	.	PUNCT
cril-1051	525	1	2015	2015	NUM
cril-1051	525	2	.	.	PUNCT
cril-1051	526	1	inflection	inflection	NOUN
cril-1051	526	2	generation	generation	NOUN
cril-1051	526	3	as	as	ADP
cril-1051	526	4	discriminative	discriminative	NOUN
cril-1051	526	5	string	string	NOUN
cril-1051	526	6	transduction	transduction	NOUN
cril-1051	526	7	.	.	PUNCT
cril-1051	527	1	in	in	ADP
cril-1051	527	2	proceedings	proceeding	NOUN
cril-1051	527	3	of	of	ADP
cril-1051	527	4	the	the	DET
cril-1051	527	5	2015	2015	NUM
cril-1051	527	6	conference	conference	NOUN
cril-1051	527	7	of	of	ADP
cril-1051	527	8	the	the	DET
cril-1051	527	9	north	north	ADJ
cril-1051	527	10	american	american	ADJ
cril-1051	527	11	chapter	chapter	NOUN
cril-1051	527	12	of	of	ADP
cril-1051	527	13	the	the	DET
cril-1051	527	14	association	association	NOUN
cril-1051	527	15	for	for	ADP
cril-1051	527	16	computational	computational	ADJ
cril-1051	527	17	linguistics	linguistic	NOUN
cril-1051	527	18	:	:	PUNCT
cril-1051	527	19	human	human	ADJ
cril-1051	527	20	language	language	NOUN
cril-1051	527	21	technologies	technology	NOUN
cril-1051	527	22	,	,	PUNCT
cril-1051	527	23	922–931	922–931	NUM
cril-1051	527	24	.	.	PUNCT
cril-1051	528	1	denver	denver	PROPN
cril-1051	528	2	,	,	PUNCT
cril-1051	528	3	colorado	colorado	PROPN
cril-1051	528	4	:	:	PUNCT
cril-1051	528	5	association	association	NOUN
cril-1051	528	6	for	for	ADP
cril-1051	528	7	computational	computational	ADJ
cril-1051	528	8	linguistics	linguistic	NOUN
cril-1051	528	9	.	.	PUNCT
cril-1051	529	1	https://doi.org/10.3115/v1/n15-1093	https://doi.org/10.3115/v1/n15-1093	NOUN
cril-1051	529	2	.	.	PUNCT
cril-1051	530	1	https://www.aclweb.org/anthology/n15-1093	https://www.aclweb.org/anthology/n15-1093	PUNCT
cril-1051	531	1	(	(	PUNCT
cril-1051	531	2	26	26	NUM
cril-1051	531	3	january	january	PROPN
cril-1051	531	4	,	,	PUNCT
cril-1051	531	5	2021	2021	NUM
cril-1051	531	6	)	)	PUNCT
cril-1051	531	7	.	.	PUNCT
cril-1051	532	1	nicolai	nicolai	PROPN
cril-1051	532	2	,	,	PUNCT
cril-1051	532	3	garrett	garrett	PROPN
cril-1051	532	4	,	,	PUNCT
cril-1051	532	5	kyle	kyle	PROPN
cril-1051	532	6	gorman	gorman	PROPN
cril-1051	532	7	&	&	CCONJ
cril-1051	532	8	ryan	ryan	PROPN
cril-1051	532	9	cotterell	cotterell	PROPN
cril-1051	532	10	(	(	PUNCT
cril-1051	532	11	eds	ed	NOUN
cril-1051	532	12	.	.	PUNCT
cril-1051	532	13	)	)	PUNCT
cril-1051	532	14	.	.	PUNCT
cril-1051	533	1	2020	2020	NUM
cril-1051	533	2	.	.	PUNCT
cril-1051	534	1	proceedings	proceeding	NOUN
cril-1051	534	2	of	of	ADP
cril-1051	534	3	the	the	DET
cril-1051	534	4	17th	17th	ADJ
cril-1051	534	5	sigmorphon	sigmorphon	NOUN
cril-1051	534	6	workshop	workshop	NOUN
cril-1051	534	7	on	on	ADP
cril-1051	534	8	computational	computational	ADJ
cril-1051	534	9	research	research	NOUN
cril-1051	534	10	in	in	ADP
cril-1051	534	11	phonetics	phonetic	NOUN
cril-1051	534	12	,	,	PUNCT
cril-1051	534	13	phonology	phonology	NOUN
cril-1051	534	14	,	,	PUNCT
cril-1051	534	15	and	and	CCONJ
cril-1051	534	16	colorado	colorado	PROPN
cril-1051	534	17	research	research	NOUN
cril-1051	534	18	in	in	ADP
cril-1051	534	19	linguistics	linguistic	NOUN
cril-1051	534	20	,	,	PUNCT
cril-1051	534	21	volume	volume	NOUN
cril-1051	534	22	25	25	NUM
cril-1051	534	23	(	(	PUNCT
cril-1051	534	24	2021	2021	NUM
cril-1051	534	25	)	)	PUNCT
cril-1051	534	26	26	26	NUM
cril-1051	534	27	morphology	morphology	NOUN
cril-1051	534	28	.	.	PUNCT
cril-1051	535	1	online	online	PROPN
cril-1051	535	2	:	:	PUNCT
cril-1051	536	1	association	association	NOUN
cril-1051	536	2	for	for	ADP
cril-1051	536	3	computational	computational	ADJ
cril-1051	536	4	linguistics	linguistic	NOUN
cril-1051	536	5	.	.	PUNCT
cril-1051	537	1	https://www.aclweb.org/anthology/2020.sigmorphon-1.0	https://www.aclweb.org/anthology/2020.sigmorphon-1.0	PROPN
cril-1051	537	2	(	(	PUNCT
cril-1051	537	3	8	8	NUM
cril-1051	537	4	april	april	PROPN
cril-1051	537	5	,	,	PUNCT
cril-1051	537	6	2021	2021	NUM
cril-1051	537	7	)	)	PUNCT
cril-1051	537	8	.	.	PUNCT
cril-1051	538	1	nicolai	nicolai	PROPN
cril-1051	538	2	,	,	PUNCT
cril-1051	538	3	garrett	garrett	PROPN
cril-1051	538	4	&	&	CCONJ
cril-1051	538	5	grzegorz	grzegorz	PROPN
cril-1051	538	6	kondrak	kondrak	PROPN
cril-1051	538	7	.	.	PUNCT
cril-1051	539	1	2017	2017	NUM
cril-1051	539	2	.	.	PUNCT
cril-1051	540	1	morphological	morphological	ADJ
cril-1051	540	2	analysis	analysis	NOUN
cril-1051	540	3	without	without	ADP
cril-1051	540	4	expert	expert	NOUN
cril-1051	540	5	annotation	annotation	NOUN
cril-1051	540	6	.	.	PUNCT
cril-1051	541	1	proceedings	proceeding	NOUN
cril-1051	541	2	of	of	ADP
cril-1051	541	3	the	the	DET
cril-1051	541	4	15th	15th	ADJ
cril-1051	541	5	conference	conference	NOUN
cril-1051	541	6	of	of	ADP
cril-1051	541	7	the	the	DET
cril-1051	541	8	european	european	ADJ
cril-1051	541	9	chapter	chapter	NOUN
cril-1051	541	10	of	of	ADP
cril-1051	541	11	the	the	DET
cril-1051	541	12	association	association	NOUN
cril-1051	541	13	for	for	ADP
cril-1051	541	14	computational	computational	ADJ
cril-1051	541	15	linguistics	linguistic	NOUN
cril-1051	541	16	:	:	PUNCT
cril-1051	541	17	volume	volume	NOUN
cril-1051	541	18	2	2	NUM
cril-1051	541	19	,	,	PUNCT
cril-1051	541	20	short	short	ADJ
cril-1051	541	21	papers	paper	NOUN
cril-1051	541	22	2	2	NUM
cril-1051	541	23	.	.	X
cril-1051	541	24	211–216	211–216	NUM
cril-1051	541	25	.	.	PUNCT
cril-1051	542	1	https://aclanthology.coli.uni-saarland.de/papers/e17-2034/e17-2034	https://aclanthology.coli.uni-saarland.de/papers/e17-2034/e17-2034	PROPN
cril-1051	542	2	(	(	PUNCT
cril-1051	542	3	16	16	NUM
cril-1051	542	4	january	january	PROPN
cril-1051	542	5	,	,	PUNCT
cril-1051	542	6	2018	2018	NUM
cril-1051	542	7	)	)	PUNCT
cril-1051	542	8	.	.	PUNCT
cril-1051	543	1	palmer	palmer	PROPN
cril-1051	543	2	,	,	PUNCT
cril-1051	543	3	alexis	alexis	PROPN
cril-1051	543	4	mary	mary	PROPN
cril-1051	543	5	.	.	PROPN
cril-1051	543	6	2009	2009	NUM
cril-1051	543	7	.	.	PUNCT
cril-1051	544	1	semi	semi	ADJ
cril-1051	544	2	-	-	ADJ
cril-1051	544	3	automated	automated	ADJ
cril-1051	544	4	annotation	annotation	NOUN
cril-1051	544	5	and	and	CCONJ
cril-1051	544	6	active	active	ADJ
cril-1051	544	7	learning	learning	NOUN
cril-1051	544	8	for	for	ADP
cril-1051	544	9	language	language	NOUN
cril-1051	544	10	documentation	documentation	NOUN
cril-1051	544	11	.	.	PUNCT
cril-1051	545	1	university	university	NOUN
cril-1051	545	2	of	of	ADP
cril-1051	545	3	texas	texas	PROPN
cril-1051	545	4	at	at	ADP
cril-1051	545	5	austin	austin	PROPN
cril-1051	545	6	phd	phd	PROPN
cril-1051	545	7	thesis	thesis	PROPN
cril-1051	545	8	.	.	PUNCT
cril-1051	546	1	palmer	palmer	PROPN
cril-1051	546	2	,	,	PUNCT
cril-1051	546	3	alexis	alexis	PROPN
cril-1051	546	4	,	,	PUNCT
cril-1051	546	5	taesun	taesun	NOUN
cril-1051	546	6	moon	moon	NOUN
cril-1051	546	7	,	,	PUNCT
cril-1051	546	8	jason	jason	PROPN
cril-1051	546	9	baldridge	baldridge	PROPN
cril-1051	546	10	,	,	PUNCT
cril-1051	546	11	katrin	katrin	PROPN
cril-1051	546	12	erk	erk	PROPN
cril-1051	546	13	,	,	PUNCT
cril-1051	546	14	eric	eric	PROPN
cril-1051	546	15	campbell	campbell	PROPN
cril-1051	546	16	&	&	CCONJ
cril-1051	546	17	telma	telma	PROPN
cril-1051	546	18	can	can	AUX
cril-1051	546	19	.	.	PUNCT
cril-1051	547	1	2010	2010	NUM
cril-1051	547	2	.	.	PUNCT
cril-1051	548	1	computational	computational	ADJ
cril-1051	548	2	strategies	strategy	NOUN
cril-1051	548	3	for	for	ADP
cril-1051	548	4	reducing	reduce	VERB
cril-1051	548	5	annotation	annotation	NOUN
cril-1051	548	6	effort	effort	NOUN
cril-1051	548	7	in	in	ADP
cril-1051	548	8	language	language	NOUN
cril-1051	548	9	documentation	documentation	NOUN
cril-1051	548	10	.	.	PUNCT
cril-1051	549	1	linguistic	linguistic	ADJ
cril-1051	549	2	issues	issue	NOUN
cril-1051	549	3	in	in	ADP
cril-1051	549	4	language	language	NOUN
cril-1051	549	5	technology	technology	NOUN
cril-1051	549	6	3(4	3(4	NUM
cril-1051	549	7	)	)	PUNCT
cril-1051	549	8	.	.	PUNCT
cril-1051	550	1	1–42	1–42	PROPN
cril-1051	550	2	.	.	PUNCT
cril-1051	551	1	http://journals.linguisticsociety.org/elanguage/lilt/article/view/663.html	http://journals.linguisticsociety.org/elanguage/lilt/article/view/663.html	PUNCT
cril-1051	551	2	(	(	PUNCT
cril-1051	551	3	24	24	NUM
cril-1051	551	4	january	january	PROPN
cril-1051	551	5	,	,	PUNCT
cril-1051	551	6	2014	2014	NUM
cril-1051	551	7	)	)	PUNCT
cril-1051	551	8	.	.	PUNCT
cril-1051	552	1	poon	poon	NOUN
cril-1051	552	2	,	,	PUNCT
cril-1051	552	3	hoifung	hoifung	NOUN
cril-1051	552	4	,	,	PUNCT
cril-1051	552	5	colin	colin	PROPN
cril-1051	552	6	cherry	cherry	PROPN
cril-1051	552	7	&	&	CCONJ
cril-1051	552	8	kristina	kristina	PROPN
cril-1051	552	9	toutanova	toutanova	PROPN
cril-1051	552	10	.	.	PROPN
cril-1051	552	11	2009	2009	NUM
cril-1051	552	12	.	.	PUNCT
cril-1051	553	1	unsupervised	unsupervised	ADJ
cril-1051	553	2	morphological	morphological	ADJ
cril-1051	553	3	segmentation	segmentation	NOUN
cril-1051	553	4	with	with	ADP
cril-1051	553	5	log	log	NOUN
cril-1051	553	6	-	-	PUNCT
cril-1051	553	7	linear	linear	NOUN
cril-1051	553	8	models	model	NOUN
cril-1051	553	9	.	.	PUNCT
cril-1051	554	1	in	in	ADP
cril-1051	554	2	proceedings	proceeding	NOUN
cril-1051	554	3	of	of	ADP
cril-1051	554	4	human	human	ADJ
cril-1051	554	5	language	language	NOUN
cril-1051	554	6	technologies	technology	NOUN
cril-1051	554	7	:	:	PUNCT
cril-1051	554	8	the	the	DET
cril-1051	554	9	2009	2009	NUM
cril-1051	554	10	annual	annual	ADJ
cril-1051	554	11	conference	conference	NOUN
cril-1051	554	12	of	of	ADP
cril-1051	554	13	the	the	DET
cril-1051	554	14	north	north	ADJ
cril-1051	554	15	american	american	ADJ
cril-1051	554	16	chapter	chapter	NOUN
cril-1051	554	17	of	of	ADP
cril-1051	554	18	the	the	DET
cril-1051	554	19	association	association	NOUN
cril-1051	554	20	for	for	ADP
cril-1051	554	21	computational	computational	ADJ
cril-1051	554	22	linguistics	linguistic	NOUN
cril-1051	554	23	,	,	PUNCT
cril-1051	554	24	209–217	209–217	NUM
cril-1051	554	25	.	.	PUNCT
cril-1051	555	1	association	association	NOUN
cril-1051	555	2	for	for	ADP
cril-1051	555	3	computational	computational	ADJ
cril-1051	555	4	linguistics	linguistic	NOUN
cril-1051	555	5	.	.	PUNCT
cril-1051	556	1	popel	popel	PROPN
cril-1051	556	2	,	,	PUNCT
cril-1051	556	3	martin	martin	PROPN
cril-1051	556	4	&	&	CCONJ
cril-1051	556	5	ondřej	ondřej	PROPN
cril-1051	556	6	bojar	bojar	PROPN
cril-1051	556	7	.	.	PUNCT
cril-1051	557	1	2018	2018	NUM
cril-1051	557	2	.	.	PUNCT
cril-1051	558	1	training	train	VERB
cril-1051	558	2	tips	tip	NOUN
cril-1051	558	3	for	for	ADP
cril-1051	558	4	the	the	DET
cril-1051	558	5	transformer	transformer	NOUN
cril-1051	558	6	model	model	NOUN
cril-1051	558	7	.	.	PUNCT
cril-1051	559	1	the	the	DET
cril-1051	559	2	prague	prague	PROPN
cril-1051	559	3	bulletin	bulletin	NOUN
cril-1051	559	4	of	of	ADP
cril-1051	559	5	mathematical	mathematical	ADJ
cril-1051	559	6	linguistics	linguistic	NOUN
cril-1051	559	7	110(1	110(1	NUM
cril-1051	559	8	)	)	PUNCT
cril-1051	559	9	.	.	PUNCT
cril-1051	560	1	43–70	43–70	NUM
cril-1051	560	2	.	.	PUNCT
cril-1051	561	1	http://content.sciendo.com/view/journals/pralin/110/1/article-p43.xml	http://content.sciendo.com/view/journals/pralin/110/1/article-p43.xml	X
cril-1051	561	2	(	(	PUNCT
cril-1051	561	3	13	13	NUM
cril-1051	561	4	may	may	NOUN
cril-1051	561	5	,	,	PUNCT
cril-1051	561	6	2020	2020	NUM
cril-1051	561	7	)	)	PUNCT
cril-1051	561	8	.	.	PUNCT
cril-1051	562	1	rice	rice	NOUN
cril-1051	562	2	,	,	PUNCT
cril-1051	562	3	sally	sally	PROPN
cril-1051	562	4	&	&	CCONJ
cril-1051	562	5	dorothy	dorothy	PROPN
cril-1051	562	6	thunder	thunder	PROPN
cril-1051	562	7	.	.	PUNCT
cril-1051	563	1	2017	2017	NUM
cril-1051	563	2	.	.	PUNCT
cril-1051	564	1	community	community	NOUN
cril-1051	564	2	-	-	PUNCT
cril-1051	564	3	based	base	VERB
cril-1051	564	4	corpus	corpus	NOUN
cril-1051	564	5	-	-	PUNCT
cril-1051	564	6	building	building	NOUN
cril-1051	564	7	:	:	PUNCT
cril-1051	564	8	three	three	NUM
cril-1051	564	9	case	case	NOUN
cril-1051	564	10	studies	study	NOUN
cril-1051	564	11	.	.	PUNCT
cril-1051	565	1	presented	present	VERB
cril-1051	565	2	at	at	ADP
cril-1051	565	3	the	the	DET
cril-1051	565	4	5th	5th	ADJ
cril-1051	565	5	international	international	ADJ
cril-1051	565	6	conference	conference	NOUN
cril-1051	565	7	on	on	ADP
cril-1051	565	8	language	language	NOUN
cril-1051	565	9	documentation	documentation	NOUN
cril-1051	565	10	and	and	CCONJ
cril-1051	565	11	conservation	conservation	NOUN
cril-1051	565	12	(	(	PUNCT
cril-1051	565	13	icldc	icldc	NOUN
cril-1051	565	14	)	)	PUNCT
cril-1051	565	15	,	,	PUNCT
cril-1051	565	16	honolulu	honolulu	PROPN
cril-1051	565	17	,	,	PUNCT
cril-1051	565	18	hi	hi	PROPN
cril-1051	565	19	.	.	PROPN
cril-1051	565	20	http://scholarspace.manoa.hawaii.edu/handle/10125/42052	http://scholarspace.manoa.hawaii.edu/handle/10125/42052	X
cril-1051	566	1	(	(	PUNCT
cril-1051	566	2	6	6	NUM
cril-1051	566	3	june	june	PROPN
cril-1051	566	4	,	,	PUNCT
cril-1051	566	5	2017	2017	NUM
cril-1051	566	6	)	)	PUNCT
cril-1051	566	7	.	.	PUNCT
cril-1051	567	1	roark	roark	PROPN
cril-1051	567	2	,	,	PUNCT
cril-1051	567	3	brian	brian	PROPN
cril-1051	567	4	&	&	CCONJ
cril-1051	567	5	richard	richard	PROPN
cril-1051	567	6	william	william	PROPN
cril-1051	567	7	sproat	sproat	PROPN
cril-1051	567	8	.	.	PUNCT
cril-1051	568	1	2007	2007	NUM
cril-1051	568	2	.	.	PUNCT
cril-1051	569	1	computational	computational	ADJ
cril-1051	569	2	approaches	approach	NOUN
cril-1051	569	3	to	to	ADP
cril-1051	569	4	morphology	morphology	NOUN
cril-1051	569	5	and	and	CCONJ
cril-1051	569	6	syntax	syntax	NOUN
cril-1051	569	7	.	.	PUNCT
cril-1051	570	1	oxford	oxford	PROPN
cril-1051	570	2	;	;	PUNCT
cril-1051	570	3	new	new	PROPN
cril-1051	570	4	york	york	PROPN
cril-1051	570	5	:	:	PUNCT
cril-1051	571	1	oxford	oxford	PROPN
cril-1051	571	2	university	university	PROPN
cril-1051	571	3	press	press	NOUN
cril-1051	571	4	.	.	PUNCT
cril-1051	572	1	rogers	rogers	PROPN
cril-1051	572	2	,	,	PUNCT
cril-1051	572	3	chris	chris	PROPN
cril-1051	572	4	.	.	PROPN
cril-1051	572	5	2010	2010	NUM
cril-1051	572	6	.	.	PUNCT
cril-1051	573	1	review	review	NOUN
cril-1051	573	2	of	of	ADP
cril-1051	573	3	fieldworks	fieldwork	NOUN
cril-1051	573	4	language	language	NOUN
cril-1051	573	5	explorer	explorer	NOUN
cril-1051	573	6	(	(	PUNCT
cril-1051	573	7	flex	flex	ADJ
cril-1051	573	8	)	)	PUNCT
cril-1051	573	9	3.0	3.0	NUM
cril-1051	573	10	.	.	PUNCT
cril-1051	574	1	language	language	NOUN
cril-1051	574	2	documentation	documentation	NOUN
cril-1051	574	3	&	&	CCONJ
cril-1051	574	4	conservation	conservation	NOUN
cril-1051	574	5	4	4	NUM
cril-1051	574	6	.	.	PUNCT
cril-1051	575	1	78–84	78–84	NOUN
cril-1051	575	2	.	.	PUNCT
cril-1051	576	1	http://scholarspace.manoa.hawaii.edu/handle/10125/4471	http://scholarspace.manoa.hawaii.edu/handle/10125/4471	X
cril-1051	576	2	(	(	PUNCT
cril-1051	576	3	13	13	NUM
cril-1051	576	4	march	march	NOUN
cril-1051	576	5	,	,	PUNCT
cril-1051	576	6	2018	2018	NUM
cril-1051	576	7	)	)	PUNCT
cril-1051	576	8	.	.	PUNCT
cril-1051	577	1	ruokolainen	ruokolainen	PROPN
cril-1051	577	2	,	,	PUNCT
cril-1051	577	3	teemu	teemu	PROPN
cril-1051	577	4	,	,	PUNCT
cril-1051	577	5	oskar	oskar	PROPN
cril-1051	577	6	kohonen	kohonen	PROPN
cril-1051	577	7	,	,	PUNCT
cril-1051	577	8	kairit	kairit	PROPN
cril-1051	577	9	sirts	sirt	NOUN
cril-1051	577	10	,	,	PUNCT
cril-1051	577	11	stig	stig	PROPN
cril-1051	577	12	-	-	PUNCT
cril-1051	577	13	arne	arne	VERB
cril-1051	577	14	grönroos	grönroos	PROPN
cril-1051	577	15	,	,	PUNCT
cril-1051	577	16	mikko	mikko	PROPN
cril-1051	577	17	kurimo	kurimo	PROPN
cril-1051	577	18	&	&	CCONJ
cril-1051	577	19	sami	sami	PROPN
cril-1051	577	20	virpioja	virpioja	PROPN
cril-1051	577	21	.	.	PUNCT
cril-1051	578	1	2016	2016	NUM
cril-1051	578	2	.	.	PUNCT
cril-1051	579	1	a	a	DET
cril-1051	579	2	comparative	comparative	ADJ
cril-1051	579	3	study	study	NOUN
cril-1051	579	4	of	of	ADP
cril-1051	579	5	minimally	minimally	ADV
cril-1051	579	6	supervised	supervised	ADJ
cril-1051	579	7	morphological	morphological	ADJ
cril-1051	579	8	computational	computational	ADJ
cril-1051	579	9	morphology	morphology	NOUN
cril-1051	579	10	for	for	ADP
cril-1051	579	11	language	language	NOUN
cril-1051	579	12	description	description	NOUN
cril-1051	579	13	and	and	CCONJ
cril-1051	579	14	documentation	documentation	NOUN
cril-1051	579	15	27	27	NUM
cril-1051	579	16	segmentation	segmentation	NOUN
cril-1051	579	17	.	.	PUNCT
cril-1051	580	1	computational	computational	ADJ
cril-1051	580	2	linguistics	linguistic	NOUN
cril-1051	580	3	42(1	42(1	NOUN
cril-1051	580	4	)	)	PUNCT
cril-1051	580	5	.	.	PUNCT
cril-1051	581	1	91–120	91–120	X
cril-1051	581	2	.	.	PUNCT
cril-1051	582	1	http://www.mitpressjournals.org/doi/10.1162/coli_a_00243	http://www.mitpressjournals.org/doi/10.1162/coli_a_00243	PROPN
cril-1051	582	2	(	(	PUNCT
cril-1051	582	3	17	17	NUM
cril-1051	582	4	january	january	PROPN
cril-1051	582	5	,	,	PUNCT
cril-1051	582	6	2018	2018	NUM
cril-1051	582	7	)	)	PUNCT
cril-1051	582	8	.	.	PUNCT
cril-1051	583	1	ruokolainen	ruokolainen	PROPN
cril-1051	583	2	,	,	PUNCT
cril-1051	583	3	teemu	teemu	PROPN
cril-1051	583	4	,	,	PUNCT
cril-1051	583	5	oskar	oskar	PROPN
cril-1051	583	6	kohonen	kohonen	PROPN
cril-1051	583	7	,	,	PUNCT
cril-1051	583	8	sami	sami	PROPN
cril-1051	583	9	virpioja	virpioja	PROPN
cril-1051	583	10	&	&	CCONJ
cril-1051	583	11	mikko	mikko	PROPN
cril-1051	583	12	kurimo	kurimo	PROPN
cril-1051	583	13	.	.	PUNCT
cril-1051	584	1	2013	2013	NUM
cril-1051	584	2	.	.	PUNCT
cril-1051	585	1	supervised	supervise	VERB
cril-1051	585	2	morphological	morphological	ADJ
cril-1051	585	3	segmentation	segmentation	NOUN
cril-1051	585	4	in	in	ADP
cril-1051	585	5	a	a	DET
cril-1051	585	6	low	low	ADJ
cril-1051	585	7	-	-	PUNCT
cril-1051	585	8	resource	resource	NOUN
cril-1051	585	9	learning	learning	NOUN
cril-1051	585	10	setting	set	VERB
cril-1051	585	11	using	use	VERB
cril-1051	585	12	conditional	conditional	ADJ
cril-1051	585	13	random	random	ADJ
cril-1051	585	14	fields	field	NOUN
cril-1051	585	15	.	.	PUNCT
cril-1051	586	1	in	in	ADP
cril-1051	586	2	conll	conll	PROPN
cril-1051	586	3	,	,	PUNCT
cril-1051	586	4	29–37	29–37	NUM
cril-1051	586	5	.	.	PUNCT
cril-1051	587	1	samardzic	samardzic	ADJ
cril-1051	587	2	,	,	PUNCT
cril-1051	587	3	tanja	tanja	PROPN
cril-1051	587	4	,	,	PUNCT
cril-1051	587	5	robert	robert	PROPN
cril-1051	587	6	schikowski	schikowski	PROPN
cril-1051	587	7	&	&	CCONJ
cril-1051	587	8	sabine	sabine	PROPN
cril-1051	587	9	stoll	stoll	PROPN
cril-1051	587	10	.	.	PROPN
cril-1051	588	1	2015	2015	NUM
cril-1051	588	2	.	.	PUNCT
cril-1051	589	1	automatic	automatic	ADJ
cril-1051	589	2	interlinear	interlinear	ADJ
cril-1051	589	3	glossing	glossing	NOUN
cril-1051	589	4	as	as	ADP
cril-1051	589	5	two	two	NUM
cril-1051	589	6	-	-	PUNCT
cril-1051	589	7	level	level	NOUN
cril-1051	589	8	sequence	sequence	NOUN
cril-1051	589	9	classification	classification	NOUN
cril-1051	589	10	.	.	PUNCT
cril-1051	590	1	in	in	ADP
cril-1051	590	2	proceedings	proceeding	NOUN
cril-1051	590	3	of	of	ADP
cril-1051	590	4	the	the	DET
cril-1051	590	5	9th	9th	ADJ
cril-1051	590	6	sighum	sighum	ADJ
cril-1051	590	7	workshop	workshop	NOUN
cril-1051	590	8	on	on	ADP
cril-1051	590	9	language	language	NOUN
cril-1051	590	10	technology	technology	NOUN
cril-1051	590	11	for	for	ADP
cril-1051	590	12	cultural	cultural	ADJ
cril-1051	590	13	herita	herita	PROPN
cril-1051	590	14	ge	ge	PROPN
cril-1051	590	15	,	,	PUNCT
cril-1051	590	16	social	social	ADJ
cril-1051	590	17	sciences	science	NOUN
cril-1051	590	18	,	,	PUNCT
cril-1051	590	19	and	and	CCONJ
cril-1051	590	20	humanities	humanity	NOUN
cril-1051	590	21	(	(	PUNCT
cril-1051	590	22	latech	latech	PROPN
cril-1051	590	23	)	)	PUNCT
cril-1051	590	24	,	,	PUNCT
cril-1051	590	25	68–72	68–72	NUM
cril-1051	590	26	.	.	PUNCT
cril-1051	591	1	beijing	beijing	PROPN
cril-1051	591	2	,	,	PUNCT
cril-1051	591	3	china	china	PROPN
cril-1051	591	4	:	:	PUNCT
cril-1051	591	5	association	association	NOUN
cril-1051	591	6	for	for	ADP
cril-1051	591	7	computational	computational	ADJ
cril-1051	591	8	linguistics	linguistic	NOUN
cril-1051	591	9	.	.	PUNCT
cril-1051	592	1	http://aclweb.org/anthology/w15-3710	http://aclweb.org/anthology/w15-3710	PROPN
cril-1051	592	2	(	(	PUNCT
cril-1051	592	3	25	25	NUM
cril-1051	592	4	may	may	PROPN
cril-1051	592	5	,	,	PUNCT
cril-1051	592	6	2020	2020	NUM
cril-1051	592	7	)	)	PUNCT
cril-1051	592	8	.	.	PUNCT
cril-1051	593	1	seifart	seifart	PROPN
cril-1051	593	2	,	,	PUNCT
cril-1051	593	3	frank	frank	PROPN
cril-1051	593	4	,	,	PUNCT
cril-1051	593	5	nicholas	nicholas	PROPN
cril-1051	593	6	evans	evans	PROPN
cril-1051	593	7	,	,	PUNCT
cril-1051	593	8	harald	harald	PROPN
cril-1051	593	9	hammarström	hammarström	PROPN
cril-1051	593	10	&	&	CCONJ
cril-1051	593	11	stephen	stephen	PROPN
cril-1051	593	12	c.	c.	PROPN
cril-1051	593	13	levinson	levinson	PROPN
cril-1051	593	14	.	.	PUNCT
cril-1051	594	1	2018	2018	NUM
cril-1051	594	2	.	.	PUNCT
cril-1051	595	1	language	language	NOUN
cril-1051	595	2	documentation	documentation	NOUN
cril-1051	595	3	twenty	twenty	NUM
cril-1051	595	4	-	-	PUNCT
cril-1051	595	5	five	five	NUM
cril-1051	595	6	years	year	NOUN
cril-1051	595	7	on	on	ADP
cril-1051	595	8	.	.	PUNCT
cril-1051	596	1	language	language	NOUN
cril-1051	596	2	94(4	94(4	NUM
cril-1051	596	3	)	)	PUNCT
cril-1051	596	4	.	.	PUNCT
cril-1051	597	1	e324	e324	PROPN
cril-1051	597	2	–	–	PUNCT
cril-1051	597	3	e345	e345	PROPN
cril-1051	597	4	.	.	PUNCT
cril-1051	598	1	https://muse.jhu.edu/article/712110	https://muse.jhu.edu/article/712110	PROPN
cril-1051	598	2	(	(	PUNCT
cril-1051	598	3	8	8	NUM
cril-1051	598	4	october	october	PROPN
cril-1051	598	5	,	,	PUNCT
cril-1051	598	6	2019	2019	NUM
cril-1051	598	7	)	)	PUNCT
cril-1051	598	8	.	.	PUNCT
cril-1051	599	1	sharma	sharma	PROPN
cril-1051	599	2	,	,	PUNCT
cril-1051	599	3	abhishek	abhishek	PROPN
cril-1051	599	4	,	,	PUNCT
cril-1051	599	5	ganesh	ganesh	PROPN
cril-1051	599	6	katrapati	katrapati	PROPN
cril-1051	599	7	&	&	CCONJ
cril-1051	599	8	dipti	dipti	PROPN
cril-1051	599	9	misra	misra	PROPN
cril-1051	599	10	sharma	sharma	PROPN
cril-1051	599	11	.	.	PUNCT
cril-1051	599	12	2018	2018	NUM
cril-1051	599	13	.	.	PUNCT
cril-1051	600	1	iit(bhu)–iiith	iit(bhu)–iiith	NOUN
cril-1051	600	2	at	at	ADP
cril-1051	600	3	conll	conll	PROPN
cril-1051	600	4	–	–	PUNCT
cril-1051	600	5	sigmorphon	sigmorphon	NOUN
cril-1051	600	6	2018	2018	NUM
cril-1051	600	7	shared	share	VERB
cril-1051	600	8	task	task	NOUN
cril-1051	600	9	on	on	ADP
cril-1051	600	10	universal	universal	ADJ
cril-1051	600	11	morphological	morphological	ADJ
cril-1051	600	12	reinflection	reinflection	NOUN
cril-1051	600	13	.	.	PUNCT
cril-1051	601	1	in	in	ADP
cril-1051	601	2	proceedings	proceeding	NOUN
cril-1051	601	3	of	of	ADP
cril-1051	601	4	the	the	DET
cril-1051	601	5	conll	conll	PROPN
cril-1051	601	6	–	–	PUNCT
cril-1051	601	7	sigmorphon	sigmorphon	NOUN
cril-1051	601	8	2018	2018	NUM
cril-1051	601	9	shared	share	VERB
cril-1051	601	10	task	task	NOUN
cril-1051	601	11	:	:	PUNCT
cril-1051	601	12	universal	universal	ADJ
cril-1051	601	13	morphological	morphological	ADJ
cril-1051	601	14	reinflection	reinflection	NOUN
cril-1051	601	15	,	,	PUNCT
cril-1051	601	16	105–111	105–111	NUM
cril-1051	601	17	.	.	PUNCT
cril-1051	602	1	brussels	brussels	PROPN
cril-1051	602	2	:	:	PUNCT
cril-1051	602	3	association	association	NOUN
cril-1051	602	4	for	for	ADP
cril-1051	602	5	computational	computational	ADJ
cril-1051	602	6	linguistics	linguistic	NOUN
cril-1051	602	7	.	.	PUNCT
cril-1051	603	1	https://www.aclweb.org/anthology/k18-3013	https://www.aclweb.org/anthology/k18-3013	PUNCT
cril-1051	603	2	(	(	PUNCT
cril-1051	603	3	26	26	NUM
cril-1051	603	4	january	january	PROPN
cril-1051	603	5	,	,	PUNCT
cril-1051	603	6	2021	2021	NUM
cril-1051	603	7	)	)	PUNCT
cril-1051	603	8	.	.	PUNCT
cril-1051	604	1	silfverberg	silfverberg	PROPN
cril-1051	604	2	,	,	PUNCT
cril-1051	604	3	miikka	miikka	PROPN
cril-1051	604	4	&	&	CCONJ
cril-1051	604	5	mans	mans	PROPN
cril-1051	604	6	hulden	hulden	VERB
cril-1051	604	7	.	.	PUNCT
cril-1051	605	1	2018	2018	NUM
cril-1051	605	2	.	.	PUNCT
cril-1051	606	1	an	an	DET
cril-1051	606	2	encoder	encoder	NOUN
cril-1051	606	3	-	-	PUNCT
cril-1051	606	4	decoder	decoder	NOUN
cril-1051	606	5	approach	approach	NOUN
cril-1051	606	6	to	to	ADP
cril-1051	606	7	the	the	DET
cril-1051	606	8	paradigm	paradigm	NOUN
cril-1051	606	9	cell	cell	NOUN
cril-1051	606	10	filling	filling	NOUN
cril-1051	606	11	problem	problem	NOUN
cril-1051	606	12	.	.	PUNCT
cril-1051	607	1	in	in	ADP
cril-1051	607	2	proceedings	proceeding	NOUN
cril-1051	607	3	of	of	ADP
cril-1051	607	4	the	the	DET
cril-1051	607	5	2018	2018	NUM
cril-1051	607	6	conference	conference	NOUN
cril-1051	607	7	on	on	ADP
cril-1051	607	8	empirical	empirical	ADJ
cril-1051	607	9	methods	method	NOUN
cril-1051	607	10	in	in	ADP
cril-1051	607	11	natural	natural	ADJ
cril-1051	607	12	language	language	NOUN
cril-1051	607	13	processing	processing	NOUN
cril-1051	607	14	,	,	PUNCT
cril-1051	607	15	2883–2889	2883–2889	NUM
cril-1051	607	16	.	.	PUNCT
cril-1051	607	17	brussels	brussels	PROPN
cril-1051	607	18	,	,	PUNCT
cril-1051	607	19	belgium	belgium	PROPN
cril-1051	607	20	:	:	PUNCT
cril-1051	607	21	association	association	NOUN
cril-1051	607	22	for	for	ADP
cril-1051	607	23	computational	computational	ADJ
cril-1051	607	24	linguistics	linguistic	NOUN
cril-1051	607	25	.	.	PUNCT
cril-1051	608	1	https://www.aclweb.org/anthology/d18-1315	https://www.aclweb.org/anthology/d18-1315	PROPN
cril-1051	608	2	(	(	PUNCT
cril-1051	608	3	25	25	NUM
cril-1051	608	4	april	april	PROPN
cril-1051	608	5	,	,	PUNCT
cril-1051	608	6	2019	2019	NUM
cril-1051	608	7	)	)	PUNCT
cril-1051	608	8	.	.	PUNCT
cril-1051	609	1	snyder	snyder	PROPN
cril-1051	609	2	,	,	PUNCT
cril-1051	609	3	benjamin	benjamin	PROPN
cril-1051	609	4	&	&	CCONJ
cril-1051	609	5	regina	regina	PROPN
cril-1051	609	6	barzilay	barzilay	PROPN
cril-1051	609	7	.	.	PUNCT
cril-1051	610	1	2008	2008	NUM
cril-1051	610	2	.	.	PUNCT
cril-1051	611	1	unsupervised	unsupervised	ADJ
cril-1051	611	2	multilingual	multilingual	ADJ
cril-1051	611	3	learning	learning	NOUN
cril-1051	611	4	for	for	ADP
cril-1051	611	5	morphological	morphological	ADJ
cril-1051	611	6	segmentation	segmentation	NOUN
cril-1051	611	7	.	.	PUNCT
cril-1051	612	1	in	in	ADP
cril-1051	612	2	acl	acl	PROPN
cril-1051	612	3	,	,	PUNCT
cril-1051	612	4	737–745	737–745	NUM
cril-1051	612	5	.	.	PUNCT
cril-1051	612	6	soricut	soricut	PROPN
cril-1051	612	7	,	,	PUNCT
cril-1051	612	8	radu	radu	PROPN
cril-1051	612	9	&	&	CCONJ
cril-1051	612	10	franz	franz	PROPN
cril-1051	612	11	och	och	PROPN
cril-1051	612	12	.	.	PROPN
cril-1051	612	13	2015	2015	NUM
cril-1051	612	14	.	.	PUNCT
cril-1051	613	1	unsupervised	unsupervised	ADJ
cril-1051	613	2	morphology	morphology	NOUN
cril-1051	613	3	induction	induction	NOUN
cril-1051	613	4	using	use	VERB
cril-1051	613	5	word	word	NOUN
cril-1051	613	6	embeddings	embedding	NOUN
cril-1051	613	7	.	.	PUNCT
cril-1051	614	1	in	in	ADP
cril-1051	614	2	proceedings	proceeding	NOUN
cril-1051	614	3	of	of	ADP
cril-1051	614	4	the	the	DET
cril-1051	614	5	2015	2015	NUM
cril-1051	614	6	conference	conference	NOUN
cril-1051	614	7	of	of	ADP
cril-1051	614	8	the	the	DET
cril-1051	614	9	north	north	ADJ
cril-1051	614	10	american	american	ADJ
cril-1051	614	11	chapter	chapter	NOUN
cril-1051	614	12	of	of	ADP
cril-1051	614	13	the	the	DET
cril-1051	614	14	association	association	NOUN
cril-1051	614	15	for	for	ADP
cril-1051	614	16	computational	computational	ADJ
cril-1051	614	17	linguistics	linguistic	NOUN
cril-1051	614	18	:	:	PUNCT
cril-1051	614	19	human	human	ADJ
cril-1051	614	20	language	language	NOUN
cril-1051	614	21	technologies	technology	NOUN
cril-1051	614	22	,	,	PUNCT
cril-1051	614	23	1627	1627	NUM
cril-1051	614	24	–	–	PUNCT
cril-1051	614	25	1637	1637	NUM
cril-1051	614	26	.	.	PUNCT
cril-1051	615	1	denver	denver	PROPN
cril-1051	615	2	,	,	PUNCT
cril-1051	615	3	colorado	colorado	PROPN
cril-1051	615	4	.	.	PUNCT
cril-1051	615	5	sudhakar	sudhakar	PROPN
cril-1051	615	6	,	,	PUNCT
cril-1051	615	7	akhilesh	akhilesh	PROPN
cril-1051	615	8	&	&	CCONJ
cril-1051	615	9	anil	anil	PROPN
cril-1051	615	10	kumar	kumar	PROPN
cril-1051	615	11	singh	singh	PROPN
cril-1051	615	12	.	.	PUNCT
cril-1051	616	1	2017	2017	NUM
cril-1051	616	2	.	.	PUNCT
cril-1051	617	1	experiments	experiment	NOUN
cril-1051	617	2	on	on	ADP
cril-1051	617	3	morphological	morphological	ADJ
cril-1051	617	4	reinflection	reinflection	NOUN
cril-1051	617	5	:	:	PUNCT
cril-1051	617	6	conll-2017	conll-2017	ADJ
cril-1051	617	7	shared	shared	ADJ
cril-1051	617	8	task	task	NOUN
cril-1051	617	9	.	.	PUNCT
cril-1051	618	1	proceedings	proceeding	NOUN
cril-1051	618	2	of	of	ADP
cril-1051	618	3	the	the	DET
cril-1051	618	4	conll	conll	PROPN
cril-1051	618	5	sigmorphon	sigmorphon	PROPN
cril-1051	618	6	2017	2017	NUM
cril-1051	618	7	shared	share	VERB
cril-1051	618	8	colorado	colorado	NOUN
cril-1051	618	9	research	research	NOUN
cril-1051	618	10	in	in	ADP
cril-1051	618	11	linguistics	linguistic	NOUN
cril-1051	618	12	,	,	PUNCT
cril-1051	618	13	volume	volume	NOUN
cril-1051	618	14	25	25	NUM
cril-1051	618	15	(	(	PUNCT
cril-1051	618	16	2021	2021	NUM
cril-1051	618	17	)	)	PUNCT
cril-1051	618	18	28	28	NUM
cril-1051	618	19	task	task	NOUN
cril-1051	618	20	:	:	PUNCT
cril-1051	618	21	universal	universal	ADJ
cril-1051	618	22	morphological	morphological	ADJ
cril-1051	618	23	reinflection	reinflection	NOUN
cril-1051	618	24	71–78	71–78	NUM
cril-1051	618	25	.	.	PUNCT
cril-1051	619	1	https://doi.org/10.18653/v1/k172007	https://doi.org/10.18653/v1/k172007	PROPN
cril-1051	619	2	.	.	PUNCT
cril-1051	620	1	https://aclanthology.coli.uni-saarland.de/papers/k17-2007/k17-2007	https://aclanthology.coli.uni-saarland.de/papers/k17-2007/k17-2007	PROPN
cril-1051	620	2	(	(	PUNCT
cril-1051	620	3	16	16	NUM
cril-1051	620	4	january	january	PROPN
cril-1051	620	5	,	,	PUNCT
cril-1051	620	6	2018	2018	NUM
cril-1051	620	7	)	)	PUNCT
cril-1051	620	8	.	.	PUNCT
cril-1051	621	1	vallejos	vallejos	PROPN
cril-1051	621	2	,	,	PUNCT
cril-1051	621	3	rosa	rosa	PROPN
cril-1051	621	4	.	.	PROPN
cril-1051	621	5	2014	2014	NUM
cril-1051	621	6	.	.	PUNCT
cril-1051	622	1	integrating	integrate	VERB
cril-1051	622	2	language	language	NOUN
cril-1051	622	3	documentation	documentation	NOUN
cril-1051	622	4	,	,	PUNCT
cril-1051	622	5	language	language	NOUN
cril-1051	622	6	preservation	preservation	NOUN
cril-1051	622	7	,	,	PUNCT
cril-1051	622	8	and	and	CCONJ
cril-1051	622	9	linguistic	linguistic	ADJ
cril-1051	622	10	research	research	NOUN
cril-1051	622	11	:	:	PUNCT
cril-1051	622	12	working	work	VERB
cril-1051	622	13	with	with	ADP
cril-1051	622	14	the	the	DET
cril-1051	622	15	kokamas	kokama	NOUN
cril-1051	622	16	from	from	ADP
cril-1051	622	17	the	the	DET
cril-1051	622	18	amazon	amazon	NOUN
cril-1051	622	19	.	.	PUNCT
cril-1051	623	1	language	language	NOUN
cril-1051	623	2	documentation	documentation	NOUN
cril-1051	623	3	&	&	CCONJ
cril-1051	623	4	conservation	conservation	NOUN
cril-1051	623	5	8	8	NUM
cril-1051	623	6	.	.	PUNCT
cril-1051	623	7	38–65	38–65	NUM
cril-1051	623	8	.	.	PUNCT
cril-1051	624	1	http://scholarspace.manoa.hawaii.edu/bitstream/handle/10125/4618/vallejos.pdf?sequenc	http://scholarspace.manoa.hawaii.edu/bitstream/handle/10125/4618/vallejos.pdf?sequenc	PRON
cril-1051	624	2	e=1	e=1	PROPN
cril-1051	624	3	.	.	PUNCT
cril-1051	624	4	virpioja	virpioja	PROPN
cril-1051	624	5	,	,	PUNCT
cril-1051	624	6	sami	sami	PROPN
cril-1051	624	7	,	,	PUNCT
cril-1051	624	8	ville	ville	PROPN
cril-1051	624	9	turunen	turunen	PROPN
cril-1051	624	10	,	,	PUNCT
cril-1051	624	11	sebastian	sebastian	PROPN
cril-1051	624	12	spiegler	spiegler	NOUN
cril-1051	624	13	,	,	PUNCT
cril-1051	624	14	oskar	oskar	PROPN
cril-1051	624	15	kohonen	kohonen	PROPN
cril-1051	624	16	&	&	CCONJ
cril-1051	624	17	mikko	mikko	PROPN
cril-1051	624	18	kurimo	kurimo	PROPN
cril-1051	624	19	.	.	PUNCT
cril-1051	625	1	2011	2011	NUM
cril-1051	625	2	.	.	PUNCT
cril-1051	626	1	empirical	empirical	ADJ
cril-1051	626	2	comparison	comparison	NOUN
cril-1051	626	3	of	of	ADP
cril-1051	626	4	evaluation	evaluation	NOUN
cril-1051	626	5	methods	method	NOUN
cril-1051	626	6	for	for	ADP
cril-1051	626	7	unsupervised	unsupervised	ADJ
cril-1051	626	8	learning	learning	NOUN
cril-1051	626	9	of	of	ADP
cril-1051	626	10	morphology	morphology	NOUN
cril-1051	626	11	.	.	PUNCT
cril-1051	627	1	trait	trait	PROPN
cril-1051	627	2	.	.	PUNCT
cril-1051	628	1	autom	autom	PROPN
cril-1051	628	2	.	.	PUNCT
cril-1051	629	1	des	des	PROPN
cril-1051	629	2	langues	langues	PROPN
cril-1051	629	3	52(2	52(2	NUM
cril-1051	629	4	)	)	PUNCT
cril-1051	629	5	.	.	PUNCT
cril-1051	630	1	45–90	45–90	NUM
cril-1051	630	2	.	.	PUNCT
cril-1051	631	1	vylomova	vylomova	PROPN
cril-1051	631	2	,	,	PUNCT
cril-1051	631	3	ekaterina	ekaterina	PROPN
cril-1051	631	4	,	,	PUNCT
cril-1051	631	5	jennifer	jennifer	PROPN
cril-1051	631	6	white	white	PROPN
cril-1051	631	7	,	,	PUNCT
cril-1051	631	8	elizabeth	elizabeth	PROPN
cril-1051	631	9	salesky	salesky	PROPN
cril-1051	631	10	,	,	PUNCT
cril-1051	631	11	sabrina	sabrina	PROPN
cril-1051	631	12	j.	j.	PROPN
cril-1051	631	13	mielke	mielke	PROPN
cril-1051	631	14	,	,	PUNCT
cril-1051	631	15	shijie	shijie	PROPN
cril-1051	631	16	wu	wu	PROPN
cril-1051	631	17	,	,	PUNCT
cril-1051	631	18	edoardo	edoardo	PROPN
cril-1051	631	19	maria	maria	PROPN
cril-1051	631	20	ponti	ponti	PROPN
cril-1051	631	21	,	,	PUNCT
cril-1051	631	22	rowan	rowan	PROPN
cril-1051	631	23	hall	hall	PROPN
cril-1051	631	24	maudslay	maudslay	PROPN
cril-1051	631	25	,	,	PUNCT
cril-1051	631	26	et	et	PROPN
cril-1051	631	27	al	al	PROPN
cril-1051	631	28	.	.	PROPN
cril-1051	631	29	2020	2020	NUM
cril-1051	631	30	.	.	PUNCT
cril-1051	632	1	sigmorphon	sigmorphon	PROPN
cril-1051	632	2	2020	2020	NUM
cril-1051	632	3	shared	share	VERB
cril-1051	632	4	task	task	NOUN
cril-1051	632	5	0	0	NUM
cril-1051	632	6	:	:	PUNCT
cril-1051	632	7	typologically	typologically	ADV
cril-1051	632	8	diverse	diverse	ADJ
cril-1051	632	9	morphological	morphological	ADJ
cril-1051	632	10	inflection	inflection	NOUN
cril-1051	632	11	.	.	PUNCT
cril-1051	633	1	in	in	ADP
cril-1051	633	2	proceedings	proceeding	NOUN
cril-1051	633	3	of	of	ADP
cril-1051	633	4	the	the	DET
cril-1051	633	5	17th	17th	ADJ
cril-1051	633	6	sigmorphon	sigmorphon	NOUN
cril-1051	633	7	workshop	workshop	NOUN
cril-1051	633	8	on	on	ADP
cril-1051	633	9	computational	computational	ADJ
cril-1051	633	10	research	research	NOUN
cril-1051	633	11	in	in	ADP
cril-1051	633	12	phonetics	phonetic	NOUN
cril-1051	633	13	,	,	PUNCT
cril-1051	633	14	phonology	phonology	NOUN
cril-1051	633	15	,	,	PUNCT
cril-1051	633	16	and	and	CCONJ
cril-1051	633	17	morphology	morphology	NOUN
cril-1051	633	18	,	,	PUNCT
cril-1051	633	19	1–39	1–39	PROPN
cril-1051	633	20	.	.	PUNCT
cril-1051	634	1	online	online	PROPN
cril-1051	634	2	:	:	PUNCT
cril-1051	635	1	association	association	NOUN
cril-1051	635	2	for	for	ADP
cril-1051	635	3	computational	computational	ADJ
cril-1051	635	4	linguistics	linguistic	NOUN
cril-1051	635	5	.	.	PUNCT
cril-1051	636	1	https://www.aclweb.org/anthology/2020.sigmorphon-1.1	https://www.aclweb.org/anthology/2020.sigmorphon-1.1	NOUN
cril-1051	636	2	(	(	PUNCT
cril-1051	636	3	27	27	NUM
cril-1051	636	4	april	april	PROPN
cril-1051	636	5	,	,	PUNCT
cril-1051	636	6	2021	2021	NUM
cril-1051	636	7	)	)	PUNCT
cril-1051	636	8	.	.	PUNCT
cril-1051	637	1	wang	wang	PROPN
cril-1051	637	2	,	,	PUNCT
cril-1051	637	3	linlin	linlin	PROPN
cril-1051	637	4	,	,	PUNCT
cril-1051	637	5	zhu	zhu	PROPN
cril-1051	637	6	cao	cao	PROPN
cril-1051	637	7	,	,	PUNCT
cril-1051	637	8	yu	yu	PROPN
cril-1051	637	9	xia	xia	PROPN
cril-1051	637	10	&	&	CCONJ
cril-1051	637	11	gerard	gerard	PROPN
cril-1051	637	12	de	de	PROPN
cril-1051	637	13	melo	melo	PROPN
cril-1051	637	14	.	.	PUNCT
cril-1051	638	1	2016	2016	NUM
cril-1051	638	2	.	.	PUNCT
cril-1051	639	1	morphological	morphological	ADJ
cril-1051	639	2	segmentation	segmentation	NOUN
cril-1051	639	3	with	with	ADP
cril-1051	639	4	window	window	NOUN
cril-1051	639	5	lstm	lstm	PROPN
cril-1051	639	6	neural	neural	ADJ
cril-1051	639	7	networks	network	NOUN
cril-1051	639	8	.	.	PUNCT
cril-1051	640	1	in	in	ADP
cril-1051	640	2	aaai’16	aaai’16	NOUN
cril-1051	640	3	:	:	PUNCT
cril-1051	640	4	proceedings	proceeding	NOUN
cril-1051	640	5	of	of	ADP
cril-1051	640	6	the	the	DET
cril-1051	640	7	thirtieth	thirtieth	ADJ
cril-1051	640	8	aaai	aaai	PROPN
cril-1051	640	9	conference	conference	NOUN
cril-1051	640	10	on	on	ADP
cril-1051	640	11	artificial	artificial	ADJ
cril-1051	640	12	intelligence	intelligence	NOUN
cril-1051	640	13	,	,	PUNCT
cril-1051	640	14	2842–2848	2842–2848	NUM
cril-1051	640	15	.	.	PUNCT
cril-1051	640	16	wax	wax	PROPN
cril-1051	640	17	,	,	PUNCT
cril-1051	640	18	david	david	PROPN
cril-1051	640	19	allen	allen	PROPN
cril-1051	640	20	.	.	PROPN
cril-1051	640	21	2014	2014	NUM
cril-1051	640	22	.	.	PUNCT
cril-1051	641	1	automated	automate	VERB
cril-1051	641	2	grammar	grammar	NOUN
cril-1051	641	3	engineering	engineering	NOUN
cril-1051	641	4	for	for	ADP
cril-1051	641	5	verbal	verbal	ADJ
cril-1051	641	6	morphology	morphology	NOUN
cril-1051	641	7	.	.	PUNCT
cril-1051	642	1	thesis	thesis	NOUN
cril-1051	642	2	.	.	PUNCT
cril-1051	643	1	https://digital.lib.washington.edu:443/researchworks/handle/1773/25373	https://digital.lib.washington.edu:443/researchworks/handle/1773/25373	ADP
cril-1051	643	2	(	(	PUNCT
cril-1051	643	3	6	6	NUM
cril-1051	643	4	april	april	PROPN
cril-1051	643	5	,	,	PUNCT
cril-1051	643	6	2020	2020	NUM
cril-1051	643	7	)	)	PUNCT
cril-1051	643	8	.	.	PUNCT
cril-1051	644	1	woodbury	woodbury	PROPN
cril-1051	644	2	,	,	PUNCT
cril-1051	644	3	tony	tony	PROPN
cril-1051	644	4	.	.	PROPN
cril-1051	644	5	2003	2003	NUM
cril-1051	644	6	.	.	PUNCT
cril-1051	645	1	defining	define	VERB
cril-1051	645	2	documentary	documentary	NOUN
cril-1051	645	3	linguistics	linguistic	NOUN
cril-1051	645	4	.	.	PUNCT
cril-1051	646	1	language	language	NOUN
cril-1051	646	2	documentation	documentation	NOUN
cril-1051	646	3	and	and	CCONJ
cril-1051	646	4	description	description	NOUN
cril-1051	646	5	1	1	NUM
cril-1051	646	6	.	.	PUNCT
cril-1051	647	1	35–51	35–51	NUM
cril-1051	647	2	.	.	PUNCT
cril-1051	648	1	wu	wu	PROPN
cril-1051	648	2	,	,	PUNCT
cril-1051	648	3	shijie	shijie	PROPN
cril-1051	648	4	&	&	CCONJ
cril-1051	648	5	ryan	ryan	PROPN
cril-1051	648	6	cotterell	cotterell	PROPN
cril-1051	648	7	.	.	PUNCT
cril-1051	649	1	2019	2019	NUM
cril-1051	649	2	.	.	PUNCT
cril-1051	650	1	exact	exact	ADJ
cril-1051	650	2	hard	hard	ADJ
cril-1051	650	3	monotonic	monotonic	ADJ
cril-1051	650	4	attention	attention	NOUN
cril-1051	650	5	for	for	ADP
cril-1051	650	6	character	character	NOUN
cril-1051	650	7	-	-	PUNCT
cril-1051	650	8	level	level	NOUN
cril-1051	650	9	transduction	transduction	NOUN
cril-1051	650	10	.	.	PUNCT
cril-1051	651	1	in	in	ADP
cril-1051	651	2	proceedings	proceeding	NOUN
cril-1051	651	3	of	of	ADP
cril-1051	651	4	the	the	DET
cril-1051	651	5	57th	57th	ADJ
cril-1051	651	6	annual	annual	ADJ
cril-1051	651	7	meeting	meeting	NOUN
cril-1051	651	8	of	of	ADP
cril-1051	651	9	the	the	DET
cril-1051	651	10	association	association	NOUN
cril-1051	651	11	for	for	ADP
cril-1051	651	12	computational	computational	ADJ
cril-1051	651	13	linguistics	linguistic	NOUN
cril-1051	651	14	,	,	PUNCT
cril-1051	651	15	1530–1537	1530–1537	NUM
cril-1051	651	16	.	.	PUNCT
cril-1051	651	17	florence	florence	PROPN
cril-1051	651	18	,	,	PUNCT
cril-1051	651	19	italy	italy	PROPN
cril-1051	651	20	:	:	PUNCT
cril-1051	651	21	association	association	NOUN
cril-1051	651	22	for	for	ADP
cril-1051	651	23	computational	computational	ADJ
cril-1051	651	24	linguistics	linguistic	NOUN
cril-1051	651	25	.	.	PUNCT
cril-1051	652	1	https://www.aclweb.org/anthology/p19-1148	https://www.aclweb.org/anthology/p19-1148	X
cril-1051	652	2	(	(	PUNCT
cril-1051	652	3	26	26	NUM
cril-1051	652	4	january	january	PROPN
cril-1051	652	5	,	,	PUNCT
cril-1051	652	6	2021	2021	NUM
cril-1051	652	7	)	)	PUNCT
cril-1051	652	8	.	.	PUNCT
cril-1051	653	1	wu	wu	PROPN
cril-1051	653	2	,	,	PUNCT
cril-1051	653	3	shijie	shijie	PROPN
cril-1051	653	4	,	,	PUNCT
cril-1051	653	5	ryan	ryan	PROPN
cril-1051	653	6	cotterell	cotterell	PROPN
cril-1051	653	7	&	&	CCONJ
cril-1051	653	8	mans	mans	PROPN
cril-1051	653	9	hulden	hulden	VERB
cril-1051	653	10	.	.	PUNCT
cril-1051	654	1	2021	2021	NUM
cril-1051	654	2	.	.	PUNCT
cril-1051	655	1	applying	apply	VERB
cril-1051	655	2	the	the	DET
cril-1051	655	3	transformer	transformer	NOUN
cril-1051	655	4	to	to	ADP
cril-1051	655	5	character	character	NOUN
cril-1051	655	6	-	-	PUNCT
cril-1051	655	7	level	level	NOUN
cril-1051	655	8	transduction	transduction	NOUN
cril-1051	655	9	.	.	PUNCT
cril-1051	656	1	in	in	ADP
cril-1051	656	2	proceedings	proceeding	NOUN
cril-1051	656	3	of	of	ADP
cril-1051	656	4	the	the	DET
cril-1051	656	5	16th	16th	ADJ
cril-1051	656	6	conference	conference	NOUN
cril-1051	656	7	of	of	ADP
cril-1051	656	8	the	the	DET
cril-1051	656	9	european	european	ADJ
cril-1051	656	10	chapter	chapter	NOUN
cril-1051	656	11	of	of	ADP
cril-1051	656	12	the	the	DET
cril-1051	656	13	computational	computational	ADJ
cril-1051	656	14	morphology	morphology	NOUN
cril-1051	656	15	for	for	ADP
cril-1051	656	16	language	language	NOUN
cril-1051	656	17	description	description	NOUN
cril-1051	656	18	and	and	CCONJ
cril-1051	656	19	documentation	documentation	NOUN
cril-1051	656	20	29	29	NUM
cril-1051	656	21	association	association	NOUN
cril-1051	656	22	for	for	ADP
cril-1051	656	23	computational	computational	ADJ
cril-1051	656	24	linguistics	linguistic	NOUN
cril-1051	656	25	:	:	PUNCT
cril-1051	656	26	main	main	ADJ
cril-1051	656	27	volume	volume	NOUN
cril-1051	656	28	,	,	PUNCT
cril-1051	656	29	1901	1901	NUM
cril-1051	656	30	-	-	SYM
cril-1051	656	31	-1907	-1907	PROPN
cril-1051	656	32	.	.	PUNCT
cril-1051	656	33	association	association	PROPN
cril-1051	656	34	for	for	ADP
cril-1051	656	35	computational	computational	ADJ
cril-1051	656	36	linguistics	linguistic	NOUN
cril-1051	656	37	.	.	PUNCT
cril-1051	656	38	https://aclanthology.org/2021.eacl-main.163	https://aclanthology.org/2021.eacl-main.163	PROPN
cril-1051	656	39	.	.	PUNCT
cril-1051	657	1	xia	xia	PROPN
cril-1051	657	2	,	,	PUNCT
cril-1051	657	3	fei	fei	PROPN
cril-1051	657	4	,	,	PUNCT
cril-1051	657	5	william	william	PROPN
cril-1051	657	6	d.	d.	PROPN
cril-1051	657	7	lewis	lewis	PROPN
cril-1051	657	8	,	,	PUNCT
cril-1051	657	9	michael	michael	PROPN
cril-1051	657	10	wayne	wayne	PROPN
cril-1051	657	11	goodman	goodman	PROPN
cril-1051	657	12	,	,	PUNCT
cril-1051	657	13	glenn	glenn	PROPN
cril-1051	657	14	slayden	slayden	PROPN
cril-1051	657	15	,	,	PUNCT
cril-1051	657	16	ryan	ryan	PROPN
cril-1051	657	17	georgi	georgi	PROPN
cril-1051	657	18	,	,	PUNCT
cril-1051	657	19	joshua	joshua	PROPN
cril-1051	657	20	crowgey	crowgey	PROPN
cril-1051	657	21	&	&	CCONJ
cril-1051	657	22	emily	emily	PROPN
cril-1051	657	23	m.	m.	PROPN
cril-1051	657	24	bender	bender	PROPN
cril-1051	657	25	.	.	PUNCT
cril-1051	658	1	2016	2016	NUM
cril-1051	658	2	.	.	PUNCT
cril-1051	659	1	enriching	enrich	VERB
cril-1051	659	2	a	a	DET
cril-1051	659	3	massively	massively	ADV
cril-1051	659	4	multilingual	multilingual	ADJ
cril-1051	659	5	database	database	NOUN
cril-1051	659	6	of	of	ADP
cril-1051	659	7	interlinear	interlinear	ADJ
cril-1051	659	8	glossed	glossed	ADJ
cril-1051	659	9	text	text	NOUN
cril-1051	659	10	.	.	PUNCT
cril-1051	660	1	language	language	NOUN
cril-1051	660	2	resources	resource	NOUN
cril-1051	660	3	and	and	CCONJ
cril-1051	660	4	evaluation	evaluation	NOUN
cril-1051	660	5	50(2	50(2	NUM
cril-1051	660	6	)	)	PUNCT
cril-1051	660	7	.	.	PUNCT
cril-1051	661	1	321–349	321–349	NUM
cril-1051	661	2	.	.	PUNCT
cril-1051	662	1	https://doi.org/10.1007/s10579-015-9325-4	https://doi.org/10.1007/s10579-015-9325-4	PROPN
cril-1051	662	2	(	(	PUNCT
cril-1051	662	3	14	14	NUM
cril-1051	662	4	january	january	PROPN
cril-1051	662	5	,	,	PUNCT
cril-1051	662	6	2020	2020	NUM
cril-1051	662	7	)	)	PUNCT
cril-1051	662	8	.	.	PUNCT
cril-1051	663	1	zmigrod	zmigrod	NOUN
cril-1051	663	2	,	,	PUNCT
cril-1051	663	3	ran	run	VERB
cril-1051	663	4	,	,	PUNCT
cril-1051	663	5	sabrina	sabrina	PROPN
cril-1051	663	6	j.	j.	PROPN
cril-1051	663	7	mielke	mielke	PROPN
cril-1051	663	8	,	,	PUNCT
cril-1051	663	9	hanna	hanna	PROPN
cril-1051	663	10	wallach	wallach	PROPN
cril-1051	663	11	&	&	CCONJ
cril-1051	663	12	ryan	ryan	PROPN
cril-1051	663	13	cotterell	cotterell	PROPN
cril-1051	663	14	.	.	PUNCT
cril-1051	664	1	2019	2019	NUM
cril-1051	664	2	.	.	PUNCT
cril-1051	665	1	counterfactual	counterfactual	ADJ
cril-1051	665	2	data	datum	NOUN
cril-1051	665	3	augmentation	augmentation	NOUN
cril-1051	665	4	for	for	ADP
cril-1051	665	5	mitigating	mitigate	VERB
cril-1051	665	6	gender	gender	NOUN
cril-1051	665	7	stereotypes	stereotype	NOUN
cril-1051	665	8	in	in	ADP
cril-1051	665	9	languages	language	NOUN
cril-1051	665	10	with	with	ADP
cril-1051	665	11	rich	rich	ADJ
cril-1051	665	12	morphology	morphology	NOUN
cril-1051	665	13	.	.	PUNCT
cril-1051	666	1	in	in	ADP
cril-1051	666	2	proceedings	proceeding	NOUN
cril-1051	666	3	of	of	ADP
cril-1051	666	4	the	the	DET
cril-1051	666	5	57th	57th	ADJ
cril-1051	666	6	annual	annual	ADJ
cril-1051	666	7	meeting	meeting	NOUN
cril-1051	666	8	of	of	ADP
cril-1051	666	9	the	the	DET
cril-1051	666	10	association	association	NOUN
cril-1051	666	11	for	for	ADP
cril-1051	666	12	computational	computational	ADJ
cril-1051	666	13	linguistics	linguistic	NOUN
cril-1051	666	14	,	,	PUNCT
cril-1051	666	15	1651–1661	1651–1661	NUM
cril-1051	666	16	.	.	PUNCT
cril-1051	666	17	florence	florence	NOUN
cril-1051	666	18	,	,	PUNCT
cril-1051	666	19	italy	italy	PROPN
cril-1051	666	20	:	:	PUNCT
cril-1051	666	21	association	association	NOUN
cril-1051	666	22	for	for	ADP
cril-1051	666	23	computational	computational	ADJ
cril-1051	666	24	linguistics	linguistic	NOUN
cril-1051	666	25	.	.	PUNCT
cril-1051	667	1	https://www.aclweb.org/anthology/p19-1161	https://www.aclweb.org/anthology/p19-1161	PUNCT
cril-1051	667	2	(	(	PUNCT
cril-1051	667	3	27	27	NUM
cril-1051	667	4	april	april	PROPN
cril-1051	667	5	,	,	PUNCT
cril-1051	667	6	2021	2021	NUM
cril-1051	667	7	)	)	PUNCT
cril-1051	667	8	.	.	PUNCT
cril-1051	668	1	colorado	colorado	PROPN
cril-1051	668	2	research	research	PROPN
cril-1051	668	3	in	in	ADP
cril-1051	668	4	linguistics	linguistic	NOUN
cril-1051	668	5	,	,	PUNCT
cril-1051	668	6	volume	volume	NOUN
cril-1051	668	7	25	25	NUM
cril-1051	668	8	(	(	PUNCT
cril-1051	668	9	2021	2021	NUM
cril-1051	668	10	)	)	PUNCT
cril-1051	668	11	30	30	NUM
cril-1051	668	12	endnotes	endnote	NOUN
cril-1051	668	13	1	1	NUM
cril-1051	668	14	according	accord	VERB
cril-1051	668	15	to	to	ADP
cril-1051	668	16	lorelei	lorelei	PROPN
cril-1051	668	17	(	(	PUNCT
cril-1051	668	18	https://www.darpa.mil/program/low-resource-languages-for-emergentincidents	https://www.darpa.mil/program/low-resource-languages-for-emergentincident	NOUN
cril-1051	668	19	)	)	PUNCT
cril-1051	668	20	,	,	PUNCT
cril-1051	668	21	`	`	PUNCT
cril-1051	668	22	`	`	PUNCT
cril-1051	668	23	low	low	ADJ
cril-1051	668	24	-	-	PUNCT
cril-1051	668	25	resource	resource	NOUN
cril-1051	668	26	''	''	PUNCT
cril-1051	668	27	refers	refer	VERB
cril-1051	668	28	to	to	ADP
cril-1051	668	29	languages	language	NOUN
cril-1051	668	30	for	for	ADP
cril-1051	668	31	which	which	PRON
cril-1051	668	32	no	no	DET
cril-1051	668	33	automated	automate	VERB
cril-1051	668	34	human	human	ADJ
cril-1051	668	35	language	language	NOUN
cril-1051	668	36	technology	technology	NOUN
cril-1051	668	37	exists	exist	VERB
cril-1051	668	38	.	.	PUNCT
cril-1051	669	1	this	this	PRON
cril-1051	669	2	is	be	AUX
cril-1051	669	3	due	due	ADJ
cril-1051	669	4	to	to	ADP
cril-1051	669	5	a	a	DET
cril-1051	669	6	lack	lack	NOUN
cril-1051	669	7	of	of	ADP
cril-1051	669	8	linguistic	linguistic	ADJ
cril-1051	669	9	resources	resource	NOUN
cril-1051	669	10	.	.	PUNCT
cril-1051	670	1	\citet{szymanski_morphological_2012	\citet{szymanski_morphological_2012	NOUN
cril-1051	670	2	}	}	PUNCT
cril-1051	670	3	estimates	estimate	VERB
cril-1051	670	4	that	that	SCONJ
cril-1051	670	5	99\%	99\%	NOUN
cril-1051	670	6	of	of	ADP
cril-1051	670	7	the	the	DET
cril-1051	670	8	world	world	NOUN
cril-1051	670	9	's	's	PART
cril-1051	670	10	languages	language	NOUN
cril-1051	670	11	are	be	AUX
cril-1051	670	12	`	`	PUNCT
cril-1051	670	13	`	`	PUNCT
cril-1051	670	14	resource	resource	NOUN
cril-1051	670	15	-	-	PUNCT
cril-1051	670	16	poor	poor	ADJ
cril-1051	670	17	''	''	PUNCT
cril-1051	670	18	.	.	PUNCT
cril-1051	671	1	in	in	ADP
cril-1051	671	2	linguistics	linguistic	NOUN
cril-1051	671	3	,	,	PUNCT
cril-1051	671	4	it	it	PRON
cril-1051	671	5	is	be	AUX
cril-1051	671	6	more	more	ADV
cril-1051	671	7	common	common	ADJ
cril-1051	671	8	to	to	PART
cril-1051	671	9	hear	hear	VERB
cril-1051	671	10	other	other	ADJ
cril-1051	671	11	terms	term	NOUN
cril-1051	671	12	.	.	PUNCT
cril-1051	672	1	the	the	DET
cril-1051	672	2	current	current	ADJ
cril-1051	672	3	work	work	NOUN
cril-1051	672	4	uses	use	VERB
cril-1051	672	5	the	the	DET
cril-1051	672	6	term	term	NOUN
cril-1051	672	7	`	`	PUNCT
cril-1051	672	8	`	`	PUNCT
cril-1051	672	9	under	under	ADV
cril-1051	672	10	-	-	PUNCT
cril-1051	672	11	described	describe	VERB
cril-1051	672	12	languages	language	NOUN
cril-1051	672	13	''	''	PUNCT
cril-1051	672	14	to	to	PART
cril-1051	672	15	refer	refer	VERB
cril-1051	672	16	to	to	ADP
cril-1051	672	17	languages	language	NOUN
cril-1051	672	18	with	with	ADP
cril-1051	672	19	minimal	minimal	ADJ
cril-1051	672	20	published	publish	VERB
cril-1051	672	21	linguistic	linguistic	ADJ
cril-1051	672	22	resources	resource	NOUN
cril-1051	672	23	;	;	PUNCT
cril-1051	672	24	these	these	PRON
cril-1051	672	25	have	have	AUX
cril-1051	672	26	been	be	AUX
cril-1051	672	27	called	call	VERB
cril-1051	672	28	`	`	PUNCT
cril-1051	672	29	`	`	PUNCT
cril-1051	672	30	very	very	ADV
cril-1051	672	31	scarce	scarce	ADJ
cril-1051	672	32	-	-	PUNCT
cril-1051	672	33	resource	resource	NOUN
cril-1051	672	34	language	language	NOUN
cril-1051	672	35	''	''	PUNCT
cril-1051	672	36	\citep{duong_natural_2017	\citep{duong_natural_2017	ADV
cril-1051	672	37	}	}	PUNCT
cril-1051	672	38	.	.	PUNCT
cril-1051	673	1	the	the	DET
cril-1051	673	2	term	term	NOUN
cril-1051	673	3	`	`	PUNCT
cril-1051	673	4	`	`	PUNCT
cril-1051	673	5	under	under	ADV
cril-1051	673	6	-	-	PUNCT
cril-1051	673	7	documented	document	VERB
cril-1051	673	8	languages	language	NOUN
cril-1051	673	9	''	''	PUNCT
cril-1051	673	10	(	(	PUNCT
cril-1051	673	11	duong	duong	PROPN
cril-1051	673	12	's	's	PART
cril-1051	673	13	`	`	PUNCT
cril-1051	673	14	`	`	PUNCT
cril-1051	673	15	extremely	extremely	ADV
cril-1051	673	16	scarce	scarce	ADJ
cril-1051	673	17	-	-	PUNCT
cril-1051	673	18	resource	resource	NOUN
cril-1051	673	19	languages	language	NOUN
cril-1051	673	20	''	''	PUNCT
cril-1051	673	21	)	)	PUNCT
cril-1051	673	22	refer	refer	VERB
cril-1051	673	23	to	to	ADP
cril-1051	673	24	languages	language	NOUN
cril-1051	673	25	that	that	PRON
cril-1051	673	26	lack	lack	VERB
cril-1051	673	27	sufficient	sufficient	ADJ
cril-1051	673	28	raw	raw	ADJ
cril-1051	673	29	or	or	CCONJ
cril-1051	673	30	annotated	annotated	ADJ
cril-1051	673	31	data	datum	NOUN
cril-1051	673	32	to	to	PART
cril-1051	673	33	write	write	VERB
cril-1051	673	34	a	a	DET
cril-1051	673	35	full	full	ADJ
cril-1051	673	36	reference	reference	NOUN
cril-1051	673	37	grammar	grammar	NOUN
cril-1051	673	38	.	.	PUNCT
cril-1051	674	1	the	the	DET
cril-1051	674	2	term	term	NOUN
cril-1051	674	3	`	`	PUNCT
cril-1051	674	4	`	`	PUNCT
cril-1051	674	5	endangered	endangered	PROPN
cril-1051	674	6	languages	language	NOUN
cril-1051	674	7	''	''	PUNCT
cril-1051	674	8	refers	refer	VERB
cril-1051	674	9	to	to	ADP
cril-1051	674	10	languages	language	NOUN
cril-1051	674	11	that	that	PRON
cril-1051	674	12	are	be	AUX
cril-1051	674	13	predicted	predict	VERB
cril-1051	674	14	to	to	PART
cril-1051	674	15	have	have	VERB
cril-1051	674	16	no	no	DET
cril-1051	674	17	native	native	ADJ
cril-1051	674	18	speakers	speaker	NOUN
cril-1051	674	19	within	within	ADP
cril-1051	674	20	a	a	DET
cril-1051	674	21	generation	generation	NOUN
cril-1051	674	22	or	or	CCONJ
cril-1051	674	23	two	two	NUM
cril-1051	674	24	.	.	PUNCT
cril-1051	675	1	most	most	ADJ
cril-1051	675	2	endangered	endanger	VERB
cril-1051	675	3	languages	language	NOUN
cril-1051	675	4	are	be	AUX
cril-1051	675	5	under	under	ADV
cril-1051	675	6	-	-	PUNCT
cril-1051	675	7	documented	document	VERB
cril-1051	675	8	and/or	and/or	CCONJ
cril-1051	675	9	under	under	ADV
cril-1051	675	10	-	-	PUNCT
cril-1051	675	11	described	describe	VERB
cril-1051	675	12	,	,	PUNCT
cril-1051	675	13	as	as	ADV
cril-1051	675	14	well	well	ADV
cril-1051	675	15	as	as	ADP
cril-1051	675	16	fitting	fit	VERB
cril-1051	675	17	the	the	DET
cril-1051	675	18	definition	definition	NOUN
cril-1051	675	19	of	of	ADP
cril-1051	675	20	low	low	ADJ
cril-1051	675	21	-	-	PUNCT
cril-1051	675	22	resource	resource	NOUN
cril-1051	675	23	languages	language	NOUN
cril-1051	675	24	.	.	PUNCT
cril-1051	676	1	the	the	DET
cril-1051	676	2	distinctions	distinction	NOUN
cril-1051	676	3	between	between	ADP
cril-1051	676	4	the	the	DET
cril-1051	676	5	terms	term	NOUN
cril-1051	676	6	are	be	AUX
cril-1051	676	7	rarely	rarely	ADV
cril-1051	676	8	crucial	crucial	ADJ
cril-1051	676	9	in	in	ADP
cril-1051	676	10	the	the	DET
cril-1051	676	11	current	current	ADJ
cril-1051	676	12	work	work	NOUN
cril-1051	676	13	.	.	PUNCT
cril-1051	677	1	in	in	ADP
cril-1051	677	2	practice	practice	NOUN
cril-1051	677	3	,	,	PUNCT
cril-1051	677	4	these	these	DET
cril-1051	677	5	terms	term	NOUN
cril-1051	677	6	can	can	AUX
cril-1051	677	7	be	be	AUX
cril-1051	677	8	used	use	VERB
cril-1051	677	9	almost	almost	ADV
cril-1051	677	10	interchangeably	interchangeably	ADV
cril-1051	677	11	.	.	PUNCT
cril-1051	678	1	2	2	NUM
cril-1051	678	2	deep	deep	ADJ
cril-1051	678	3	learning	learn	VERB
cril-1051	678	4	morphological	morphological	ADJ
cril-1051	678	5	segmentation	segmentation	NOUN
cril-1051	678	6	has	have	AUX
cril-1051	678	7	been	be	AUX
cril-1051	678	8	performed	perform	VERB
cril-1051	678	9	on	on	ADP
cril-1051	678	10	unsupervised	unsupervised	ADJ
cril-1051	678	11	texts	text	NOUN
cril-1051	678	12	with	with	ADP
cril-1051	678	13	some	some	DET
cril-1051	678	14	success	success	NOUN
cril-1051	678	15	(	(	PUNCT
cril-1051	678	16	wang	wang	PROPN
cril-1051	678	17	et	et	PROPN
cril-1051	678	18	al	al	PROPN
cril-1051	678	19	.	.	PROPN
cril-1051	678	20	2016	2016	NUM
cril-1051	678	21	)	)	PUNCT
cril-1051	678	22	.	.	PUNCT
cril-1051	679	1	3nicolai	3nicolai	NUM
cril-1051	679	2	and	and	CCONJ
cril-1051	679	3	kondrak	kondrak	PROPN
cril-1051	679	4	(	(	PUNCT
cril-1051	679	5	2017	2017	NUM
cril-1051	679	6	)	)	PUNCT
cril-1051	679	7	subdivide	subdivide	VERB
cril-1051	679	8	morphological	morphological	ADJ
cril-1051	679	9	analysis	analysis	NOUN
cril-1051	679	10	slightly	slightly	ADV
cril-1051	679	11	differently	differently	ADV
cril-1051	679	12	,	,	PUNCT
cril-1051	679	13	making	make	VERB
cril-1051	679	14	a	a	DET
cril-1051	679	15	distinction	distinction	NOUN
cril-1051	679	16	between	between	ADP
cril-1051	679	17	morphological	morphological	ADJ
cril-1051	679	18	“	"	PUNCT
cril-1051	679	19	analysis	analysis	NOUN
cril-1051	679	20	”	"	PUNCT
cril-1051	679	21	and	and	CCONJ
cril-1051	679	22	morphological	morphological	ADJ
cril-1051	679	23	tagging	tagging	NOUN
cril-1051	679	24	.	.	PUNCT
cril-1051	680	1	they	they	PRON
cril-1051	680	2	describe	describe	VERB
cril-1051	680	3	morphological	morphological	ADJ
cril-1051	680	4	analysis	analysis	NOUN
cril-1051	680	5	as	as	ADP
cril-1051	680	6	a	a	DET
cril-1051	680	7	combination	combination	NOUN
cril-1051	680	8	of	of	ADP
cril-1051	680	9	segmentation	segmentation	NOUN
cril-1051	680	10	and	and	CCONJ
cril-1051	680	11	labeling	labeling	NOUN
cril-1051	680	12	,	,	PUNCT
cril-1051	680	13	though	though	SCONJ
cril-1051	680	14	they	they	PRON
cril-1051	680	15	later	later	ADV
cril-1051	680	16	state	state	VERB
cril-1051	680	17	that	that	SCONJ
cril-1051	680	18	“	"	PUNCT
cril-1051	680	19	morphological	morphological	ADJ
cril-1051	680	20	tagging	tagging	NOUN
cril-1051	680	21	can	can	AUX
cril-1051	680	22	be	be	AUX
cril-1051	680	23	performed	perform	VERB
cril-1051	680	24	as	as	ADP
cril-1051	680	25	a	a	DET
cril-1051	680	26	downstream	downstream	ADJ
cril-1051	680	27	application	application	NOUN
cril-1051	680	28	of	of	ADP
cril-1051	680	29	morphological	morphological	ADJ
cril-1051	680	30	analysis	analysis	NOUN
cril-1051	680	31	”	"	PUNCT
cril-1051	680	32	(	(	PUNCT
cril-1051	680	33	p.	p.	NOUN
cril-1051	680	34	211	211	NUM
cril-1051	680	35	)	)	PUNCT
cril-1051	680	36	,	,	PUNCT
cril-1051	680	37	thereby	thereby	ADV
cril-1051	680	38	adhering	adhere	VERB
cril-1051	680	39	to	to	ADP
cril-1051	680	40	the	the	DET
cril-1051	680	41	same	same	ADJ
cril-1051	680	42	two	two	NUM
cril-1051	680	43	distinctions	distinction	NOUN
cril-1051	680	44	described	describe	VERB
cril-1051	680	45	in	in	ADP
cril-1051	680	46	this	this	DET
cril-1051	680	47	section	section	NOUN
cril-1051	680	48	.	.	PUNCT