id	sid	tid	token	lemma	pos
bracis-33550	1	1	a	a	DET
bracis-33550	1	2	large	large	ADJ
bracis-33550	1	3	dataset	dataset	NOUN
bracis-33550	1	4	of	of	ADP
bracis-33550	1	5	 	 	SPACE
bracis-33550	1	6	spontaneous	spontaneous	ADJ
bracis-33550	1	7	speech	speech	NOUN
bracis-33550	1	8	with	with	ADP
bracis-33550	1	9	 	 	SPACE
bracis-33550	1	10	the	the	DET
bracis-33550	1	11	 	 	SPACE
bracis-33550	1	12	accent	accent	NOUN
bracis-33550	1	13	spoken	speak	VERB
bracis-33550	1	14	in	in	ADP
bracis-33550	1	15	 	 	SPACE
bracis-33550	1	16	são	são	PROPN
bracis-33550	1	17	paulo	paulo	PROPN
bracis-33550	1	18	for	for	ADP
bracis-33550	1	19	 	 	SPACE
bracis-33550	1	20	automatic	automatic	ADJ
bracis-33550	1	21	speech	speech	NOUN
bracis-33550	1	22	recognition	recognition	NOUN
bracis-33550	1	23	evaluation	evaluation	NOUN
bracis-33550	1	24	|	|	NOUN
bracis-33550	1	25	springer	springer	NOUN
bracis-33550	1	26	nature	nature	PROPN
bracis-33550	1	27	link	link	PROPN
bracis-33550	1	28	(	(	PUNCT
bracis-33550	1	29	formerly	formerly	ADV
bracis-33550	1	30	springerlink	springerlink	NOUN
bracis-33550	1	31	)	)	PUNCT
bracis-33550	1	32	skip	skip	VERB
bracis-33550	1	33	to	to	ADP
bracis-33550	1	34	main	main	ADJ
bracis-33550	1	35	content	content	NOUN
bracis-33550	1	36	advertisement	advertisement	NOUN
bracis-33550	1	37	log	log	NOUN
bracis-33550	1	38	in	in	ADP
bracis-33550	1	39	menu	menu	NOUN
bracis-33550	1	40	find	find	VERB
bracis-33550	1	41	a	a	DET
bracis-33550	1	42	journal	journal	NOUN
bracis-33550	1	43	publish	publish	VERB
bracis-33550	1	44	with	with	ADP
bracis-33550	1	45	us	we	PRON
bracis-33550	1	46	track	track	VERB
bracis-33550	1	47	your	your	PRON
bracis-33550	1	48	research	research	NOUN
bracis-33550	1	49	search	search	NOUN
bracis-33550	1	50	cart	cart	NOUN
bracis-33550	1	51	home	home	NOUN
bracis-33550	1	52	intelligent	intelligent	ADJ
bracis-33550	1	53	systems	system	NOUN
bracis-33550	1	54	conference	conference	NOUN
bracis-33550	1	55	paper	paper	NOUN
bracis-33550	1	56	a	a	DET
bracis-33550	1	57	large	large	ADJ
bracis-33550	1	58	dataset	dataset	NOUN
bracis-33550	1	59	of	of	ADP
bracis-33550	1	60	 	 	SPACE
bracis-33550	1	61	spontaneous	spontaneous	ADJ
bracis-33550	1	62	speech	speech	NOUN
bracis-33550	1	63	with	with	ADP
bracis-33550	1	64	 	 	SPACE
bracis-33550	1	65	the	the	DET
bracis-33550	1	66	 	 	SPACE
bracis-33550	1	67	accent	accent	NOUN
bracis-33550	1	68	spoken	speak	VERB
bracis-33550	1	69	in	in	ADP
bracis-33550	1	70	 	 	SPACE
bracis-33550	1	71	são	são	PROPN
bracis-33550	1	72	paulo	paulo	PROPN
bracis-33550	1	73	for	for	ADP
bracis-33550	1	74	 	 	SPACE
bracis-33550	1	75	automatic	automatic	ADJ
bracis-33550	1	76	speech	speech	NOUN
bracis-33550	1	77	recognition	recognition	NOUN
bracis-33550	1	78	evaluation	evaluation	NOUN
bracis-33550	1	79	conference	conference	NOUN
bracis-33550	1	80	paper	paper	NOUN
bracis-33550	1	81	first	first	ADV
bracis-33550	1	82	online	online	ADV
bracis-33550	1	83	:	:	PUNCT
bracis-33550	1	84	30	30	NUM
bracis-33550	1	85	january	january	NOUN
bracis-33550	1	86	2025	2025	NUM
bracis-33550	1	87	pp	pp	ADP
bracis-33550	1	88	33–47	33–47	NUM
bracis-33550	1	89	cite	cite	VERB
bracis-33550	1	90	this	this	DET
bracis-33550	1	91	conference	conference	NOUN
bracis-33550	1	92	paper	paper	NOUN
bracis-33550	1	93	access	access	NOUN
bracis-33550	1	94	provided	provide	VERB
bracis-33550	1	95	by	by	ADP
bracis-33550	1	96	university	university	PROPN
bracis-33550	1	97	of	of	ADP
bracis-33550	1	98	notre	notre	PROPN
bracis-33550	1	99	dame	dame	PROPN
bracis-33550	1	100	hesburgh	hesburgh	PROPN
bracis-33550	1	101	library	library	PROPN
bracis-33550	1	102	download	download	PROPN
bracis-33550	1	103	book	book	NOUN
bracis-33550	1	104	pdf	pdf	PROPN
bracis-33550	1	105	download	download	NOUN
bracis-33550	1	106	book	book	NOUN
bracis-33550	1	107	epub	epub	PROPN
bracis-33550	1	108	intelligent	intelligent	ADJ
bracis-33550	1	109	systems	system	NOUN
bracis-33550	1	110	(	(	PUNCT
bracis-33550	1	111	bracis	bracis	NOUN
bracis-33550	1	112	2024	2024	NUM
bracis-33550	1	113	)	)	PUNCT
bracis-33550	1	114	a	a	DET
bracis-33550	1	115	large	large	ADJ
bracis-33550	1	116	dataset	dataset	NOUN
bracis-33550	1	117	of	of	ADP
bracis-33550	1	118	 	 	SPACE
bracis-33550	1	119	spontaneous	spontaneous	ADJ
bracis-33550	1	120	speech	speech	NOUN
bracis-33550	1	121	with	with	ADP
bracis-33550	1	122	 	 	SPACE
bracis-33550	1	123	the	the	DET
bracis-33550	1	124	 	 	SPACE
bracis-33550	1	125	accent	accent	NOUN
bracis-33550	1	126	spoken	speak	VERB
bracis-33550	1	127	in	in	ADP
bracis-33550	1	128	 	 	SPACE
bracis-33550	1	129	são	são	PROPN
bracis-33550	1	130	paulo	paulo	PROPN
bracis-33550	1	131	for	for	ADP
bracis-33550	1	132	 	 	SPACE
bracis-33550	1	133	automatic	automatic	ADJ
bracis-33550	1	134	speech	speech	NOUN
bracis-33550	1	135	recognition	recognition	NOUN
bracis-33550	1	136	evaluation	evaluation	NOUN
bracis-33550	1	137	download	download	NOUN
bracis-33550	1	138	book	book	NOUN
bracis-33550	1	139	pdf	pdf	PROPN
bracis-33550	1	140	download	download	NOUN
bracis-33550	1	141	book	book	PROPN
bracis-33550	1	142	epub	epub	PROPN
bracis-33550	1	143	rodrigo	rodrigo	PROPN
bracis-33550	1	144	lima	lima	PROPN
bracis-33550	1	145	  	  	SPACE
bracis-33550	1	146	orcid	orcid	NOUN
bracis-33550	1	147	:	:	PUNCT
bracis-33550	1	148	orcid.org/0009-0009-4344-11099	orcid.org/0009-0009-4344-11099	NOUN
bracis-33550	1	149	,	,	PUNCT
bracis-33550	1	150	sidney	sidney	PROPN
bracis-33550	1	151	e.	e.	PROPN
bracis-33550	1	152	leal	leal	PROPN
bracis-33550	1	153	  	  	SPACE
bracis-33550	1	154	orcid	orcid	NOUN
bracis-33550	1	155	:	:	PUNCT
bracis-33550	1	156	orcid.org/0000-0002-8817-20639	orcid.org/0000-0002-8817-20639	PROPN
bracis-33550	1	157	,	,	PUNCT
bracis-33550	1	158	arnaldo	arnaldo	PROPN
bracis-33550	1	159	candido	candido	PROPN
bracis-33550	1	160	junior	junior	PROPN
bracis-33550	1	161	  	  	SPACE
bracis-33550	1	162	orcid	orcid	PROPN
bracis-33550	1	163	:	:	PUNCT
bracis-33550	1	164	orcid.org/0000-0002-5647-089110	orcid.org/0000-0002-5647-089110	PROPN
bracis-33550	1	165	&	&	CCONJ
bracis-33550	1	166	…	…	PUNCT
bracis-33550	1	167	sandra	sandra	PROPN
bracis-33550	1	168	m.	m.	PROPN
bracis-33550	1	169	aluísio	aluísio	PROPN
bracis-33550	1	170	  	  	SPACE
bracis-33550	1	171	orcid	orcid	NOUN
bracis-33550	1	172	:	:	PUNCT
bracis-33550	1	173	orcid.org/0000-0001-5108-26309	orcid.org/0000-0001-5108-26309	NOUN
bracis-33550	1	174	  	  	SPACE
bracis-33550	1	175	show	show	VERB
bracis-33550	1	176	authors	author	NOUN
bracis-33550	1	177	part	part	NOUN
bracis-33550	1	178	of	of	ADP
bracis-33550	1	179	the	the	DET
bracis-33550	1	180	book	book	NOUN
bracis-33550	1	181	series	series	NOUN
bracis-33550	1	182	:	:	PUNCT
bracis-33550	1	183	lecture	lecture	NOUN
bracis-33550	1	184	notes	note	NOUN
bracis-33550	1	185	in	in	ADP
bracis-33550	1	186	computer	computer	NOUN
bracis-33550	1	187	science	science	NOUN
bracis-33550	1	188	(	(	PUNCT
bracis-33550	1	189	(	(	PUNCT
bracis-33550	1	190	lnai	lnai	ADJ
bracis-33550	1	191	,	,	PUNCT
bracis-33550	1	192	volume	volume	NOUN
bracis-33550	1	193	15412	15412	NUM
bracis-33550	1	194	)	)	PUNCT
bracis-33550	1	195	)	)	PUNCT
bracis-33550	1	196	included	include	VERB
bracis-33550	1	197	in	in	ADP
bracis-33550	1	198	the	the	DET
bracis-33550	1	199	following	follow	VERB
bracis-33550	1	200	conference	conference	NOUN
bracis-33550	1	201	series	series	NOUN
bracis-33550	1	202	:	:	PUNCT
bracis-33550	1	203	brazilian	brazilian	ADJ
bracis-33550	1	204	conference	conference	NOUN
bracis-33550	1	205	on	on	ADP
bracis-33550	1	206	intelligent	intelligent	ADJ
bracis-33550	1	207	systems	system	NOUN
bracis-33550	1	208	514	514	NUM
bracis-33550	1	209	accesses	access	NOUN
bracis-33550	1	210	abstract	abstract	ADV
bracis-33550	1	211	we	we	PRON
bracis-33550	1	212	present	present	VERB
bracis-33550	1	213	a	a	DET
bracis-33550	1	214	freely	freely	ADV
bracis-33550	1	215	available	available	ADJ
bracis-33550	1	216	spontaneous	spontaneous	ADJ
bracis-33550	1	217	speech	speech	NOUN
bracis-33550	1	218	corpus	corpus	NOUN
bracis-33550	1	219	for	for	ADP
bracis-33550	1	220	the	the	DET
bracis-33550	1	221	brazilian	brazilian	ADJ
bracis-33550	1	222	portuguese	portuguese	ADJ
bracis-33550	1	223	language	language	NOUN
bracis-33550	1	224	and	and	CCONJ
bracis-33550	1	225	report	report	VERB
bracis-33550	1	226	preliminary	preliminary	ADJ
bracis-33550	1	227	automatic	automatic	ADJ
bracis-33550	1	228	speech	speech	NOUN
bracis-33550	1	229	recognition	recognition	NOUN
bracis-33550	1	230	(	(	PUNCT
bracis-33550	1	231	asr	asr	NOUN
bracis-33550	1	232	)	)	PUNCT
bracis-33550	1	233	results	result	NOUN
bracis-33550	1	234	,	,	PUNCT
bracis-33550	1	235	using	use	VERB
bracis-33550	1	236	both	both	CCONJ
bracis-33550	1	237	the	the	DET
bracis-33550	1	238	wav2vec2	wav2vec2	PROPN
bracis-33550	1	239	-	-	PUNCT
bracis-33550	1	240	xlsr-53	xlsr-53	PROPN
bracis-33550	1	241	and	and	CCONJ
bracis-33550	1	242	distil	distil	NOUN
bracis-33550	1	243	-	-	PUNCT
bracis-33550	1	244	whisper	whisper	NOUN
bracis-33550	1	245	models	model	NOUN
bracis-33550	1	246	fine	fine	ADV
bracis-33550	1	247	-	-	PUNCT
bracis-33550	1	248	tuned	tune	VERB
bracis-33550	1	249	and	and	CCONJ
bracis-33550	1	250	trained	train	VERB
bracis-33550	1	251	on	on	ADP
bracis-33550	1	252	our	our	PRON
bracis-33550	1	253	corpus	corpus	NOUN
bracis-33550	1	254	.	.	PUNCT
bracis-33550	2	1	the	the	DET
bracis-33550	2	2	nurc	nurc	ADJ
bracis-33550	2	3	-	-	PUNCT
bracis-33550	2	4	sp	sp	NOUN
bracis-33550	2	5	audio	audio	NOUN
bracis-33550	2	6	corpus	corpus	NOUN
bracis-33550	2	7	comprises	comprise	VERB
bracis-33550	2	8	401	401	NUM
bracis-33550	2	9	different	different	ADJ
bracis-33550	2	10	speakers	speaker	NOUN
bracis-33550	2	11	(	(	PUNCT
bracis-33550	2	12	204	204	NUM
bracis-33550	2	13	females	female	NOUN
bracis-33550	2	14	,	,	PUNCT
bracis-33550	2	15	197	197	NUM
bracis-33550	2	16	males	male	NOUN
bracis-33550	2	17	)	)	PUNCT
bracis-33550	2	18	with	with	ADP
bracis-33550	2	19	a	a	DET
bracis-33550	2	20	total	total	NOUN
bracis-33550	2	21	of	of	ADP
bracis-33550	2	22	239.30	239.30	NUM
bracis-33550	2	23	 	 	SPACE
bracis-33550	2	24	h	h	NOUN
bracis-33550	2	25	of	of	ADP
bracis-33550	2	26	transcribed	transcribe	VERB
bracis-33550	2	27	audio	audio	ADJ
bracis-33550	2	28	recordings	recording	NOUN
bracis-33550	2	29	.	.	PUNCT
bracis-33550	3	1	to	to	ADP
bracis-33550	3	2	the	the	DET
bracis-33550	3	3	best	good	ADJ
bracis-33550	3	4	of	of	ADP
bracis-33550	3	5	our	our	PRON
bracis-33550	3	6	knowledge	knowledge	NOUN
bracis-33550	3	7	,	,	PUNCT
bracis-33550	3	8	this	this	PRON
bracis-33550	3	9	is	be	AUX
bracis-33550	3	10	the	the	DET
bracis-33550	3	11	first	first	ADJ
bracis-33550	3	12	large	large	ADJ
bracis-33550	3	13	paulistano	paulistano	NOUN
bracis-33550	3	14	accented	accent	VERB
bracis-33550	3	15	spontaneous	spontaneous	ADJ
bracis-33550	3	16	speech	speech	NOUN
bracis-33550	3	17	corpus	corpus	NOUN
bracis-33550	3	18	dedicated	dedicate	VERB
bracis-33550	3	19	to	to	ADP
bracis-33550	3	20	the	the	DET
bracis-33550	3	21	asr	asr	PROPN
bracis-33550	3	22	task	task	NOUN
bracis-33550	3	23	in	in	ADP
bracis-33550	3	24	portuguese	portuguese	PROPN
bracis-33550	3	25	.	.	PUNCT
bracis-33550	4	1	we	we	PRON
bracis-33550	4	2	first	first	ADV
bracis-33550	4	3	present	present	VERB
bracis-33550	4	4	the	the	DET
bracis-33550	4	5	design	design	NOUN
bracis-33550	4	6	and	and	CCONJ
bracis-33550	4	7	development	development	NOUN
bracis-33550	4	8	procedures	procedure	NOUN
bracis-33550	4	9	of	of	ADP
bracis-33550	4	10	the	the	DET
bracis-33550	4	11	nurc	nurc	ADJ
bracis-33550	4	12	-	-	PUNCT
bracis-33550	4	13	sp	sp	NOUN
bracis-33550	4	14	audio	audio	NOUN
bracis-33550	4	15	corpus	corpus	NOUN
bracis-33550	4	16	,	,	PUNCT
bracis-33550	4	17	and	and	CCONJ
bracis-33550	4	18	then	then	ADV
bracis-33550	4	19	describe	describe	VERB
bracis-33550	4	20	four	four	NUM
bracis-33550	4	21	asr	asr	NOUN
bracis-33550	4	22	experiments	experiment	NOUN
bracis-33550	4	23	in	in	ADP
bracis-33550	4	24	detail	detail	NOUN
bracis-33550	4	25	.	.	PUNCT
bracis-33550	5	1	the	the	DET
bracis-33550	5	2	experiments	experiment	NOUN
bracis-33550	5	3	demonstrated	demonstrate	VERB
bracis-33550	5	4	promising	promising	ADJ
bracis-33550	5	5	results	result	NOUN
bracis-33550	5	6	for	for	ADP
bracis-33550	5	7	the	the	DET
bracis-33550	5	8	applicability	applicability	NOUN
bracis-33550	5	9	of	of	ADP
bracis-33550	5	10	the	the	DET
bracis-33550	5	11	corpus	corpus	NOUN
bracis-33550	5	12	for	for	ADP
bracis-33550	5	13	asr	asr	PROPN
bracis-33550	5	14	.	.	PUNCT
bracis-33550	6	1	specifically	specifically	ADV
bracis-33550	6	2	,	,	PUNCT
bracis-33550	6	3	we	we	PRON
bracis-33550	6	4	fine	fine	ADV
bracis-33550	6	5	-	-	PUNCT
bracis-33550	6	6	tuned	tune	VERB
bracis-33550	6	7	two	two	NUM
bracis-33550	6	8	versions	version	NOUN
bracis-33550	6	9	of	of	ADP
bracis-33550	6	10	wav2vec2	wav2vec2	PROPN
bracis-33550	6	11	-	-	PUNCT
bracis-33550	6	12	xlsr-53	xlsr-53	PROPN
bracis-33550	6	13	model	model	NOUN
bracis-33550	6	14	,	,	PUNCT
bracis-33550	6	15	trained	train	VERB
bracis-33550	6	16	a	a	DET
bracis-33550	6	17	distil	distil	NOUN
bracis-33550	6	18	-	-	PUNCT
bracis-33550	6	19	whisper	whisper	NOUN
bracis-33550	6	20	model	model	NOUN
bracis-33550	6	21	using	use	VERB
bracis-33550	6	22	our	our	PRON
bracis-33550	6	23	dataset	dataset	NOUN
bracis-33550	6	24	with	with	ADP
bracis-33550	6	25	labels	label	NOUN
bracis-33550	6	26	determined	determine	VERB
bracis-33550	6	27	by	by	ADP
bracis-33550	6	28	whisper	whisper	NOUN
bracis-33550	6	29	large	large	ADJ
bracis-33550	6	30	-	-	PUNCT
bracis-33550	6	31	v3	v3	NOUN
bracis-33550	6	32	model	model	NOUN
bracis-33550	6	33	,	,	PUNCT
bracis-33550	6	34	and	and	CCONJ
bracis-33550	6	35	fine	fine	ADV
bracis-33550	6	36	-	-	PUNCT
bracis-33550	6	37	tuned	tune	VERB
bracis-33550	6	38	this	this	DET
bracis-33550	6	39	distil	distil	NOUN
bracis-33550	6	40	-	-	PUNCT
bracis-33550	6	41	whisper	whisper	NOUN
bracis-33550	6	42	model	model	NOUN
bracis-33550	6	43	with	with	ADP
bracis-33550	6	44	our	our	PRON
bracis-33550	6	45	corpus	corpus	NOUN
bracis-33550	6	46	.	.	PUNCT
bracis-33550	7	1	our	our	PRON
bracis-33550	7	2	best	good	ADJ
bracis-33550	7	3	results	result	NOUN
bracis-33550	7	4	were	be	AUX
bracis-33550	7	5	the	the	DET
bracis-33550	7	6	distil	distil	NOUN
bracis-33550	7	7	-	-	PUNCT
bracis-33550	7	8	whisper	whisper	NOUN
bracis-33550	7	9	fine	fine	ADV
bracis-33550	7	10	-	-	PUNCT
bracis-33550	7	11	tuned	tune	VERB
bracis-33550	7	12	over	over	ADP
bracis-33550	7	13	nurc	nurc	ADV
bracis-33550	7	14	-	-	PUNCT
bracis-33550	7	15	sp	sp	NOUN
bracis-33550	7	16	audio	audio	NOUN
bracis-33550	7	17	corpus	corpus	NOUN
bracis-33550	7	18	with	with	ADP
bracis-33550	7	19	a	a	DET
bracis-33550	7	20	wer	wer	NOUN
bracis-33550	7	21	of	of	ADP
bracis-33550	7	22	24.22	24.22	NUM
bracis-33550	7	23	%	%	NOUN
bracis-33550	7	24	followed	follow	VERB
bracis-33550	7	25	by	by	ADP
bracis-33550	7	26	a	a	DET
bracis-33550	7	27	fine	fine	ADV
bracis-33550	7	28	-	-	PUNCT
bracis-33550	7	29	tuned	tune	VERB
bracis-33550	7	30	versions	version	NOUN
bracis-33550	7	31	of	of	ADP
bracis-33550	7	32	wav2vec2	wav2vec2	PROPN
bracis-33550	7	33	-	-	PUNCT
bracis-33550	7	34	xlsr-53	xlsr-53	PROPN
bracis-33550	7	35	model	model	NOUN
bracis-33550	7	36	with	with	ADP
bracis-33550	7	37	a	a	DET
bracis-33550	7	38	wer	wer	NOUN
bracis-33550	7	39	of	of	ADP
bracis-33550	7	40	33.73	33.73	NUM
bracis-33550	7	41	%	%	NOUN
bracis-33550	7	42	,	,	PUNCT
bracis-33550	7	43	that	that	PRON
bracis-33550	7	44	is	be	AUX
bracis-33550	7	45	almost	almost	ADV
bracis-33550	7	46	10	10	NUM
bracis-33550	7	47	%	%	NOUN
bracis-33550	7	48	point	point	NOUN
bracis-33550	7	49	worse	bad	ADJ
bracis-33550	7	50	than	than	ADP
bracis-33550	7	51	distil	distil	NOUN
bracis-33550	7	52	-	-	PUNCT
bracis-33550	7	53	whisper	whisper	NOUN
bracis-33550	7	54	’s	’s	NOUN
bracis-33550	7	55	.	.	PUNCT
bracis-33550	8	1	to	to	PART
bracis-33550	8	2	enable	enable	VERB
bracis-33550	8	3	experiment	experiment	NOUN
bracis-33550	8	4	reproducibility	reproducibility	NOUN
bracis-33550	8	5	,	,	PUNCT
bracis-33550	8	6	we	we	PRON
bracis-33550	8	7	share	share	VERB
bracis-33550	8	8	the	the	DET
bracis-33550	8	9	nurc	nurc	ADJ
bracis-33550	8	10	-	-	PUNCT
bracis-33550	8	11	sp	sp	NOUN
bracis-33550	8	12	audio	audio	NOUN
bracis-33550	8	13	corpus	corpus	NOUN
bracis-33550	8	14	dataset	dataset	NOUN
bracis-33550	8	15	,	,	PUNCT
bracis-33550	8	16	pre	pre	ADJ
bracis-33550	8	17	-	-	ADJ
bracis-33550	8	18	trained	train	VERB
bracis-33550	8	19	models	model	NOUN
bracis-33550	8	20	,	,	PUNCT
bracis-33550	8	21	and	and	CCONJ
bracis-33550	8	22	training	training	NOUN
bracis-33550	8	23	recipes	recipe	NOUN
bracis-33550	8	24	in	in	ADP
bracis-33550	8	25	hugging	hugging	ADJ
bracis-33550	8	26	-	-	PUNCT
bracis-33550	8	27	face	face	NOUN
bracis-33550	8	28	and	and	CCONJ
bracis-33550	8	29	github	github	PROPN
bracis-33550	8	30	repositories	repository	NOUN
bracis-33550	8	31	.	.	PUNCT
bracis-33550	9	1	access	access	NOUN
bracis-33550	9	2	provided	provide	VERB
bracis-33550	9	3	by	by	ADP
bracis-33550	9	4	university	university	PROPN
bracis-33550	9	5	of	of	ADP
bracis-33550	9	6	notre	notre	PROPN
bracis-33550	9	7	dame	dame	PROPN
bracis-33550	9	8	hesburgh	hesburgh	PROPN
bracis-33550	9	9	library	library	PROPN
bracis-33550	9	10	.	.	PUNCT
bracis-33550	10	1	download	download	PROPN
bracis-33550	10	2	conference	conference	NOUN
bracis-33550	10	3	paper	paper	NOUN
bracis-33550	10	4	pdf	pdf	NOUN
bracis-33550	10	5	similar	similar	ADJ
bracis-33550	10	6	content	content	NOUN
bracis-33550	10	7	being	be	AUX
bracis-33550	10	8	viewed	view	VERB
bracis-33550	10	9	by	by	ADP
bracis-33550	10	10	others	other	NOUN
bracis-33550	10	11	incorporation	incorporation	NOUN
bracis-33550	10	12	of	of	ADP
bracis-33550	10	13	iterative	iterative	NOUN
bracis-33550	10	14	self	self	NOUN
bracis-33550	10	15	-	-	PUNCT
bracis-33550	10	16	supervised	supervise	VERB
bracis-33550	10	17	pre	pre	NOUN
bracis-33550	10	18	-	-	NOUN
bracis-33550	10	19	training	training	NOUN
bracis-33550	10	20	in	in	ADP
bracis-33550	10	21	the	the	DET
bracis-33550	10	22	creation	creation	NOUN
bracis-33550	10	23	of	of	ADP
bracis-33550	10	24	the	the	DET
bracis-33550	10	25	asr	asr	NOUN
bracis-33550	10	26	system	system	NOUN
bracis-33550	10	27	for	for	ADP
bracis-33550	10	28	the	the	DET
bracis-33550	10	29	tatar	tatar	NOUN
bracis-33550	10	30	language	language	NOUN
bracis-33550	10	31	chapter	chapter	NOUN
bracis-33550	10	32	©	©	PROPN
bracis-33550	10	33	2021	2021	NUM
bracis-33550	10	34	brazilian	brazilian	ADJ
bracis-33550	10	35	portuguese	portuguese	ADJ
bracis-33550	10	36	speech	speech	NOUN
bracis-33550	10	37	recognition	recognition	NOUN
bracis-33550	10	38	using	use	VERB
bracis-33550	10	39	wav2vec	wav2vec	PROPN
bracis-33550	10	40	2.0	2.0	NUM
bracis-33550	10	41	chapter	chapter	NOUN
bracis-33550	10	42	©	©	PROPN
bracis-33550	10	43	2022	2022	NUM
bracis-33550	10	44	fine	fine	ADV
bracis-33550	10	45	-	-	PUNCT
bracis-33550	10	46	tuning	tune	VERB
bracis-33550	10	47	xlsr	xlsr	PROPN
bracis-33550	10	48	-	-	PUNCT
bracis-33550	10	49	wav2vec2	wav2vec2	PROPN
bracis-33550	10	50	for	for	ADP
bracis-33550	10	51	sanskrit	sanskrit	PROPN
bracis-33550	10	52	asr	asr	PROPN
bracis-33550	10	53	with	with	ADP
bracis-33550	10	54	transformer	transformer	NOUN
bracis-33550	10	55	chapter	chapter	NOUN
bracis-33550	10	56	©	©	PROPN
bracis-33550	10	57	2026	2026	NUM
bracis-33550	10	58	1	1	NUM
bracis-33550	10	59	introduction	introduction	NOUN
bracis-33550	10	60	public	public	ADJ
bracis-33550	10	61	or	or	CCONJ
bracis-33550	10	62	open	open	ADJ
bracis-33550	10	63	datasets	dataset	NOUN
bracis-33550	10	64	for	for	ADP
bracis-33550	10	65	training	training	NOUN
bracis-33550	10	66	and	and	CCONJ
bracis-33550	10	67	evaluating	evaluate	VERB
bracis-33550	10	68	automatic	automatic	ADJ
bracis-33550	10	69	speech	speech	NOUN
bracis-33550	10	70	recognizers	recognizer	NOUN
bracis-33550	10	71	(	(	PUNCT
bracis-33550	10	72	asr	asr	NOUN
bracis-33550	10	73	)	)	PUNCT
bracis-33550	10	74	in	in	ADP
bracis-33550	10	75	brazilian	brazilian	ADJ
bracis-33550	10	76	portuguese	portuguese	PROPN
bracis-33550	10	77	(	(	PUNCT
bracis-33550	10	78	bp	bp	PROPN
bracis-33550	10	79	)	)	PUNCT
bracis-33550	10	80	have	have	AUX
bracis-33550	10	81	increased	increase	VERB
bracis-33550	10	82	in	in	ADP
bracis-33550	10	83	number	number	NOUN
bracis-33550	10	84	and	and	CCONJ
bracis-33550	10	85	hours	hour	NOUN
bracis-33550	10	86	since	since	SCONJ
bracis-33550	10	87	mid-2020	mid-2020	NOUN
bracis-33550	10	88	,	,	PUNCT
bracis-33550	10	89	when	when	SCONJ
bracis-33550	10	90	there	there	PRON
bracis-33550	10	91	were	be	VERB
bracis-33550	10	92	approximately	approximately	ADV
bracis-33550	10	93	60	60	NUM
bracis-33550	10	94	 	 	SPACE
bracis-33550	10	95	h	h	NOUN
bracis-33550	10	96	available	available	ADJ
bracis-33550	10	97	,	,	PUNCT
bracis-33550	10	98	divided	divide	VERB
bracis-33550	10	99	into	into	ADP
bracis-33550	10	100	four	four	NUM
bracis-33550	10	101	small	small	ADJ
bracis-33550	10	102	datasets	dataset	NOUN
bracis-33550	10	103	of	of	ADP
bracis-33550	10	104	read	read	NOUN
bracis-33550	10	105	speech	speech	NOUN
bracis-33550	10	106	.	.	PUNCT
bracis-33550	11	1	in	in	ADP
bracis-33550	11	2	2024	2024	NUM
bracis-33550	11	3	,	,	PUNCT
bracis-33550	11	4	more	more	ADJ
bracis-33550	11	5	than	than	ADP
bracis-33550	11	6	8	8	NUM
bracis-33550	11	7	thousand	thousand	NUM
bracis-33550	11	8	hours	hour	NOUN
bracis-33550	11	9	are	be	AUX
bracis-33550	11	10	available	available	ADJ
bracis-33550	11	11	for	for	ADP
bracis-33550	11	12	training	train	VERB
bracis-33550	11	13	asr	asr	NOUN
bracis-33550	11	14	models	model	NOUN
bracis-33550	11	15	,	,	PUNCT
bracis-33550	11	16	either	either	CCONJ
bracis-33550	11	17	from	from	ADP
bracis-33550	11	18	datasets	dataset	NOUN
bracis-33550	11	19	automatically	automatically	ADV
bracis-33550	11	20	labelled	label	VERB
bracis-33550	11	21	or	or	CCONJ
bracis-33550	11	22	manually	manually	ADV
bracis-33550	11	23	revised	revise	VERB
bracis-33550	11	24	(	(	PUNCT
bracis-33550	11	25	see	see	VERB
bracis-33550	11	26	table	table	NOUN
bracis-33550	11	27	 	 	SPACE
bracis-33550	11	28	1	1	NUM
bracis-33550	11	29	for	for	ADP
bracis-33550	11	30	a	a	DET
bracis-33550	11	31	list	list	NOUN
bracis-33550	11	32	of	of	ADP
bracis-33550	11	33	datasets	dataset	NOUN
bracis-33550	11	34	)	)	PUNCT
bracis-33550	11	35	.	.	PUNCT
bracis-33550	12	1	these	these	DET
bracis-33550	12	2	bp	bp	PROPN
bracis-33550	12	3	resources	resource	NOUN
bracis-33550	12	4	allow	allow	VERB
bracis-33550	12	5	the	the	DET
bracis-33550	12	6	training	training	NOUN
bracis-33550	12	7	of	of	ADP
bracis-33550	12	8	state	state	NOUN
bracis-33550	12	9	-	-	PUNCT
bracis-33550	12	10	of	of	ADP
bracis-33550	12	11	-	-	PUNCT
bracis-33550	12	12	the	the	DET
bracis-33550	12	13	-	-	PUNCT
bracis-33550	12	14	art	art	NOUN
bracis-33550	12	15	asr	asr	NOUN
bracis-33550	12	16	models	model	NOUN
bracis-33550	12	17	,	,	PUNCT
bracis-33550	12	18	such	such	ADJ
bracis-33550	12	19	as	as	ADP
bracis-33550	12	20	wav2vec2	wav2vec2	PROPN
bracis-33550	12	21	-	-	PUNCT
bracis-33550	12	22	xlsr-53	xlsr-53	PROPN
bracis-33550	13	1	[	[	X
bracis-33550	13	2	8	8	NUM
bracis-33550	13	3	]	]	PUNCT
bracis-33550	13	4	and	and	CCONJ
bracis-33550	13	5	distil	distil	NOUN
bracis-33550	13	6	-	-	PUNCT
bracis-33550	13	7	whisper	whisper	NOUN
bracis-33550	14	1	[	[	X
bracis-33550	14	2	11	11	NUM
bracis-33550	14	3	]	]	PUNCT
bracis-33550	14	4	.	.	PUNCT
bracis-33550	15	1	specifically	specifically	ADV
bracis-33550	15	2	,	,	PUNCT
bracis-33550	15	3	both	both	DET
bracis-33550	15	4	models	model	NOUN
bracis-33550	15	5	are	be	AUX
bracis-33550	15	6	suitable	suitable	ADJ
bracis-33550	15	7	for	for	ADP
bracis-33550	15	8	lowor	lowor	ADJ
bracis-33550	15	9	medium	medium	ADJ
bracis-33550	15	10	-	-	PUNCT
bracis-33550	15	11	resource	resource	NOUN
bracis-33550	15	12	languages	language	NOUN
bracis-33550	15	13	,	,	PUNCT
bracis-33550	15	14	providing	provide	VERB
bracis-33550	15	15	consistent	consistent	ADJ
bracis-33550	15	16	results	result	NOUN
bracis-33550	15	17	without	without	ADP
bracis-33550	15	18	the	the	DET
bracis-33550	15	19	need	need	NOUN
bracis-33550	15	20	for	for	ADP
bracis-33550	15	21	training	training	NOUN
bracis-33550	15	22	sets	set	NOUN
bracis-33550	15	23	having	have	VERB
bracis-33550	15	24	tens	ten	NOUN
bracis-33550	15	25	of	of	ADP
bracis-33550	15	26	thousand	thousand	NUM
bracis-33550	15	27	of	of	ADP
bracis-33550	15	28	hours	hour	NOUN
bracis-33550	15	29	.	.	PUNCT
bracis-33550	16	1	wav2vec2	wav2vec2	PROPN
bracis-33550	16	2	-	-	PUNCT
bracis-33550	16	3	xlsr-53	xlsr-53	PROPN
bracis-33550	16	4	is	be	AUX
bracis-33550	16	5	pre	pre	ADJ
bracis-33550	16	6	-	-	VERB
bracis-33550	16	7	trained	train	VERB
bracis-33550	16	8	in	in	ADP
bracis-33550	16	9	a	a	DET
bracis-33550	16	10	large	large	ADJ
bracis-33550	16	11	dataset	dataset	NOUN
bracis-33550	16	12	of	of	ADP
bracis-33550	16	13	unlabeled	unlabeled	ADJ
bracis-33550	16	14	speech	speech	NOUN
bracis-33550	16	15	encompassing	encompass	VERB
bracis-33550	16	16	53	53	NUM
bracis-33550	16	17	languages	language	NOUN
bracis-33550	16	18	,	,	PUNCT
bracis-33550	16	19	including	include	VERB
bracis-33550	16	20	bp	bp	PROPN
bracis-33550	16	21	,	,	PUNCT
bracis-33550	16	22	through	through	ADP
bracis-33550	16	23	auto	auto	NOUN
bracis-33550	16	24	-	-	PUNCT
bracis-33550	16	25	supervised	supervise	VERB
bracis-33550	16	26	learning	learning	NOUN
bracis-33550	16	27	.	.	PUNCT
bracis-33550	17	1	the	the	DET
bracis-33550	17	2	pre	pre	ADJ
bracis-33550	17	3	-	-	ADJ
bracis-33550	17	4	training	training	ADJ
bracis-33550	17	5	process	process	NOUN
bracis-33550	17	6	allows	allow	VERB
bracis-33550	17	7	for	for	ADP
bracis-33550	17	8	efficient	efficient	ADJ
bracis-33550	17	9	fine	fine	NOUN
bracis-33550	17	10	-	-	PUNCT
bracis-33550	17	11	tuning	tuning	NOUN
bracis-33550	17	12	on	on	ADP
bracis-33550	17	13	available	available	ADJ
bracis-33550	17	14	bp	bp	PROPN
bracis-33550	17	15	datasets	dataset	NOUN
bracis-33550	17	16	,	,	PUNCT
bracis-33550	17	17	generating	generate	VERB
bracis-33550	17	18	competitive	competitive	ADJ
bracis-33550	17	19	asr	asr	NOUN
bracis-33550	17	20	results	result	NOUN
bracis-33550	17	21	.	.	PUNCT
bracis-33550	18	1	distil	distil	NOUN
bracis-33550	18	2	-	-	PUNCT
bracis-33550	18	3	whisper	whisper	NOUN
bracis-33550	18	4	is	be	AUX
bracis-33550	18	5	a	a	DET
bracis-33550	18	6	distilled	distil	VERB
bracis-33550	18	7	version	version	NOUN
bracis-33550	18	8	of	of	ADP
bracis-33550	18	9	whisper	whisper	NOUN
bracis-33550	18	10	[	[	X
bracis-33550	18	11	19	19	NUM
bracis-33550	18	12	]	]	PUNCT
bracis-33550	18	13	.	.	PUNCT
bracis-33550	19	1	the	the	DET
bracis-33550	19	2	latter	latter	ADJ
bracis-33550	19	3	,	,	PUNCT
bracis-33550	19	4	trained	train	VERB
bracis-33550	19	5	on	on	ADP
bracis-33550	19	6	680,000	680,000	NUM
bracis-33550	19	7	 	 	SPACE
bracis-33550	19	8	h	h	NOUN
bracis-33550	19	9	of	of	ADP
bracis-33550	19	10	multilingual	multilingual	ADJ
bracis-33550	19	11	data	datum	NOUN
bracis-33550	19	12	in	in	ADP
bracis-33550	19	13	a	a	DET
bracis-33550	19	14	multitask	multitask	ADJ
bracis-33550	19	15	approach	approach	NOUN
bracis-33550	19	16	,	,	PUNCT
bracis-33550	19	17	has	have	AUX
bracis-33550	19	18	shown	show	VERB
bracis-33550	19	19	superior	superior	ADJ
bracis-33550	19	20	performance	performance	NOUN
bracis-33550	19	21	across	across	ADP
bracis-33550	19	22	multiple	multiple	ADJ
bracis-33550	19	23	datasets	dataset	NOUN
bracis-33550	19	24	and	and	CCONJ
bracis-33550	19	25	domains	domain	NOUN
bracis-33550	19	26	when	when	SCONJ
bracis-33550	19	27	compared	compare	VERB
bracis-33550	19	28	to	to	ADP
bracis-33550	19	29	speech	speech	NOUN
bracis-33550	19	30	recognizers	recognizer	NOUN
bracis-33550	19	31	using	use	VERB
bracis-33550	19	32	the	the	DET
bracis-33550	19	33	wav2vec2	wav2vec2	PROPN
bracis-33550	19	34	model	model	NOUN
bracis-33550	19	35	.	.	PUNCT
bracis-33550	20	1	a	a	DET
bracis-33550	20	2	disadvantage	disadvantage	NOUN
bracis-33550	20	3	of	of	ADP
bracis-33550	20	4	the	the	DET
bracis-33550	20	5	various	various	ADJ
bracis-33550	20	6	large	large	ADJ
bracis-33550	20	7	whisper	whisper	NOUN
bracis-33550	20	8	models	model	NOUN
bracis-33550	20	9	is	be	AUX
bracis-33550	20	10	their	their	PRON
bracis-33550	20	11	requirement	requirement	NOUN
bracis-33550	20	12	for	for	ADP
bracis-33550	20	13	a	a	DET
bracis-33550	20	14	large	large	ADJ
bracis-33550	20	15	dataset	dataset	NOUN
bracis-33550	20	16	for	for	ADP
bracis-33550	20	17	increased	increase	VERB
bracis-33550	20	18	performance	performance	NOUN
bracis-33550	20	19	and	and	CCONJ
bracis-33550	20	20	the	the	DET
bracis-33550	20	21	inability	inability	NOUN
bracis-33550	20	22	to	to	PART
bracis-33550	20	23	use	use	VERB
bracis-33550	20	24	them	they	PRON
bracis-33550	20	25	in	in	ADP
bracis-33550	20	26	computationally	computationally	ADV
bracis-33550	20	27	limited	limit	VERB
bracis-33550	20	28	environments	environment	NOUN
bracis-33550	20	29	.	.	PUNCT
bracis-33550	21	1	distil	distil	NOUN
bracis-33550	21	2	-	-	PUNCT
bracis-33550	21	3	whisper	whisper	NOUN
bracis-33550	21	4	,	,	PUNCT
bracis-33550	21	5	on	on	ADP
bracis-33550	21	6	the	the	DET
bracis-33550	21	7	other	other	ADJ
bracis-33550	21	8	hand	hand	NOUN
bracis-33550	21	9	,	,	PUNCT
bracis-33550	21	10	is	be	AUX
bracis-33550	21	11	a	a	DET
bracis-33550	21	12	small	small	ADJ
bracis-33550	21	13	variant	variant	NOUN
bracis-33550	21	14	based	base	VERB
bracis-33550	21	15	on	on	ADP
bracis-33550	21	16	knowledge	knowledge	NOUN
bracis-33550	21	17	distillation	distillation	NOUN
bracis-33550	21	18	focused	focus	VERB
bracis-33550	21	19	on	on	ADP
bracis-33550	21	20	efficiency	efficiency	NOUN
bracis-33550	21	21	which	which	PRON
bracis-33550	21	22	is	be	AUX
bracis-33550	21	23	6	6	NUM
bracis-33550	21	24	times	time	NOUN
bracis-33550	21	25	faster	fast	ADJ
bracis-33550	21	26	,	,	PUNCT
bracis-33550	21	27	49	49	NUM
bracis-33550	21	28	%	%	NOUN
bracis-33550	21	29	smaller	small	ADJ
bracis-33550	21	30	,	,	PUNCT
bracis-33550	21	31	and	and	CCONJ
bracis-33550	21	32	performs	perform	VERB
bracis-33550	21	33	within	within	ADP
bracis-33550	21	34	1	1	NUM
bracis-33550	21	35	%	%	NOUN
bracis-33550	21	36	word	word	NOUN
bracis-33550	21	37	error	error	NOUN
bracis-33550	21	38	rate	rate	NOUN
bracis-33550	21	39	(	(	PUNCT
bracis-33550	21	40	wer	wer	NOUN
bracis-33550	21	41	)	)	PUNCT
bracis-33550	21	42	on	on	ADP
bracis-33550	21	43	out	out	ADV
bracis-33550	21	44	-	-	PUNCT
bracis-33550	21	45	of	of	ADP
bracis-33550	21	46	-	-	PUNCT
bracis-33550	21	47	distribution	distribution	NOUN
bracis-33550	21	48	evaluation	evaluation	NOUN
bracis-33550	21	49	sets	set	NOUN
bracis-33550	21	50	.	.	PUNCT
bracis-33550	22	1	to	to	ADP
bracis-33550	22	2	the	the	DET
bracis-33550	22	3	best	good	ADJ
bracis-33550	22	4	of	of	ADP
bracis-33550	22	5	our	our	PRON
bracis-33550	22	6	knowledge	knowledge	NOUN
bracis-33550	22	7	,	,	PUNCT
bracis-33550	22	8	distil	distil	NOUN
bracis-33550	22	9	-	-	PUNCT
bracis-33550	22	10	whisper	whisper	NOUN
bracis-33550	22	11	has	have	AUX
bracis-33550	22	12	never	never	ADV
bracis-33550	22	13	been	be	AUX
bracis-33550	22	14	evaluated	evaluate	VERB
bracis-33550	22	15	for	for	ADP
bracis-33550	22	16	bp	bp	PROPN
bracis-33550	22	17	datasets	dataset	NOUN
bracis-33550	22	18	.	.	PUNCT
bracis-33550	23	1	regarding	regard	VERB
bracis-33550	23	2	the	the	DET
bracis-33550	23	3	data	datum	NOUN
bracis-33550	23	4	used	use	VERB
bracis-33550	23	5	for	for	ADP
bracis-33550	23	6	training	training	NOUN
bracis-33550	23	7	speech	speech	NOUN
bracis-33550	23	8	recognizers	recognizer	NOUN
bracis-33550	23	9	,	,	PUNCT
bracis-33550	23	10	[	[	X
bracis-33550	23	11	10	10	NUM
bracis-33550	23	12	]	]	PUNCT
bracis-33550	23	13	characterizes	characterize	VERB
bracis-33550	23	14	them	they	PRON
bracis-33550	23	15	in	in	ADP
bracis-33550	23	16	two	two	NUM
bracis-33550	23	17	dimensions	dimension	NOUN
bracis-33550	23	18	:	:	PUNCT
bracis-33550	23	19	(	(	PUNCT
bracis-33550	23	20	i	i	NOUN
bracis-33550	23	21	)	)	PUNCT
bracis-33550	23	22	production	production	NOUN
bracis-33550	23	23	style	style	NOUN
bracis-33550	23	24	,	,	PUNCT
bracis-33550	23	25	which	which	PRON
bracis-33550	23	26	describes	describe	VERB
bracis-33550	23	27	a	a	DET
bracis-33550	23	28	continuum	continuum	NOUN
bracis-33550	23	29	of	of	ADP
bracis-33550	23	30	spontaneity	spontaneity	NOUN
bracis-33550	23	31	,	,	PUNCT
bracis-33550	23	32	coming	come	VERB
bracis-33550	23	33	from	from	ADP
bracis-33550	23	34	one	one	NUM
bracis-33550	23	35	side	side	NOUN
bracis-33550	23	36	,	,	PUNCT
bracis-33550	23	37	which	which	PRON
bracis-33550	23	38	is	be	AUX
bracis-33550	23	39	planned	plan	VERB
bracis-33550	23	40	speech	speech	NOUN
bracis-33550	23	41	,	,	PUNCT
bracis-33550	23	42	to	to	ADP
bracis-33550	23	43	the	the	DET
bracis-33550	23	44	other	other	ADJ
bracis-33550	23	45	,	,	PUNCT
bracis-33550	23	46	which	which	PRON
bracis-33550	23	47	would	would	AUX
bracis-33550	23	48	be	be	AUX
bracis-33550	23	49	unplanned	unplanned	ADJ
bracis-33550	23	50	speech	speech	NOUN
bracis-33550	23	51	;	;	PUNCT
bracis-33550	23	52	(	(	PUNCT
bracis-33550	23	53	ii	ii	NOUN
bracis-33550	23	54	)	)	PUNCT
bracis-33550	23	55	mode	mode	NOUN
bracis-33550	23	56	,	,	PUNCT
bracis-33550	23	57	which	which	PRON
bracis-33550	23	58	characterizes	characterize	VERB
bracis-33550	23	59	the	the	DET
bracis-33550	23	60	elicitation	elicitation	NOUN
bracis-33550	23	61	process	process	NOUN
bracis-33550	23	62	of	of	ADP
bracis-33550	23	63	the	the	DET
bracis-33550	23	64	training	training	NOUN
bracis-33550	23	65	pairs	pair	NOUN
bracis-33550	23	66	,	,	PUNCT
bracis-33550	23	67	in	in	ADP
bracis-33550	23	68	which	which	PRON
bracis-33550	23	69	on	on	ADP
bracis-33550	23	70	the	the	DET
bracis-33550	23	71	one	one	NUM
bracis-33550	23	72	hand	hand	NOUN
bracis-33550	23	73	the	the	DET
bracis-33550	23	74	text	text	NOUN
bracis-33550	23	75	is	be	AUX
bracis-33550	23	76	used	use	VERB
bracis-33550	23	77	as	as	ADP
bracis-33550	23	78	a	a	DET
bracis-33550	23	79	stimulus	stimulus	NOUN
bracis-33550	23	80	for	for	ADP
bracis-33550	23	81	speech	speech	NOUN
bracis-33550	23	82	(	(	PUNCT
bracis-33550	23	83	leading	lead	VERB
bracis-33550	23	84	to	to	ADP
bracis-33550	23	85	a	a	DET
bracis-33550	23	86	read	read	NOUN
bracis-33550	23	87	speech	speech	NOUN
bracis-33550	23	88	)	)	PUNCT
bracis-33550	23	89	and	and	CCONJ
bracis-33550	23	90	on	on	ADP
bracis-33550	23	91	the	the	DET
bracis-33550	23	92	other	other	ADJ
bracis-33550	23	93	hand	hand	NOUN
bracis-33550	23	94	it	it	PRON
bracis-33550	23	95	is	be	AUX
bracis-33550	23	96	the	the	DET
bracis-33550	23	97	speech	speech	NOUN
bracis-33550	23	98	that	that	PRON
bracis-33550	23	99	is	be	AUX
bracis-33550	23	100	transcribed	transcribe	VERB
bracis-33550	23	101	into	into	ADP
bracis-33550	23	102	text	text	NOUN
bracis-33550	23	103	.	.	PUNCT
bracis-33550	24	1	thus	thus	ADV
bracis-33550	24	2	,	,	PUNCT
bracis-33550	24	3	read	read	VERB
bracis-33550	24	4	speech	speech	NOUN
bracis-33550	24	5	is	be	AUX
bracis-33550	24	6	planned	plan	VERB
bracis-33550	24	7	speech	speech	NOUN
bracis-33550	24	8	that	that	PRON
bracis-33550	24	9	comes	come	VERB
bracis-33550	24	10	from	from	ADP
bracis-33550	24	11	a	a	DET
bracis-33550	24	12	recited	recite	VERB
bracis-33550	24	13	text	text	NOUN
bracis-33550	24	14	and	and	CCONJ
bracis-33550	24	15	,	,	PUNCT
bracis-33550	24	16	at	at	ADP
bracis-33550	24	17	the	the	DET
bracis-33550	24	18	other	other	ADJ
bracis-33550	24	19	extreme	extreme	ADJ
bracis-33550	24	20	,	,	PUNCT
bracis-33550	24	21	spontaneous	spontaneous	ADJ
bracis-33550	24	22	/	/	SYM
bracis-33550	24	23	conversational	conversational	ADJ
bracis-33550	24	24	speech	speech	NOUN
bracis-33550	24	25	is	be	AUX
bracis-33550	24	26	characterized	characterize	VERB
bracis-33550	24	27	as	as	ADP
bracis-33550	24	28	speech	speech	NOUN
bracis-33550	24	29	transcribed	transcribe	VERB
bracis-33550	24	30	in	in	ADP
bracis-33550	24	31	an	an	DET
bracis-33550	24	32	unplanned	unplanned	ADJ
bracis-33550	24	33	style	style	NOUN
bracis-33550	24	34	.	.	PUNCT
bracis-33550	25	1	as	as	ADP
bracis-33550	25	2	a	a	DET
bracis-33550	25	3	compromise	compromise	NOUN
bracis-33550	25	4	,	,	PUNCT
bracis-33550	25	5	we	we	PRON
bracis-33550	25	6	have	have	VERB
bracis-33550	25	7	an	an	DET
bracis-33550	25	8	speech	speech	NOUN
bracis-33550	25	9	prepared	prepare	VERB
bracis-33550	25	10	to	to	PART
bracis-33550	25	11	be	be	AUX
bracis-33550	25	12	spoken	speak	VERB
bracis-33550	25	13	later	later	ADV
bracis-33550	25	14	(	(	PUNCT
bracis-33550	25	15	e.g.	e.g.	ADV
bracis-33550	25	16	the	the	DET
bracis-33550	25	17	ted	ted	NOUN
bracis-33550	25	18	talksfootnote	talksfootnote	VERB
bracis-33550	25	19	1	1	NUM
bracis-33550	25	20	,	,	PUNCT
bracis-33550	25	21	in	in	ADP
bracis-33550	25	22	which	which	PRON
bracis-33550	25	23	exponents	exponent	NOUN
bracis-33550	25	24	from	from	ADP
bracis-33550	25	25	all	all	ADV
bracis-33550	25	26	over	over	ADP
bracis-33550	25	27	the	the	DET
bracis-33550	25	28	world	world	NOUN
bracis-33550	25	29	present	present	VERB
bracis-33550	25	30	a	a	DET
bracis-33550	25	31	talk	talk	NOUN
bracis-33550	25	32	in	in	ADP
bracis-33550	25	33	18	18	NUM
bracis-33550	25	34	 	 	SPACE
bracis-33550	25	35	min	min	NOUN
bracis-33550	25	36	or	or	CCONJ
bracis-33550	25	37	less	less	ADJ
bracis-33550	25	38	,	,	PUNCT
bracis-33550	25	39	resulting	result	VERB
bracis-33550	25	40	in	in	ADP
bracis-33550	25	41	the	the	DET
bracis-33550	25	42	need	need	NOUN
bracis-33550	25	43	to	to	PART
bracis-33550	25	44	use	use	VERB
bracis-33550	25	45	a	a	DET
bracis-33550	25	46	planned	plan	VERB
bracis-33550	25	47	text	text	NOUN
bracis-33550	25	48	to	to	PART
bracis-33550	25	49	transmit	transmit	VERB
bracis-33550	25	50	the	the	DET
bracis-33550	25	51	message	message	NOUN
bracis-33550	25	52	in	in	ADP
bracis-33550	25	53	the	the	DET
bracis-33550	25	54	short	short	ADJ
bracis-33550	25	55	time	time	NOUN
bracis-33550	25	56	)	)	PUNCT
bracis-33550	25	57	.	.	PUNCT
bracis-33550	26	1	spontaneous	spontaneous	ADJ
bracis-33550	26	2	speech	speech	NOUN
bracis-33550	26	3	has	have	VERB
bracis-33550	26	4	phenomena	phenomenon	NOUN
bracis-33550	26	5	that	that	PRON
bracis-33550	26	6	make	make	VERB
bracis-33550	26	7	its	its	PRON
bracis-33550	26	8	recognition	recognition	NOUN
bracis-33550	26	9	more	more	ADV
bracis-33550	26	10	complex	complex	ADJ
bracis-33550	26	11	than	than	ADP
bracis-33550	26	12	that	that	PRON
bracis-33550	26	13	of	of	ADP
bracis-33550	26	14	read	read	NOUN
bracis-33550	26	15	or	or	CCONJ
bracis-33550	26	16	prepared	prepared	ADJ
bracis-33550	26	17	speech	speech	NOUN
bracis-33550	26	18	.	.	PUNCT
bracis-33550	27	1	as	as	ADP
bracis-33550	27	2	a	a	DET
bracis-33550	27	3	result	result	NOUN
bracis-33550	27	4	,	,	PUNCT
bracis-33550	27	5	datasets	dataset	NOUN
bracis-33550	27	6	with	with	ADP
bracis-33550	27	7	read	read	NOUN
bracis-33550	27	8	/	/	SYM
bracis-33550	27	9	prepared	prepare	VERB
bracis-33550	27	10	speech	speech	NOUN
bracis-33550	27	11	were	be	AUX
bracis-33550	27	12	first	first	ADV
bracis-33550	27	13	and	and	CCONJ
bracis-33550	27	14	widely	widely	ADV
bracis-33550	27	15	disseminated	disseminate	VERB
bracis-33550	27	16	for	for	ADP
bracis-33550	27	17	training	training	NOUN
bracis-33550	27	18	speech	speech	NOUN
bracis-33550	27	19	recognizers	recognizer	NOUN
bracis-33550	27	20	or	or	CCONJ
bracis-33550	27	21	speech	speech	NOUN
bracis-33550	27	22	synthesizers	synthesizer	NOUN
bracis-33550	27	23	,	,	PUNCT
bracis-33550	27	24	mainly	mainly	ADV
bracis-33550	27	25	for	for	ADP
bracis-33550	27	26	the	the	DET
bracis-33550	27	27	english	english	ADJ
bracis-33550	27	28	language	language	NOUN
bracis-33550	27	29	(	(	PUNCT
bracis-33550	27	30	see	see	VERB
bracis-33550	27	31	[	[	X
bracis-33550	27	32	17	17	NUM
bracis-33550	27	33	,	,	PUNCT
bracis-33550	27	34	24	24	NUM
bracis-33550	27	35	]	]	PUNCT
bracis-33550	27	36	)	)	PUNCT
bracis-33550	27	37	,	,	PUNCT
bracis-33550	27	38	since	since	SCONJ
bracis-33550	27	39	the	the	DET
bracis-33550	27	40	words	word	NOUN
bracis-33550	27	41	in	in	ADP
bracis-33550	27	42	the	the	DET
bracis-33550	27	43	audios	audio	NOUN
bracis-33550	27	44	correspond	correspond	VERB
bracis-33550	27	45	directly	directly	ADV
bracis-33550	27	46	to	to	ADP
bracis-33550	27	47	the	the	DET
bracis-33550	27	48	words	word	NOUN
bracis-33550	27	49	of	of	ADP
bracis-33550	27	50	the	the	DET
bracis-33550	27	51	transcript	transcript	NOUN
bracis-33550	27	52	,	,	PUNCT
bracis-33550	27	53	as	as	SCONJ
bracis-33550	27	54	they	they	PRON
bracis-33550	27	55	were	be	AUX
bracis-33550	27	56	read	read	VERB
bracis-33550	27	57	.	.	PUNCT
bracis-33550	28	1	however	however	ADV
bracis-33550	28	2	,	,	PUNCT
bracis-33550	28	3	more	more	ADV
bracis-33550	28	4	naturally	naturally	ADV
bracis-33550	28	5	occurring	occur	VERB
bracis-33550	28	6	conversational	conversational	ADJ
bracis-33550	28	7	and	and	CCONJ
bracis-33550	28	8	spontaneous	spontaneous	ADJ
bracis-33550	28	9	speech	speech	NOUN
bracis-33550	28	10	contains	contain	VERB
bracis-33550	28	11	many	many	ADJ
bracis-33550	28	12	prosodic	prosodic	ADJ
bracis-33550	28	13	phenomena	phenomenon	NOUN
bracis-33550	28	14	that	that	PRON
bracis-33550	28	15	are	be	AUX
bracis-33550	28	16	not	not	PART
bracis-33550	28	17	present	present	ADJ
bracis-33550	28	18	in	in	ADP
bracis-33550	28	19	read	read	NOUN
bracis-33550	28	20	speech	speech	NOUN
bracis-33550	28	21	,	,	PUNCT
bracis-33550	28	22	such	such	ADJ
bracis-33550	28	23	as	as	ADP
bracis-33550	28	24	higher	high	ADJ
bracis-33550	28	25	degrees	degree	NOUN
bracis-33550	28	26	of	of	ADP
bracis-33550	28	27	segmental	segmental	ADJ
bracis-33550	28	28	reductionfootnote	reductionfootnote	ADJ
bracis-33550	28	29	2	2	NUM
bracis-33550	28	30	and	and	CCONJ
bracis-33550	28	31	more	more	ADJ
bracis-33550	28	32	complex	complex	ADJ
bracis-33550	28	33	forms	form	NOUN
bracis-33550	28	34	of	of	ADP
bracis-33550	28	35	variation	variation	NOUN
bracis-33550	28	36	(	(	PUNCT
bracis-33550	28	37	fillers	filler	NOUN
bracis-33550	28	38	,	,	PUNCT
bracis-33550	28	39	self	self	NOUN
bracis-33550	28	40	-	-	PUNCT
bracis-33550	28	41	corrections	correction	NOUN
bracis-33550	28	42	and	and	CCONJ
bracis-33550	28	43	repetitions	repetition	NOUN
bracis-33550	28	44	)	)	PUNCT
bracis-33550	28	45	affected	affect	VERB
bracis-33550	28	46	by	by	ADP
bracis-33550	28	47	speech	speech	NOUN
bracis-33550	28	48	rate	rate	NOUN
bracis-33550	28	49	.	.	PUNCT
bracis-33550	29	1	thus	thus	ADV
bracis-33550	29	2	,	,	PUNCT
bracis-33550	29	3	the	the	DET
bracis-33550	29	4	various	various	ADJ
bracis-33550	29	5	types	type	NOUN
bracis-33550	29	6	of	of	ADP
bracis-33550	29	7	spontaneous	spontaneous	ADJ
bracis-33550	29	8	speech	speech	NOUN
bracis-33550	29	9	that	that	PRON
bracis-33550	29	10	are	be	AUX
bracis-33550	29	11	common	common	ADJ
bracis-33550	29	12	in	in	ADP
bracis-33550	29	13	everyday	everyday	ADJ
bracis-33550	29	14	life	life	NOUN
bracis-33550	29	15	,	,	PUNCT
bracis-33550	29	16	such	such	ADJ
bracis-33550	29	17	as	as	ADP
bracis-33550	29	18	classes	class	NOUN
bracis-33550	29	19	,	,	PUNCT
bracis-33550	29	20	conversations	conversation	NOUN
bracis-33550	29	21	between	between	ADP
bracis-33550	29	22	two	two	NUM
bracis-33550	29	23	or	or	CCONJ
bracis-33550	29	24	more	more	ADJ
bracis-33550	29	25	people	people	NOUN
bracis-33550	29	26	,	,	PUNCT
bracis-33550	29	27	and	and	CCONJ
bracis-33550	29	28	interviews	interview	NOUN
bracis-33550	29	29	,	,	PUNCT
bracis-33550	29	30	are	be	AUX
bracis-33550	29	31	important	important	ADJ
bracis-33550	29	32	materials	material	NOUN
bracis-33550	29	33	for	for	ADP
bracis-33550	29	34	the	the	DET
bracis-33550	29	35	datasets	dataset	NOUN
bracis-33550	29	36	used	use	VERB
bracis-33550	29	37	to	to	PART
bracis-33550	29	38	train	train	VERB
bracis-33550	29	39	asrs	asrs	ADV
bracis-33550	29	40	,	,	PUNCT
bracis-33550	29	41	as	as	SCONJ
bracis-33550	29	42	they	they	PRON
bracis-33550	29	43	bring	bring	VERB
bracis-33550	29	44	natural	natural	ADJ
bracis-33550	29	45	intonation	intonation	NOUN
bracis-33550	29	46	to	to	ADP
bracis-33550	29	47	questions	question	NOUN
bracis-33550	29	48	,	,	PUNCT
bracis-33550	29	49	statements	statement	NOUN
bracis-33550	29	50	,	,	PUNCT
bracis-33550	29	51	and	and	CCONJ
bracis-33550	29	52	expressions	expression	NOUN
bracis-33550	29	53	of	of	ADP
bracis-33550	29	54	emotions	emotion	NOUN
bracis-33550	29	55	such	such	ADJ
bracis-33550	29	56	as	as	ADP
bracis-33550	29	57	surprise	surprise	NOUN
bracis-33550	29	58	,	,	PUNCT
bracis-33550	29	59	admiration	admiration	NOUN
bracis-33550	29	60	,	,	PUNCT
bracis-33550	29	61	indignation	indignation	NOUN
bracis-33550	29	62	,	,	PUNCT
bracis-33550	29	63	anger	anger	NOUN
bracis-33550	29	64	,	,	PUNCT
bracis-33550	29	65	astonishment	astonishment	ADJ
bracis-33550	29	66	,	,	PUNCT
bracis-33550	29	67	fright	fright	ADJ
bracis-33550	29	68	,	,	PUNCT
bracis-33550	29	69	exaltation	exaltation	NOUN
bracis-33550	29	70	,	,	PUNCT
bracis-33550	29	71	enthusiasm	enthusiasm	NOUN
bracis-33550	29	72	,	,	PUNCT
bracis-33550	29	73	among	among	ADP
bracis-33550	29	74	others	other	NOUN
bracis-33550	29	75	[	[	X
bracis-33550	29	76	4	4	NUM
bracis-33550	29	77	,	,	PUNCT
bracis-33550	29	78	22	22	NUM
bracis-33550	29	79	]	]	PUNCT
bracis-33550	29	80	.	.	PUNCT
bracis-33550	30	1	furthermore	furthermore	ADV
bracis-33550	30	2	,	,	PUNCT
bracis-33550	30	3	they	they	PRON
bracis-33550	30	4	bring	bring	VERB
bracis-33550	30	5	linguistic	linguistic	ADJ
bracis-33550	30	6	phenomena	phenomenon	NOUN
bracis-33550	30	7	such	such	ADJ
bracis-33550	30	8	as	as	ADP
bracis-33550	30	9	filled	fill	VERB
bracis-33550	30	10	pauses	pause	NOUN
bracis-33550	30	11	,	,	PUNCT
bracis-33550	30	12	generally	generally	ADV
bracis-33550	30	13	written	write	VERB
bracis-33550	30	14	as	as	ADP
bracis-33550	30	15	“	"	PUNCT
bracis-33550	30	16	eh	eh	INTJ
bracis-33550	30	17	”	"	PUNCT
bracis-33550	30	18	,	,	PUNCT
bracis-33550	30	19	“	"	PUNCT
bracis-33550	30	20	ah	ah	INTJ
bracis-33550	30	21	”	"	PUNCT
bracis-33550	30	22	,	,	PUNCT
bracis-33550	30	23	“	"	PUNCT
bracis-33550	30	24	ahh	ahh	INTJ
bracis-33550	30	25	”	"	PUNCT
bracis-33550	30	26	,	,	PUNCT
bracis-33550	30	27	“	"	PUNCT
bracis-33550	30	28	mm	mm	INTJ
bracis-33550	30	29	”	"	PUNCT
bracis-33550	30	30	,	,	PUNCT
bracis-33550	30	31	“	"	PUNCT
bracis-33550	30	32	uhn	uhn	PROPN
bracis-33550	30	33	”	"	PUNCT
bracis-33550	30	34	and	and	CCONJ
bracis-33550	30	35	editing	edit	VERB
bracis-33550	30	36	disfluencies	disfluency	NOUN
bracis-33550	30	37	(	(	PUNCT
bracis-33550	30	38	repetitions	repetition	NOUN
bracis-33550	30	39	of	of	ADP
bracis-33550	30	40	words	word	NOUN
bracis-33550	30	41	or	or	CCONJ
bracis-33550	30	42	parts	part	NOUN
bracis-33550	30	43	of	of	ADP
bracis-33550	30	44	words	word	NOUN
bracis-33550	30	45	,	,	PUNCT
bracis-33550	30	46	revisions	revision	NOUN
bracis-33550	30	47	of	of	ADP
bracis-33550	30	48	what	what	PRON
bracis-33550	30	49	is	be	AUX
bracis-33550	30	50	intended	intend	VERB
bracis-33550	30	51	to	to	PART
bracis-33550	30	52	be	be	AUX
bracis-33550	30	53	said	say	VERB
bracis-33550	30	54	,	,	PUNCT
bracis-33550	30	55	with	with	ADP
bracis-33550	30	56	speech	speech	NOUN
bracis-33550	30	57	restarts	restart	NOUN
bracis-33550	30	58	)	)	PUNCT
bracis-33550	31	1	[	[	X
bracis-33550	31	2	16	16	NUM
bracis-33550	31	3	]	]	PUNCT
bracis-33550	31	4	.	.	PUNCT
bracis-33550	32	1	consequently	consequently	ADV
bracis-33550	32	2	,	,	PUNCT
bracis-33550	32	3	speech	speech	NOUN
bracis-33550	32	4	recognizers	recognizer	NOUN
bracis-33550	32	5	are	be	AUX
bracis-33550	32	6	expected	expect	VERB
bracis-33550	32	7	to	to	PART
bracis-33550	32	8	perform	perform	VERB
bracis-33550	32	9	worse	bad	ADJ
bracis-33550	32	10	with	with	ADP
bracis-33550	32	11	spontaneous	spontaneous	ADJ
bracis-33550	32	12	speech	speech	NOUN
bracis-33550	32	13	than	than	ADP
bracis-33550	32	14	when	when	SCONJ
bracis-33550	32	15	applied	apply	VERB
bracis-33550	32	16	to	to	PART
bracis-33550	32	17	read	read	VERB
bracis-33550	32	18	speech	speech	NOUN
bracis-33550	32	19	.	.	PUNCT
bracis-33550	33	1	[	[	X
bracis-33550	33	2	19	19	NUM
bracis-33550	33	3	]	]	PUNCT
bracis-33550	33	4	shows	show	VERB
bracis-33550	33	5	the	the	DET
bracis-33550	33	6	application	application	NOUN
bracis-33550	33	7	of	of	ADP
bracis-33550	33	8	asr	asr	PROPN
bracis-33550	33	9	whisper	whisper	NOUN
bracis-33550	33	10	to	to	ADP
bracis-33550	33	11	14	14	NUM
bracis-33550	33	12	datasets	dataset	NOUN
bracis-33550	33	13	of	of	ADP
bracis-33550	33	14	read	read	NOUN
bracis-33550	33	15	and	and	CCONJ
bracis-33550	33	16	spontaneous	spontaneous	ADJ
bracis-33550	33	17	speech	speech	NOUN
bracis-33550	33	18	in	in	ADP
bracis-33550	33	19	the	the	DET
bracis-33550	33	20	english	english	ADJ
bracis-33550	33	21	language	language	NOUN
bracis-33550	33	22	.	.	PUNCT
bracis-33550	34	1	the	the	DET
bracis-33550	34	2	highest	high	ADJ
bracis-33550	34	3	wer	wer	NOUN
bracis-33550	34	4	(	(	PUNCT
bracis-33550	34	5	word	word	NOUN
bracis-33550	34	6	error	error	NOUN
bracis-33550	34	7	rate	rate	NOUN
bracis-33550	34	8	)	)	PUNCT
bracis-33550	34	9	values	value	NOUN
bracis-33550	34	10	occur	occur	VERB
bracis-33550	34	11	for	for	ADP
bracis-33550	34	12	the	the	DET
bracis-33550	34	13	spontaneous	spontaneous	ADJ
bracis-33550	34	14	speech	speech	NOUN
bracis-33550	34	15	datasets	dataset	NOUN
bracis-33550	34	16	,	,	PUNCT
bracis-33550	34	17	for	for	ADP
bracis-33550	34	18	example	example	NOUN
bracis-33550	34	19	chime6footnote	chime6footnote	VERB
bracis-33550	34	20	3	3	NUM
bracis-33550	34	21	has	have	VERB
bracis-33550	34	22	a	a	DET
bracis-33550	34	23	wer	wer	NOUN
bracis-33550	34	24	of	of	ADP
bracis-33550	34	25	25.5	25.5	NUM
bracis-33550	34	26	and	and	CCONJ
bracis-33550	34	27	ami	ami	VERB
bracis-33550	35	1	sdm1footnote	sdm1footnote	PROPN
bracis-33550	35	2	4	4	NUM
bracis-33550	35	3	has	have	VERB
bracis-33550	35	4	a	a	DET
bracis-33550	35	5	wer	wer	NOUN
bracis-33550	35	6	of	of	ADP
bracis-33550	35	7	36.4	36.4	NUM
bracis-33550	35	8	while	while	SCONJ
bracis-33550	35	9	common	common	ADJ
bracis-33550	35	10	voicefootnote	voicefootnote	ADJ
bracis-33550	35	11	5	5	NUM
bracis-33550	35	12	has	have	VERB
bracis-33550	35	13	a	a	DET
bracis-33550	35	14	wer	wer	NOUN
bracis-33550	35	15	of	of	ADP
bracis-33550	35	16	9.0	9.0	NUM
bracis-33550	35	17	and	and	CCONJ
bracis-33550	35	18	tedliumfootnote	tedliumfootnote	NOUN
bracis-33550	35	19	6	6	NUM
bracis-33550	35	20	has	have	VERB
bracis-33550	35	21	a	a	DET
bracis-33550	35	22	wer	wer	NOUN
bracis-33550	35	23	of	of	ADP
bracis-33550	35	24	4.0	4.0	NUM
bracis-33550	35	25	,	,	PUNCT
bracis-33550	35	26	using	use	VERB
bracis-33550	35	27	the	the	DET
bracis-33550	35	28	large	large	ADJ
bracis-33550	35	29	-	-	PUNCT
bracis-33550	35	30	v2	v2	NOUN
bracis-33550	35	31	model	model	NOUN
bracis-33550	35	32	.	.	PUNCT
bracis-33550	36	1	for	for	ADP
bracis-33550	36	2	brazilian	brazilian	ADJ
bracis-33550	36	3	portuguese	portuguese	NOUN
bracis-33550	36	4	,	,	PUNCT
bracis-33550	36	5	[	[	X
bracis-33550	36	6	15	15	NUM
bracis-33550	36	7	]	]	X
bracis-33550	36	8	reports	report	VERB
bracis-33550	36	9	a	a	DET
bracis-33550	36	10	wer	wer	NOUN
bracis-33550	36	11	of	of	ADP
bracis-33550	36	12	14.50	14.50	NUM
bracis-33550	36	13	%	%	NOUN
bracis-33550	36	14	for	for	ADP
bracis-33550	36	15	the	the	DET
bracis-33550	36	16	whisper	whisper	NOUN
bracis-33550	36	17	large	large	ADJ
bracis-33550	36	18	-	-	PUNCT
bracis-33550	36	19	v2	v2	NOUN
bracis-33550	36	20	model	model	NOUN
bracis-33550	36	21	applied	apply	VERB
bracis-33550	36	22	to	to	ADP
bracis-33550	36	23	a	a	DET
bracis-33550	36	24	portuguese	portuguese	ADJ
bracis-33550	36	25	dataset	dataset	NOUN
bracis-33550	36	26	of	of	ADP
bracis-33550	36	27	approximately	approximately	ADV
bracis-33550	36	28	17	17	NUM
bracis-33550	36	29	 	 	SPACE
bracis-33550	36	30	h	h	NOUN
bracis-33550	36	31	of	of	ADP
bracis-33550	36	32	spontaneous	spontaneous	ADJ
bracis-33550	36	33	speech	speech	NOUN
bracis-33550	36	34	in	in	ADP
bracis-33550	36	35	interviews	interview	NOUN
bracis-33550	36	36	about	about	ADP
bracis-33550	36	37	life	life	NOUN
bracis-33550	36	38	histories	history	NOUN
bracis-33550	36	39	.	.	PUNCT
bracis-33550	37	1	this	this	DET
bracis-33550	37	2	paper	paper	NOUN
bracis-33550	37	3	introduces	introduce	VERB
bracis-33550	37	4	a	a	DET
bracis-33550	37	5	new	new	ADJ
bracis-33550	37	6	corpus	corpus	NOUN
bracis-33550	37	7	of	of	ADP
bracis-33550	37	8	spontaneous	spontaneous	ADJ
bracis-33550	37	9	bp	bp	PROPN
bracis-33550	37	10	speech	speech	NOUN
bracis-33550	37	11	,	,	PUNCT
bracis-33550	37	12	particularly	particularly	ADV
bracis-33550	37	13	of	of	ADP
bracis-33550	37	14	paulistano	paulistano	NOUN
bracis-33550	37	15	accent	accent	ADJ
bracis-33550	37	16	(	(	PUNCT
bracis-33550	37	17	são	são	PROPN
bracis-33550	37	18	paulo	paulo	PROPN
bracis-33550	37	19	city	city	PROPN
bracis-33550	37	20	)	)	PUNCT
bracis-33550	37	21	,	,	PUNCT
bracis-33550	37	22	suitable	suitable	ADJ
bracis-33550	37	23	for	for	ADP
bracis-33550	37	24	training	training	NOUN
bracis-33550	37	25	and	and	CCONJ
bracis-33550	37	26	evaluating	evaluate	VERB
bracis-33550	37	27	asr	asr	NOUN
bracis-33550	37	28	systems	system	NOUN
bracis-33550	37	29	.	.	PUNCT
bracis-33550	38	1	the	the	DET
bracis-33550	38	2	nurc	nurc	ADJ
bracis-33550	38	3	-	-	PUNCT
bracis-33550	38	4	sp	sp	NOUN
bracis-33550	38	5	audio	audio	NOUN
bracis-33550	38	6	corpus	corpus	NOUN
bracis-33550	38	7	is	be	AUX
bracis-33550	38	8	part	part	NOUN
bracis-33550	38	9	of	of	ADP
bracis-33550	38	10	são	são	PROPN
bracis-33550	38	11	paulo	paulo	PROPN
bracis-33550	38	12	division	division	NOUN
bracis-33550	38	13	of	of	ADP
bracis-33550	38	14	the	the	DET
bracis-33550	38	15	nurc	nurc	ADJ
bracis-33550	38	16	project	project	NOUN
bracis-33550	38	17	(	(	PUNCT
bracis-33550	38	18	norma	norma	PROPN
bracis-33550	38	19	urbana	urbana	PROPN
bracis-33550	38	20	linguística	linguística	PROPN
bracis-33550	38	21	culta	culta	PROPN
bracis-33550	38	22	)	)	PUNCT
bracis-33550	38	23	,	,	PUNCT
bracis-33550	38	24	a	a	DET
bracis-33550	38	25	project	project	NOUN
bracis-33550	38	26	designed	design	VERB
bracis-33550	38	27	to	to	PART
bracis-33550	38	28	document	document	VERB
bracis-33550	38	29	and	and	CCONJ
bracis-33550	38	30	study	study	NOUN
bracis-33550	38	31	portuguese	portuguese	PROPN
bracis-33550	38	32	spoken	speak	VERB
bracis-33550	38	33	language	language	NOUN
bracis-33550	38	34	by	by	ADP
bracis-33550	38	35	people	people	NOUN
bracis-33550	38	36	with	with	ADP
bracis-33550	38	37	a	a	DET
bracis-33550	38	38	high	high	ADJ
bracis-33550	38	39	degree	degree	NOUN
bracis-33550	38	40	of	of	ADP
bracis-33550	38	41	formal	formal	ADJ
bracis-33550	38	42	education	education	NOUN
bracis-33550	38	43	in	in	ADP
bracis-33550	38	44	five	five	NUM
bracis-33550	38	45	brazilian	brazilian	ADJ
bracis-33550	38	46	capitals	capital	NOUN
bracis-33550	38	47	(	(	PUNCT
bracis-33550	38	48	see	see	VERB
bracis-33550	38	49	details	detail	NOUN
bracis-33550	38	50	of	of	ADP
bracis-33550	38	51	this	this	DET
bracis-33550	38	52	corpus	corpus	NOUN
bracis-33550	38	53	in	in	ADP
bracis-33550	38	54	sect	sect	NOUN
bracis-33550	38	55	.	.	PUNCT
bracis-33550	38	56	 	 	SPACE
bracis-33550	38	57	3	3	NUM
bracis-33550	38	58	)	)	PUNCT
bracis-33550	38	59	.	.	PUNCT
bracis-33550	39	1	it	it	PRON
bracis-33550	39	2	contains	contain	VERB
bracis-33550	39	3	239.30	239.30	NUM
bracis-33550	39	4	 	 	SPACE
bracis-33550	39	5	h	h	NOUN
bracis-33550	39	6	of	of	ADP
bracis-33550	39	7	audios	audio	NOUN
bracis-33550	39	8	sampled	sample	VERB
bracis-33550	39	9	at	at	ADP
bracis-33550	39	10	16	16	NUM
bracis-33550	39	11	khz	khz	NOUN
bracis-33550	39	12	and	and	CCONJ
bracis-33550	39	13	their	their	PRON
bracis-33550	39	14	respective	respective	ADJ
bracis-33550	39	15	transcriptions	transcription	NOUN
bracis-33550	39	16	,	,	PUNCT
bracis-33550	39	17	totalling	total	VERB
bracis-33550	39	18	170k	170k	NUM
bracis-33550	39	19	segmented	segment	VERB
bracis-33550	39	20	audios	audio	NOUN
bracis-33550	39	21	.	.	PUNCT
bracis-33550	40	1	the	the	DET
bracis-33550	40	2	audios	audio	NOUN
bracis-33550	40	3	were	be	AUX
bracis-33550	40	4	automatically	automatically	ADV
bracis-33550	40	5	transcribed	transcribe	VERB
bracis-33550	40	6	for	for	ADP
bracis-33550	40	7	the	the	DET
bracis-33550	40	8	first	first	ADJ
bracis-33550	40	9	time	time	NOUN
bracis-33550	40	10	and	and	CCONJ
bracis-33550	40	11	manually	manually	ADV
bracis-33550	40	12	revised	revise	VERB
bracis-33550	40	13	aiming	aim	VERB
bracis-33550	40	14	at	at	ADP
bracis-33550	40	15	the	the	DET
bracis-33550	40	16	asr	asr	PROPN
bracis-33550	40	17	task	task	NOUN
bracis-33550	40	18	.	.	PUNCT
bracis-33550	41	1	therefore	therefore	ADV
bracis-33550	41	2	,	,	PUNCT
bracis-33550	41	3	this	this	DET
bracis-33550	41	4	new	new	ADJ
bracis-33550	41	5	corpus	corpus	NOUN
bracis-33550	41	6	adds	add	VERB
bracis-33550	41	7	239	239	NUM
bracis-33550	41	8	 	 	SPACE
bracis-33550	41	9	h	h	NOUN
bracis-33550	41	10	to	to	ADP
bracis-33550	41	11	the	the	DET
bracis-33550	41	12	amount	amount	NOUN
bracis-33550	41	13	available	available	ADJ
bracis-33550	41	14	for	for	ADP
bracis-33550	41	15	bp	bp	PROPN
bracis-33550	41	16	(	(	PUNCT
bracis-33550	41	17	see	see	VERB
bracis-33550	41	18	table	table	NOUN
bracis-33550	41	19	 	 	SPACE
bracis-33550	41	20	1	1	NUM
bracis-33550	41	21	)	)	PUNCT
bracis-33550	41	22	,	,	PUNCT
bracis-33550	41	23	totaling	total	VERB
bracis-33550	41	24	8,598	8,598	NUM
bracis-33550	41	25	 	 	SPACE
bracis-33550	41	26	h	h	NOUN
bracis-33550	41	27	for	for	ADP
bracis-33550	41	28	training	training	NOUN
bracis-33550	41	29	and	and	CCONJ
bracis-33550	41	30	evaluating	evaluate	VERB
bracis-33550	41	31	speech	speech	NOUN
bracis-33550	41	32	recognition	recognition	NOUN
bracis-33550	41	33	systems	system	NOUN
bracis-33550	41	34	.	.	PUNCT
bracis-33550	42	1	in	in	ADP
bracis-33550	42	2	order	order	NOUN
bracis-33550	42	3	to	to	PART
bracis-33550	42	4	compare	compare	VERB
bracis-33550	42	5	the	the	DET
bracis-33550	42	6	quality	quality	NOUN
bracis-33550	42	7	of	of	ADP
bracis-33550	42	8	our	our	PRON
bracis-33550	42	9	corpus	corpus	NOUN
bracis-33550	42	10	,	,	PUNCT
bracis-33550	42	11	four	four	NUM
bracis-33550	42	12	asr	asr	NOUN
bracis-33550	42	13	models	model	NOUN
bracis-33550	42	14	are	be	AUX
bracis-33550	42	15	made	make	VERB
bracis-33550	42	16	available	available	ADJ
bracis-33550	42	17	in	in	ADP
bracis-33550	42	18	this	this	DET
bracis-33550	42	19	work	work	NOUN
bracis-33550	42	20	:	:	PUNCT
bracis-33550	42	21	(	(	PUNCT
bracis-33550	42	22	i	i	NOUN
bracis-33550	42	23	)	)	PUNCT
bracis-33550	42	24	a	a	DET
bracis-33550	42	25	fine	fine	ADV
bracis-33550	42	26	-	-	PUNCT
bracis-33550	42	27	tuned	tune	VERB
bracis-33550	42	28	version	version	NOUN
bracis-33550	42	29	of	of	ADP
bracis-33550	42	30	wav2vec2.0	wav2vec2.0	PROPN
bracis-33550	42	31	xlsr-53	xlsr-53	PROPN
bracis-33550	42	32	with	with	ADP
bracis-33550	42	33	the	the	DET
bracis-33550	42	34	train	train	NOUN
bracis-33550	42	35	and	and	CCONJ
bracis-33550	42	36	validation	validation	NOUN
bracis-33550	42	37	subsets	subset	NOUN
bracis-33550	42	38	of	of	ADP
bracis-33550	42	39	nurc	nurc	ADJ
bracis-33550	42	40	-	-	PUNCT
bracis-33550	42	41	sp	sp	NOUN
bracis-33550	42	42	audio	audio	NOUN
bracis-33550	42	43	corpus	corpus	NOUN
bracis-33550	42	44	,	,	PUNCT
bracis-33550	42	45	(	(	PUNCT
bracis-33550	42	46	ii	ii	NOUN
bracis-33550	42	47	)	)	PUNCT
bracis-33550	42	48	the	the	DET
bracis-33550	42	49	same	same	ADJ
bracis-33550	42	50	as	as	ADP
bracis-33550	42	51	the	the	DET
bracis-33550	42	52	first	first	ADJ
bracis-33550	42	53	,	,	PUNCT
bracis-33550	42	54	but	but	CCONJ
bracis-33550	42	55	using	use	VERB
bracis-33550	42	56	as	as	ADP
bracis-33550	42	57	start	start	NOUN
bracis-33550	42	58	point	point	VERB
bracis-33550	42	59	the	the	DET
bracis-33550	42	60	model	model	NOUN
bracis-33550	42	61	that	that	SCONJ
bracis-33550	43	1	[	[	X
bracis-33550	43	2	6	6	NUM
bracis-33550	43	3	]	]	PUNCT
bracis-33550	43	4	trained	train	VERB
bracis-33550	43	5	for	for	ADP
bracis-33550	43	6	coraa	coraa	NOUN
bracis-33550	43	7	-	-	PUNCT
bracis-33550	43	8	asr	asr	NOUN
bracis-33550	43	9	v1.1footnote	v1.1footnote	NOUN
bracis-33550	43	10	7	7	NUM
bracis-33550	43	11	,	,	PUNCT
bracis-33550	43	12	(	(	PUNCT
bracis-33550	43	13	iii	iii	X
bracis-33550	43	14	)	)	PUNCT
bracis-33550	43	15	a	a	DET
bracis-33550	43	16	distilled	distil	VERB
bracis-33550	43	17	version	version	NOUN
bracis-33550	43	18	of	of	ADP
bracis-33550	43	19	whisper	whisper	NOUN
bracis-33550	43	20	large	large	ADJ
bracis-33550	43	21	-	-	PUNCT
bracis-33550	43	22	v3	v3	NOUN
bracis-33550	43	23	model	model	NOUN
bracis-33550	43	24	[	[	X
bracis-33550	43	25	19	19	NUM
bracis-33550	43	26	]	]	X
bracis-33550	43	27	,	,	PUNCT
bracis-33550	43	28	a	a	DET
bracis-33550	43	29	model	model	NOUN
bracis-33550	43	30	that	that	PRON
bracis-33550	43	31	has	have	VERB
bracis-33550	43	32	support	support	NOUN
bracis-33550	43	33	for	for	ADP
bracis-33550	43	34	the	the	DET
bracis-33550	43	35	portuguese	portuguese	ADJ
bracis-33550	43	36	language	language	NOUN
bracis-33550	43	37	,	,	PUNCT
bracis-33550	43	38	trained	train	VERB
bracis-33550	43	39	using	use	VERB
bracis-33550	43	40	our	our	PRON
bracis-33550	43	41	dataset	dataset	NOUN
bracis-33550	43	42	with	with	ADP
bracis-33550	43	43	labels	label	NOUN
bracis-33550	43	44	determined	determine	VERB
bracis-33550	43	45	by	by	ADP
bracis-33550	43	46	whisper	whisper	NOUN
bracis-33550	43	47	large	large	ADJ
bracis-33550	43	48	-	-	PUNCT
bracis-33550	43	49	v3	v3	NOUN
bracis-33550	43	50	,	,	PUNCT
bracis-33550	43	51	and	and	CCONJ
bracis-33550	43	52	(	(	PUNCT
bracis-33550	43	53	iv	iv	X
bracis-33550	43	54	)	)	PUNCT
bracis-33550	43	55	a	a	DET
bracis-33550	43	56	fine	fine	ADV
bracis-33550	43	57	-	-	PUNCT
bracis-33550	43	58	tuned	tune	VERB
bracis-33550	43	59	version	version	NOUN
bracis-33550	43	60	of	of	ADP
bracis-33550	43	61	the	the	DET
bracis-33550	43	62	third	third	ADJ
bracis-33550	43	63	model	model	NOUN
bracis-33550	43	64	with	with	ADP
bracis-33550	43	65	the	the	DET
bracis-33550	43	66	train	train	NOUN
bracis-33550	43	67	and	and	CCONJ
bracis-33550	43	68	validation	validation	NOUN
bracis-33550	43	69	subsets	subset	NOUN
bracis-33550	43	70	of	of	ADP
bracis-33550	43	71	nurc	nurc	ADJ
bracis-33550	43	72	-	-	PUNCT
bracis-33550	43	73	sp	sp	NOUN
bracis-33550	43	74	audio	audio	NOUN
bracis-33550	43	75	corpus	corpus	NOUN
bracis-33550	43	76	.	.	PUNCT
bracis-33550	44	1	the	the	DET
bracis-33550	44	2	corpus	corpus	NOUN
bracis-33550	44	3	and	and	CCONJ
bracis-33550	44	4	trained	train	VERB
bracis-33550	44	5	models	model	NOUN
bracis-33550	44	6	are	be	AUX
bracis-33550	44	7	publicly	publicly	ADV
bracis-33550	44	8	available	available	ADJ
bracis-33550	44	9	in	in	ADP
bracis-33550	44	10	our	our	PRON
bracis-33550	44	11	github	github	NOUN
bracis-33550	44	12	repositoryfootnote	repositoryfootnote	VERB
bracis-33550	44	13	8	8	NUM
bracis-33550	44	14	under	under	ADP
bracis-33550	44	15	the	the	DET
bracis-33550	44	16	cc	cc	PROPN
bracis-33550	44	17	by	by	ADP
bracis-33550	44	18	-	-	PUNCT
bracis-33550	44	19	nc	nc	VERB
bracis-33550	44	20	-	-	PROPN
bracis-33550	44	21	nd	nd	ADV
bracis-33550	44	22	4.0	4.0	NUM
bracis-33550	44	23	license	license	NOUN
bracis-33550	44	24	.	.	PUNCT
bracis-33550	45	1	the	the	DET
bracis-33550	45	2	main	main	ADJ
bracis-33550	45	3	contributions	contribution	NOUN
bracis-33550	45	4	made	make	VERB
bracis-33550	45	5	in	in	ADP
bracis-33550	45	6	this	this	DET
bracis-33550	45	7	paper	paper	NOUN
bracis-33550	45	8	are	be	AUX
bracis-33550	45	9	summarised	summarise	VERB
bracis-33550	45	10	as	as	SCONJ
bracis-33550	45	11	follows	follow	VERB
bracis-33550	45	12	:	:	PUNCT
bracis-33550	45	13	1	1	X
bracis-33550	45	14	.	.	X
bracis-33550	45	15	a	a	DET
bracis-33550	45	16	large	large	ADJ
bracis-33550	45	17	bp	bp	PROPN
bracis-33550	45	18	corpus	corpus	NOUN
bracis-33550	45	19	of	of	ADP
bracis-33550	45	20	human	human	ADJ
bracis-33550	45	21	validated	validate	VERB
bracis-33550	45	22	audio	audio	ADJ
bracis-33550	45	23	-	-	PUNCT
bracis-33550	45	24	transcription	transcription	NOUN
bracis-33550	45	25	pairs	pair	NOUN
bracis-33550	45	26	containing	contain	VERB
bracis-33550	45	27	239.30	239.30	NUM
bracis-33550	45	28	 	 	SPACE
bracis-33550	45	29	h	h	NOUN
bracis-33550	45	30	of	of	ADP
bracis-33550	45	31	spontaneous	spontaneous	ADJ
bracis-33550	45	32	speech	speech	NOUN
bracis-33550	45	33	.	.	PUNCT
bracis-33550	46	1	2	2	X
bracis-33550	46	2	.	.	X
bracis-33550	46	3	the	the	DET
bracis-33550	46	4	first	first	ADJ
bracis-33550	46	5	corpus	corpus	NOUN
bracis-33550	46	6	,	,	PUNCT
bracis-33550	46	7	according	accord	VERB
bracis-33550	46	8	to	to	ADP
bracis-33550	46	9	our	our	PRON
bracis-33550	46	10	knowledge	knowledge	NOUN
bracis-33550	46	11	,	,	PUNCT
bracis-33550	46	12	tackling	tackle	VERB
bracis-33550	46	13	a	a	DET
bracis-33550	46	14	large	large	ADJ
bracis-33550	46	15	amount	amount	NOUN
bracis-33550	46	16	of	of	ADP
bracis-33550	46	17	accented	accent	VERB
bracis-33550	46	18	paulistano	paulistano	NOUN
bracis-33550	46	19	speech	speech	NOUN
bracis-33550	46	20	for	for	ADP
bracis-33550	46	21	asr	asr	NOUN
bracis-33550	46	22	in	in	ADP
bracis-33550	46	23	bp	bp	PROPN
bracis-33550	46	24	(	(	PUNCT
bracis-33550	46	25	coraa	coraa	PROPN
bracis-33550	46	26	asr	asr	NOUN
bracis-33550	46	27	brings	bring	VERB
bracis-33550	46	28	31.14	31.14	NUM
bracis-33550	46	29	 	 	SPACE
bracis-33550	46	30	h	h	NOUN
bracis-33550	46	31	of	of	ADP
bracis-33550	46	32	speech	speech	NOUN
bracis-33550	46	33	from	from	ADP
bracis-33550	46	34	são	são	PROPN
bracis-33550	46	35	paulo	paulo	PROPN
bracis-33550	46	36	capital	capital	PROPN
bracis-33550	46	37	)	)	PUNCT
bracis-33550	46	38	3	3	NUM
bracis-33550	46	39	.	.	X
bracis-33550	46	40	four	four	NUM
bracis-33550	46	41	asr	asr	NOUN
bracis-33550	46	42	models	model	NOUN
bracis-33550	46	43	,	,	PUNCT
bracis-33550	46	44	publicly	publicly	ADV
bracis-33550	46	45	available	available	ADJ
bracis-33550	46	46	,	,	PUNCT
bracis-33550	46	47	based	base	VERB
bracis-33550	46	48	on	on	ADP
bracis-33550	46	49	the	the	DET
bracis-33550	46	50	presented	present	VERB
bracis-33550	46	51	corpus	corpus	NOUN
bracis-33550	46	52	.	.	PUNCT
bracis-33550	47	1	2	2	NUM
bracis-33550	47	2	related	relate	VERB
bracis-33550	47	3	work	work	NOUN
bracis-33550	47	4	on	on	ADP
bracis-33550	47	5	 	 	SPACE
bracis-33550	47	6	datasets	dataset	NOUN
bracis-33550	47	7	of	of	ADP
bracis-33550	47	8	 	 	SPACE
bracis-33550	47	9	brazilian	brazilian	ADJ
bracis-33550	47	10	portuguese	portuguese	NOUN
bracis-33550	47	11	for	for	ADP
bracis-33550	47	12	 	 	SPACE
bracis-33550	47	13	asr	asr	NOUN
bracis-33550	47	14	table	table	NOUN
bracis-33550	47	15	 	 	SPACE
bracis-33550	47	16	1	1	NUM
bracis-33550	47	17	presents	present	VERB
bracis-33550	47	18	large	large	ADJ
bracis-33550	47	19	corpora	corpora	NOUN
bracis-33550	47	20	for	for	ADP
bracis-33550	47	21	building	build	VERB
bracis-33550	47	22	asr	asr	NOUN
bracis-33550	47	23	systems	system	NOUN
bracis-33550	47	24	focused	focus	VERB
bracis-33550	47	25	on	on	ADP
bracis-33550	47	26	the	the	DET
bracis-33550	47	27	portuguese	portuguese	ADJ
bracis-33550	47	28	language	language	NOUN
bracis-33550	47	29	.	.	PUNCT
bracis-33550	48	1	some	some	DET
bracis-33550	48	2	resources	resource	NOUN
bracis-33550	48	3	are	be	AUX
bracis-33550	48	4	multilingual	multilingual	ADJ
bracis-33550	48	5	,	,	PUNCT
bracis-33550	48	6	however	however	ADV
bracis-33550	48	7	table	table	NOUN
bracis-33550	48	8	 	 	SPACE
bracis-33550	48	9	1	1	NUM
bracis-33550	48	10	specifically	specifically	ADV
bracis-33550	48	11	details	detail	VERB
bracis-33550	48	12	the	the	DET
bracis-33550	48	13	statistics	statistic	NOUN
bracis-33550	48	14	for	for	ADP
bracis-33550	48	15	the	the	DET
bracis-33550	48	16	portuguese	portuguese	ADJ
bracis-33550	48	17	language	language	NOUN
bracis-33550	48	18	.	.	PUNCT
bracis-33550	49	1	among	among	ADP
bracis-33550	49	2	the	the	DET
bracis-33550	49	3	resources	resource	NOUN
bracis-33550	49	4	presented	present	VERB
bracis-33550	49	5	,	,	PUNCT
bracis-33550	49	6	there	there	PRON
bracis-33550	49	7	is	be	VERB
bracis-33550	49	8	a	a	DET
bracis-33550	49	9	slightly	slightly	ADV
bracis-33550	49	10	greater	great	ADJ
bracis-33550	49	11	preponderance	preponderance	NOUN
bracis-33550	49	12	of	of	ADP
bracis-33550	49	13	the	the	DET
bracis-33550	49	14	brazilian	brazilian	ADJ
bracis-33550	49	15	variant	variant	NOUN
bracis-33550	49	16	in	in	ADP
bracis-33550	49	17	the	the	DET
bracis-33550	49	18	existing	exist	VERB
bracis-33550	49	19	resources	resource	NOUN
bracis-33550	49	20	,	,	PUNCT
bracis-33550	49	21	although	although	SCONJ
bracis-33550	49	22	the	the	DET
bracis-33550	49	23	european	european	ADJ
bracis-33550	49	24	portuguese	portuguese	ADJ
bracis-33550	49	25	variant	variant	NOUN
bracis-33550	49	26	is	be	AUX
bracis-33550	49	27	also	also	ADV
bracis-33550	49	28	included	include	VERB
bracis-33550	49	29	in	in	ADP
bracis-33550	49	30	some	some	PRON
bracis-33550	49	31	of	of	ADP
bracis-33550	49	32	them	they	PRON
bracis-33550	49	33	.	.	PUNCT
bracis-33550	50	1	table	table	NOUN
bracis-33550	50	2	1	1	NUM
bracis-33550	50	3	.	.	PUNCT
bracis-33550	51	1	statistics	statistic	NOUN
bracis-33550	51	2	of	of	ADP
bracis-33550	51	3	the	the	DET
bracis-33550	51	4	main	main	ADJ
bracis-33550	51	5	datasets	dataset	NOUN
bracis-33550	51	6	available	available	ADJ
bracis-33550	51	7	for	for	ADP
bracis-33550	51	8	asr	asr	PROPN
bracis-33550	51	9	in	in	ADP
bracis-33550	51	10	portuguese	portuguese	PROPN
bracis-33550	51	11	.	.	PUNCT
bracis-33550	52	1	some	some	DET
bracis-33550	52	2	datasets	dataset	NOUN
bracis-33550	52	3	are	be	AUX
bracis-33550	52	4	multilingual	multilingual	ADJ
bracis-33550	52	5	,	,	PUNCT
bracis-33550	52	6	however	however	ADV
bracis-33550	52	7	,	,	PUNCT
bracis-33550	52	8	the	the	DET
bracis-33550	52	9	numbers	number	NOUN
bracis-33550	52	10	shown	show	VERB
bracis-33550	52	11	are	be	AUX
bracis-33550	52	12	for	for	ADP
bracis-33550	52	13	the	the	DET
bracis-33550	52	14	portuguese	portuguese	ADJ
bracis-33550	52	15	language.full	language.full	ADJ
bracis-33550	52	16	size	size	NOUN
bracis-33550	52	17	table	table	NOUN
bracis-33550	52	18	one	one	NUM
bracis-33550	52	19	of	of	ADP
bracis-33550	52	20	the	the	DET
bracis-33550	52	21	best	well	ADV
bracis-33550	52	22	-	-	PUNCT
bracis-33550	52	23	known	know	VERB
bracis-33550	52	24	projects	project	NOUN
bracis-33550	52	25	dealing	deal	VERB
bracis-33550	52	26	with	with	ADP
bracis-33550	52	27	read	read	NOUN
bracis-33550	52	28	/	/	SYM
bracis-33550	52	29	prepared	prepared	ADJ
bracis-33550	52	30	speech	speech	NOUN
bracis-33550	52	31	in	in	ADP
bracis-33550	52	32	english	english	PROPN
bracis-33550	52	33	is	be	AUX
bracis-33550	52	34	librivoxfootnote	librivoxfootnote	NOUN
bracis-33550	52	35	9	9	NUM
bracis-33550	52	36	,	,	PUNCT
bracis-33550	52	37	which	which	PRON
bracis-33550	52	38	distributes	distribute	VERB
bracis-33550	52	39	public	public	ADJ
bracis-33550	52	40	domain	domain	NOUN
bracis-33550	52	41	books	book	NOUN
bracis-33550	52	42	in	in	ADP
bracis-33550	52	43	audio	audio	ADJ
bracis-33550	52	44	format	format	NOUN
bracis-33550	52	45	.	.	PUNCT
bracis-33550	53	1	these	these	DET
bracis-33550	53	2	audios	audio	NOUN
bracis-33550	53	3	were	be	AUX
bracis-33550	53	4	used	use	VERB
bracis-33550	53	5	in	in	ADP
bracis-33550	53	6	several	several	ADJ
bracis-33550	53	7	projects	project	NOUN
bracis-33550	53	8	to	to	PART
bracis-33550	53	9	create	create	VERB
bracis-33550	53	10	resources	resource	NOUN
bracis-33550	53	11	for	for	ADP
bracis-33550	53	12	processing	processing	NOUN
bracis-33550	53	13	speech	speech	NOUN
bracis-33550	53	14	in	in	ADP
bracis-33550	53	15	english	english	PROPN
bracis-33550	53	16	,	,	PUNCT
bracis-33550	53	17	such	such	ADJ
bracis-33550	53	18	as	as	ADP
bracis-33550	53	19	librispeech	librispeech	PROPN
bracis-33550	53	20	asr	asr	PROPN
bracis-33550	53	21	corpusfootnote	corpusfootnote	NOUN
bracis-33550	53	22	10	10	NUM
bracis-33550	53	23	and	and	CCONJ
bracis-33550	53	24	librittsfootnote	librittsfootnote	VERB
bracis-33550	53	25	11	11	NUM
bracis-33550	53	26	,	,	PUNCT
bracis-33550	53	27	both	both	PRON
bracis-33550	53	28	hosted	host	VERB
bracis-33550	53	29	in	in	ADP
bracis-33550	53	30	the	the	DET
bracis-33550	53	31	open	open	ADJ
bracis-33550	53	32	speech	speech	NOUN
bracis-33550	53	33	and	and	CCONJ
bracis-33550	53	34	language	language	NOUN
bracis-33550	53	35	resources	resource	NOUN
bracis-33550	53	36	repository	repository	NOUN
bracis-33550	53	37	.	.	PUNCT
bracis-33550	54	1	the	the	DET
bracis-33550	54	2	multilingual	multilingual	ADJ
bracis-33550	54	3	librispeech	librispeech	NOUN
bracis-33550	54	4	(	(	PUNCT
bracis-33550	54	5	mls	mls	PROPN
bracis-33550	54	6	)	)	PUNCT
bracis-33550	54	7	corpus	corpus	NOUN
bracis-33550	55	1	[	[	X
bracis-33550	55	2	18	18	NUM
bracis-33550	55	3	]	]	PUNCT
bracis-33550	55	4	is	be	AUX
bracis-33550	55	5	a	a	DET
bracis-33550	55	6	large	large	ADJ
bracis-33550	55	7	multilingual	multilingual	ADJ
bracis-33550	55	8	corpus	corpus	NOUN
bracis-33550	55	9	suitable	suitable	ADJ
bracis-33550	55	10	for	for	ADP
bracis-33550	55	11	speech	speech	NOUN
bracis-33550	55	12	research	research	NOUN
bracis-33550	55	13	,	,	PUNCT
bracis-33550	55	14	derived	derive	VERB
bracis-33550	55	15	from	from	ADP
bracis-33550	55	16	read	read	NOUN
bracis-33550	55	17	audiobooks	audiobook	NOUN
bracis-33550	55	18	from	from	ADP
bracis-33550	55	19	librivox	librivox	PROPN
bracis-33550	55	20	and	and	CCONJ
bracis-33550	55	21	consists	consist	VERB
bracis-33550	55	22	of	of	ADP
bracis-33550	55	23	8	8	NUM
bracis-33550	55	24	languages	language	NOUN
bracis-33550	55	25	,	,	PUNCT
bracis-33550	55	26	including	include	VERB
bracis-33550	55	27	about	about	ADP
bracis-33550	55	28	32k	32k	NOUN
bracis-33550	55	29	hours	hour	NOUN
bracis-33550	55	30	of	of	ADP
bracis-33550	55	31	english	english	PROPN
bracis-33550	55	32	and	and	CCONJ
bracis-33550	55	33	a	a	DET
bracis-33550	55	34	total	total	NOUN
bracis-33550	55	35	of	of	ADP
bracis-33550	55	36	4.5k	4.5k	NUM
bracis-33550	55	37	hours	hour	NOUN
bracis-33550	55	38	for	for	ADP
bracis-33550	55	39	other	other	ADJ
bracis-33550	55	40	languages	language	NOUN
bracis-33550	55	41	.	.	PUNCT
bracis-33550	56	1	for	for	ADP
bracis-33550	56	2	portuguese	portuguese	ADJ
bracis-33550	56	3	,	,	PUNCT
bracis-33550	56	4	there	there	PRON
bracis-33550	56	5	are	be	VERB
bracis-33550	56	6	131	131	NUM
bracis-33550	56	7	 	 	SPACE
bracis-33550	56	8	h	h	NOUN
bracis-33550	56	9	and	and	CCONJ
bracis-33550	56	10	54	54	NUM
bracis-33550	56	11	speakers	speaker	NOUN
bracis-33550	56	12	.	.	PUNCT
bracis-33550	57	1	specifically	specifically	ADV
bracis-33550	57	2	for	for	ADP
bracis-33550	57	3	the	the	DET
bracis-33550	57	4	recognition	recognition	NOUN
bracis-33550	57	5	task	task	NOUN
bracis-33550	57	6	,	,	PUNCT
bracis-33550	57	7	it	it	PRON
bracis-33550	57	8	can	can	AUX
bracis-33550	57	9	be	be	AUX
bracis-33550	57	10	combined	combine	VERB
bracis-33550	57	11	with	with	ADP
bracis-33550	57	12	other	other	ADJ
bracis-33550	57	13	resources	resource	NOUN
bracis-33550	57	14	,	,	PUNCT
bracis-33550	57	15	as	as	SCONJ
bracis-33550	57	16	it	it	PRON
bracis-33550	57	17	has	have	VERB
bracis-33550	57	18	relatively	relatively	ADV
bracis-33550	57	19	few	few	ADJ
bracis-33550	57	20	speakers	speaker	NOUN
bracis-33550	57	21	(	(	PUNCT
bracis-33550	57	22	audiobook	audiobook	NOUN
bracis-33550	57	23	speakers	speaker	NOUN
bracis-33550	57	24	)	)	PUNCT
bracis-33550	57	25	.	.	PUNCT
bracis-33550	58	1	these	these	DET
bracis-33550	58	2	resources	resource	NOUN
bracis-33550	58	3	consist	consist	VERB
bracis-33550	58	4	of	of	ADP
bracis-33550	58	5	cleaner	clean	ADJ
bracis-33550	58	6	audio	audio	NOUN
bracis-33550	58	7	,	,	PUNCT
bracis-33550	58	8	generally	generally	ADV
bracis-33550	58	9	in	in	ADP
bracis-33550	58	10	studio	studio	NOUN
bracis-33550	58	11	quality	quality	NOUN
bracis-33550	58	12	.	.	PUNCT
bracis-33550	59	1	because	because	SCONJ
bracis-33550	59	2	of	of	ADP
bracis-33550	59	3	this	this	PRON
bracis-33550	59	4	,	,	PUNCT
bracis-33550	59	5	models	model	NOUN
bracis-33550	59	6	built	build	VERB
bracis-33550	59	7	solely	solely	ADV
bracis-33550	59	8	on	on	ADP
bracis-33550	59	9	this	this	DET
bracis-33550	59	10	type	type	NOUN
bracis-33550	59	11	of	of	ADP
bracis-33550	59	12	audio	audio	NOUN
bracis-33550	59	13	are	be	AUX
bracis-33550	59	14	only	only	ADV
bracis-33550	59	15	suitable	suitable	ADJ
bracis-33550	59	16	for	for	ADP
bracis-33550	59	17	speech	speech	NOUN
bracis-33550	59	18	recognition	recognition	NOUN
bracis-33550	59	19	in	in	ADP
bracis-33550	59	20	low	low	ADJ
bracis-33550	59	21	-	-	PUNCT
bracis-33550	59	22	noise	noise	NOUN
bracis-33550	59	23	scenarios	scenario	NOUN
bracis-33550	59	24	.	.	PUNCT
bracis-33550	60	1	to	to	PART
bracis-33550	60	2	overcome	overcome	VERB
bracis-33550	60	3	this	this	DET
bracis-33550	60	4	feature	feature	NOUN
bracis-33550	60	5	,	,	PUNCT
bracis-33550	60	6	one	one	NUM
bracis-33550	60	7	solution	solution	NOUN
bracis-33550	60	8	would	would	AUX
bracis-33550	60	9	be	be	AUX
bracis-33550	60	10	to	to	PART
bracis-33550	60	11	inject	inject	VERB
bracis-33550	60	12	noise	noise	NOUN
bracis-33550	60	13	into	into	ADP
bracis-33550	60	14	the	the	DET
bracis-33550	60	15	audio	audio	NOUN
bracis-33550	60	16	or	or	CCONJ
bracis-33550	60	17	combine	combine	VERB
bracis-33550	60	18	it	it	PRON
bracis-33550	60	19	with	with	ADP
bracis-33550	60	20	audio	audio	NOUN
bracis-33550	60	21	from	from	ADP
bracis-33550	60	22	other	other	ADJ
bracis-33550	60	23	projects	project	NOUN
bracis-33550	60	24	at	at	ADP
bracis-33550	60	25	different	different	ADJ
bracis-33550	60	26	quality	quality	NOUN
bracis-33550	60	27	levels	level	NOUN
bracis-33550	60	28	.	.	PUNCT
bracis-33550	61	1	the	the	DET
bracis-33550	61	2	multilingual	multilingual	ADJ
bracis-33550	61	3	tedx	tedx	ADJ
bracis-33550	61	4	corpus	corpus	NOUN
bracis-33550	61	5	[	[	X
bracis-33550	61	6	21	21	NUM
bracis-33550	61	7	]	]	PUNCT
bracis-33550	61	8	was	be	AUX
bracis-33550	61	9	proposed	propose	VERB
bracis-33550	61	10	to	to	PART
bracis-33550	61	11	enable	enable	VERB
bracis-33550	61	12	research	research	NOUN
bracis-33550	61	13	in	in	ADP
bracis-33550	61	14	the	the	DET
bracis-33550	61	15	areas	area	NOUN
bracis-33550	61	16	of	of	ADP
bracis-33550	61	17	automatic	automatic	ADJ
bracis-33550	61	18	speech	speech	NOUN
bracis-33550	61	19	recognition	recognition	NOUN
bracis-33550	61	20	and	and	CCONJ
bracis-33550	61	21	speech	speech	NOUN
bracis-33550	61	22	-	-	PUNCT
bracis-33550	61	23	to	to	ADP
bracis-33550	61	24	-	-	PUNCT
bracis-33550	61	25	text	text	NOUN
bracis-33550	61	26	translation	translation	NOUN
bracis-33550	61	27	.	.	PUNCT
bracis-33550	62	1	for	for	ADP
bracis-33550	62	2	portuguese	portuguese	ADJ
bracis-33550	62	3	,	,	PUNCT
bracis-33550	62	4	there	there	PRON
bracis-33550	62	5	are	be	VERB
bracis-33550	62	6	164	164	NUM
bracis-33550	62	7	 	 	SPACE
bracis-33550	62	8	h	h	NOUN
bracis-33550	62	9	available	available	ADJ
bracis-33550	62	10	in	in	ADP
bracis-33550	62	11	93k	93k	PROPN
bracis-33550	62	12	audios	audio	NOUN
bracis-33550	62	13	.	.	PUNCT
bracis-33550	63	1	the	the	DET
bracis-33550	63	2	corpus	corpus	NOUN
bracis-33550	63	3	is	be	AUX
bracis-33550	63	4	made	make	VERB
bracis-33550	63	5	up	up	ADP
bracis-33550	63	6	of	of	ADP
bracis-33550	63	7	talks	talk	NOUN
bracis-33550	63	8	on	on	ADP
bracis-33550	63	9	a	a	DET
bracis-33550	63	10	wide	wide	ADJ
bracis-33550	63	11	range	range	NOUN
bracis-33550	63	12	of	of	ADP
bracis-33550	63	13	subjects	subject	NOUN
bracis-33550	63	14	,	,	PUNCT
bracis-33550	63	15	being	be	AUX
bracis-33550	63	16	managed	manage	VERB
bracis-33550	63	17	within	within	ADP
bracis-33550	63	18	the	the	DET
bracis-33550	63	19	scope	scope	NOUN
bracis-33550	63	20	of	of	ADP
bracis-33550	63	21	the	the	DET
bracis-33550	63	22	tedx	tedx	ADJ
bracis-33550	63	23	project	project	NOUN
bracis-33550	63	24	,	,	PUNCT
bracis-33550	63	25	linked	link	VERB
bracis-33550	63	26	to	to	ADP
bracis-33550	63	27	the	the	DET
bracis-33550	63	28	ted	ted	PROPN
bracis-33550	63	29	group	group	NOUN
bracis-33550	63	30	(	(	PUNCT
bracis-33550	63	31	technology	technology	NOUN
bracis-33550	63	32	,	,	PUNCT
bracis-33550	63	33	entertainment	entertainment	NOUN
bracis-33550	63	34	and	and	CCONJ
bracis-33550	63	35	design	design	NOUN
bracis-33550	63	36	)	)	PUNCT
bracis-33550	63	37	.	.	PUNCT
bracis-33550	64	1	in	in	ADP
bracis-33550	64	2	the	the	DET
bracis-33550	64	3	case	case	NOUN
bracis-33550	64	4	of	of	ADP
bracis-33550	64	5	portuguese	portuguese	ADJ
bracis-33550	64	6	,	,	PUNCT
bracis-33550	64	7	there	there	PRON
bracis-33550	64	8	are	be	VERB
bracis-33550	64	9	also	also	ADV
bracis-33550	64	10	translations	translation	NOUN
bracis-33550	64	11	of	of	ADP
bracis-33550	64	12	the	the	DET
bracis-33550	64	13	transcriptions	transcription	NOUN
bracis-33550	64	14	into	into	ADP
bracis-33550	64	15	english	english	PROPN
bracis-33550	64	16	and	and	CCONJ
bracis-33550	64	17	spanish	spanish	ADJ
bracis-33550	64	18	.	.	PUNCT
bracis-33550	65	1	in	in	ADP
bracis-33550	65	2	addition	addition	NOUN
bracis-33550	65	3	,	,	PUNCT
bracis-33550	65	4	audios	audio	NOUN
bracis-33550	65	5	in	in	ADP
bracis-33550	65	6	spanish	spanish	ADJ
bracis-33550	65	7	and	and	CCONJ
bracis-33550	65	8	french	french	NOUN
bracis-33550	65	9	also	also	ADV
bracis-33550	65	10	have	have	VERB
bracis-33550	65	11	translations	translation	NOUN
bracis-33550	65	12	into	into	ADP
bracis-33550	65	13	portuguese	portuguese	PROPN
bracis-33550	65	14	.	.	PUNCT
bracis-33550	66	1	the	the	DET
bracis-33550	66	2	spotify	spotify	NOUN
bracis-33550	66	3	corpus	corpus	NOUN
bracis-33550	66	4	[	[	X
bracis-33550	66	5	7	7	NUM
bracis-33550	66	6	]	]	PUNCT
bracis-33550	66	7	was	be	AUX
bracis-33550	66	8	first	first	ADV
bracis-33550	66	9	released	release	VERB
bracis-33550	66	10	in	in	ADP
bracis-33550	66	11	the	the	DET
bracis-33550	66	12	english	english	ADJ
bracis-33550	66	13	language	language	NOUN
bracis-33550	66	14	.	.	PUNCT
bracis-33550	67	1	in	in	ADP
bracis-33550	67	2	2022	2022	NUM
bracis-33550	67	3	,	,	PUNCT
bracis-33550	67	4	the	the	DET
bracis-33550	67	5	company	company	NOUN
bracis-33550	67	6	launched	launch	VERB
bracis-33550	67	7	a	a	DET
bracis-33550	67	8	new	new	ADJ
bracis-33550	67	9	versionfootnote	versionfootnote	NOUN
bracis-33550	67	10	12	12	NUM
bracis-33550	67	11	incorporating	incorporate	VERB
bracis-33550	67	12	portuguese	portuguese	NOUN
bracis-33550	67	13	[	[	X
bracis-33550	67	14	12	12	NUM
bracis-33550	67	15	]	]	PUNCT
bracis-33550	67	16	,	,	PUNCT
bracis-33550	67	17	offering	offer	VERB
bracis-33550	67	18	several	several	ADJ
bracis-33550	67	19	audios	audio	NOUN
bracis-33550	67	20	for	for	ADP
bracis-33550	67	21	the	the	DET
bracis-33550	67	22	portuguese	portuguese	ADJ
bracis-33550	67	23	language	language	NOUN
bracis-33550	67	24	coming	come	VERB
bracis-33550	67	25	mainly	mainly	ADV
bracis-33550	67	26	from	from	ADP
bracis-33550	67	27	podcasts	podcast	NOUN
bracis-33550	67	28	available	available	ADJ
bracis-33550	67	29	on	on	ADP
bracis-33550	67	30	the	the	DET
bracis-33550	67	31	platform	platform	NOUN
bracis-33550	67	32	.	.	PUNCT
bracis-33550	68	1	in	in	ADP
bracis-33550	68	2	total	total	ADJ
bracis-33550	68	3	,	,	PUNCT
bracis-33550	68	4	76k	76k	ADJ
bracis-33550	68	5	hours	hour	NOUN
bracis-33550	68	6	of	of	ADP
bracis-33550	68	7	audio	audio	NOUN
bracis-33550	68	8	were	be	AUX
bracis-33550	68	9	made	make	VERB
bracis-33550	68	10	available	available	ADJ
bracis-33550	68	11	from	from	ADP
bracis-33550	68	12	123k	123k	NUM
bracis-33550	68	13	episodes	episode	NOUN
bracis-33550	68	14	of	of	ADP
bracis-33550	68	15	shows	show	NOUN
bracis-33550	68	16	on	on	ADP
bracis-33550	68	17	the	the	DET
bracis-33550	68	18	platform	platform	NOUN
bracis-33550	68	19	.	.	PUNCT
bracis-33550	69	1	transcripts	transcript	NOUN
bracis-33550	69	2	were	be	AUX
bracis-33550	69	3	automatically	automatically	ADV
bracis-33550	69	4	generated	generate	VERB
bracis-33550	69	5	and	and	CCONJ
bracis-33550	69	6	are	be	AUX
bracis-33550	69	7	subject	subject	ADJ
bracis-33550	69	8	to	to	ADP
bracis-33550	69	9	transcription	transcription	NOUN
bracis-33550	69	10	errors	error	NOUN
bracis-33550	69	11	.	.	PUNCT
bracis-33550	70	1	it	it	PRON
bracis-33550	70	2	has	have	VERB
bracis-33550	70	3	a	a	DET
bracis-33550	70	4	free	free	ADJ
bracis-33550	70	5	license	license	NOUN
bracis-33550	70	6	for	for	ADP
bracis-33550	70	7	academic	academic	ADJ
bracis-33550	70	8	use	use	NOUN
bracis-33550	70	9	,	,	PUNCT
bracis-33550	70	10	but	but	CCONJ
bracis-33550	70	11	researchers	researcher	NOUN
bracis-33550	70	12	interested	interested	ADJ
bracis-33550	70	13	in	in	ADP
bracis-33550	70	14	accessing	access	VERB
bracis-33550	70	15	the	the	DET
bracis-33550	70	16	audios	audio	NOUN
bracis-33550	70	17	must	must	AUX
bracis-33550	70	18	submit	submit	VERB
bracis-33550	70	19	a	a	DET
bracis-33550	70	20	request	request	NOUN
bracis-33550	70	21	for	for	ADP
bracis-33550	70	22	access	access	NOUN
bracis-33550	70	23	on	on	ADP
bracis-33550	70	24	the	the	DET
bracis-33550	70	25	organizers	organizer	NOUN
bracis-33550	70	26	’	'	PUNCT
bracis-33550	70	27	website	website	NOUN
bracis-33550	70	28	.	.	PUNCT
bracis-33550	71	1	coraa	coraa	PROPN
bracis-33550	71	2	asr	asr	PROPN
bracis-33550	71	3	is	be	AUX
bracis-33550	71	4	the	the	DET
bracis-33550	71	5	combination	combination	NOUN
bracis-33550	71	6	of	of	ADP
bracis-33550	71	7	five	five	NUM
bracis-33550	71	8	independent	independent	ADJ
bracis-33550	71	9	projects	project	NOUN
bracis-33550	71	10	dealing	deal	VERB
bracis-33550	71	11	with	with	ADP
bracis-33550	71	12	speech	speech	NOUN
bracis-33550	71	13	in	in	ADP
bracis-33550	71	14	the	the	DET
bracis-33550	71	15	interior	interior	NOUN
bracis-33550	71	16	of	of	ADP
bracis-33550	71	17	são	são	PROPN
bracis-33550	71	18	paulo	paulo	PROPN
bracis-33550	71	19	(	(	PUNCT
bracis-33550	71	20	with	with	ADP
bracis-33550	71	21	35.96	35.96	NUM
bracis-33550	71	22	hours	hour	NOUN
bracis-33550	71	23	–	–	PUNCT
bracis-33550	71	24	alip	alip	NOUN
bracis-33550	71	25	project	project	NOUN
bracis-33550	71	26	)	)	PUNCT
bracis-33550	71	27	,	,	PUNCT
bracis-33550	71	28	minas	minas	PROPN
bracis-33550	71	29	gerais	gerais	PROPN
bracis-33550	71	30	(	(	PUNCT
bracis-33550	71	31	with	with	ADP
bracis-33550	71	32	9.64	9.64	NUM
bracis-33550	71	33	hours	hour	NOUN
bracis-33550	71	34	—	—	PUNCT
bracis-33550	71	35	c	c	X
bracis-33550	71	36	-	-	PUNCT
bracis-33550	71	37	oral	oral	ADJ
bracis-33550	71	38	brasil	brasil	PROPN
bracis-33550	71	39	project	project	NOUN
bracis-33550	71	40	)	)	PUNCT
bracis-33550	71	41	,	,	PUNCT
bracis-33550	71	42	recife	recife	NOUN
bracis-33550	71	43	(	(	PUNCT
bracis-33550	71	44	141.31	141.31	NUM
bracis-33550	71	45	hours	hour	NOUN
bracis-33550	71	46	–	–	PUNCT
bracis-33550	71	47	nurc	nurc	ADJ
bracis-33550	71	48	-	-	PUNCT
bracis-33550	71	49	recife	recife	NOUN
bracis-33550	71	50	project	project	NOUN
bracis-33550	71	51	)	)	PUNCT
bracis-33550	71	52	and	and	CCONJ
bracis-33550	71	53	são	são	PROPN
bracis-33550	71	54	paulo	paulo	PROPN
bracis-33550	71	55	capital	capital	PROPN
bracis-33550	71	56	(	(	PUNCT
bracis-33550	71	57	31.14	31.14	NUM
bracis-33550	71	58	hours	hour	NOUN
bracis-33550	71	59	–	–	PUNCT
bracis-33550	71	60	sp2010	sp2010	NOUN
bracis-33550	71	61	project	project	NOUN
bracis-33550	71	62	)	)	PUNCT
bracis-33550	71	63	,	,	PUNCT
bracis-33550	71	64	in	in	ADP
bracis-33550	71	65	addition	addition	NOUN
bracis-33550	71	66	to	to	ADP
bracis-33550	71	67	the	the	DET
bracis-33550	71	68	prepared	prepared	ADJ
bracis-33550	71	69	speech	speech	NOUN
bracis-33550	71	70	of	of	ADP
bracis-33550	71	71	tedx	tedx	ADJ
bracis-33550	71	72	talks	talk	NOUN
bracis-33550	71	73	in	in	ADP
bracis-33550	71	74	brazilian	brazilian	ADJ
bracis-33550	71	75	and	and	CCONJ
bracis-33550	71	76	european	european	ADJ
bracis-33550	71	77	portuguese	portuguese	PROPN
bracis-33550	71	78	(	(	PUNCT
bracis-33550	71	79	72.74	72.74	NUM
bracis-33550	71	80	hours	hour	NOUN
bracis-33550	71	81	)	)	PUNCT
bracis-33550	71	82	,	,	PUNCT
bracis-33550	71	83	totaling	total	VERB
bracis-33550	71	84	290	290	NUM
bracis-33550	71	85	hours	hour	NOUN
bracis-33550	71	86	and	and	CCONJ
bracis-33550	71	87	402k	402k	PROPN
bracis-33550	71	88	audios	audio	NOUN
bracis-33550	71	89	.	.	PUNCT
bracis-33550	72	1	the	the	DET
bracis-33550	72	2	common	common	ADJ
bracis-33550	72	3	voice	voice	NOUN
bracis-33550	72	4	corpus	corpus	NOUN
bracis-33550	72	5	[	[	X
bracis-33550	72	6	1	1	X
bracis-33550	72	7	]	]	PUNCT
bracis-33550	72	8	is	be	AUX
bracis-33550	72	9	an	an	DET
bracis-33550	72	10	open	open	ADJ
bracis-33550	72	11	-	-	PUNCT
bracis-33550	72	12	use	use	NOUN
bracis-33550	72	13	project	project	NOUN
bracis-33550	72	14	created	create	VERB
bracis-33550	72	15	by	by	ADP
bracis-33550	72	16	the	the	DET
bracis-33550	72	17	mozilla	mozilla	PROPN
bracis-33550	72	18	foundation	foundation	PROPN
bracis-33550	72	19	.	.	PUNCT
bracis-33550	73	1	the	the	DET
bracis-33550	73	2	project	project	NOUN
bracis-33550	73	3	is	be	AUX
bracis-33550	73	4	a	a	DET
bracis-33550	73	5	response	response	NOUN
bracis-33550	73	6	to	to	ADP
bracis-33550	73	7	the	the	DET
bracis-33550	73	8	lack	lack	NOUN
bracis-33550	73	9	of	of	ADP
bracis-33550	73	10	resources	resource	NOUN
bracis-33550	73	11	for	for	ADP
bracis-33550	73	12	several	several	ADJ
bracis-33550	73	13	languages	language	NOUN
bracis-33550	73	14	,	,	PUNCT
bracis-33550	73	15	including	include	VERB
bracis-33550	73	16	portuguese	portuguese	NOUN
bracis-33550	73	17	.	.	PUNCT
bracis-33550	74	1	in	in	ADP
bracis-33550	74	2	the	the	DET
bracis-33550	74	3	project	project	NOUN
bracis-33550	74	4	,	,	PUNCT
bracis-33550	74	5	users	user	NOUN
bracis-33550	74	6	can	can	AUX
bracis-33550	74	7	simultaneously	simultaneously	ADV
bracis-33550	74	8	contribute	contribute	VERB
bracis-33550	74	9	to	to	ADP
bracis-33550	74	10	the	the	DET
bracis-33550	74	11	growth	growth	NOUN
bracis-33550	74	12	of	of	ADP
bracis-33550	74	13	the	the	DET
bracis-33550	74	14	base	base	NOUN
bracis-33550	74	15	and	and	CCONJ
bracis-33550	74	16	access	access	NOUN
bracis-33550	74	17	other	other	ADJ
bracis-33550	74	18	people	people	NOUN
bracis-33550	74	19	’s	’s	PART
bracis-33550	74	20	audio	audio	NOUN
bracis-33550	74	21	.	.	PUNCT
bracis-33550	75	1	to	to	PART
bracis-33550	75	2	collaborate	collaborate	VERB
bracis-33550	75	3	with	with	ADP
bracis-33550	75	4	the	the	DET
bracis-33550	75	5	project	project	NOUN
bracis-33550	75	6	,	,	PUNCT
bracis-33550	75	7	users	user	NOUN
bracis-33550	75	8	can	can	AUX
bracis-33550	75	9	donate	donate	VERB
bracis-33550	75	10	audio	audio	NOUN
bracis-33550	75	11	in	in	ADP
bracis-33550	75	12	their	their	PRON
bracis-33550	75	13	own	own	ADJ
bracis-33550	75	14	voices	voice	NOUN
bracis-33550	75	15	and	and	CCONJ
bracis-33550	75	16	review	review	VERB
bracis-33550	75	17	donations	donation	NOUN
bracis-33550	75	18	from	from	ADP
bracis-33550	75	19	other	other	ADJ
bracis-33550	75	20	users	user	NOUN
bracis-33550	75	21	.	.	PUNCT
bracis-33550	76	1	the	the	DET
bracis-33550	76	2	project	project	NOUN
bracis-33550	76	3	has	have	VERB
bracis-33550	76	4	tools	tool	NOUN
bracis-33550	76	5	for	for	ADP
bracis-33550	76	6	collection	collection	NOUN
bracis-33550	76	7	,	,	PUNCT
bracis-33550	76	8	validation	validation	NOUN
bracis-33550	76	9	and	and	CCONJ
bracis-33550	76	10	internationalization	internationalization	NOUN
bracis-33550	76	11	(	(	PUNCT
bracis-33550	76	12	adaptation	adaptation	NOUN
bracis-33550	76	13	to	to	ADP
bracis-33550	76	14	different	different	ADJ
bracis-33550	76	15	languages	language	NOUN
bracis-33550	76	16	)	)	PUNCT
bracis-33550	76	17	.	.	PUNCT
bracis-33550	77	1	the	the	DET
bracis-33550	77	2	permissive	permissive	ADJ
bracis-33550	77	3	use	use	NOUN
bracis-33550	77	4	license	license	NOUN
bracis-33550	77	5	of	of	ADP
bracis-33550	77	6	this	this	DET
bracis-33550	77	7	project	project	NOUN
bracis-33550	77	8	allows	allow	VERB
bracis-33550	77	9	the	the	DET
bracis-33550	77	10	exploration	exploration	NOUN
bracis-33550	77	11	of	of	ADP
bracis-33550	77	12	the	the	DET
bracis-33550	77	13	corpus	corpus	NOUN
bracis-33550	77	14	including	include	VERB
bracis-33550	77	15	for	for	ADP
bracis-33550	77	16	commercial	commercial	ADJ
bracis-33550	77	17	purposes	purpose	NOUN
bracis-33550	77	18	.	.	PUNCT
bracis-33550	78	1	in	in	ADP
bracis-33550	78	2	version	version	NOUN
bracis-33550	78	3	17	17	NUM
bracis-33550	78	4	,	,	PUNCT
bracis-33550	78	5	the	the	DET
bracis-33550	78	6	subcorpus	subcorpus	NOUN
bracis-33550	78	7	for	for	ADP
bracis-33550	78	8	the	the	DET
bracis-33550	78	9	portuguese	portuguese	ADJ
bracis-33550	78	10	language	language	NOUN
bracis-33550	78	11	has	have	VERB
bracis-33550	78	12	211	211	NUM
bracis-33550	78	13	 	 	SPACE
bracis-33550	78	14	h	h	NOUN
bracis-33550	78	15	of	of	ADP
bracis-33550	78	16	audio	audio	NOUN
bracis-33550	78	17	and	and	CCONJ
bracis-33550	78	18	transcriptions	transcription	NOUN
bracis-33550	78	19	,	,	PUNCT
bracis-33550	78	20	of	of	ADP
bracis-33550	78	21	which	which	PRON
bracis-33550	78	22	175	175	NUM
bracis-33550	78	23	were	be	AUX
bracis-33550	78	24	validated	validate	VERB
bracis-33550	78	25	.	.	PUNCT
bracis-33550	79	1	the	the	DET
bracis-33550	79	2	nurc	nurc	ADJ
bracis-33550	79	3	-	-	PUNCT
bracis-33550	79	4	sp	sp	NOUN
bracis-33550	79	5	audio	audio	NOUN
bracis-33550	79	6	corpus	corpus	NOUN
bracis-33550	79	7	,	,	PUNCT
bracis-33550	79	8	the	the	DET
bracis-33550	79	9	focus	focus	NOUN
bracis-33550	79	10	of	of	ADP
bracis-33550	79	11	this	this	DET
bracis-33550	79	12	paper	paper	NOUN
bracis-33550	79	13	,	,	PUNCT
bracis-33550	79	14	is	be	AUX
bracis-33550	79	15	described	describe	VERB
bracis-33550	79	16	in	in	ADP
bracis-33550	79	17	detail	detail	NOUN
bracis-33550	79	18	in	in	ADP
bracis-33550	79	19	sect	sect	NOUN
bracis-33550	79	20	.	.	PUNCT
bracis-33550	79	21	 	 	SPACE
bracis-33550	80	1	3	3	NUM
bracis-33550	80	2	.	.	NOUN
bracis-33550	80	3	3	3	NUM
bracis-33550	80	4	the	the	DET
bracis-33550	80	5	nurc	nurc	ADJ
bracis-33550	80	6	-	-	PUNCT
bracis-33550	80	7	sp	sp	NOUN
bracis-33550	80	8	corpus	corpus	NOUN
bracis-33550	80	9	nurc	nurc	NOUN
bracis-33550	80	10	-	-	PUNCT
bracis-33550	80	11	sp	sp	NOUN
bracis-33550	80	12	was	be	AUX
bracis-33550	80	13	the	the	DET
bracis-33550	80	14	são	são	PROPN
bracis-33550	80	15	paulo	paulo	PROPN
bracis-33550	80	16	division	division	NOUN
bracis-33550	80	17	of	of	ADP
bracis-33550	80	18	the	the	DET
bracis-33550	80	19	nurc	nurc	NOUN
bracis-33550	80	20	(	(	PUNCT
bracis-33550	80	21	norma	norma	PROPN
bracis-33550	80	22	urbana	urbana	PROPN
bracis-33550	80	23	linguística	linguística	PROPN
bracis-33550	80	24	culta	culta	PROPN
bracis-33550	80	25	)	)	PUNCT
bracis-33550	80	26	project	project	NOUN
bracis-33550	80	27	.	.	PUNCT
bracis-33550	81	1	nurc	nurc	ADJ
bracis-33550	81	2	-	-	PUNCT
bracis-33550	81	3	sp	sp	NOUN
bracis-33550	81	4	collected	collect	VERB
bracis-33550	81	5	more	more	ADJ
bracis-33550	81	6	than	than	ADP
bracis-33550	81	7	300	300	NUM
bracis-33550	81	8	 	 	SPACE
bracis-33550	81	9	h	h	NOUN
bracis-33550	81	10	of	of	ADP
bracis-33550	81	11	são	são	PROPN
bracis-33550	81	12	paulo	paulo	PROPN
bracis-33550	81	13	speakers	speaker	NOUN
bracis-33550	81	14	throughout	throughout	ADP
bracis-33550	81	15	the	the	DET
bracis-33550	81	16	1970s	1970s	NUM
bracis-33550	81	17	.	.	PUNCT
bracis-33550	82	1	the	the	DET
bracis-33550	82	2	nurc	nurc	ADJ
bracis-33550	82	3	-	-	PUNCT
bracis-33550	82	4	sp	sp	NOUN
bracis-33550	82	5	corpus	corpus	NOUN
bracis-33550	82	6	is	be	AUX
bracis-33550	82	7	made	make	VERB
bracis-33550	82	8	up	up	ADP
bracis-33550	82	9	of	of	ADP
bracis-33550	82	10	375	375	NUM
bracis-33550	82	11	audio	audio	ADJ
bracis-33550	82	12	recordings	recording	NOUN
bracis-33550	82	13	of	of	ADP
bracis-33550	82	14	three	three	NUM
bracis-33550	82	15	genre	genre	NOUN
bracis-33550	82	16	types	type	NOUN
bracis-33550	82	17	:	:	PUNCT
bracis-33550	83	1	1	1	X
bracis-33550	83	2	.	.	PUNCT
bracis-33550	83	3	(	(	PUNCT
bracis-33550	83	4	did	do	AUX
bracis-33550	83	5	)	)	PUNCT
bracis-33550	83	6	documenter	documenter	NOUN
bracis-33550	83	7	and	and	CCONJ
bracis-33550	83	8	participant	participant	ADJ
bracis-33550	83	9	dialogue	dialogue	NOUN
bracis-33550	83	10	carried	carry	VERB
bracis-33550	83	11	out	out	ADP
bracis-33550	83	12	between	between	ADP
bracis-33550	83	13	a	a	DET
bracis-33550	83	14	documenter	documenter	NOUN
bracis-33550	83	15	and	and	CCONJ
bracis-33550	83	16	a	a	DET
bracis-33550	83	17	participant	participant	NOUN
bracis-33550	83	18	directly	directly	ADV
bracis-33550	83	19	;	;	PUNCT
bracis-33550	83	20	2	2	X
bracis-33550	83	21	.	.	X
bracis-33550	83	22	(	(	PUNCT
bracis-33550	83	23	d2	d2	PROPN
bracis-33550	83	24	)	)	PUNCT
bracis-33550	83	25	participant	participant	NOUN
bracis-33550	83	26	and	and	CCONJ
bracis-33550	83	27	participant	participant	ADJ
bracis-33550	83	28	dialogue	dialogue	NOUN
bracis-33550	83	29	between	between	ADP
bracis-33550	83	30	two	two	NUM
bracis-33550	83	31	participants	participant	NOUN
bracis-33550	83	32	mediated	mediate	VERB
bracis-33550	83	33	by	by	ADP
bracis-33550	83	34	a	a	DET
bracis-33550	83	35	documenter	documenter	NOUN
bracis-33550	83	36	;	;	PUNCT
bracis-33550	83	37	and	and	CCONJ
bracis-33550	83	38	3	3	X
bracis-33550	83	39	.	.	PUNCT
bracis-33550	83	40	(	(	PUNCT
bracis-33550	83	41	ef	ef	PROPN
bracis-33550	83	42	)	)	PUNCT
bracis-33550	83	43	participant	participant	ADJ
bracis-33550	83	44	lectures	lecture	NOUN
bracis-33550	83	45	,	,	PUNCT
bracis-33550	83	46	seminars	seminar	NOUN
bracis-33550	83	47	,	,	PUNCT
bracis-33550	83	48	classes	class	NOUN
bracis-33550	83	49	,	,	PUNCT
bracis-33550	83	50	speeches	speech	NOUN
bracis-33550	83	51	in	in	ADP
bracis-33550	83	52	general	general	ADJ
bracis-33550	83	53	,	,	PUNCT
bracis-33550	83	54	given	give	VERB
bracis-33550	83	55	by	by	ADP
bracis-33550	83	56	a	a	DET
bracis-33550	83	57	participant	participant	NOUN
bracis-33550	83	58	in	in	ADP
bracis-33550	83	59	a	a	DET
bracis-33550	83	60	formal	formal	ADJ
bracis-33550	83	61	context	context	NOUN
bracis-33550	83	62	.	.	PUNCT
bracis-33550	84	1	which	which	PRON
bracis-33550	84	2	are	be	AUX
bracis-33550	84	3	divided	divide	VERB
bracis-33550	84	4	into	into	ADP
bracis-33550	84	5	three	three	NUM
bracis-33550	84	6	subcorpora	subcorpora	NOUN
bracis-33550	84	7	:	:	PUNCT
bracis-33550	84	8	the	the	DET
bracis-33550	84	9	minimum	minimum	PROPN
bracis-33550	84	10	corpus	corpus	NOUN
bracis-33550	84	11	(	(	PUNCT
bracis-33550	84	12	21	21	NUM
bracis-33550	84	13	audio	audio	ADJ
bracis-33550	84	14	recordings	recording	NOUN
bracis-33550	84	15	)	)	PUNCT
bracis-33550	84	16	,	,	PUNCT
bracis-33550	84	17	the	the	DET
bracis-33550	84	18	corpus	corpus	NOUN
bracis-33550	84	19	of	of	ADP
bracis-33550	84	20	non	non	ADJ
bracis-33550	84	21	-	-	ADJ
bracis-33550	84	22	aligned	aligned	ADJ
bracis-33550	84	23	audios	audio	NOUN
bracis-33550	84	24	and	and	CCONJ
bracis-33550	84	25	transcriptions	transcription	NOUN
bracis-33550	84	26	(	(	PUNCT
bracis-33550	84	27	26	26	NUM
bracis-33550	84	28	audio	audio	ADJ
bracis-33550	84	29	recordings	recording	NOUN
bracis-33550	84	30	)	)	PUNCT
bracis-33550	84	31	,	,	PUNCT
bracis-33550	84	32	and	and	CCONJ
bracis-33550	84	33	the	the	DET
bracis-33550	84	34	audio	audio	ADJ
bracis-33550	84	35	corpus	corpus	NOUN
bracis-33550	84	36	.	.	PUNCT
bracis-33550	85	1	nurc	nurc	ADJ
bracis-33550	85	2	-	-	PUNCT
bracis-33550	85	3	sp	sp	NOUN
bracis-33550	85	4	had	have	VERB
bracis-33550	85	5	its	its	PRON
bracis-33550	85	6	original	original	ADJ
bracis-33550	85	7	analog	analog	NOUN
bracis-33550	85	8	audios	audio	NOUN
bracis-33550	85	9	digitized	digitize	VERB
bracis-33550	85	10	by	by	ADP
bracis-33550	85	11	the	the	DET
bracis-33550	85	12	alexandre	alexandre	NOUN
bracis-33550	85	13	eulalio	eulalio	VERB
bracis-33550	85	14	cultural	cultural	ADJ
bracis-33550	85	15	documentation	documentation	NOUN
bracis-33550	85	16	center	center	NOUN
bracis-33550	85	17	(	(	PUNCT
bracis-33550	85	18	cedae	cedae	NOUN
bracis-33550	85	19	/	/	SYM
bracis-33550	85	20	unicamp	unicamp	NOUN
bracis-33550	85	21	)	)	PUNCT
bracis-33550	85	22	and	and	CCONJ
bracis-33550	85	23	in	in	ADP
bracis-33550	85	24	december	december	PROPN
bracis-33550	85	25	2020	2020	NUM
bracis-33550	85	26	it	it	PRON
bracis-33550	85	27	was	be	AUX
bracis-33550	85	28	made	make	VERB
bracis-33550	85	29	available	available	ADJ
bracis-33550	85	30	to	to	ADP
bracis-33550	85	31	the	the	DET
bracis-33550	85	32	tarsila	tarsila	ADJ
bracis-33550	85	33	projectfootnote	projectfootnote	NOUN
bracis-33550	85	34	13	13	NUM
bracis-33550	85	35	as	as	ADP
bracis-33550	85	36	a	a	DET
bracis-33550	85	37	resource	resource	NOUN
bracis-33550	85	38	to	to	AUX
bracis-33550	85	39	:	:	PUNCT
bracis-33550	85	40	(	(	PUNCT
bracis-33550	85	41	i	i	NOUN
bracis-33550	85	42	)	)	PUNCT
bracis-33550	85	43	build	build	VERB
bracis-33550	85	44	training	training	NOUN
bracis-33550	85	45	data	datum	NOUN
bracis-33550	85	46	sets	set	NOUN
bracis-33550	85	47	for	for	ADP
bracis-33550	85	48	spontaneous	spontaneous	ADJ
bracis-33550	85	49	speech	speech	NOUN
bracis-33550	85	50	recognition	recognition	NOUN
bracis-33550	85	51	systems	system	NOUN
bracis-33550	85	52	,	,	PUNCT
bracis-33550	85	53	and	and	CCONJ
bracis-33550	85	54	(	(	PUNCT
bracis-33550	85	55	ii	ii	NOUN
bracis-33550	85	56	)	)	PUNCT
bracis-33550	85	57	facilitate	facilitate	VERB
bracis-33550	85	58	future	future	ADJ
bracis-33550	85	59	linguistic	linguistic	ADJ
bracis-33550	85	60	studies	study	NOUN
bracis-33550	85	61	.	.	PUNCT
bracis-33550	86	1	the	the	DET
bracis-33550	86	2	three	three	NUM
bracis-33550	86	3	subcorpora	subcorpora	NOUN
bracis-33550	86	4	that	that	PRON
bracis-33550	86	5	integrate	integrate	VERB
bracis-33550	86	6	the	the	DET
bracis-33550	86	7	nurc	nurc	ADJ
bracis-33550	86	8	-	-	PUNCT
bracis-33550	86	9	sp	sp	NOUN
bracis-33550	86	10	digital	digital	ADJ
bracis-33550	86	11	repository	repository	NOUN
bracis-33550	86	12	were	be	AUX
bracis-33550	86	13	made	make	VERB
bracis-33550	86	14	available	available	ADJ
bracis-33550	86	15	in	in	ADP
bracis-33550	86	16	the	the	DET
bracis-33550	86	17	portal	portal	ADJ
bracis-33550	86	18	nurc	nurc	ADJ
bracis-33550	86	19	-	-	PUNCT
bracis-33550	86	20	sp	sp	NOUN
bracis-33550	86	21	digital	digital	NOUN
bracis-33550	86	22	[	[	X
bracis-33550	86	23	20	20	NUM
bracis-33550	86	24	]	]	PUNCT
bracis-33550	86	25	for	for	ADP
bracis-33550	86	26	researchers	researcher	NOUN
bracis-33550	86	27	from	from	ADP
bracis-33550	86	28	fields	field	NOUN
bracis-33550	86	29	of	of	ADP
bracis-33550	86	30	linguistics	linguistic	NOUN
bracis-33550	86	31	as	as	ADV
bracis-33550	86	32	well	well	ADV
bracis-33550	86	33	for	for	ADP
bracis-33550	86	34	the	the	DET
bracis-33550	86	35	general	general	ADJ
bracis-33550	86	36	public	public	NOUN
bracis-33550	86	37	,	,	PUNCT
bracis-33550	86	38	due	due	ADP
bracis-33550	86	39	to	to	ADP
bracis-33550	86	40	easy	easy	ADJ
bracis-33550	86	41	access	access	NOUN
bracis-33550	86	42	and	and	CCONJ
bracis-33550	86	43	filtering	filtering	NOUN
bracis-33550	86	44	tools	tool	NOUN
bracis-33550	86	45	to	to	PART
bracis-33550	86	46	use	use	VERB
bracis-33550	86	47	the	the	DET
bracis-33550	86	48	material	material	NOUN
bracis-33550	86	49	.	.	PUNCT
bracis-33550	87	1	here	here	ADV
bracis-33550	87	2	,	,	PUNCT
bracis-33550	87	3	in	in	ADP
bracis-33550	87	4	this	this	DET
bracis-33550	87	5	paper	paper	NOUN
bracis-33550	87	6	,	,	PUNCT
bracis-33550	87	7	we	we	PRON
bracis-33550	87	8	focus	focus	VERB
bracis-33550	87	9	on	on	ADP
bracis-33550	87	10	the	the	DET
bracis-33550	87	11	nurc	nurc	ADJ
bracis-33550	87	12	-	-	PUNCT
bracis-33550	87	13	sp	sp	NOUN
bracis-33550	87	14	audio	audio	NOUN
bracis-33550	87	15	corpus	corpus	NOUN
bracis-33550	87	16	.	.	PUNCT
bracis-33550	88	1	it	it	PRON
bracis-33550	88	2	was	be	AUX
bracis-33550	88	3	originally	originally	ADV
bracis-33550	88	4	composed	compose	VERB
bracis-33550	88	5	of	of	ADP
bracis-33550	88	6	328	328	NUM
bracis-33550	88	7	audio	audio	NOUN
bracis-33550	88	8	recordings	recording	NOUN
bracis-33550	88	9	without	without	ADP
bracis-33550	88	10	transcriptions	transcription	NOUN
bracis-33550	88	11	,	,	PUNCT
bracis-33550	88	12	which	which	PRON
bracis-33550	88	13	have	have	AUX
bracis-33550	88	14	been	be	AUX
bracis-33550	88	15	automatically	automatically	ADV
bracis-33550	88	16	transcribed	transcribe	VERB
bracis-33550	88	17	by	by	ADP
bracis-33550	88	18	whisperx	whisperx	NOUN
bracis-33550	89	1	[	[	X
bracis-33550	89	2	3	3	NUM
bracis-33550	89	3	]	]	PUNCT
bracis-33550	89	4	.	.	PUNCT
bracis-33550	90	1	the	the	DET
bracis-33550	90	2	revision	revision	NOUN
bracis-33550	90	3	of	of	ADP
bracis-33550	90	4	automated	automate	VERB
bracis-33550	90	5	whisperx	whisperx	NOUN
bracis-33550	90	6	transcription	transcription	NOUN
bracis-33550	90	7	segments	segment	NOUN
bracis-33550	90	8	were	be	AUX
bracis-33550	90	9	performed	perform	VERB
bracis-33550	90	10	from	from	ADP
bracis-33550	90	11	june	june	PROPN
bracis-33550	90	12	2023	2023	NUM
bracis-33550	90	13	to	to	ADP
bracis-33550	90	14	december	december	PROPN
bracis-33550	90	15	2023	2023	NUM
bracis-33550	90	16	by	by	ADP
bracis-33550	90	17	14	14	NUM
bracis-33550	90	18	native	native	ADJ
bracis-33550	90	19	speakers	speaker	NOUN
bracis-33550	90	20	of	of	ADP
bracis-33550	90	21	bp	bp	PROPN
bracis-33550	90	22	.	.	PUNCT
bracis-33550	91	1	the	the	DET
bracis-33550	91	2	revision	revision	NOUN
bracis-33550	91	3	process	process	NOUN
bracis-33550	91	4	was	be	AUX
bracis-33550	91	5	based	base	VERB
bracis-33550	91	6	on	on	ADP
bracis-33550	91	7	an	an	DET
bracis-33550	91	8	annotation	annotation	NOUN
bracis-33550	91	9	guideline	guideline	NOUN
bracis-33550	91	10	designed	design	VERB
bracis-33550	91	11	to	to	ADP
bracis-33550	91	12	:	:	PUNCT
bracis-33550	91	13	(	(	PUNCT
bracis-33550	91	14	i	i	NOUN
bracis-33550	91	15	)	)	PUNCT
bracis-33550	91	16	help	help	VERB
bracis-33550	91	17	making	make	VERB
bracis-33550	91	18	the	the	DET
bracis-33550	91	19	revision	revision	NOUN
bracis-33550	91	20	uniform	uniform	NOUN
bracis-33550	91	21	and	and	CCONJ
bracis-33550	91	22	(	(	PUNCT
bracis-33550	91	23	ii	ii	NOUN
bracis-33550	91	24	)	)	PUNCT
bracis-33550	91	25	remove	remove	VERB
bracis-33550	91	26	segments	segment	NOUN
bracis-33550	91	27	with	with	ADP
bracis-33550	91	28	high	high	ADJ
bracis-33550	91	29	amount	amount	NOUN
bracis-33550	91	30	of	of	ADP
bracis-33550	91	31	noise	noise	NOUN
bracis-33550	91	32	and	and	CCONJ
bracis-33550	91	33	overlapping	overlap	VERB
bracis-33550	91	34	voices	voice	NOUN
bracis-33550	91	35	.	.	PUNCT
bracis-33550	92	1	the	the	DET
bracis-33550	92	2	guideline	guideline	NOUN
bracis-33550	92	3	contains	contain	VERB
bracis-33550	92	4	11	11	NUM
bracis-33550	92	5	rules	rule	NOUN
bracis-33550	92	6	,	,	PUNCT
bracis-33550	92	7	dealing	deal	VERB
bracis-33550	92	8	with	with	ADP
bracis-33550	92	9	(	(	PUNCT
bracis-33550	92	10	i	i	NOUN
bracis-33550	92	11	)	)	PUNCT
bracis-33550	92	12	orality	orality	NOUN
bracis-33550	92	13	marks	mark	NOUN
bracis-33550	92	14	,	,	PUNCT
bracis-33550	92	15	(	(	PUNCT
bracis-33550	92	16	ii	ii	NOUN
bracis-33550	92	17	)	)	PUNCT
bracis-33550	92	18	how	how	SCONJ
bracis-33550	92	19	to	to	PART
bracis-33550	92	20	transcribe	transcribe	VERB
bracis-33550	92	21	filled	fill	VERB
bracis-33550	92	22	pauses	pause	NOUN
bracis-33550	92	23	,	,	PUNCT
bracis-33550	92	24	(	(	PUNCT
bracis-33550	92	25	iii	iii	X
bracis-33550	92	26	)	)	PUNCT
bracis-33550	92	27	repetitive	repetitive	ADJ
bracis-33550	92	28	hesitations	hesitation	NOUN
bracis-33550	92	29	,	,	PUNCT
bracis-33550	92	30	(	(	PUNCT
bracis-33550	92	31	iv	iv	X
bracis-33550	92	32	)	)	PUNCT
bracis-33550	92	33	numbers	number	NOUN
bracis-33550	92	34	,	,	PUNCT
bracis-33550	92	35	(	(	PUNCT
bracis-33550	92	36	v	v	NOUN
bracis-33550	92	37	)	)	PUNCT
bracis-33550	92	38	individual	individual	ADJ
bracis-33550	92	39	letters	letter	NOUN
bracis-33550	92	40	,	,	PUNCT
bracis-33550	92	41	(	(	PUNCT
bracis-33550	92	42	vi	vi	NOUN
bracis-33550	92	43	)	)	PUNCT
bracis-33550	92	44	acronyms	acronym	NOUN
bracis-33550	92	45	,	,	PUNCT
bracis-33550	92	46	(	(	PUNCT
bracis-33550	92	47	vii	vii	PROPN
bracis-33550	92	48	)	)	PUNCT
bracis-33550	92	49	foreign	foreign	ADJ
bracis-33550	92	50	terms	term	NOUN
bracis-33550	92	51	,	,	PUNCT
bracis-33550	92	52	(	(	PUNCT
bracis-33550	92	53	viii	viii	NOUN
bracis-33550	92	54	)	)	PUNCT
bracis-33550	92	55	punctuation	punctuation	NOUN
bracis-33550	92	56	and	and	CCONJ
bracis-33550	92	57	capitalization	capitalization	NOUN
bracis-33550	92	58	,	,	PUNCT
bracis-33550	92	59	(	(	PUNCT
bracis-33550	92	60	iv	iv	X
bracis-33550	92	61	)	)	PUNCT
bracis-33550	92	62	emotion	emotion	NOUN
bracis-33550	92	63	sounds	sound	VERB
bracis-33550	92	64	(	(	PUNCT
bracis-33550	92	65	ex	ex	NOUN
bracis-33550	92	66	.	.	NOUN
bracis-33550	92	67	laughter	laughter	NOUN
bracis-33550	92	68	)	)	PUNCT
bracis-33550	92	69	which	which	PRON
bracis-33550	92	70	were	be	AUX
bracis-33550	92	71	annotated	annotate	VERB
bracis-33550	92	72	in	in	ADP
bracis-33550	92	73	parentheses	parenthesis	NOUN
bracis-33550	92	74	,	,	PUNCT
bracis-33550	92	75	(	(	PUNCT
bracis-33550	92	76	x	x	X
bracis-33550	92	77	)	)	PUNCT
bracis-33550	92	78	misunderstanding	misunderstanding	NOUN
bracis-33550	92	79	of	of	ADP
bracis-33550	92	80	words	word	NOUN
bracis-33550	92	81	or	or	CCONJ
bracis-33550	92	82	passages	passage	NOUN
bracis-33550	92	83	and	and	CCONJ
bracis-33550	92	84	(	(	PUNCT
bracis-33550	92	85	xi	xi	X
bracis-33550	92	86	)	)	PUNCT
bracis-33550	92	87	how	how	SCONJ
bracis-33550	92	88	to	to	PART
bracis-33550	92	89	deal	deal	VERB
bracis-33550	92	90	with	with	ADP
bracis-33550	92	91	automatic	automatic	ADJ
bracis-33550	92	92	segmentation	segmentation	NOUN
bracis-33550	92	93	failures	failure	NOUN
bracis-33550	92	94	.	.	PUNCT
bracis-33550	93	1	3.1	3.1	NUM
bracis-33550	93	2	pre	pre	ADJ
bracis-33550	93	3	-	-	ADJ
bracis-33550	93	4	processing	processing	ADJ
bracis-33550	93	5	and	and	CCONJ
bracis-33550	93	6	 	 	SPACE
bracis-33550	93	7	filtering	filtering	NOUN
bracis-33550	93	8	of	of	ADP
bracis-33550	93	9	 	 	SPACE
bracis-33550	93	10	nurc	nurc	ADJ
bracis-33550	93	11	-	-	PUNCT
bracis-33550	93	12	sp	sp	NOUN
bracis-33550	93	13	audio	audio	NOUN
bracis-33550	93	14	corpus	corpus	NOUN
bracis-33550	93	15	for	for	ADP
bracis-33550	93	16	 	 	SPACE
bracis-33550	93	17	evaluating	evaluate	VERB
bracis-33550	93	18	asr	asr	NOUN
bracis-33550	93	19	models	model	NOUN
bracis-33550	93	20	audio	audio	ADJ
bracis-33550	93	21	quality	quality	NOUN
bracis-33550	93	22	was	be	AUX
bracis-33550	93	23	a	a	DET
bracis-33550	93	24	very	very	ADV
bracis-33550	93	25	important	important	ADJ
bracis-33550	93	26	characteristic	characteristic	ADJ
bracis-33550	93	27	when	when	SCONJ
bracis-33550	93	28	preparing	prepare	VERB
bracis-33550	93	29	the	the	DET
bracis-33550	93	30	dataset	dataset	NOUN
bracis-33550	93	31	for	for	ADP
bracis-33550	93	32	evaluating	evaluate	VERB
bracis-33550	93	33	asr	asr	NOUN
bracis-33550	93	34	models	model	NOUN
bracis-33550	93	35	.	.	PUNCT
bracis-33550	94	1	in	in	ADP
bracis-33550	94	2	this	this	DET
bracis-33550	94	3	section	section	NOUN
bracis-33550	94	4	,	,	PUNCT
bracis-33550	94	5	we	we	PRON
bracis-33550	94	6	present	present	VERB
bracis-33550	94	7	the	the	DET
bracis-33550	94	8	processing	processing	NOUN
bracis-33550	94	9	steps	step	NOUN
bracis-33550	94	10	of	of	ADP
bracis-33550	94	11	the	the	DET
bracis-33550	94	12	nurc	nurc	ADJ
bracis-33550	94	13	-	-	PUNCT
bracis-33550	94	14	sp	sp	NOUN
bracis-33550	94	15	audio	audio	NOUN
bracis-33550	94	16	corpus	corpus	NOUN
bracis-33550	94	17	to	to	PART
bracis-33550	94	18	generate	generate	VERB
bracis-33550	94	19	the	the	DET
bracis-33550	94	20	version	version	NOUN
bracis-33550	94	21	used	use	VERB
bracis-33550	94	22	to	to	PART
bracis-33550	94	23	evaluate	evaluate	VERB
bracis-33550	94	24	the	the	DET
bracis-33550	94	25	four	four	NUM
bracis-33550	94	26	asr	asr	NOUN
bracis-33550	94	27	models	model	NOUN
bracis-33550	94	28	in	in	ADP
bracis-33550	94	29	this	this	DET
bracis-33550	94	30	paper	paper	NOUN
bracis-33550	94	31	:	:	PUNCT
bracis-33550	94	32	1	1	X
bracis-33550	94	33	.	.	X
bracis-33550	94	34	removal	removal	NOUN
bracis-33550	94	35	of	of	ADP
bracis-33550	94	36	those	those	DET
bracis-33550	94	37	audios	audio	NOUN
bracis-33550	94	38	(	(	PUNCT
bracis-33550	94	39	and	and	CCONJ
bracis-33550	94	40	all	all	PRON
bracis-33550	94	41	of	of	ADP
bracis-33550	94	42	its	its	PRON
bracis-33550	94	43	segments	segment	NOUN
bracis-33550	94	44	)	)	PUNCT
bracis-33550	94	45	with	with	ADP
bracis-33550	94	46	poor	poor	ADJ
bracis-33550	94	47	/	/	SYM
bracis-33550	94	48	fair	fair	ADJ
bracis-33550	94	49	quality	quality	NOUN
bracis-33550	94	50	(	(	PUNCT
bracis-33550	94	51	with	with	ADP
bracis-33550	94	52	high	high	ADJ
bracis-33550	94	53	number	number	NOUN
bracis-33550	94	54	of	of	ADP
bracis-33550	94	55	problems	problem	NOUN
bracis-33550	94	56	with	with	ADP
bracis-33550	94	57	voice	voice	NOUN
bracis-33550	94	58	distortion	distortion	NOUN
bracis-33550	94	59	,	,	PUNCT
bracis-33550	94	60	hissing	hissing	NOUN
bracis-33550	94	61	,	,	PUNCT
bracis-33550	94	62	background	background	NOUN
bracis-33550	94	63	noise	noise	NOUN
bracis-33550	94	64	and	and	CCONJ
bracis-33550	94	65	interruptions	interruption	NOUN
bracis-33550	94	66	)	)	PUNCT
bracis-33550	94	67	;	;	PUNCT
bracis-33550	94	68	2	2	X
bracis-33550	94	69	.	.	X
bracis-33550	94	70	annotation	annotation	NOUN
bracis-33550	94	71	of	of	ADP
bracis-33550	94	72	the	the	DET
bracis-33550	94	73	segments	segment	NOUN
bracis-33550	94	74	using	use	VERB
bracis-33550	94	75	“	"	PUNCT
bracis-33550	94	76	high	high	ADJ
bracis-33550	94	77	/	/	SYM
bracis-33550	94	78	low	low	ADJ
bracis-33550	94	79	”	"	PUNCT
bracis-33550	94	80	quality	quality	NOUN
bracis-33550	94	81	labels	label	NOUN
bracis-33550	94	82	for	for	ADP
bracis-33550	94	83	all	all	DET
bracis-33550	94	84	remaining	remain	VERB
bracis-33550	94	85	audios	audio	NOUN
bracis-33550	94	86	,	,	PUNCT
bracis-33550	94	87	based	base	VERB
bracis-33550	94	88	on	on	ADP
bracis-33550	94	89	what	what	PRON
bracis-33550	94	90	was	be	AUX
bracis-33550	94	91	reported	report	VERB
bracis-33550	94	92	by	by	ADP
bracis-33550	94	93	human	human	ADJ
bracis-33550	94	94	annotators	annotator	NOUN
bracis-33550	94	95	during	during	ADP
bracis-33550	94	96	the	the	DET
bracis-33550	94	97	automatic	automatic	ADJ
bracis-33550	94	98	transcription	transcription	NOUN
bracis-33550	94	99	revision	revision	NOUN
bracis-33550	94	100	;	;	PUNCT
bracis-33550	94	101	3	3	X
bracis-33550	94	102	.	.	X
bracis-33550	94	103	removal	removal	NOUN
bracis-33550	94	104	of	of	ADP
bracis-33550	94	105	part	part	NOUN
bracis-33550	94	106	of	of	ADP
bracis-33550	94	107	the	the	DET
bracis-33550	94	108	low	low	ADJ
bracis-33550	94	109	quality	quality	NOUN
bracis-33550	94	110	segments	segment	NOUN
bracis-33550	94	111	,	,	PUNCT
bracis-33550	94	112	as	as	SCONJ
bracis-33550	94	113	detailed	detail	VERB
bracis-33550	94	114	further	far	ADV
bracis-33550	94	115	in	in	ADP
bracis-33550	94	116	this	this	DET
bracis-33550	94	117	section	section	NOUN
bracis-33550	94	118	,	,	PUNCT
bracis-33550	94	119	in	in	ADP
bracis-33550	94	120	order	order	NOUN
bracis-33550	94	121	to	to	PART
bracis-33550	94	122	generate	generate	VERB
bracis-33550	94	123	the	the	DET
bracis-33550	94	124	final	final	ADJ
bracis-33550	94	125	dataset	dataset	NOUN
bracis-33550	94	126	for	for	ADP
bracis-33550	94	127	training	train	VERB
bracis-33550	94	128	asr	asr	NOUN
bracis-33550	94	129	models	model	NOUN
bracis-33550	94	130	;	;	PUNCT
bracis-33550	94	131	and	and	CCONJ
bracis-33550	94	132	4	4	X
bracis-33550	94	133	.	.	X
bracis-33550	94	134	generation	generation	NOUN
bracis-33550	94	135	of	of	ADP
bracis-33550	94	136	both	both	DET
bracis-33550	94	137	statistics	statistic	NOUN
bracis-33550	94	138	of	of	ADP
bracis-33550	94	139	nurc	nurc	ADJ
bracis-33550	94	140	-	-	PUNCT
bracis-33550	94	141	sp	sp	NOUN
bracis-33550	94	142	audio	audio	NOUN
bracis-33550	94	143	corpus	corpus	NOUN
bracis-33550	94	144	and	and	CCONJ
bracis-33550	94	145	the	the	DET
bracis-33550	94	146	train	train	NOUN
bracis-33550	94	147	,	,	PUNCT
bracis-33550	94	148	validation	validation	NOUN
bracis-33550	94	149	and	and	CCONJ
bracis-33550	94	150	test	test	NOUN
bracis-33550	94	151	sets	set	NOUN
bracis-33550	94	152	.	.	PUNCT
bracis-33550	95	1	of	of	ADP
bracis-33550	95	2	the	the	DET
bracis-33550	95	3	328	328	NUM
bracis-33550	95	4	audios	audio	NOUN
bracis-33550	95	5	,	,	PUNCT
bracis-33550	95	6	four	four	NUM
bracis-33550	95	7	of	of	ADP
bracis-33550	95	8	them	they	PRON
bracis-33550	95	9	had	have	VERB
bracis-33550	95	10	poor	poor	ADJ
bracis-33550	95	11	quality	quality	NOUN
bracis-33550	95	12	,	,	PUNCT
bracis-33550	95	13	92	92	NUM
bracis-33550	95	14	had	have	VERB
bracis-33550	95	15	fair	fair	ADJ
bracis-33550	95	16	quality	quality	NOUN
bracis-33550	95	17	,	,	PUNCT
bracis-33550	95	18	52	52	NUM
bracis-33550	95	19	had	have	VERB
bracis-33550	95	20	good	good	ADJ
bracis-33550	95	21	quality	quality	NOUN
bracis-33550	95	22	and	and	CCONJ
bracis-33550	95	23	180	180	NUM
bracis-33550	95	24	had	have	VERB
bracis-33550	95	25	excellent	excellent	ADJ
bracis-33550	95	26	quality	quality	NOUN
bracis-33550	95	27	.	.	PUNCT
bracis-33550	96	1	therefore	therefore	ADV
bracis-33550	96	2	,	,	PUNCT
bracis-33550	96	3	of	of	ADP
bracis-33550	96	4	these	these	DET
bracis-33550	96	5	328	328	NUM
bracis-33550	96	6	audios	audio	NOUN
bracis-33550	96	7	,	,	PUNCT
bracis-33550	96	8	five	five	NUM
bracis-33550	96	9	of	of	ADP
bracis-33550	96	10	them	they	PRON
bracis-33550	96	11	were	be	AUX
bracis-33550	96	12	not	not	PART
bracis-33550	96	13	used	use	VERB
bracis-33550	96	14	for	for	ADP
bracis-33550	96	15	training	train	VERB
bracis-33550	96	16	the	the	DET
bracis-33550	96	17	asr	asr	NOUN
bracis-33550	96	18	models	model	NOUN
bracis-33550	96	19	in	in	ADP
bracis-33550	96	20	this	this	DET
bracis-33550	96	21	paper	paper	NOUN
bracis-33550	96	22	:	:	PUNCT
bracis-33550	96	23	sp_ef_395	sp_ef_395	NOUN
bracis-33550	96	24	,	,	PUNCT
bracis-33550	96	25	sp_did_283	sp_did_283	PROPN
bracis-33550	96	26	,	,	PUNCT
bracis-33550	96	27	sp_did_194	sp_did_194	PROPN
bracis-33550	96	28	,	,	PUNCT
bracis-33550	96	29	sp_d2_397	sp_d2_397	PROPN
bracis-33550	96	30	,	,	PUNCT
bracis-33550	96	31	sp_d2_337	sp_d2_337	PROPN
bracis-33550	96	32	,	,	PUNCT
bracis-33550	96	33	as	as	SCONJ
bracis-33550	96	34	they	they	PRON
bracis-33550	96	35	had	have	VERB
bracis-33550	96	36	problems	problem	NOUN
bracis-33550	96	37	with	with	ADP
bracis-33550	96	38	voice	voice	NOUN
bracis-33550	96	39	distortion	distortion	NOUN
bracis-33550	96	40	,	,	PUNCT
bracis-33550	96	41	hissing	hissing	NOUN
bracis-33550	96	42	,	,	PUNCT
bracis-33550	96	43	background	background	NOUN
bracis-33550	96	44	noise	noise	NOUN
bracis-33550	96	45	and	and	CCONJ
bracis-33550	96	46	interruptions	interruption	NOUN
bracis-33550	96	47	,	,	PUNCT
bracis-33550	96	48	as	as	SCONJ
bracis-33550	96	49	they	they	PRON
bracis-33550	96	50	were	be	AUX
bracis-33550	96	51	responsible	responsible	ADJ
bracis-33550	96	52	for	for	ADP
bracis-33550	96	53	generating	generate	VERB
bracis-33550	96	54	transcriptions	transcription	NOUN
bracis-33550	96	55	with	with	ADP
bracis-33550	96	56	a	a	DET
bracis-33550	96	57	high	high	ADJ
bracis-33550	96	58	rate	rate	NOUN
bracis-33550	96	59	of	of	ADP
bracis-33550	96	60	errors	error	NOUN
bracis-33550	96	61	.	.	PUNCT
bracis-33550	97	1	the	the	DET
bracis-33550	97	2	transcriptions	transcription	NOUN
bracis-33550	97	3	were	be	AUX
bracis-33550	97	4	processed	process	VERB
bracis-33550	97	5	to	to	PART
bracis-33550	97	6	remove	remove	VERB
bracis-33550	97	7	the	the	DET
bracis-33550	97	8	labels	label	NOUN
bracis-33550	97	9	used	use	VERB
bracis-33550	97	10	in	in	ADP
bracis-33550	97	11	the	the	DET
bracis-33550	97	12	automatic	automatic	ADJ
bracis-33550	97	13	transcription	transcription	NOUN
bracis-33550	97	14	revision	revision	NOUN
bracis-33550	97	15	:	:	PUNCT
bracis-33550	97	16	(	(	PUNCT
bracis-33550	97	17	i	i	NOUN
bracis-33550	97	18	)	)	PUNCT
bracis-33550	97	19	“	"	PUNCT
bracis-33550	97	20	#	#	SYM
bracis-33550	97	21	#	#	SYM
bracis-33550	97	22	#	#	NOUN
bracis-33550	97	23	”	"	PUNCT
bracis-33550	97	24	was	be	AUX
bracis-33550	97	25	used	use	VERB
bracis-33550	97	26	to	to	PART
bracis-33550	97	27	indicate	indicate	VERB
bracis-33550	97	28	segments	segment	NOUN
bracis-33550	97	29	with	with	ADP
bracis-33550	97	30	high	high	ADJ
bracis-33550	97	31	background	background	NOUN
bracis-33550	97	32	noise	noise	NOUN
bracis-33550	97	33	,	,	PUNCT
bracis-33550	97	34	very	very	ADV
bracis-33550	97	35	low	low	ADJ
bracis-33550	97	36	audio	audio	NOUN
bracis-33550	97	37	,	,	PUNCT
bracis-33550	97	38	overlapping	overlap	VERB
bracis-33550	97	39	voices	voice	NOUN
bracis-33550	97	40	and	and	CCONJ
bracis-33550	97	41	music	music	NOUN
bracis-33550	97	42	;	;	PUNCT
bracis-33550	97	43	(	(	PUNCT
bracis-33550	97	44	ii	ii	NOUN
bracis-33550	97	45	)	)	PUNCT
bracis-33550	97	46	paralinguistic	paralinguistic	NOUN
bracis-33550	97	47	sounds	sound	VERB
bracis-33550	97	48	such	such	ADJ
bracis-33550	97	49	as	as	ADP
bracis-33550	97	50	laughter	laughter	NOUN
bracis-33550	97	51	or	or	CCONJ
bracis-33550	97	52	cough	cough	NOUN
bracis-33550	97	53	were	be	AUX
bracis-33550	97	54	annotated	annotate	VERB
bracis-33550	97	55	in	in	ADP
bracis-33550	97	56	parentheses—(laughter	parentheses—(laughter	NOUN
bracis-33550	97	57	)	)	PUNCT
bracis-33550	97	58	,	,	PUNCT
bracis-33550	97	59	(	(	PUNCT
bracis-33550	97	60	cough	cough	NOUN
bracis-33550	97	61	)	)	PUNCT
bracis-33550	97	62	;	;	PUNCT
bracis-33550	97	63	(	(	PUNCT
bracis-33550	97	64	iii	iii	X
bracis-33550	97	65	)	)	PUNCT
bracis-33550	97	66	misunderstanding	misunderstanding	NOUN
bracis-33550	97	67	of	of	ADP
bracis-33550	97	68	words	word	NOUN
bracis-33550	97	69	or	or	CCONJ
bracis-33550	97	70	passages	passage	NOUN
bracis-33550	97	71	were	be	AUX
bracis-33550	97	72	also	also	ADV
bracis-33550	97	73	marked	mark	VERB
bracis-33550	97	74	with	with	ADP
bracis-33550	97	75	parentheses	parenthesis	NOUN
bracis-33550	97	76	“	"	PUNCT
bracis-33550	97	77	(	(	PUNCT
bracis-33550	97	78	)	)	PUNCT
bracis-33550	97	79	”	"	PUNCT
bracis-33550	97	80	;	;	PUNCT
bracis-33550	97	81	and	and	CCONJ
bracis-33550	97	82	(	(	PUNCT
bracis-33550	97	83	iv	iv	X
bracis-33550	97	84	)	)	PUNCT
bracis-33550	97	85	words	word	NOUN
bracis-33550	97	86	truncated	truncate	VERB
bracis-33550	97	87	at	at	ADP
bracis-33550	97	88	the	the	DET
bracis-33550	97	89	end	end	NOUN
bracis-33550	97	90	or	or	CCONJ
bracis-33550	97	91	the	the	DET
bracis-33550	97	92	beginning	beginning	NOUN
bracis-33550	97	93	of	of	ADP
bracis-33550	97	94	the	the	DET
bracis-33550	97	95	audio	audio	NOUN
bracis-33550	97	96	due	due	ADJ
bracis-33550	97	97	to	to	ADP
bracis-33550	97	98	automatic	automatic	ADJ
bracis-33550	97	99	segmentation	segmentation	NOUN
bracis-33550	97	100	failure	failure	NOUN
bracis-33550	97	101	were	be	AUX
bracis-33550	97	102	partially	partially	ADV
bracis-33550	97	103	transcribed	transcribe	VERB
bracis-33550	97	104	with	with	ADP
bracis-33550	97	105	the	the	DET
bracis-33550	97	106	help	help	NOUN
bracis-33550	97	107	of	of	ADP
bracis-33550	97	108	“	"	PUNCT
bracis-33550	97	109	>	>	PUNCT
bracis-33550	97	110	”	"	PUNCT
bracis-33550	97	111	and	and	CCONJ
bracis-33550	97	112	“	"	PUNCT
bracis-33550	97	113	<	<	X
bracis-33550	97	114	”	"	PUNCT
bracis-33550	97	115	(	(	PUNCT
bracis-33550	97	116	for	for	ADP
bracis-33550	97	117	example	example	NOUN
bracis-33550	97	118	:	:	PUNCT
bracis-33550	97	119	if	if	SCONJ
bracis-33550	97	120	the	the	DET
bracis-33550	97	121	word	word	NOUN
bracis-33550	97	122	“	"	PUNCT
bracis-33550	97	123	casa	casa	PROPN
bracis-33550	97	124	”	"	PUNCT
bracis-33550	97	125	–	–	PUNCT
bracis-33550	97	126	house	house	NOUN
bracis-33550	97	127	)	)	PUNCT
bracis-33550	97	128	where	where	SCONJ
bracis-33550	97	129	truncated	truncate	VERB
bracis-33550	97	130	at	at	ADP
bracis-33550	97	131	the	the	DET
bracis-33550	97	132	end	end	NOUN
bracis-33550	97	133	of	of	ADP
bracis-33550	97	134	the	the	DET
bracis-33550	97	135	segment	segment	NOUN
bracis-33550	97	136	,	,	PUNCT
bracis-33550	97	137	it	it	PRON
bracis-33550	97	138	was	be	AUX
bracis-33550	97	139	annotated	annotate	VERB
bracis-33550	97	140	as	as	ADP
bracis-33550	97	141	“	"	PUNCT
bracis-33550	97	142	ca	ca	NOUN
bracis-33550	97	143	>	>	PUNCT
bracis-33550	97	144	”	"	PUNCT
bracis-33550	97	145	and	and	CCONJ
bracis-33550	97	146	if	if	SCONJ
bracis-33550	97	147	it	it	PRON
bracis-33550	97	148	was	be	AUX
bracis-33550	97	149	at	at	ADP
bracis-33550	97	150	the	the	DET
bracis-33550	97	151	beginning	beginning	NOUN
bracis-33550	97	152	,	,	PUNCT
bracis-33550	97	153	it	it	PRON
bracis-33550	97	154	was	be	AUX
bracis-33550	97	155	annotated	annotate	VERB
bracis-33550	97	156	as	as	ADP
bracis-33550	97	157	“	"	PUNCT
bracis-33550	97	158	<	<	X
bracis-33550	97	159	sa	sa	X
bracis-33550	97	160	”	"	PUNCT
bracis-33550	97	161	.	.	PUNCT
bracis-33550	98	1	segments	segment	NOUN
bracis-33550	98	2	marked	mark	VERB
bracis-33550	98	3	with	with	ADP
bracis-33550	98	4	“	"	PUNCT
bracis-33550	98	5	#	#	SYM
bracis-33550	98	6	#	#	SYM
bracis-33550	98	7	#	#	NOUN
bracis-33550	98	8	”	"	PUNCT
bracis-33550	98	9	are	be	AUX
bracis-33550	98	10	removed	remove	VERB
bracis-33550	98	11	completely	completely	ADV
bracis-33550	98	12	,	,	PUNCT
bracis-33550	98	13	as	as	ADV
bracis-33550	98	14	well	well	ADV
bracis-33550	98	15	as	as	ADP
bracis-33550	98	16	segments	segment	NOUN
bracis-33550	98	17	having	have	VERB
bracis-33550	98	18	only	only	ADV
bracis-33550	98	19	paralinguistic	paralinguistic	ADJ
bracis-33550	98	20	sounds	sound	NOUN
bracis-33550	98	21	(	(	PUNCT
bracis-33550	98	22	e.g.	e.g.	ADV
bracis-33550	98	23	,	,	PUNCT
bracis-33550	98	24	laughter	laughter	NOUN
bracis-33550	98	25	)	)	PUNCT
bracis-33550	98	26	or	or	CCONJ
bracis-33550	98	27	containing	contain	VERB
bracis-33550	98	28	only	only	ADV
bracis-33550	98	29	truncated	truncate	VERB
bracis-33550	98	30	words	word	NOUN
bracis-33550	98	31	.	.	PUNCT
bracis-33550	99	1	segments	segment	NOUN
bracis-33550	99	2	having	having	AUX
bracis-33550	99	3	truncated	truncate	VERB
bracis-33550	99	4	and	and	CCONJ
bracis-33550	99	5	non	non	ADJ
bracis-33550	99	6	-	-	ADJ
bracis-33550	99	7	truncated	truncated	ADJ
bracis-33550	99	8	words	word	NOUN
bracis-33550	99	9	(	(	PUNCT
bracis-33550	99	10	marked	mark	VERB
bracis-33550	99	11	with	with	ADP
bracis-33550	99	12	“	"	PUNCT
bracis-33550	99	13	>	>	PUNCT
bracis-33550	99	14	”	"	PUNCT
bracis-33550	99	15	or	or	CCONJ
bracis-33550	99	16	“	"	PUNCT
bracis-33550	99	17	<	<	X
bracis-33550	99	18	”	"	PUNCT
bracis-33550	99	19	)	)	PUNCT
bracis-33550	99	20	had	have	VERB
bracis-33550	99	21	the	the	DET
bracis-33550	99	22	truncated	truncated	ADJ
bracis-33550	99	23	word	word	NOUN
bracis-33550	99	24	removed	remove	VERB
bracis-33550	99	25	and	and	CCONJ
bracis-33550	99	26	were	be	AUX
bracis-33550	99	27	marked	mark	VERB
bracis-33550	99	28	as	as	ADP
bracis-33550	99	29	low	low	ADJ
bracis-33550	99	30	quality	quality	NOUN
bracis-33550	99	31	.	.	PUNCT
bracis-33550	100	1	pairs	pair	NOUN
bracis-33550	100	2	that	that	PRON
bracis-33550	100	3	have	have	VERB
bracis-33550	100	4	words	word	NOUN
bracis-33550	100	5	or	or	CCONJ
bracis-33550	100	6	passages	passage	NOUN
bracis-33550	100	7	marked	mark	VERB
bracis-33550	100	8	as	as	ADP
bracis-33550	100	9	misunderstood	misunderstood	NOUN
bracis-33550	100	10	,	,	PUNCT
bracis-33550	100	11	using	use	VERB
bracis-33550	100	12	the	the	DET
bracis-33550	100	13	“	"	PUNCT
bracis-33550	100	14	(	(	PUNCT
bracis-33550	100	15	)	)	PUNCT
bracis-33550	100	16	”	"	PUNCT
bracis-33550	100	17	tags	tag	NOUN
bracis-33550	100	18	,	,	PUNCT
bracis-33550	100	19	had	have	VERB
bracis-33550	100	20	the	the	DET
bracis-33550	100	21	tags	tag	NOUN
bracis-33550	100	22	removed	remove	VERB
bracis-33550	100	23	and	and	CCONJ
bracis-33550	100	24	the	the	DET
bracis-33550	100	25	pairs	pair	NOUN
bracis-33550	100	26	were	be	AUX
bracis-33550	100	27	marked	mark	VERB
bracis-33550	100	28	as	as	ADP
bracis-33550	100	29	“	"	PUNCT
bracis-33550	100	30	low	low	ADJ
bracis-33550	100	31	quality	quality	NOUN
bracis-33550	100	32	”	"	PUNCT
bracis-33550	100	33	.	.	PUNCT
bracis-33550	101	1	audio	audio	ADJ
bracis-33550	101	2	-	-	PUNCT
bracis-33550	101	3	transcription	transcription	NOUN
bracis-33550	101	4	pairs	pair	NOUN
bracis-33550	101	5	with	with	ADP
bracis-33550	101	6	paralinguistic	paralinguistic	PROPN
bracis-33550	101	7	sounds	sound	VERB
bracis-33550	101	8	tags	tag	NOUN
bracis-33550	101	9	only	only	ADV
bracis-33550	101	10	had	have	VERB
bracis-33550	101	11	the	the	DET
bracis-33550	101	12	tags	tag	NOUN
bracis-33550	101	13	removed	remove	VERB
bracis-33550	101	14	and	and	CCONJ
bracis-33550	101	15	were	be	AUX
bracis-33550	101	16	marked	mark	VERB
bracis-33550	101	17	as	as	ADP
bracis-33550	101	18	“	"	PUNCT
bracis-33550	101	19	high	high	ADJ
bracis-33550	101	20	quality	quality	NOUN
bracis-33550	101	21	”	"	PUNCT
bracis-33550	101	22	.	.	PUNCT
bracis-33550	102	1	the	the	DET
bracis-33550	102	2	dataset	dataset	NOUN
bracis-33550	102	3	available	available	ADJ
bracis-33550	102	4	in	in	ADP
bracis-33550	102	5	this	this	DET
bracis-33550	102	6	work	work	NOUN
bracis-33550	102	7	has	have	VERB
bracis-33550	102	8	323	323	NUM
bracis-33550	102	9	audio	audio	ADJ
bracis-33550	102	10	-	-	PUNCT
bracis-33550	102	11	transcription	transcription	NOUN
bracis-33550	102	12	pairs	pair	NOUN
bracis-33550	102	13	,	,	PUNCT
bracis-33550	102	14	which	which	PRON
bracis-33550	102	15	were	be	AUX
bracis-33550	102	16	divided	divide	VERB
bracis-33550	102	17	into	into	ADP
bracis-33550	102	18	177,224	177,224	NUM
bracis-33550	102	19	segments	segment	NOUN
bracis-33550	102	20	transcribed	transcribe	VERB
bracis-33550	102	21	by	by	ADP
bracis-33550	102	22	whisperx	whisperx	NOUN
bracis-33550	102	23	,	,	PUNCT
bracis-33550	102	24	and	and	CCONJ
bracis-33550	102	25	revised	revise	VERB
bracis-33550	102	26	by	by	ADP
bracis-33550	102	27	native	native	ADJ
bracis-33550	102	28	bp	bp	PROPN
bracis-33550	102	29	speakers	speaker	NOUN
bracis-33550	102	30	(	(	PUNCT
bracis-33550	102	31	see	see	VERB
bracis-33550	102	32	more	more	ADJ
bracis-33550	102	33	details	detail	NOUN
bracis-33550	102	34	in	in	ADP
bracis-33550	102	35	table	table	NOUN
bracis-33550	102	36	 	 	SPACE
bracis-33550	102	37	2	2	NUM
bracis-33550	102	38	)	)	PUNCT
bracis-33550	102	39	.	.	PUNCT
bracis-33550	103	1	3.2	3.2	NUM
bracis-33550	103	2	statistics	statistic	NOUN
bracis-33550	103	3	of	of	ADP
bracis-33550	103	4	 	 	SPACE
bracis-33550	103	5	nurc	nurc	ADJ
bracis-33550	103	6	-	-	PUNCT
bracis-33550	103	7	sp	sp	NOUN
bracis-33550	103	8	audio	audio	NOUN
bracis-33550	103	9	corpus	corpus	NOUN
bracis-33550	103	10	table	table	NOUN
bracis-33550	103	11	 	 	SPACE
bracis-33550	103	12	2	2	NUM
bracis-33550	103	13	presents	present	NOUN
bracis-33550	103	14	statistics	statistic	NOUN
bracis-33550	103	15	of	of	ADP
bracis-33550	103	16	nurc	nurc	ADJ
bracis-33550	103	17	-	-	PUNCT
bracis-33550	103	18	sp	sp	NOUN
bracis-33550	103	19	audio	audio	NOUN
bracis-33550	103	20	corpus	corpus	NOUN
bracis-33550	103	21	dataset	dataset	NOUN
bracis-33550	103	22	.	.	PUNCT
bracis-33550	104	1	overall	overall	ADJ
bracis-33550	104	2	,	,	PUNCT
bracis-33550	104	3	almost	almost	ADV
bracis-33550	104	4	240	240	NUM
bracis-33550	104	5	 	 	SPACE
bracis-33550	104	6	h	h	NOUN
bracis-33550	104	7	of	of	ADP
bracis-33550	104	8	speech	speech	NOUN
bracis-33550	104	9	were	be	AUX
bracis-33550	104	10	generated	generate	VERB
bracis-33550	104	11	in	in	ADP
bracis-33550	104	12	the	the	DET
bracis-33550	104	13	final	final	ADJ
bracis-33550	104	14	export	export	NOUN
bracis-33550	104	15	,	,	PUNCT
bracis-33550	104	16	distributed	distribute	VERB
bracis-33550	104	17	in	in	ADP
bracis-33550	104	18	around	around	ADP
bracis-33550	104	19	177	177	NUM
bracis-33550	104	20	thousand	thousand	NUM
bracis-33550	104	21	segments	segment	NOUN
bracis-33550	104	22	with	with	ADP
bracis-33550	104	23	an	an	DET
bracis-33550	104	24	average	average	ADJ
bracis-33550	104	25	duration	duration	NOUN
bracis-33550	104	26	of	of	ADP
bracis-33550	104	27	approximately	approximately	ADV
bracis-33550	104	28	5	5	NUM
bracis-33550	104	29	 	 	SPACE
bracis-33550	104	30	s.	s.	PROPN
bracis-33550	104	31	in	in	ADP
bracis-33550	104	32	total	total	ADJ
bracis-33550	104	33	,	,	PUNCT
bracis-33550	104	34	more	more	ADJ
bracis-33550	104	35	than	than	ADP
bracis-33550	104	36	2	2	NUM
bracis-33550	104	37	million	million	NUM
bracis-33550	104	38	tokens	token	NOUN
bracis-33550	104	39	were	be	AUX
bracis-33550	104	40	transcribed	transcribe	VERB
bracis-33550	104	41	,	,	PUNCT
bracis-33550	104	42	resulting	result	VERB
bracis-33550	104	43	in	in	ADP
bracis-33550	104	44	approximately	approximately	ADV
bracis-33550	104	45	12	12	NUM
bracis-33550	104	46	tokens	token	NOUN
bracis-33550	104	47	per	per	ADP
bracis-33550	104	48	segment	segment	NOUN
bracis-33550	104	49	.	.	PUNCT
bracis-33550	105	1	it	it	PRON
bracis-33550	105	2	should	should	AUX
bracis-33550	105	3	be	be	AUX
bracis-33550	105	4	noted	note	VERB
bracis-33550	105	5	that	that	SCONJ
bracis-33550	105	6	the	the	DET
bracis-33550	105	7	sum	sum	NOUN
bracis-33550	105	8	of	of	ADP
bracis-33550	105	9	audios	audio	NOUN
bracis-33550	105	10	with	with	ADP
bracis-33550	105	11	females	female	NOUN
bracis-33550	105	12	and	and	CCONJ
bracis-33550	105	13	males	male	NOUN
bracis-33550	105	14	is	be	AUX
bracis-33550	105	15	higher	high	ADJ
bracis-33550	105	16	than	than	ADP
bracis-33550	105	17	the	the	DET
bracis-33550	105	18	total	total	ADJ
bracis-33550	105	19	number	number	NOUN
bracis-33550	105	20	of	of	ADP
bracis-33550	105	21	audios	audio	NOUN
bracis-33550	105	22	,	,	PUNCT
bracis-33550	105	23	because	because	SCONJ
bracis-33550	105	24	a	a	DET
bracis-33550	105	25	given	give	VERB
bracis-33550	105	26	audio	audio	NOUN
bracis-33550	105	27	can	can	AUX
bracis-33550	105	28	have	have	VERB
bracis-33550	105	29	simultaneously	simultaneously	ADV
bracis-33550	105	30	both	both	DET
bracis-33550	105	31	types	type	NOUN
bracis-33550	105	32	of	of	ADP
bracis-33550	105	33	voices	voice	NOUN
bracis-33550	105	34	.	.	PUNCT
bracis-33550	106	1	table	table	NOUN
bracis-33550	106	2	2	2	NUM
bracis-33550	106	3	.	.	PUNCT
bracis-33550	107	1	nurc	nurc	ADJ
bracis-33550	107	2	-	-	PUNCT
bracis-33550	107	3	sp	sp	NOUN
bracis-33550	107	4	audio	audio	NOUN
bracis-33550	107	5	corpus	corpus	NOUN
bracis-33550	107	6	statisticsfull	statisticsfull	NOUN
bracis-33550	107	7	size	size	NOUN
bracis-33550	107	8	table	table	NOUN
bracis-33550	107	9	figure	figure	NOUN
bracis-33550	107	10	 	 	SPACE
bracis-33550	107	11	1	1	NUM
bracis-33550	107	12	shows	show	VERB
bracis-33550	107	13	the	the	DET
bracis-33550	107	14	duration	duration	NOUN
bracis-33550	107	15	of	of	ADP
bracis-33550	107	16	audio	audio	ADJ
bracis-33550	107	17	segments	segment	NOUN
bracis-33550	107	18	.	.	PUNCT
bracis-33550	108	1	it	it	PRON
bracis-33550	108	2	can	can	AUX
bracis-33550	108	3	be	be	AUX
bracis-33550	108	4	noted	note	VERB
bracis-33550	108	5	that	that	SCONJ
bracis-33550	108	6	audios	audio	NOUN
bracis-33550	108	7	smaller	small	ADJ
bracis-33550	108	8	than	than	ADP
bracis-33550	108	9	5	5	NUM
bracis-33550	108	10	are	be	AUX
bracis-33550	108	11	predominant	predominant	ADJ
bracis-33550	108	12	,	,	PUNCT
bracis-33550	108	13	with	with	ADP
bracis-33550	108	14	a	a	DET
bracis-33550	108	15	peak	peak	NOUN
bracis-33550	108	16	in	in	ADP
bracis-33550	108	17	approximately	approximately	ADV
bracis-33550	108	18	3	3	NUM
bracis-33550	108	19	 	 	SPACE
bracis-33550	108	20	s	s	NOUN
bracis-33550	108	21	,	,	PUNCT
bracis-33550	108	22	while	while	SCONJ
bracis-33550	108	23	the	the	DET
bracis-33550	108	24	interval	interval	NOUN
bracis-33550	108	25	from	from	ADP
bracis-33550	108	26	5	5	NUM
bracis-33550	108	27	to	to	PART
bracis-33550	108	28	20	20	NUM
bracis-33550	108	29	 	 	SPACE
bracis-33550	108	30	s	s	VERB
bracis-33550	108	31	is	be	AUX
bracis-33550	108	32	also	also	ADV
bracis-33550	108	33	well	well	ADV
bracis-33550	108	34	represented	represent	VERB
bracis-33550	108	35	.	.	PUNCT
bracis-33550	109	1	there	there	PRON
bracis-33550	109	2	are	be	VERB
bracis-33550	109	3	also	also	ADV
bracis-33550	109	4	audios	audio	NOUN
bracis-33550	109	5	longer	long	ADV
bracis-33550	109	6	than	than	ADP
bracis-33550	109	7	20	20	NUM
bracis-33550	109	8	 	 	SPACE
bracis-33550	109	9	s	s	PROPN
bracis-33550	109	10	,	,	PUNCT
bracis-33550	109	11	but	but	CCONJ
bracis-33550	109	12	their	their	PRON
bracis-33550	109	13	occurrence	occurrence	NOUN
bracis-33550	109	14	is	be	AUX
bracis-33550	109	15	less	less	ADV
bracis-33550	109	16	common	common	ADJ
bracis-33550	109	17	.	.	PUNCT
bracis-33550	110	1	figure	figure	NOUN
bracis-33550	110	2	 	 	SPACE
bracis-33550	110	3	2	2	NUM
bracis-33550	110	4	shows	show	VERB
bracis-33550	110	5	the	the	DET
bracis-33550	110	6	number	number	NOUN
bracis-33550	110	7	of	of	ADP
bracis-33550	110	8	audios	audio	NOUN
bracis-33550	110	9	by	by	ADP
bracis-33550	110	10	age	age	NOUN
bracis-33550	110	11	and	and	CCONJ
bracis-33550	110	12	gender	gender	NOUN
bracis-33550	110	13	.	.	PUNCT
bracis-33550	111	1	the	the	DET
bracis-33550	111	2	number	number	NOUN
bracis-33550	111	3	of	of	ADP
bracis-33550	111	4	audios	audio	NOUN
bracis-33550	111	5	of	of	ADP
bracis-33550	111	6	females	female	NOUN
bracis-33550	111	7	and	and	CCONJ
bracis-33550	111	8	males	male	NOUN
bracis-33550	111	9	are	be	AUX
bracis-33550	111	10	similar	similar	ADJ
bracis-33550	111	11	,	,	PUNCT
bracis-33550	111	12	resulting	result	VERB
bracis-33550	111	13	reasonable	reasonable	ADJ
bracis-33550	111	14	balancing	balancing	NOUN
bracis-33550	111	15	regarding	regard	VERB
bracis-33550	111	16	gender	gender	NOUN
bracis-33550	111	17	,	,	PUNCT
bracis-33550	111	18	which	which	PRON
bracis-33550	111	19	can	can	AUX
bracis-33550	111	20	also	also	ADV
bracis-33550	111	21	be	be	AUX
bracis-33550	111	22	seem	seem	VERB
bracis-33550	111	23	in	in	ADP
bracis-33550	111	24	male	male	ADJ
bracis-33550	111	25	/	/	SYM
bracis-33550	111	26	female	female	ADJ
bracis-33550	111	27	ratio	ratio	NOUN
bracis-33550	111	28	in	in	ADP
bracis-33550	111	29	table	table	NOUN
bracis-33550	111	30	 	 	SPACE
bracis-33550	111	31	2	2	NUM
bracis-33550	111	32	.	.	PUNCT
bracis-33550	111	33	regarding	regard	VERB
bracis-33550	111	34	age	age	NOUN
bracis-33550	111	35	,	,	PUNCT
bracis-33550	111	36	there	there	PRON
bracis-33550	111	37	are	be	VERB
bracis-33550	111	38	more	more	ADJ
bracis-33550	111	39	speakers	speaker	NOUN
bracis-33550	111	40	in	in	ADP
bracis-33550	111	41	age	age	PROPN
bracis-33550	111	42	group	group	PROPN
bracis-33550	111	43	ii	ii	PROPN
bracis-33550	111	44	(	(	PUNCT
bracis-33550	111	45	from	from	ADP
bracis-33550	111	46	36	36	NUM
bracis-33550	111	47	to	to	PART
bracis-33550	111	48	55	55	NUM
bracis-33550	111	49	 	 	SPACE
bracis-33550	111	50	years	year	NOUN
bracis-33550	111	51	old	old	ADJ
bracis-33550	111	52	)	)	PUNCT
bracis-33550	111	53	than	than	ADP
bracis-33550	111	54	the	the	DET
bracis-33550	111	55	others	other	NOUN
bracis-33550	111	56	groups	group	NOUN
bracis-33550	111	57	(	(	PUNCT
bracis-33550	111	58	i	i	PRON
bracis-33550	111	59	 	 	SPACE
bracis-33550	111	60	=	=	NOUN
bracis-33550	111	61	 	 	SPACE
bracis-33550	111	62	25–35	25–35	NUM
bracis-33550	111	63	and	and	CCONJ
bracis-33550	111	64	iii	iii	NOUN
bracis-33550	111	65	 	 	SPACE
bracis-33550	111	66	=	=	NOUN
bracis-33550	111	67	 	 	SPACE
bracis-33550	111	68	36–55	36–55	NUM
bracis-33550	111	69	)	)	PUNCT
bracis-33550	111	70	.	.	PUNCT
bracis-33550	112	1	there	there	PRON
bracis-33550	112	2	also	also	ADV
bracis-33550	112	3	a	a	DET
bracis-33550	112	4	small	small	ADJ
bracis-33550	112	5	number	number	NOUN
bracis-33550	112	6	of	of	ADP
bracis-33550	112	7	audios	audio	NOUN
bracis-33550	112	8	with	with	ADP
bracis-33550	112	9	speakers	speaker	NOUN
bracis-33550	112	10	of	of	ADP
bracis-33550	112	11	unknown	unknown	ADJ
bracis-33550	112	12	age	age	NOUN
bracis-33550	112	13	.	.	PUNCT
bracis-33550	113	1	figure	figure	NOUN
bracis-33550	113	2	 	 	SPACE
bracis-33550	113	3	3	3	NUM
bracis-33550	113	4	shows	show	VERB
bracis-33550	113	5	the	the	DET
bracis-33550	113	6	number	number	NOUN
bracis-33550	113	7	of	of	ADP
bracis-33550	113	8	audios	audio	NOUN
bracis-33550	113	9	by	by	ADP
bracis-33550	113	10	speech	speech	NOUN
bracis-33550	113	11	gender	gender	NOUN
bracis-33550	113	12	(	(	PUNCT
bracis-33550	113	13	lectures	lecture	NOUN
bracis-33550	113	14	and	and	CCONJ
bracis-33550	113	15	talks—(ef	talks—(ef	NOUN
bracis-33550	113	16	)	)	PUNCT
bracis-33550	113	17	,	,	PUNCT
bracis-33550	113	18	interviews	interview	NOUN
bracis-33550	113	19	(	(	PUNCT
bracis-33550	113	20	did	do	AUX
bracis-33550	113	21	)	)	PUNCT
bracis-33550	113	22	and	and	CCONJ
bracis-33550	113	23	dialogs	dialog	NOUN
bracis-33550	113	24	(	(	PUNCT
bracis-33550	113	25	d2	d2	PROPN
bracis-33550	113	26	)	)	PUNCT
bracis-33550	113	27	,	,	PUNCT
bracis-33550	113	28	respectively	respectively	ADV
bracis-33550	113	29	)	)	PUNCT
bracis-33550	113	30	.	.	PUNCT
bracis-33550	114	1	interviews	interview	NOUN
bracis-33550	114	2	and	and	CCONJ
bracis-33550	114	3	dialogues	dialogue	NOUN
bracis-33550	114	4	are	be	AUX
bracis-33550	114	5	much	much	ADV
bracis-33550	114	6	more	more	ADV
bracis-33550	114	7	common	common	ADJ
bracis-33550	114	8	than	than	ADP
bracis-33550	114	9	lectures	lecture	NOUN
bracis-33550	114	10	,	,	PUNCT
bracis-33550	114	11	totalling	total	VERB
bracis-33550	114	12	90	90	NUM
bracis-33550	114	13	%	%	NOUN
bracis-33550	114	14	of	of	ADP
bracis-33550	114	15	the	the	DET
bracis-33550	114	16	audios	audio	NOUN
bracis-33550	114	17	.	.	PUNCT
bracis-33550	115	1	thus	thus	ADV
bracis-33550	115	2	,	,	PUNCT
bracis-33550	115	3	nurc	nurc	NOUN
bracis-33550	115	4	-	-	PUNCT
bracis-33550	115	5	sp	sp	NOUN
bracis-33550	115	6	is	be	AUX
bracis-33550	115	7	a	a	DET
bracis-33550	115	8	corpus	corpus	NOUN
bracis-33550	115	9	predominantly	predominantly	ADV
bracis-33550	115	10	made	make	VERB
bracis-33550	115	11	of	of	ADP
bracis-33550	115	12	spontaneous	spontaneous	ADJ
bracis-33550	115	13	speech	speech	NOUN
bracis-33550	115	14	.	.	PUNCT
bracis-33550	116	1	fig	fig	NOUN
bracis-33550	116	2	.	.	PUNCT
bracis-33550	117	1	1	1	NUM
bracis-33550	117	2	.	.	X
bracis-33550	117	3	audio	audio	ADJ
bracis-33550	117	4	segments	segment	NOUN
bracis-33550	117	5	duration	duration	NOUN
bracis-33550	117	6	full	full	ADJ
bracis-33550	117	7	size	size	NOUN
bracis-33550	117	8	image	image	NOUN
bracis-33550	117	9	fig	fig	NOUN
bracis-33550	117	10	.	.	PUNCT
bracis-33550	118	1	2	2	X
bracis-33550	118	2	.	.	X
bracis-33550	118	3	(	(	PUNCT
bracis-33550	118	4	a	a	X
bracis-33550	118	5	)	)	PUNCT
bracis-33550	118	6	audios	audio	NOUN
bracis-33550	118	7	by	by	ADP
bracis-33550	118	8	speaker	speaker	NOUN
bracis-33550	118	9	’s	’s	PART
bracis-33550	118	10	age	age	NOUN
bracis-33550	118	11	range	range	NOUN
bracis-33550	118	12	and	and	CCONJ
bracis-33550	118	13	(	(	PUNCT
bracis-33550	118	14	b	b	NOUN
bracis-33550	118	15	)	)	PUNCT
bracis-33550	118	16	gender	gender	NOUN
bracis-33550	118	17	(	(	PUNCT
bracis-33550	118	18	m	m	NOUN
bracis-33550	118	19	/	/	SYM
bracis-33550	118	20	f	f	NOUN
bracis-33550	118	21	)	)	PUNCT
bracis-33550	118	22	.	.	PUNCT
bracis-33550	119	1	speakers	speaker	NOUN
bracis-33550	119	2	are	be	AUX
bracis-33550	119	3	distributed	distribute	VERB
bracis-33550	119	4	into	into	ADP
bracis-33550	119	5	three	three	NUM
bracis-33550	119	6	age	age	NOUN
bracis-33550	119	7	groups	group	NOUN
bracis-33550	119	8	(	(	PUNCT
bracis-33550	119	9	i	i	NOUN
bracis-33550	119	10	=	=	NOUN
bracis-33550	119	11	25–35	25–35	NUM
bracis-33550	119	12	,	,	PUNCT
bracis-33550	119	13	ii	ii	PROPN
bracis-33550	119	14	=	=	SYM
bracis-33550	119	15	36–55	36–55	NUM
bracis-33550	119	16	,	,	PUNCT
bracis-33550	119	17	and	and	CCONJ
bracis-33550	119	18	iii	iii	X
bracis-33550	119	19	=	=	SYM
bracis-33550	119	20	56	56	NUM
bracis-33550	119	21	onwards	onward	NOUN
bracis-33550	119	22	)	)	PUNCT
bracis-33550	119	23	.	.	PUNCT
bracis-33550	120	1	full	full	ADJ
bracis-33550	120	2	size	size	NOUN
bracis-33550	120	3	image	image	NOUN
bracis-33550	120	4	fig	fig	NOUN
bracis-33550	120	5	.	.	PUNCT
bracis-33550	121	1	3	3	X
bracis-33550	121	2	.	.	X
bracis-33550	121	3	audios	audio	NOUN
bracis-33550	121	4	by	by	ADP
bracis-33550	121	5	speech	speech	NOUN
bracis-33550	121	6	genre	genre	NOUN
bracis-33550	121	7	.	.	PUNCT
bracis-33550	122	1	lectures	lecture	NOUN
bracis-33550	122	2	and	and	CCONJ
bracis-33550	122	3	talks	talk	NOUN
bracis-33550	122	4	receive	receive	VERB
bracis-33550	122	5	the	the	DET
bracis-33550	122	6	acronym	acronym	NOUN
bracis-33550	122	7	ef	ef	NOUN
bracis-33550	122	8	,	,	PUNCT
bracis-33550	122	9	interviews	interview	NOUN
bracis-33550	122	10	are	be	AUX
bracis-33550	122	11	called	call	VERB
bracis-33550	122	12	did	do	AUX
bracis-33550	122	13	and	and	CCONJ
bracis-33550	122	14	dialogues	dialogue	NOUN
bracis-33550	122	15	between	between	ADP
bracis-33550	122	16	two	two	NUM
bracis-33550	122	17	participants	participant	NOUN
bracis-33550	122	18	are	be	AUX
bracis-33550	122	19	named	name	VERB
bracis-33550	122	20	d2	d2	PROPN
bracis-33550	122	21	.	.	PUNCT
bracis-33550	123	1	full	full	ADJ
bracis-33550	123	2	size	size	NOUN
bracis-33550	123	3	image	image	NOUN
bracis-33550	123	4	4	4	NUM
bracis-33550	123	5	experiments	experiment	NOUN
bracis-33550	123	6	on	on	ADP
bracis-33550	123	7	 	 	SPACE
bracis-33550	123	8	nurc	nurc	ADJ
bracis-33550	123	9	-	-	PUNCT
bracis-33550	123	10	sp	sp	NOUN
bracis-33550	123	11	audio	audio	NOUN
bracis-33550	123	12	corpus	corpus	NOUN
bracis-33550	123	13	we	we	PRON
bracis-33550	123	14	performed	perform	VERB
bracis-33550	123	15	experiments	experiment	NOUN
bracis-33550	123	16	over	over	ADP
bracis-33550	123	17	nurc	nurc	ADJ
bracis-33550	123	18	-	-	PUNCT
bracis-33550	123	19	sp	sp	NOUN
bracis-33550	123	20	audio	audio	NOUN
bracis-33550	123	21	corpus	corpus	NOUN
bracis-33550	123	22	in	in	ADP
bracis-33550	123	23	order	order	NOUN
bracis-33550	123	24	to	to	PART
bracis-33550	123	25	assess	assess	VERB
bracis-33550	123	26	the	the	DET
bracis-33550	123	27	dataset	dataset	ADJ
bracis-33550	123	28	quality	quality	NOUN
bracis-33550	123	29	and	and	CCONJ
bracis-33550	123	30	limitations	limitation	NOUN
bracis-33550	123	31	.	.	PUNCT
bracis-33550	124	1	we	we	PRON
bracis-33550	124	2	based	base	VERB
bracis-33550	124	3	our	our	PRON
bracis-33550	124	4	experiments	experiment	NOUN
bracis-33550	124	5	in	in	ADP
bracis-33550	124	6	two	two	NUM
bracis-33550	124	7	main	main	ADJ
bracis-33550	124	8	architectures	architecture	NOUN
bracis-33550	124	9	:	:	PUNCT
bracis-33550	125	1	wav2vec2	wav2vec2	PROPN
bracis-33550	126	1	[	[	X
bracis-33550	126	2	8	8	NUM
bracis-33550	126	3	]	]	PUNCT
bracis-33550	126	4	and	and	CCONJ
bracis-33550	126	5	distil	distil	NOUN
bracis-33550	126	6	-	-	PUNCT
bracis-33550	126	7	whisper	whisper	NOUN
bracis-33550	127	1	[	[	X
bracis-33550	127	2	11	11	NUM
bracis-33550	127	3	]	]	PUNCT
bracis-33550	127	4	.	.	PUNCT
bracis-33550	128	1	wav2vec2	wav2vec2	PROPN
bracis-33550	128	2	has	have	VERB
bracis-33550	128	3	the	the	DET
bracis-33550	128	4	advantage	advantage	NOUN
bracis-33550	128	5	of	of	ADP
bracis-33550	128	6	presenting	present	VERB
bracis-33550	128	7	a	a	DET
bracis-33550	128	8	good	good	ADJ
bracis-33550	128	9	performance	performance	NOUN
bracis-33550	128	10	in	in	ADP
bracis-33550	128	11	low	low	ADJ
bracis-33550	128	12	and	and	CCONJ
bracis-33550	128	13	middle	middle	ADJ
bracis-33550	128	14	resource	resource	NOUN
bracis-33550	128	15	languages	language	NOUN
bracis-33550	128	16	.	.	PUNCT
bracis-33550	129	1	wa2vec2	wa2vec2	PROPN
bracis-33550	129	2	is	be	AUX
bracis-33550	129	3	an	an	DET
bracis-33550	129	4	auto	auto	NOUN
bracis-33550	129	5	-	-	PUNCT
bracis-33550	129	6	supervised	supervised	ADJ
bracis-33550	129	7	pretraining	pretraine	VERB
bracis-33550	129	8	in	in	ADP
bracis-33550	129	9	different	different	ADJ
bracis-33550	129	10	languages	language	NOUN
bracis-33550	129	11	,	,	PUNCT
bracis-33550	129	12	which	which	PRON
bracis-33550	129	13	speeds	speed	VERB
bracis-33550	129	14	up	up	ADP
bracis-33550	129	15	the	the	DET
bracis-33550	129	16	training	training	NOUN
bracis-33550	129	17	process	process	NOUN
bracis-33550	129	18	for	for	ADP
bracis-33550	129	19	asr	asr	NOUN
bracis-33550	129	20	,	,	PUNCT
bracis-33550	129	21	allowing	allow	VERB
bracis-33550	129	22	robustness	robustness	NOUN
bracis-33550	129	23	against	against	ADP
bracis-33550	129	24	noises	noise	NOUN
bracis-33550	129	25	that	that	PRON
bracis-33550	129	26	are	be	AUX
bracis-33550	129	27	common	common	ADJ
bracis-33550	129	28	in	in	ADP
bracis-33550	129	29	spontaneous	spontaneous	ADJ
bracis-33550	129	30	speech	speech	NOUN
bracis-33550	129	31	.	.	PUNCT
bracis-33550	130	1	distil	distil	NOUN
bracis-33550	130	2	-	-	PUNCT
bracis-33550	130	3	whisper	whisper	NOUN
bracis-33550	131	1	[	[	X
bracis-33550	131	2	11	11	NUM
bracis-33550	131	3	]	]	PUNCT
bracis-33550	131	4	is	be	AUX
bracis-33550	131	5	a	a	DET
bracis-33550	131	6	distilled	distil	VERB
bracis-33550	131	7	version	version	NOUN
bracis-33550	131	8	of	of	ADP
bracis-33550	131	9	whisper	whisper	NOUN
bracis-33550	131	10	,	,	PUNCT
bracis-33550	131	11	the	the	DET
bracis-33550	131	12	state	state	NOUN
bracis-33550	131	13	-	-	PUNCT
bracis-33550	131	14	of	of	ADP
bracis-33550	131	15	-	-	PUNCT
bracis-33550	131	16	the	the	DET
bracis-33550	131	17	-	-	PUNCT
bracis-33550	131	18	art	art	NOUN
bracis-33550	131	19	in	in	ADP
bracis-33550	131	20	asr	asr	NOUN
bracis-33550	131	21	in	in	ADP
bracis-33550	131	22	several	several	ADJ
bracis-33550	131	23	languages	language	NOUN
bracis-33550	131	24	.	.	PUNCT
bracis-33550	132	1	we	we	PRON
bracis-33550	132	2	opted	opt	VERB
bracis-33550	132	3	for	for	ADP
bracis-33550	132	4	the	the	DET
bracis-33550	132	5	distilled	distil	VERB
bracis-33550	132	6	version	version	NOUN
bracis-33550	132	7	because	because	SCONJ
bracis-33550	132	8	original	original	ADJ
bracis-33550	132	9	whisper	whisper	NOUN
bracis-33550	132	10	is	be	AUX
bracis-33550	132	11	costly	costly	ADJ
bracis-33550	132	12	,	,	PUNCT
bracis-33550	132	13	being	be	AUX
bracis-33550	132	14	originally	originally	ADV
bracis-33550	132	15	trained	train	VERB
bracis-33550	132	16	over	over	ADP
bracis-33550	132	17	thousands	thousand	NOUN
bracis-33550	132	18	hours	hour	NOUN
bracis-33550	132	19	of	of	ADP
bracis-33550	132	20	labelled	label	VERB
bracis-33550	132	21	speech	speech	NOUN
bracis-33550	132	22	,	,	PUNCT
bracis-33550	132	23	while	while	SCONJ
bracis-33550	132	24	distil	distil	NOUN
bracis-33550	132	25	-	-	PUNCT
bracis-33550	132	26	whisper	whisper	NOUN
bracis-33550	132	27	can	can	AUX
bracis-33550	132	28	perform	perform	VERB
bracis-33550	132	29	efficiently	efficiently	ADV
bracis-33550	132	30	with	with	ADP
bracis-33550	132	31	less	less	ADJ
bracis-33550	132	32	training	training	NOUN
bracis-33550	132	33	data	datum	NOUN
bracis-33550	132	34	.	.	PUNCT
bracis-33550	133	1	distil	distil	NOUN
bracis-33550	133	2	-	-	PUNCT
bracis-33550	133	3	whisper	whisper	NOUN
bracis-33550	133	4	is	be	AUX
bracis-33550	133	5	smaller	small	ADJ
bracis-33550	133	6	and	and	CCONJ
bracis-33550	133	7	faster	fast	ADV
bracis-33550	133	8	,	,	PUNCT
bracis-33550	133	9	while	while	SCONJ
bracis-33550	133	10	preserving	preserve	VERB
bracis-33550	133	11	many	many	ADJ
bracis-33550	133	12	of	of	ADP
bracis-33550	133	13	original	original	ADJ
bracis-33550	133	14	whisper	whisper	NOUN
bracis-33550	133	15	’s	’s	PART
bracis-33550	133	16	strengths	strength	NOUN
bracis-33550	133	17	.	.	PUNCT
bracis-33550	134	1	wav2vec2	wav2vec2	PROPN
bracis-33550	134	2	architecture	architecture	NOUN
bracis-33550	134	3	operates	operate	VERB
bracis-33550	134	4	directly	directly	ADV
bracis-33550	134	5	on	on	ADP
bracis-33550	134	6	the	the	DET
bracis-33550	134	7	wave	wave	NOUN
bracis-33550	134	8	signal	signal	NOUN
bracis-33550	134	9	,	,	PUNCT
bracis-33550	134	10	without	without	ADP
bracis-33550	134	11	the	the	DET
bracis-33550	134	12	need	need	NOUN
bracis-33550	134	13	of	of	ADP
bracis-33550	134	14	intermediary	intermediary	ADJ
bracis-33550	134	15	representations	representation	NOUN
bracis-33550	134	16	such	such	ADJ
bracis-33550	134	17	as	as	ADP
bracis-33550	134	18	spectrograms	spectrogram	NOUN
bracis-33550	134	19	.	.	PUNCT
bracis-33550	135	1	furthermore	furthermore	ADV
bracis-33550	135	2	,	,	PUNCT
bracis-33550	135	3	there	there	PRON
bracis-33550	135	4	are	be	VERB
bracis-33550	135	5	two	two	NUM
bracis-33550	135	6	main	main	ADJ
bracis-33550	135	7	layer	layer	NOUN
bracis-33550	135	8	types	type	NOUN
bracis-33550	135	9	:	:	PUNCT
bracis-33550	135	10	convolutional	convolutional	ADJ
bracis-33550	135	11	layers	layer	NOUN
bracis-33550	135	12	process	process	VERB
bracis-33550	135	13	the	the	DET
bracis-33550	135	14	raw	raw	ADJ
bracis-33550	135	15	wave	wave	NOUN
bracis-33550	135	16	signal	signal	NOUN
bracis-33550	135	17	;	;	PUNCT
bracis-33550	135	18	transformer	transformer	NOUN
bracis-33550	135	19	[	[	X
bracis-33550	135	20	23	23	NUM
bracis-33550	135	21	]	]	PUNCT
bracis-33550	135	22	layers	layer	NOUN
bracis-33550	135	23	process	process	VERB
bracis-33550	135	24	the	the	DET
bracis-33550	135	25	result	result	NOUN
bracis-33550	135	26	of	of	ADP
bracis-33550	135	27	the	the	DET
bracis-33550	135	28	previous	previous	ADJ
bracis-33550	135	29	step	step	NOUN
bracis-33550	135	30	.	.	PUNCT
bracis-33550	136	1	the	the	DET
bracis-33550	136	2	model	model	NOUN
bracis-33550	136	3	can	can	AUX
bracis-33550	136	4	be	be	AUX
bracis-33550	136	5	trained	train	VERB
bracis-33550	136	6	in	in	ADP
bracis-33550	136	7	an	an	DET
bracis-33550	136	8	auto	auto	NOUN
bracis-33550	136	9	-	-	PUNCT
bracis-33550	136	10	supervised	supervise	VERB
bracis-33550	136	11	manner	manner	NOUN
bracis-33550	136	12	.	.	PUNCT
bracis-33550	137	1	the	the	DET
bracis-33550	137	2	signal	signal	NOUN
bracis-33550	137	3	is	be	AUX
bracis-33550	137	4	also	also	ADV
bracis-33550	137	5	quantized	quantize	VERB
bracis-33550	137	6	in	in	ADP
bracis-33550	137	7	the	the	DET
bracis-33550	137	8	after	after	ADP
bracis-33550	137	9	passing	pass	VERB
bracis-33550	137	10	through	through	ADP
bracis-33550	137	11	convolutional	convolutional	ADJ
bracis-33550	137	12	layers	layer	NOUN
bracis-33550	137	13	and	and	CCONJ
bracis-33550	137	14	masked	mask	VERB
bracis-33550	137	15	language	language	NOUN
bracis-33550	137	16	modelling	modelling	NOUN
bracis-33550	137	17	is	be	AUX
bracis-33550	137	18	used	use	VERB
bracis-33550	137	19	in	in	ADP
bracis-33550	137	20	the	the	DET
bracis-33550	137	21	transformer	transformer	NOUN
bracis-33550	137	22	layers	layer	NOUN
bracis-33550	137	23	.	.	PUNCT
bracis-33550	138	1	a	a	DET
bracis-33550	138	2	rich	rich	ADJ
bracis-33550	138	3	loss	loss	NOUN
bracis-33550	138	4	takes	take	VERB
bracis-33550	138	5	into	into	ADP
bracis-33550	138	6	account	account	NOUN
bracis-33550	138	7	quantized	quantize	VERB
bracis-33550	138	8	representations	representation	NOUN
bracis-33550	138	9	and	and	CCONJ
bracis-33550	138	10	the	the	DET
bracis-33550	138	11	transformer	transformer	NOUN
bracis-33550	138	12	layer	layer	NOUN
bracis-33550	138	13	predictions	prediction	NOUN
bracis-33550	138	14	in	in	ADP
bracis-33550	138	15	order	order	NOUN
bracis-33550	138	16	to	to	PART
bracis-33550	138	17	produce	produce	VERB
bracis-33550	138	18	diversified	diversify	VERB
bracis-33550	138	19	quantized	quantize	VERB
bracis-33550	138	20	representations	representation	NOUN
bracis-33550	138	21	that	that	PRON
bracis-33550	138	22	represent	represent	VERB
bracis-33550	138	23	small	small	ADJ
bracis-33550	138	24	windows	window	NOUN
bracis-33550	138	25	in	in	ADP
bracis-33550	138	26	the	the	DET
bracis-33550	138	27	original	original	ADJ
bracis-33550	138	28	audio	audio	NOUN
bracis-33550	138	29	.	.	PUNCT
bracis-33550	139	1	the	the	DET
bracis-33550	139	2	learned	learn	VERB
bracis-33550	139	3	representations	representation	NOUN
bracis-33550	139	4	can	can	AUX
bracis-33550	139	5	then	then	ADV
bracis-33550	139	6	be	be	AUX
bracis-33550	139	7	fine	fine	ADV
bracis-33550	139	8	-	-	PUNCT
bracis-33550	139	9	tuned	tune	VERB
bracis-33550	139	10	to	to	PART
bracis-33550	139	11	represent	represent	VERB
bracis-33550	139	12	,	,	PUNCT
bracis-33550	139	13	for	for	ADP
bracis-33550	139	14	example	example	NOUN
bracis-33550	139	15	,	,	PUNCT
bracis-33550	139	16	phonemes	phoneme	NOUN
bracis-33550	139	17	in	in	ADP
bracis-33550	139	18	the	the	DET
bracis-33550	139	19	process	process	NOUN
bracis-33550	139	20	of	of	ADP
bracis-33550	139	21	asr	asr	NOUN
bracis-33550	139	22	.	.	PUNCT
bracis-33550	140	1	in	in	ADP
bracis-33550	140	2	this	this	DET
bracis-33550	140	3	case	case	NOUN
bracis-33550	140	4	,	,	PUNCT
bracis-33550	140	5	it	it	PRON
bracis-33550	140	6	is	be	AUX
bracis-33550	140	7	used	use	VERB
bracis-33550	140	8	a	a	DET
bracis-33550	140	9	classical	classical	ADJ
bracis-33550	140	10	seq2seq	seq2seq	NOUN
bracis-33550	140	11	loss	loss	NOUN
bracis-33550	140	12	for	for	ADP
bracis-33550	140	13	asr	asr	NOUN
bracis-33550	140	14	generation	generation	NOUN
bracis-33550	140	15	,	,	PUNCT
bracis-33550	140	16	namely	namely	ADV
bracis-33550	140	17	the	the	DET
bracis-33550	140	18	ctc	ctc	NOUN
bracis-33550	140	19	loss	loss	NOUN
bracis-33550	140	20	[	[	X
bracis-33550	140	21	13	13	NUM
bracis-33550	140	22	]	]	PUNCT
bracis-33550	140	23	.	.	PUNCT
bracis-33550	141	1	whisper	whisper	PROPN
bracis-33550	141	2	proposes	propose	VERB
bracis-33550	141	3	an	an	DET
bracis-33550	141	4	off	off	ADP
bracis-33550	141	5	-	-	PUNCT
bracis-33550	141	6	the	the	DET
bracis-33550	141	7	-	-	PUNCT
bracis-33550	141	8	shelf	shelf	NOUN
bracis-33550	141	9	approach	approach	NOUN
bracis-33550	141	10	to	to	ADP
bracis-33550	141	11	its	its	PRON
bracis-33550	141	12	architecture	architecture	NOUN
bracis-33550	141	13	,	,	PUNCT
bracis-33550	141	14	consisting	consist	VERB
bracis-33550	141	15	on	on	ADP
bracis-33550	141	16	a	a	DET
bracis-33550	141	17	traditional	traditional	ADJ
bracis-33550	141	18	transformer	transformer	NOUN
bracis-33550	141	19	,	,	PUNCT
bracis-33550	141	20	with	with	ADP
bracis-33550	141	21	encoding	encode	VERB
bracis-33550	141	22	and	and	CCONJ
bracis-33550	141	23	decoding	decode	VERB
bracis-33550	141	24	blocks	block	NOUN
bracis-33550	141	25	,	,	PUNCT
bracis-33550	141	26	and	and	CCONJ
bracis-33550	141	27	focusing	focus	VERB
bracis-33550	141	28	more	more	ADV
bracis-33550	141	29	on	on	ADP
bracis-33550	141	30	the	the	DET
bracis-33550	141	31	training	training	NOUN
bracis-33550	141	32	data	datum	NOUN
bracis-33550	141	33	than	than	ADP
bracis-33550	141	34	on	on	ADP
bracis-33550	141	35	architectural	architectural	ADJ
bracis-33550	141	36	adjustments	adjustment	NOUN
bracis-33550	141	37	.	.	PUNCT
bracis-33550	142	1	the	the	DET
bracis-33550	142	2	input	input	NOUN
bracis-33550	142	3	must	must	AUX
bracis-33550	142	4	be	be	AUX
bracis-33550	142	5	converted	convert	VERB
bracis-33550	142	6	to	to	ADP
bracis-33550	142	7	log	log	NOUN
bracis-33550	142	8	-	-	PUNCT
bracis-33550	142	9	mel	mel	NOUN
bracis-33550	142	10	spectrogram	spectrogram	NOUN
bracis-33550	142	11	format	format	NOUN
bracis-33550	142	12	.	.	PUNCT
bracis-33550	143	1	distil	distil	NOUN
bracis-33550	143	2	-	-	PUNCT
bracis-33550	143	3	whisper	whisper	NOUN
bracis-33550	143	4	is	be	AUX
bracis-33550	143	5	the	the	DET
bracis-33550	143	6	distilled	distil	VERB
bracis-33550	143	7	version	version	NOUN
bracis-33550	143	8	of	of	ADP
bracis-33550	143	9	whisper	whisper	NOUN
bracis-33550	143	10	.	.	PUNCT
bracis-33550	144	1	model	model	NOUN
bracis-33550	144	2	distillation	distillation	NOUN
bracis-33550	144	3	is	be	AUX
bracis-33550	144	4	a	a	DET
bracis-33550	144	5	technique	technique	NOUN
bracis-33550	144	6	that	that	PRON
bracis-33550	144	7	allows	allow	VERB
bracis-33550	144	8	for	for	ADP
bracis-33550	144	9	a	a	DET
bracis-33550	144	10	smaller	small	ADJ
bracis-33550	144	11	model	model	NOUN
bracis-33550	144	12	(	(	PUNCT
bracis-33550	144	13	the	the	DET
bracis-33550	144	14	student	student	NOUN
bracis-33550	144	15	)	)	PUNCT
bracis-33550	144	16	to	to	PART
bracis-33550	144	17	be	be	AUX
bracis-33550	144	18	trained	train	VERB
bracis-33550	144	19	on	on	ADP
bracis-33550	144	20	both	both	DET
bracis-33550	144	21	data	datum	NOUN
bracis-33550	144	22	labels	label	NOUN
bracis-33550	144	23	as	as	ADV
bracis-33550	144	24	well	well	ADV
bracis-33550	144	25	as	as	ADP
bracis-33550	144	26	predictions	prediction	NOUN
bracis-33550	144	27	from	from	ADP
bracis-33550	144	28	a	a	DET
bracis-33550	144	29	bigger	big	ADJ
bracis-33550	144	30	model	model	NOUN
bracis-33550	144	31	(	(	PUNCT
bracis-33550	144	32	the	the	DET
bracis-33550	144	33	professor	professor	NOUN
bracis-33550	144	34	)	)	PUNCT
bracis-33550	144	35	in	in	ADP
bracis-33550	144	36	a	a	DET
bracis-33550	144	37	given	give	VERB
bracis-33550	144	38	prediction	prediction	NOUN
bracis-33550	144	39	task	task	NOUN
bracis-33550	144	40	.	.	PUNCT
bracis-33550	145	1	the	the	DET
bracis-33550	145	2	main	main	ADJ
bracis-33550	145	3	idea	idea	NOUN
bracis-33550	145	4	behind	behind	ADP
bracis-33550	145	5	distillation	distillation	NOUN
bracis-33550	145	6	is	be	AUX
bracis-33550	145	7	that	that	SCONJ
bracis-33550	145	8	a	a	DET
bracis-33550	145	9	model	model	NOUN
bracis-33550	145	10	’s	’s	PART
bracis-33550	145	11	predictions	prediction	NOUN
bracis-33550	145	12	have	have	VERB
bracis-33550	145	13	richer	rich	ADJ
bracis-33550	145	14	information	information	NOUN
bracis-33550	145	15	about	about	ADP
bracis-33550	145	16	the	the	DET
bracis-33550	145	17	data	datum	NOUN
bracis-33550	145	18	distribution	distribution	NOUN
bracis-33550	145	19	than	than	ADP
bracis-33550	145	20	original	original	ADJ
bracis-33550	145	21	labels	label	NOUN
bracis-33550	145	22	.	.	PUNCT
bracis-33550	146	1	therefore	therefore	ADV
bracis-33550	146	2	,	,	PUNCT
bracis-33550	146	3	a	a	DET
bracis-33550	146	4	bigger	big	ADJ
bracis-33550	146	5	model	model	NOUN
bracis-33550	146	6	can	can	AUX
bracis-33550	146	7	capture	capture	VERB
bracis-33550	146	8	data	datum	NOUN
bracis-33550	146	9	distribution	distribution	NOUN
bracis-33550	146	10	and	and	CCONJ
bracis-33550	146	11	transmit	transmit	VERB
bracis-33550	146	12	it	it	PRON
bracis-33550	146	13	to	to	ADP
bracis-33550	146	14	the	the	DET
bracis-33550	146	15	student	student	NOUN
bracis-33550	146	16	.	.	PUNCT
bracis-33550	147	1	both	both	PRON
bracis-33550	147	2	whisper	whisper	VERB
bracis-33550	147	3	and	and	CCONJ
bracis-33550	147	4	distil	distil	NOUN
bracis-33550	147	5	-	-	PUNCT
bracis-33550	147	6	whisper	whisper	NOUN
bracis-33550	147	7	are	be	AUX
bracis-33550	147	8	available	available	ADJ
bracis-33550	147	9	in	in	ADP
bracis-33550	147	10	different	different	ADJ
bracis-33550	147	11	versions	version	NOUN
bracis-33550	147	12	.	.	PUNCT
bracis-33550	148	1	comparing	compare	VERB
bracis-33550	148	2	whisper	whisper	NOUN
bracis-33550	148	3	large	large	ADJ
bracis-33550	148	4	-	-	PUNCT
bracis-33550	148	5	v3	v3	NOUN
bracis-33550	148	6	with	with	ADP
bracis-33550	148	7	distil	distil	NOUN
bracis-33550	148	8	-	-	PUNCT
bracis-33550	148	9	whisper	whisper	NOUN
bracis-33550	148	10	large	large	ADJ
bracis-33550	148	11	-	-	PUNCT
bracis-33550	148	12	v3	v3	NOUN
bracis-33550	148	13	(	(	PUNCT
bracis-33550	148	14	used	use	VERB
bracis-33550	148	15	in	in	ADP
bracis-33550	148	16	this	this	DET
bracis-33550	148	17	work	work	NOUN
bracis-33550	148	18	)	)	PUNCT
bracis-33550	148	19	,	,	PUNCT
bracis-33550	148	20	it	it	PRON
bracis-33550	148	21	can	can	AUX
bracis-33550	148	22	be	be	AUX
bracis-33550	148	23	observed	observe	VERB
bracis-33550	148	24	an	an	DET
bracis-33550	148	25	improvement	improvement	NOUN
bracis-33550	148	26	of	of	ADP
bracis-33550	148	27	5.8	5.8	NUM
bracis-33550	148	28	times	time	NOUN
bracis-33550	148	29	in	in	ADP
bracis-33550	148	30	inference	inference	NOUN
bracis-33550	148	31	time	time	NOUN
bracis-33550	148	32	while	while	SCONJ
bracis-33550	148	33	using	use	VERB
bracis-33550	148	34	51	51	NUM
bracis-33550	148	35	%	%	NOUN
bracis-33550	148	36	fewer	few	ADJ
bracis-33550	148	37	parameters	parameter	NOUN
bracis-33550	148	38	in	in	ADP
bracis-33550	148	39	the	the	DET
bracis-33550	148	40	distilled	distil	VERB
bracis-33550	148	41	version	version	NOUN
bracis-33550	148	42	.	.	PUNCT
bracis-33550	149	1	table	table	NOUN
bracis-33550	149	2	 	 	SPACE
bracis-33550	149	3	3	3	NUM
bracis-33550	149	4	presents	present	VERB
bracis-33550	149	5	statistics	statistic	NOUN
bracis-33550	149	6	of	of	ADP
bracis-33550	149	7	training	training	NOUN
bracis-33550	149	8	,	,	PUNCT
bracis-33550	149	9	validation	validation	NOUN
bracis-33550	149	10	and	and	CCONJ
bracis-33550	149	11	test	test	NOUN
bracis-33550	149	12	partitions	partition	NOUN
bracis-33550	149	13	of	of	ADP
bracis-33550	149	14	each	each	DET
bracis-33550	149	15	nurc	nurc	ADJ
bracis-33550	149	16	-	-	PUNCT
bracis-33550	149	17	sp	sp	NOUN
bracis-33550	149	18	audio	audio	NOUN
bracis-33550	149	19	corpus	corpus	NOUN
bracis-33550	149	20	subset	subset	NOUN
bracis-33550	149	21	.	.	PUNCT
bracis-33550	150	1	4.1	4.1	NUM
bracis-33550	150	2	baseline	baseline	PROPN
bracis-33550	150	3	models	model	NOUN
bracis-33550	150	4	development	development	NOUN
bracis-33550	150	5	to	to	PART
bracis-33550	150	6	assess	assess	VERB
bracis-33550	150	7	the	the	DET
bracis-33550	150	8	dataset	dataset	NOUN
bracis-33550	150	9	quality	quality	NOUN
bracis-33550	150	10	,	,	PUNCT
bracis-33550	150	11	we	we	PRON
bracis-33550	150	12	trained	train	VERB
bracis-33550	150	13	four	four	NUM
bracis-33550	150	14	new	new	ADJ
bracis-33550	150	15	asr	asr	NOUN
bracis-33550	150	16	models	model	NOUN
bracis-33550	150	17	and	and	CCONJ
bracis-33550	150	18	compared	compare	VERB
bracis-33550	150	19	the	the	DET
bracis-33550	150	20	results	result	NOUN
bracis-33550	150	21	with	with	ADP
bracis-33550	150	22	previous	previous	ADJ
bracis-33550	150	23	related	related	ADJ
bracis-33550	150	24	works	work	NOUN
bracis-33550	150	25	.	.	PUNCT
bracis-33550	151	1	two	two	NUM
bracis-33550	151	2	of	of	ADP
bracis-33550	151	3	them	they	PRON
bracis-33550	151	4	are	be	AUX
bracis-33550	151	5	fine	fine	ADV
bracis-33550	151	6	-	-	PUNCT
bracis-33550	151	7	tuned	tune	VERB
bracis-33550	151	8	versions	version	NOUN
bracis-33550	151	9	of	of	ADP
bracis-33550	151	10	wav2vec2	wav2vec2	PROPN
bracis-33550	151	11	and	and	CCONJ
bracis-33550	151	12	the	the	DET
bracis-33550	151	13	third	third	ADJ
bracis-33550	151	14	and	and	CCONJ
bracis-33550	151	15	fourth	fourth	ADJ
bracis-33550	151	16	models	model	NOUN
bracis-33550	151	17	are	be	AUX
bracis-33550	151	18	approaches	approach	NOUN
bracis-33550	151	19	with	with	ADP
bracis-33550	151	20	distil	distil	NOUN
bracis-33550	151	21	-	-	PUNCT
bracis-33550	151	22	whisper	whisper	NOUN
bracis-33550	151	23	.	.	PUNCT
bracis-33550	152	1	all	all	DET
bracis-33550	152	2	four	four	NUM
bracis-33550	152	3	models	model	NOUN
bracis-33550	152	4	are	be	AUX
bracis-33550	152	5	described	describe	VERB
bracis-33550	152	6	below	below	ADP
bracis-33550	152	7	.	.	PUNCT
bracis-33550	153	1	wav2vec2	wav2vec2	PROPN
bracis-33550	153	2	-	-	PUNCT
bracis-33550	153	3	nurc	nurc	PROPN
bracis-33550	153	4	-	-	PUNCT
bracis-33550	153	5	sp-1	sp-1	NOUN
bracis-33550	153	6	.	.	PUNCT
bracis-33550	154	1	this	this	DET
bracis-33550	154	2	model	model	NOUN
bracis-33550	154	3	is	be	AUX
bracis-33550	154	4	a	a	DET
bracis-33550	154	5	fine	fine	ADV
bracis-33550	154	6	-	-	PUNCT
bracis-33550	154	7	tuned	tune	VERB
bracis-33550	154	8	version	version	NOUN
bracis-33550	154	9	of	of	ADP
bracis-33550	154	10	wav2vec	wav2vec	PROPN
bracis-33550	154	11	2.0	2.0	NUM
bracis-33550	154	12	xlsr-53	xlsr-53	X
bracis-33550	154	13	[	[	X
bracis-33550	154	14	2	2	NUM
bracis-33550	154	15	]	]	X
bracis-33550	154	16	[	[	X
bracis-33550	154	17	9	9	NUM
bracis-33550	154	18	]	]	PUNCT
bracis-33550	154	19	,	,	PUNCT
bracis-33550	154	20	pre	pre	VERB
bracis-33550	154	21	-	-	VERB
bracis-33550	154	22	trained	train	VERB
bracis-33550	154	23	over	over	ADP
bracis-33550	154	24	53	53	NUM
bracis-33550	154	25	languages	language	NOUN
bracis-33550	154	26	(	(	PUNCT
bracis-33550	154	27	portuguese	portuguese	PROPN
bracis-33550	154	28	included	include	VERB
bracis-33550	154	29	)	)	PUNCT
bracis-33550	154	30	.	.	PUNCT
bracis-33550	155	1	the	the	DET
bracis-33550	155	2	pre	pre	ADJ
bracis-33550	155	3	-	-	ADJ
bracis-33550	155	4	trained	train	VERB
bracis-33550	155	5	model	model	NOUN
bracis-33550	155	6	was	be	AUX
bracis-33550	155	7	fine	fine	ADV
bracis-33550	155	8	-	-	PUNCT
bracis-33550	155	9	tuned	tune	VERB
bracis-33550	155	10	with	with	ADP
bracis-33550	155	11	our	our	PRON
bracis-33550	155	12	train	train	NOUN
bracis-33550	155	13	and	and	CCONJ
bracis-33550	155	14	validation	validation	NOUN
bracis-33550	155	15	subsets	subset	NOUN
bracis-33550	155	16	of	of	ADP
bracis-33550	155	17	(	(	PUNCT
bracis-33550	155	18	nurc	nurc	ADJ
bracis-33550	155	19	-	-	PUNCT
bracis-33550	155	20	sp	sp	NOUN
bracis-33550	155	21	audio	audio	NOUN
bracis-33550	155	22	corpus	corpus	NOUN
bracis-33550	155	23	)	)	PUNCT
bracis-33550	155	24	nurc	nurc	PROPN
bracis-33550	155	25	-	-	PUNCT
bracis-33550	155	26	sp	sp	NOUN
bracis-33550	155	27	-	-	PUNCT
bracis-33550	155	28	ac	ac	NOUN
bracis-33550	155	29	in	in	ADP
bracis-33550	155	30	one	one	NUM
bracis-33550	155	31	gpu	gpu	NOUN
bracis-33550	155	32	nvidia	nvidia	PROPN
bracis-33550	155	33	dgx	dgx	PROPN
bracis-33550	155	34	a100	a100	PROPN
bracis-33550	155	35	80	80	NUM
bracis-33550	155	36	gb	gb	NOUN
bracis-33550	155	37	for	for	ADP
bracis-33550	155	38	16	16	NUM
bracis-33550	155	39	epochs	epoch	NOUN
bracis-33550	155	40	,	,	PUNCT
bracis-33550	155	41	with	with	ADP
bracis-33550	155	42	early	early	ADJ
bracis-33550	155	43	stop	stop	NOUN
bracis-33550	155	44	of	of	ADP
bracis-33550	155	45	10	10	NUM
bracis-33550	155	46	.	.	PUNCT
bracis-33550	156	1	the	the	DET
bracis-33550	156	2	other	other	ADJ
bracis-33550	156	3	settings	setting	NOUN
bracis-33550	156	4	were	be	AUX
bracis-33550	156	5	the	the	DET
bracis-33550	156	6	same	same	ADJ
bracis-33550	156	7	from	from	ADP
bracis-33550	156	8	[	[	X
bracis-33550	156	9	6	6	NUM
bracis-33550	156	10	]	]	PUNCT
bracis-33550	156	11	,	,	PUNCT
bracis-33550	156	12	training	training	NOUN
bracis-33550	156	13	code	code	NOUN
bracis-33550	156	14	is	be	AUX
bracis-33550	156	15	available	available	ADJ
bracis-33550	156	16	at	at	ADP
bracis-33550	156	17	githubfootnote	githubfootnote	PROPN
bracis-33550	156	18	14	14	NUM
bracis-33550	156	19	.	.	PUNCT
bracis-33550	157	1	wav2vec2	wav2vec2	PROPN
bracis-33550	157	2	-	-	PUNCT
bracis-33550	157	3	nurc	nurc	PROPN
bracis-33550	157	4	-	-	PUNCT
bracis-33550	157	5	sp-2	sp-2	NOUN
bracis-33550	157	6	.	.	PUNCT
bracis-33550	158	1	the	the	DET
bracis-33550	158	2	second	second	ADJ
bracis-33550	158	3	model	model	NOUN
bracis-33550	158	4	is	be	AUX
bracis-33550	158	5	almost	almost	ADV
bracis-33550	158	6	the	the	DET
bracis-33550	158	7	same	same	ADJ
bracis-33550	158	8	as	as	ADP
bracis-33550	158	9	the	the	DET
bracis-33550	158	10	first	first	ADJ
bracis-33550	158	11	one	one	NUM
bracis-33550	158	12	,	,	PUNCT
bracis-33550	158	13	but	but	CCONJ
bracis-33550	158	14	using	use	VERB
bracis-33550	158	15	as	as	ADP
bracis-33550	158	16	start	start	NOUN
bracis-33550	158	17	point	point	VERB
bracis-33550	158	18	the	the	DET
bracis-33550	158	19	model	model	NOUN
bracis-33550	158	20	that	that	SCONJ
bracis-33550	159	1	[	[	X
bracis-33550	159	2	6	6	NUM
bracis-33550	159	3	]	]	PUNCT
bracis-33550	159	4	trained	train	VERB
bracis-33550	159	5	for	for	ADP
bracis-33550	159	6	coraa	coraa	NOUN
bracis-33550	159	7	-	-	PUNCT
bracis-33550	159	8	v1	v1	NOUN
bracis-33550	159	9	and	and	CCONJ
bracis-33550	159	10	made	make	VERB
bracis-33550	159	11	publicly	publicly	ADV
bracis-33550	159	12	available	available	ADJ
bracis-33550	159	13	at	at	ADP
bracis-33550	159	14	huggingfacefootnote	huggingfacefootnote	NOUN
bracis-33550	159	15	15	15	NUM
bracis-33550	159	16	.	.	PUNCT
bracis-33550	160	1	it	it	PRON
bracis-33550	160	2	also	also	ADV
bracis-33550	160	3	was	be	AUX
bracis-33550	160	4	trained	train	VERB
bracis-33550	160	5	for	for	ADP
bracis-33550	160	6	16	16	NUM
bracis-33550	160	7	epochs	epoch	NOUN
bracis-33550	160	8	and	and	CCONJ
bracis-33550	160	9	finished	finish	VERB
bracis-33550	160	10	with	with	ADP
bracis-33550	160	11	a	a	DET
bracis-33550	160	12	better	well	ADJ
bracis-33550	160	13	wer	wer	NOUN
bracis-33550	160	14	that	that	SCONJ
bracis-33550	160	15	the	the	DET
bracis-33550	160	16	first	first	ADJ
bracis-33550	160	17	model	model	NOUN
bracis-33550	160	18	.	.	PUNCT
bracis-33550	161	1	distil	distil	NOUN
bracis-33550	161	2	-	-	PUNCT
bracis-33550	161	3	whisper	whisper	NOUN
bracis-33550	161	4	-	-	PUNCT
bracis-33550	161	5	nurc	nurc	NOUN
bracis-33550	161	6	-	-	PUNCT
bracis-33550	161	7	sp	sp	NOUN
bracis-33550	161	8	.	.	PUNCT
bracis-33550	162	1	the	the	DET
bracis-33550	162	2	third	third	ADJ
bracis-33550	162	3	model	model	NOUN
bracis-33550	162	4	is	be	AUX
bracis-33550	162	5	a	a	DET
bracis-33550	162	6	distilled	distil	VERB
bracis-33550	162	7	version	version	NOUN
bracis-33550	162	8	of	of	ADP
bracis-33550	162	9	whisper	whisper	NOUN
bracis-33550	162	10	large	large	ADJ
bracis-33550	162	11	-	-	PUNCT
bracis-33550	162	12	v3	v3	NOUN
bracis-33550	162	13	[	[	X
bracis-33550	162	14	19	19	NUM
bracis-33550	162	15	]	]	X
bracis-33550	162	16	,	,	PUNCT
bracis-33550	162	17	a	a	DET
bracis-33550	162	18	model	model	NOUN
bracis-33550	162	19	that	that	PRON
bracis-33550	162	20	has	have	VERB
bracis-33550	162	21	support	support	NOUN
bracis-33550	162	22	for	for	ADP
bracis-33550	162	23	the	the	DET
bracis-33550	162	24	portuguese	portuguese	ADJ
bracis-33550	162	25	language	language	NOUN
bracis-33550	162	26	,	,	PUNCT
bracis-33550	162	27	trained	train	VERB
bracis-33550	162	28	using	use	VERB
bracis-33550	162	29	our	our	PRON
bracis-33550	162	30	dataset	dataset	NOUN
bracis-33550	162	31	with	with	ADP
bracis-33550	162	32	labels	label	NOUN
bracis-33550	162	33	determined	determine	VERB
bracis-33550	162	34	by	by	ADP
bracis-33550	162	35	whisper	whisper	NOUN
bracis-33550	162	36	large	large	ADJ
bracis-33550	162	37	-	-	PUNCT
bracis-33550	162	38	v3	v3	NOUN
bracis-33550	162	39	,	,	PUNCT
bracis-33550	162	40	with	with	ADP
bracis-33550	162	41	the	the	DET
bracis-33550	162	42	reason	reason	NOUN
bracis-33550	162	43	being	be	AUX
bracis-33550	162	44	that	that	PRON
bracis-33550	162	45	more	more	ADJ
bracis-33550	162	46	knowledge	knowledge	NOUN
bracis-33550	162	47	can	can	AUX
bracis-33550	162	48	be	be	AUX
bracis-33550	162	49	passed	pass	VERB
bracis-33550	162	50	from	from	ADP
bracis-33550	162	51	the	the	DET
bracis-33550	162	52	teacher	teacher	NOUN
bracis-33550	162	53	model	model	NOUN
bracis-33550	162	54	to	to	ADP
bracis-33550	162	55	the	the	DET
bracis-33550	162	56	student	student	NOUN
bracis-33550	162	57	model	model	NOUN
bracis-33550	162	58	this	this	DET
bracis-33550	162	59	way	way	NOUN
bracis-33550	162	60	[	[	X
bracis-33550	162	61	11	11	NUM
bracis-33550	162	62	]	]	PUNCT
bracis-33550	162	63	.	.	PUNCT
bracis-33550	163	1	the	the	DET
bracis-33550	163	2	model	model	NOUN
bracis-33550	163	3	was	be	AUX
bracis-33550	163	4	trained	train	VERB
bracis-33550	163	5	with	with	ADP
bracis-33550	163	6	our	our	PRON
bracis-33550	163	7	train	train	NOUN
bracis-33550	163	8	and	and	CCONJ
bracis-33550	163	9	validation	validation	NOUN
bracis-33550	163	10	subsets	subset	NOUN
bracis-33550	163	11	of	of	ADP
bracis-33550	163	12	nurc	nurc	ADJ
bracis-33550	163	13	-	-	PUNCT
bracis-33550	163	14	sp	sp	NOUN
bracis-33550	163	15	-	-	PUNCT
bracis-33550	163	16	ac	ac	NOUN
bracis-33550	163	17	in	in	ADP
bracis-33550	163	18	one	one	NUM
bracis-33550	163	19	gpu	gpu	NOUN
bracis-33550	163	20	nvidia	nvidia	PROPN
bracis-33550	163	21	dgx	dgx	PROPN
bracis-33550	163	22	a100	a100	PROPN
bracis-33550	163	23	80	80	NUM
bracis-33550	163	24	gb	gb	NOUN
bracis-33550	163	25	for	for	ADP
bracis-33550	163	26	48	48	NUM
bracis-33550	163	27	epochs	epoch	NOUN
bracis-33550	163	28	,	,	PUNCT
bracis-33550	163	29	following	follow	VERB
bracis-33550	163	30	to	to	ADP
bracis-33550	163	31	the	the	DET
bracis-33550	163	32	steps	step	NOUN
bracis-33550	163	33	given	give	VERB
bracis-33550	163	34	by	by	ADP
bracis-33550	163	35	the	the	DET
bracis-33550	163	36	distil	distil	PROPN
bracis-33550	163	37	-	-	PUNCT
bracis-33550	163	38	whisper	whisper	NOUN
bracis-33550	163	39	githubfootnote	githubfootnote	VERB
bracis-33550	163	40	16	16	NUM
bracis-33550	163	41	.	.	PUNCT
bracis-33550	164	1	table	table	NOUN
bracis-33550	164	2	3	3	NUM
bracis-33550	164	3	.	.	PUNCT
bracis-33550	165	1	statistics	statistic	NOUN
bracis-33550	165	2	of	of	ADP
bracis-33550	165	3	training	training	NOUN
bracis-33550	165	4	,	,	PUNCT
bracis-33550	165	5	validation	validation	NOUN
bracis-33550	165	6	and	and	CCONJ
bracis-33550	165	7	test	test	NOUN
bracis-33550	165	8	partitions	partition	NOUN
bracis-33550	165	9	of	of	ADP
bracis-33550	165	10	each	each	DET
bracis-33550	165	11	nurc	nurc	ADJ
bracis-33550	165	12	-	-	PUNCT
bracis-33550	165	13	sp	sp	NOUN
bracis-33550	165	14	audio	audio	NOUN
bracis-33550	165	15	corpus	corpus	NOUN
bracis-33550	165	16	subset	subset	NOUN
bracis-33550	165	17	:	:	PUNCT
bracis-33550	165	18	the	the	DET
bracis-33550	165	19	total	total	ADJ
bracis-33550	165	20	duration	duration	NOUN
bracis-33550	165	21	in	in	ADP
bracis-33550	165	22	hours	hour	NOUN
bracis-33550	165	23	and	and	CCONJ
bracis-33550	165	24	number	number	NOUN
bracis-33550	165	25	of	of	ADP
bracis-33550	165	26	audio	audio	ADJ
bracis-33550	165	27	recordings	recording	NOUN
bracis-33550	165	28	(	(	PUNCT
bracis-33550	165	29	left	leave	VERB
bracis-33550	165	30	)	)	PUNCT
bracis-33550	165	31	and	and	CCONJ
bracis-33550	165	32	number	number	NOUN
bracis-33550	165	33	of	of	ADP
bracis-33550	165	34	speakers	speaker	NOUN
bracis-33550	165	35	in	in	ADP
bracis-33550	165	36	each	each	DET
bracis-33550	165	37	gender	gender	NOUN
bracis-33550	165	38	(	(	PUNCT
bracis-33550	165	39	right).full	right).full	NOUN
bracis-33550	165	40	size	size	NOUN
bracis-33550	165	41	table	table	NOUN
bracis-33550	165	42	distil	distil	NOUN
bracis-33550	165	43	-	-	PUNCT
bracis-33550	165	44	whisper	whisper	NOUN
bracis-33550	165	45	-	-	PUNCT
bracis-33550	165	46	nurc	nurc	ADJ
bracis-33550	165	47	-	-	PUNCT
bracis-33550	165	48	sp	sp	NOUN
bracis-33550	165	49	-	-	PUNCT
bracis-33550	165	50	fine	fine	ADV
bracis-33550	165	51	-	-	PUNCT
bracis-33550	165	52	tuned	tune	VERB
bracis-33550	165	53	.	.	PUNCT
bracis-33550	166	1	this	this	DET
bracis-33550	166	2	model	model	NOUN
bracis-33550	166	3	is	be	AUX
bracis-33550	166	4	a	a	DET
bracis-33550	166	5	fine	fine	ADV
bracis-33550	166	6	-	-	PUNCT
bracis-33550	166	7	tuning	tuning	NOUN
bracis-33550	166	8	of	of	ADP
bracis-33550	166	9	the	the	DET
bracis-33550	166	10	third	third	ADJ
bracis-33550	166	11	one	one	NOUN
bracis-33550	166	12	with	with	ADP
bracis-33550	166	13	our	our	PRON
bracis-33550	166	14	train	train	NOUN
bracis-33550	166	15	and	and	CCONJ
bracis-33550	166	16	validation	validation	NOUN
bracis-33550	166	17	subsets	subset	NOUN
bracis-33550	166	18	of	of	ADP
bracis-33550	166	19	nurc	nurc	ADJ
bracis-33550	166	20	-	-	PUNCT
bracis-33550	166	21	sp	sp	NOUN
bracis-33550	166	22	-	-	PUNCT
bracis-33550	166	23	ac	ac	NOUN
bracis-33550	166	24	,	,	PUNCT
bracis-33550	166	25	following	follow	VERB
bracis-33550	166	26	the	the	DET
bracis-33550	166	27	steps	step	NOUN
bracis-33550	166	28	recommended	recommend	VERB
bracis-33550	166	29	by	by	ADP
bracis-33550	166	30	the	the	DET
bracis-33550	166	31	distil	distil	PROPN
bracis-33550	166	32	-	-	PUNCT
bracis-33550	166	33	whisper	whisper	NOUN
bracis-33550	166	34	developersfootnote	developersfootnote	VERB
bracis-33550	166	35	17	17	NUM
bracis-33550	166	36	.	.	PUNCT
bracis-33550	167	1	it	it	PRON
bracis-33550	167	2	was	be	AUX
bracis-33550	167	3	also	also	ADV
bracis-33550	167	4	trained	train	VERB
bracis-33550	167	5	with	with	ADP
bracis-33550	167	6	48	48	NUM
bracis-33550	167	7	epochs	epoch	NOUN
bracis-33550	167	8	in	in	ADP
bracis-33550	167	9	one	one	NUM
bracis-33550	167	10	gpu	gpu	NOUN
bracis-33550	167	11	nvidia	nvidia	PROPN
bracis-33550	167	12	dgx	dgx	PROPN
bracis-33550	167	13	a100	a100	PROPN
bracis-33550	167	14	80	80	NUM
bracis-33550	167	15	gb	gb	NOUN
bracis-33550	167	16	and	and	CCONJ
bracis-33550	167	17	achieved	achieve	VERB
bracis-33550	167	18	a	a	DET
bracis-33550	167	19	better	well	ADJ
bracis-33550	167	20	wer	wer	NOUN
bracis-33550	167	21	and	and	CCONJ
bracis-33550	167	22	cer	cer	PROPN
bracis-33550	167	23	(	(	PUNCT
bracis-33550	167	24	character	character	NOUN
bracis-33550	167	25	error	error	NOUN
bracis-33550	167	26	rate)footnote	rate)footnote	NOUN
bracis-33550	167	27	18	18	NUM
bracis-33550	167	28	than	than	ADP
bracis-33550	167	29	the	the	DET
bracis-33550	167	30	third	third	ADJ
bracis-33550	167	31	model	model	NOUN
bracis-33550	167	32	.	.	PUNCT
bracis-33550	168	1	4.2	4.2	NUM
bracis-33550	168	2	normalization	normalization	NOUN
bracis-33550	168	3	we	we	PRON
bracis-33550	168	4	performed	perform	VERB
bracis-33550	168	5	the	the	DET
bracis-33550	168	6	following	follow	VERB
bracis-33550	168	7	steps	step	NOUN
bracis-33550	168	8	to	to	PART
bracis-33550	168	9	normalize	normalize	VERB
bracis-33550	168	10	the	the	DET
bracis-33550	168	11	train	train	NOUN
bracis-33550	168	12	,	,	PUNCT
bracis-33550	168	13	validation	validation	NOUN
bracis-33550	168	14	and	and	CCONJ
bracis-33550	168	15	test	test	NOUN
bracis-33550	168	16	subsets	subset	NOUN
bracis-33550	168	17	into	into	ADP
bracis-33550	168	18	a	a	DET
bracis-33550	168	19	standardized	standardized	ADJ
bracis-33550	168	20	form	form	NOUN
bracis-33550	168	21	in	in	ADP
bracis-33550	168	22	order	order	NOUN
bracis-33550	168	23	to	to	PART
bracis-33550	168	24	penalize	penalize	VERB
bracis-33550	168	25	only	only	ADV
bracis-33550	168	26	when	when	SCONJ
bracis-33550	168	27	a	a	DET
bracis-33550	168	28	word	word	NOUN
bracis-33550	168	29	error	error	NOUN
bracis-33550	168	30	is	be	AUX
bracis-33550	168	31	caused	cause	VERB
bracis-33550	168	32	by	by	ADP
bracis-33550	168	33	failure	failure	NOUN
bracis-33550	168	34	in	in	ADP
bracis-33550	168	35	transcribing	transcribe	VERB
bracis-33550	168	36	the	the	DET
bracis-33550	168	37	word	word	NOUN
bracis-33550	168	38	,	,	PUNCT
bracis-33550	168	39	and	and	CCONJ
bracis-33550	168	40	not	not	PART
bracis-33550	168	41	by	by	ADP
bracis-33550	168	42	formatting	format	VERB
bracis-33550	168	43	,	,	PUNCT
bracis-33550	168	44	punctuation	punctuation	NOUN
bracis-33550	168	45	or	or	CCONJ
bracis-33550	168	46	spontaneous	spontaneous	ADJ
bracis-33550	168	47	speech	speech	NOUN
bracis-33550	168	48	differences	difference	NOUN
bracis-33550	168	49	.	.	PUNCT
bracis-33550	169	1	the	the	DET
bracis-33550	169	2	revised	revise	VERB
bracis-33550	169	3	transcriptions	transcription	NOUN
bracis-33550	169	4	of	of	ADP
bracis-33550	169	5	nurc	nurc	ADJ
bracis-33550	169	6	-	-	PUNCT
bracis-33550	169	7	sp	sp	NOUN
bracis-33550	169	8	audio	audio	NOUN
bracis-33550	169	9	corpus	corpu	NOUN
bracis-33550	169	10	are	be	AUX
bracis-33550	169	11	of	of	ADP
bracis-33550	169	12	different	different	ADJ
bracis-33550	169	13	text	text	NOUN
bracis-33550	169	14	genders	gender	NOUN
bracis-33550	169	15	(	(	PUNCT
bracis-33550	169	16	ef	ef	PROPN
bracis-33550	169	17	,	,	PUNCT
bracis-33550	169	18	d2	d2	PROPN
bracis-33550	169	19	and	and	CCONJ
bracis-33550	169	20	did	do	VERB
bracis-33550	169	21	)	)	PUNCT
bracis-33550	169	22	and	and	CCONJ
bracis-33550	169	23	use	use	VERB
bracis-33550	169	24	upper	upper	ADJ
bracis-33550	169	25	and	and	CCONJ
bracis-33550	169	26	lower	low	ADJ
bracis-33550	169	27	case	case	NOUN
bracis-33550	169	28	letters	letter	NOUN
bracis-33550	169	29	and	and	CCONJ
bracis-33550	169	30	punctuation	punctuation	NOUN
bracis-33550	169	31	,	,	PUNCT
bracis-33550	169	32	as	as	ADV
bracis-33550	169	33	well	well	ADV
bracis-33550	169	34	as	as	ADP
bracis-33550	169	35	filled	fill	VERB
bracis-33550	169	36	pause	pause	NOUN
bracis-33550	169	37	markers	marker	NOUN
bracis-33550	169	38	such	such	ADJ
bracis-33550	169	39	as	as	ADP
bracis-33550	169	40	eh	eh	INTJ
bracis-33550	169	41	,	,	PUNCT
bracis-33550	169	42	hum	hum	ADJ
bracis-33550	169	43	,	,	PUNCT
bracis-33550	169	44	ãh	ãh	NOUN
bracis-33550	169	45	,	,	PUNCT
bracis-33550	169	46	etc	etc	X
bracis-33550	169	47	.	.	X
bracis-33550	169	48	to	to	PART
bracis-33550	169	49	simplify	simplify	VERB
bracis-33550	169	50	the	the	DET
bracis-33550	169	51	training	training	NOUN
bracis-33550	169	52	and	and	CCONJ
bracis-33550	169	53	calculation	calculation	NOUN
bracis-33550	169	54	of	of	ADP
bracis-33550	169	55	the	the	DET
bracis-33550	169	56	cer	cer	PROPN
bracis-33550	169	57	and	and	CCONJ
bracis-33550	169	58	wer	wer	NOUN
bracis-33550	169	59	metrics	metric	NOUN
bracis-33550	169	60	,	,	PUNCT
bracis-33550	169	61	the	the	DET
bracis-33550	169	62	following	follow	VERB
bracis-33550	169	63	normalization	normalization	NOUN
bracis-33550	169	64	was	be	AUX
bracis-33550	169	65	performed	perform	VERB
bracis-33550	169	66	:	:	PUNCT
bracis-33550	170	1	1	1	X
bracis-33550	170	2	.	.	PUNCT
bracis-33550	170	3	the	the	DET
bracis-33550	170	4	texts	text	NOUN
bracis-33550	170	5	were	be	AUX
bracis-33550	170	6	transformed	transform	VERB
bracis-33550	170	7	into	into	ADP
bracis-33550	170	8	lowercase	lowercase	NOUN
bracis-33550	170	9	;	;	PUNCT
bracis-33550	170	10	2	2	X
bracis-33550	170	11	.	.	X
bracis-33550	170	12	all	all	DET
bracis-33550	170	13	punctuation	punctuation	NOUN
bracis-33550	170	14	marks	mark	NOUN
bracis-33550	170	15	generated	generate	VERB
bracis-33550	170	16	by	by	ADP
bracis-33550	170	17	whisper	whisper	NOUN
bracis-33550	170	18	were	be	AUX
bracis-33550	170	19	removed	remove	VERB
bracis-33550	170	20	(	(	PUNCT
bracis-33550	170	21	ellipsis	ellipsis	NOUN
bracis-33550	170	22	,	,	PUNCT
bracis-33550	170	23	exclamation	exclamation	NOUN
bracis-33550	170	24	mark	mark	NOUN
bracis-33550	170	25	,	,	PUNCT
bracis-33550	170	26	fullstop	fullstop	NOUN
bracis-33550	170	27	,	,	PUNCT
bracis-33550	170	28	question	question	NOUN
bracis-33550	170	29	mark	mark	NOUN
bracis-33550	170	30	,	,	PUNCT
bracis-33550	170	31	and	and	CCONJ
bracis-33550	170	32	comma	comma	NOUN
bracis-33550	170	33	)	)	PUNCT
bracis-33550	170	34	;	;	PUNCT
bracis-33550	171	1	3	3	X
bracis-33550	171	2	.	.	X
bracis-33550	172	1	the	the	DET
bracis-33550	172	2	filled	fill	VERB
bracis-33550	172	3	pauses	pause	NOUN
bracis-33550	172	4	were	be	AUX
bracis-33550	172	5	standardized	standardize	VERB
bracis-33550	172	6	to	to	PART
bracis-33550	172	7	:	:	PUNCT
bracis-33550	172	8	eh	eh	INTJ
bracis-33550	172	9	,	,	PUNCT
bracis-33550	172	10	uh	uh	INTJ
bracis-33550	172	11	,	,	PUNCT
bracis-33550	172	12	ah	ah	INTJ
bracis-33550	172	13	,	,	PUNCT
bracis-33550	172	14	as	as	SCONJ
bracis-33550	172	15	follows	follow	VERB
bracis-33550	172	16	:	:	PUNCT
bracis-33550	172	17	eh	eh	INTJ
bracis-33550	172	18	=	=	X
bracis-33550	172	19	eh	eh	INTJ
bracis-33550	172	20	,	,	PUNCT
bracis-33550	172	21	éh	éh	PROPN
bracis-33550	172	22	,	,	PUNCT
bracis-33550	172	23	ehm	ehm	PROPN
bracis-33550	172	24	,	,	PUNCT
bracis-33550	172	25	ehn	ehn	NOUN
bracis-33550	172	26	;	;	PUNCT
bracis-33550	172	27	uh	uh	INTJ
bracis-33550	172	28	=	=	SYM
bracis-33550	172	29	uh	uh	INTJ
bracis-33550	172	30	,	,	PUNCT
bracis-33550	172	31	hm	hm	INTJ
bracis-33550	172	32	,	,	PUNCT
bracis-33550	172	33	uhm	uhm	PROPN
bracis-33550	172	34	,	,	PUNCT
bracis-33550	172	35	hmm	hmm	INTJ
bracis-33550	172	36	,	,	PUNCT
bracis-33550	172	37	mm	mm	PROPN
bracis-33550	172	38	,	,	PUNCT
bracis-33550	172	39	mhm	mhm	INTJ
bracis-33550	172	40	;	;	PUNCT
bracis-33550	172	41	ah	ah	INTJ
bracis-33550	172	42	=	=	X
bracis-33550	172	43	ah	ah	INTJ
bracis-33550	172	44	,	,	PUNCT
bracis-33550	172	45	huh	huh	INTJ
bracis-33550	172	46	,	,	PUNCT
bracis-33550	172	47	ãh	ãh	ADP
bracis-33550	172	48	,	,	PUNCT
bracis-33550	172	49	ã	ã	X
bracis-33550	172	50	;	;	PUNCT
bracis-33550	172	51	4	4	X
bracis-33550	172	52	.	.	X
bracis-33550	173	1	any	any	DET
bracis-33550	173	2	successive	successive	ADJ
bracis-33550	173	3	blank	blank	ADJ
bracis-33550	173	4	spaces	space	NOUN
bracis-33550	173	5	have	have	AUX
bracis-33550	173	6	been	be	AUX
bracis-33550	173	7	replaced	replace	VERB
bracis-33550	173	8	with	with	ADP
bracis-33550	173	9	one	one	NUM
bracis-33550	173	10	space	space	NOUN
bracis-33550	173	11	.	.	PUNCT
bracis-33550	174	1	4.3	4.3	NUM
bracis-33550	174	2	experiments	experiment	NOUN
bracis-33550	174	3	results	result	VERB
bracis-33550	174	4	for	for	ADP
bracis-33550	174	5	comparison	comparison	NOUN
bracis-33550	174	6	purposes	purpose	NOUN
bracis-33550	174	7	,	,	PUNCT
bracis-33550	174	8	we	we	PRON
bracis-33550	174	9	kept	keep	VERB
bracis-33550	174	10	the	the	DET
bracis-33550	174	11	tests	test	NOUN
bracis-33550	174	12	on	on	ADP
bracis-33550	174	13	the	the	DET
bracis-33550	174	14	datasets	dataset	NOUN
bracis-33550	174	15	used	use	VERB
bracis-33550	174	16	in	in	ADP
bracis-33550	174	17	previous	previous	ADJ
bracis-33550	174	18	works	work	NOUN
bracis-33550	174	19	and	and	CCONJ
bracis-33550	174	20	added	add	VERB
bracis-33550	174	21	the	the	DET
bracis-33550	174	22	new	new	ADJ
bracis-33550	174	23	dataset	dataset	NOUN
bracis-33550	174	24	made	make	VERB
bracis-33550	174	25	available	available	ADJ
bracis-33550	174	26	by	by	ADP
bracis-33550	174	27	this	this	DET
bracis-33550	174	28	work	work	NOUN
bracis-33550	174	29	.	.	PUNCT
bracis-33550	175	1	the	the	DET
bracis-33550	175	2	[	[	X
bracis-33550	175	3	14	14	NUM
bracis-33550	175	4	]	]	PUNCT
bracis-33550	175	5	and	and	CCONJ
bracis-33550	175	6	[	[	X
bracis-33550	175	7	6	6	NUM
bracis-33550	175	8	]	]	ADJ
bracis-33550	175	9	models	model	NOUN
bracis-33550	175	10	were	be	AUX
bracis-33550	175	11	rerun	rerun	VERB
bracis-33550	175	12	on	on	ADP
bracis-33550	175	13	all	all	DET
bracis-33550	175	14	datasets	dataset	NOUN
bracis-33550	175	15	,	,	PUNCT
bracis-33550	175	16	as	as	SCONJ
bracis-33550	175	17	were	be	AUX
bracis-33550	175	18	the	the	DET
bracis-33550	175	19	four	four	NUM
bracis-33550	175	20	new	new	ADJ
bracis-33550	175	21	trained	train	VERB
bracis-33550	175	22	models	model	NOUN
bracis-33550	175	23	.	.	PUNCT
bracis-33550	176	1	wer	wer	NOUN
bracis-33550	176	2	and	and	CCONJ
bracis-33550	176	3	cer	cer	PROPN
bracis-33550	176	4	metrics	metric	NOUN
bracis-33550	176	5	were	be	AUX
bracis-33550	176	6	evaluated	evaluate	VERB
bracis-33550	176	7	for	for	ADP
bracis-33550	176	8	each	each	DET
bracis-33550	176	9	run	run	NOUN
bracis-33550	176	10	.	.	PUNCT
bracis-33550	177	1	the	the	DET
bracis-33550	177	2	new	new	ADJ
bracis-33550	177	3	dataset	dataset	NOUN
bracis-33550	177	4	proved	prove	VERB
bracis-33550	177	5	to	to	PART
bracis-33550	177	6	be	be	AUX
bracis-33550	177	7	quite	quite	ADV
bracis-33550	177	8	challenging	challenging	ADJ
bracis-33550	177	9	for	for	ADP
bracis-33550	177	10	the	the	DET
bracis-33550	177	11	asr	asr	PROPN
bracis-33550	177	12	task	task	NOUN
bracis-33550	177	13	,	,	PUNCT
bracis-33550	177	14	as	as	SCONJ
bracis-33550	177	15	can	can	AUX
bracis-33550	177	16	be	be	AUX
bracis-33550	177	17	seen	see	VERB
bracis-33550	177	18	in	in	ADP
bracis-33550	177	19	table	table	NOUN
bracis-33550	177	20	 	 	SPACE
bracis-33550	177	21	4	4	NUM
bracis-33550	177	22	.	.	PUNCT
bracis-33550	177	23	table	table	NOUN
bracis-33550	177	24	4	4	NUM
bracis-33550	177	25	.	.	PUNCT
bracis-33550	178	1	nurc	nurc	ADJ
bracis-33550	178	2	-	-	PUNCT
bracis-33550	178	3	sp	sp	NOUN
bracis-33550	178	4	audio	audio	NOUN
bracis-33550	178	5	corpus	corpus	NOUN
bracis-33550	178	6	baseline	baseline	NOUN
bracis-33550	178	7	models	model	NOUN
bracis-33550	178	8	results	result	NOUN
bracis-33550	178	9	compared	compare	VERB
bracis-33550	178	10	with	with	ADP
bracis-33550	178	11	previous	previous	ADJ
bracis-33550	178	12	works	work	NOUN
bracis-33550	178	13	.	.	PUNCT
bracis-33550	179	1	the	the	DET
bracis-33550	179	2	best	good	ADJ
bracis-33550	179	3	values	value	NOUN
bracis-33550	179	4	for	for	ADP
bracis-33550	179	5	cer	cer	PROPN
bracis-33550	179	6	and	and	CCONJ
bracis-33550	179	7	wer	wer	NOUN
bracis-33550	179	8	appear	appear	VERB
bracis-33550	179	9	in	in	ADP
bracis-33550	179	10	bold.full	bold.full	NUM
bracis-33550	179	11	size	size	NOUN
bracis-33550	179	12	table	table	NOUN
bracis-33550	179	13	below	below	ADV
bracis-33550	179	14	,	,	PUNCT
bracis-33550	179	15	we	we	PRON
bracis-33550	179	16	show	show	VERB
bracis-33550	179	17	four	four	NUM
bracis-33550	179	18	examples	example	NOUN
bracis-33550	179	19	of	of	ADP
bracis-33550	179	20	result	result	NOUN
bracis-33550	179	21	instances	instance	NOUN
bracis-33550	179	22	from	from	ADP
bracis-33550	179	23	our	our	PRON
bracis-33550	179	24	best	good	ADJ
bracis-33550	179	25	model	model	NOUN
bracis-33550	179	26	(	(	PUNCT
bracis-33550	179	27	distil	distil	NOUN
bracis-33550	179	28	-	-	PUNCT
bracis-33550	179	29	whisper	whisper	NOUN
bracis-33550	179	30	nurc	nurc	ADJ
bracis-33550	179	31	-	-	PUNCT
bracis-33550	179	32	sp	sp	NOUN
bracis-33550	179	33	-	-	PUNCT
bracis-33550	179	34	fine	fine	NOUN
bracis-33550	179	35	-	-	PUNCT
bracis-33550	179	36	tuned)—the	tuned)—the	NOUN
bracis-33550	179	37	first	first	ADJ
bracis-33550	179	38	is	be	AUX
bracis-33550	179	39	the	the	DET
bracis-33550	179	40	original	original	ADJ
bracis-33550	179	41	normalized	normalize	VERB
bracis-33550	179	42	(	(	PUNCT
bracis-33550	179	43	on	on	ADP
bracis-33550	179	44	)	)	PUNCT
bracis-33550	179	45	and	and	CCONJ
bracis-33550	179	46	the	the	DET
bracis-33550	179	47	second	second	NOUN
bracis-33550	179	48	is	be	AUX
bracis-33550	179	49	the	the	DET
bracis-33550	179	50	model	model	NOUN
bracis-33550	179	51	’s	’s	PART
bracis-33550	179	52	prediction	prediction	NOUN
bracis-33550	179	53	normalized	normalize	VERB
bracis-33550	179	54	(	(	PUNCT
bracis-33550	179	55	mpn	mpn	NOUN
bracis-33550	179	56	)	)	PUNCT
bracis-33550	179	57	.	.	PUNCT
bracis-33550	180	1	in	in	ADP
bracis-33550	180	2	the	the	DET
bracis-33550	180	3	first	first	ADJ
bracis-33550	180	4	two	two	NUM
bracis-33550	180	5	,	,	PUNCT
bracis-33550	180	6	there	there	PRON
bracis-33550	180	7	are	be	VERB
bracis-33550	180	8	problems	problem	NOUN
bracis-33550	180	9	with	with	ADP
bracis-33550	180	10	named	name	VERB
bracis-33550	180	11	entities	entity	NOUN
bracis-33550	180	12	as	as	ADP
bracis-33550	180	13	the	the	DET
bracis-33550	180	14	“	"	PUNCT
bracis-33550	180	15	martinelli	martinelli	PROPN
bracis-33550	180	16	”	"	PUNCT
bracis-33550	180	17	building	building	NOUN
bracis-33550	180	18	was	be	AUX
bracis-33550	180	19	transcribed	transcribe	VERB
bracis-33550	180	20	as	as	ADP
bracis-33550	180	21	“	"	PUNCT
bracis-33550	180	22	martini	martini	NOUN
bracis-33550	180	23	”	"	PUNCT
bracis-33550	180	24	and	and	CCONJ
bracis-33550	180	25	the	the	DET
bracis-33550	180	26	name	name	NOUN
bracis-33550	180	27	“	"	PUNCT
bracis-33550	180	28	paulo	paulo	PROPN
bracis-33550	180	29	emilio	emilio	PROPN
bracis-33550	180	30	salles	salles	PROPN
bracis-33550	180	31	gomes	gomes	PROPN
bracis-33550	180	32	”	"	PUNCT
bracis-33550	180	33	had	have	VERB
bracis-33550	180	34	a	a	DET
bracis-33550	180	35	wrong	wrong	ADJ
bracis-33550	180	36	transcription	transcription	NOUN
bracis-33550	180	37	as	as	ADP
bracis-33550	180	38	“	"	PUNCT
bracis-33550	180	39	paulo	paulo	PROPN
bracis-33550	180	40	e	e	PROPN
bracis-33550	180	41	milho	milho	NOUN
bracis-33550	180	42	fales	fales	PROPN
bracis-33550	180	43	gomes	gome	NOUN
bracis-33550	180	44	”	"	PUNCT
bracis-33550	180	45	.	.	PUNCT
bracis-33550	181	1	the	the	DET
bracis-33550	181	2	model	model	NOUN
bracis-33550	181	3	used	use	VERB
bracis-33550	181	4	more	more	ADJ
bracis-33550	181	5	frequent	frequent	ADJ
bracis-33550	181	6	common	common	ADJ
bracis-33550	181	7	names	name	NOUN
bracis-33550	181	8	than	than	ADP
bracis-33550	181	9	the	the	DET
bracis-33550	181	10	proper	proper	ADJ
bracis-33550	181	11	names	name	NOUN
bracis-33550	181	12	“	"	PUNCT
bracis-33550	181	13	martinelli	martinelli	PROPN
bracis-33550	181	14	”	"	PUNCT
bracis-33550	181	15	and	and	CCONJ
bracis-33550	181	16	“	"	PUNCT
bracis-33550	181	17	emilio	emilio	PROPN
bracis-33550	181	18	salles	salles	PROPN
bracis-33550	181	19	”	"	PUNCT
bracis-33550	181	20	.	.	PUNCT
bracis-33550	182	1	rare	rare	ADJ
bracis-33550	182	2	words	word	NOUN
bracis-33550	182	3	related	relate	VERB
bracis-33550	182	4	to	to	ADP
bracis-33550	182	5	certain	certain	ADJ
bracis-33550	182	6	domains	domain	NOUN
bracis-33550	182	7	(	(	PUNCT
bracis-33550	182	8	e.g.	e.g.	ADV
bracis-33550	182	9	food	food	NOUN
bracis-33550	182	10	)	)	PUNCT
bracis-33550	182	11	or	or	CCONJ
bracis-33550	182	12	terminologies	terminology	NOUN
bracis-33550	182	13	of	of	ADP
bracis-33550	182	14	research	research	NOUN
bracis-33550	182	15	areas	area	NOUN
bracis-33550	182	16	(	(	PUNCT
bracis-33550	182	17	e.g.	e.g.	ADV
bracis-33550	182	18	linguistics	linguistic	NOUN
bracis-33550	182	19	)	)	PUNCT
bracis-33550	182	20	also	also	ADV
bracis-33550	182	21	bring	bring	VERB
bracis-33550	182	22	some	some	DET
bracis-33550	182	23	difficulties	difficulty	NOUN
bracis-33550	182	24	to	to	ADP
bracis-33550	182	25	the	the	DET
bracis-33550	182	26	model	model	NOUN
bracis-33550	182	27	(	(	PUNCT
bracis-33550	182	28	see	see	VERB
bracis-33550	182	29	third	third	ADJ
bracis-33550	182	30	and	and	CCONJ
bracis-33550	182	31	fourth	fourth	ADJ
bracis-33550	182	32	examples	example	NOUN
bracis-33550	182	33	):	):	PUNCT
bracis-33550	182	34	1	1	X
bracis-33550	182	35	.	.	PUNCT
bracis-33550	183	1	(	(	PUNCT
bracis-33550	183	2	on	on	ADP
bracis-33550	183	3	)	)	PUNCT
bracis-33550	183	4	o	o	NOUN
bracis-33550	183	5	martinelli	martinelli	PROPN
bracis-33550	183	6	ficou	ficou	PROPN
bracis-33550	183	7	célebre	célebre	NOUN
bracis-33550	183	8	em	em	PRON
bracis-33550	183	9	todo	todo	VERB
bracis-33550	183	10	o	o	PROPN
bracis-33550	183	11	exterior	exterior	NOUN
bracis-33550	183	12	do	do	AUX
bracis-33550	183	13	estado	estado	VERB
bracis-33550	183	14	no	no	DET
bracis-33550	183	15	interior	interior	NOUN
bracis-33550	183	16	do	do	AUX
bracis-33550	183	17	estado	estado	PROPN
bracis-33550	183	18	de	de	PROPN
bracis-33550	183	19	são	são	PROPN
bracis-33550	183	20	paulo	paulo	PROPN
bracis-33550	183	21	e	e	PROPN
bracis-33550	183	22	mesmo	mesmo	PROPN
bracis-33550	183	23	pelo	pelo	PROPN
bracis-33550	183	24	brasil	brasil	PROPN
bracis-33550	183	25	afora	afora	PROPN
bracis-33550	183	26	como	como	PROPN
bracis-33550	183	27	um	um	INTJ
bracis-33550	183	28	arranha	arranha	NOUN
bracis-33550	183	29	-	-	PUNCT
bracis-33550	183	30	céu	céu	ADJ
bracis-33550	183	31	notável	notável	NOUN
bracis-33550	183	32	para	para	NOUN
bracis-33550	183	33	a	a	DET
bracis-33550	183	34	época	época	PROPN
bracis-33550	183	35	;	;	PUNCT
bracis-33550	183	36	(	(	PUNCT
bracis-33550	183	37	mpn	mpn	NOUN
bracis-33550	183	38	)	)	PUNCT
bracis-33550	184	1	o	o	NOUN
bracis-33550	185	1	martini	martini	NOUN
bracis-33550	185	2	ficou	ficou	PROPN
bracis-33550	185	3	célebre	célebre	NOUN
bracis-33550	185	4	em	em	PRON
bracis-33550	185	5	todo	todo	VERB
bracis-33550	185	6	o	o	PROPN
bracis-33550	185	7	exterior	exterior	NOUN
bracis-33550	185	8	do	do	AUX
bracis-33550	185	9	estado	estado	NOUN
bracis-33550	185	10	do	do	VERB
bracis-33550	185	11	interior	interior	PROPN
bracis-33550	185	12	de	de	X
bracis-33550	185	13	estado	estado	PROPN
bracis-33550	185	14	de	de	PROPN
bracis-33550	185	15	são	são	PROPN
bracis-33550	185	16	paulo	paulo	PROPN
bracis-33550	185	17	e	e	PROPN
bracis-33550	185	18	mesmo	mesmo	PROPN
bracis-33550	185	19	pelo	pelo	PROPN
bracis-33550	185	20	brasil	brasil	PROPN
bracis-33550	185	21	afora	afora	PROPN
bracis-33550	185	22	como	como	PROPN
bracis-33550	185	23	arranha	arranha	NOUN
bracis-33550	185	24	-	-	PUNCT
bracis-33550	185	25	céu	céu	ADJ
bracis-33550	185	26	notável	notável	NOUN
bracis-33550	185	27	para	para	NOUN
bracis-33550	185	28	a	a	DET
bracis-33550	185	29	época	época	NOUN
bracis-33550	185	30	;	;	PUNCT
bracis-33550	185	31	2	2	X
bracis-33550	185	32	.	.	PUNCT
bracis-33550	185	33	(	(	PUNCT
bracis-33550	185	34	on	on	ADP
bracis-33550	185	35	)	)	PUNCT
bracis-33550	185	36	você	você	NOUN
bracis-33550	185	37	me	i	PRON
bracis-33550	185	38	falou	falou	PROPN
bracis-33550	185	39	em	em	PROPN
bracis-33550	185	40	cinema	cinema	PROPN
bracis-33550	185	41	eu	eu	PROPN
bracis-33550	185	42	lembrei	lembrei	PROPN
bracis-33550	185	43	de	de	PROPN
bracis-33550	185	44	paulo	paulo	PROPN
bracis-33550	185	45	emilio	emilio	PROPN
bracis-33550	185	46	salles	salles	PROPN
bracis-33550	185	47	gomes	gomes	PROPN
bracis-33550	185	48	que	que	PROPN
bracis-33550	185	49	foi	foi	PROPN
bracis-33550	185	50	meu	meu	PROPN
bracis-33550	185	51	colega	colega	PROPN
bracis-33550	185	52	na	na	PART
bracis-33550	185	53	faculdade	faculdade	PROPN
bracis-33550	185	54	e	e	PROPN
bracis-33550	185	55	é	é	X
bracis-33550	185	56	um	um	INTJ
bracis-33550	185	57	entendidíssimo	entendidíssimo	PROPN
bracis-33550	185	58	de	de	X
bracis-33550	185	59	cinema	cinema	PROPN
bracis-33550	185	60	né	né	INTJ
bracis-33550	185	61	;	;	PUNCT
bracis-33550	185	62	(	(	PUNCT
bracis-33550	185	63	mpn	mpn	NOUN
bracis-33550	185	64	)	)	PUNCT
bracis-33550	185	65	você	você	NOUN
bracis-33550	185	66	me	i	PRON
bracis-33550	185	67	falou	falou	PROPN
bracis-33550	186	1	em	em	PROPN
bracis-33550	186	2	cinema	cinema	NOUN
bracis-33550	186	3	eu	eu	PROPN
bracis-33550	186	4	me	me	PROPN
bracis-33550	186	5	lembrei	lembrei	PROPN
bracis-33550	186	6	de	de	PROPN
bracis-33550	186	7	paulo	paulo	PROPN
bracis-33550	186	8	e	e	PROPN
bracis-33550	186	9	milho	milho	PROPN
bracis-33550	186	10	fales	fales	PROPN
bracis-33550	186	11	gomes	gomes	PROPN
bracis-33550	186	12	que	que	PROPN
bracis-33550	186	13	foi	foi	PROPN
bracis-33550	186	14	minha	minha	PROPN
bracis-33550	186	15	colega	colega	PROPN
bracis-33550	186	16	na	na	PART
bracis-33550	186	17	faculdade	faculdade	PROPN
bracis-33550	186	18	e	e	PROPN
bracis-33550	186	19	é	é	X
bracis-33550	186	20	um	um	INTJ
bracis-33550	186	21	entendidíssimo	entendidíssimo	PROPN
bracis-33550	186	22	de	de	X
bracis-33550	186	23	cinema	cinema	PROPN
bracis-33550	186	24	né	né	X
bracis-33550	186	25	;	;	PUNCT
bracis-33550	186	26	3	3	X
bracis-33550	186	27	.	.	PUNCT
bracis-33550	186	28	(	(	PUNCT
bracis-33550	186	29	on	on	ADP
bracis-33550	186	30	)	)	PUNCT
bracis-33550	186	31	cuscuz	cuscuz	PROPN
bracis-33550	186	32	paulista	paulista	PROPN
bracis-33550	186	33	bobó	bobó	PROPN
bracis-33550	186	34	de	de	PROPN
bracis-33550	186	35	camarão	camarão	PROPN
bracis-33550	186	36	essas	essas	PROPN
bracis-33550	186	37	coisas	coisas	PROPN
bracis-33550	186	38	assim	assim	PROPN
bracis-33550	186	39	;	;	PUNCT
bracis-33550	186	40	(	(	PUNCT
bracis-33550	186	41	mpn	mpn	NOUN
bracis-33550	186	42	)	)	PUNCT
bracis-33550	186	43	cuscos	cuscos	PROPN
bracis-33550	186	44	paulista	paulista	PROPN
bracis-33550	186	45	babota	babota	PROPN
bracis-33550	186	46	de	de	PROPN
bracis-33550	186	47	camarão	camarão	PROPN
bracis-33550	186	48	essas	essas	PROPN
bracis-33550	186	49	coisas	coisas	PROPN
bracis-33550	186	50	;	;	PUNCT
bracis-33550	186	51	4	4	X
bracis-33550	186	52	.	.	PUNCT
bracis-33550	187	1	(	(	PUNCT
bracis-33550	187	2	on	on	ADP
bracis-33550	187	3	)	)	PUNCT
bracis-33550	187	4	de	de	X
bracis-33550	187	5	um	um	INTJ
bracis-33550	187	6	lado	lado	PROPN
bracis-33550	187	7	objeto	objeto	PROPN
bracis-33550	187	8	direto	direto	PROPN
bracis-33550	187	9	do	do	AUX
bracis-33550	187	10	outro	outro	VERB
bracis-33550	187	11	adjunto	adjunto	PROPN
bracis-33550	187	12	;	;	PUNCT
bracis-33550	187	13	(	(	PUNCT
bracis-33550	187	14	mpn	mpn	NOUN
bracis-33550	187	15	)	)	PUNCT
bracis-33550	187	16	de	de	PROPN
bracis-33550	187	17	um	um	INTJ
bracis-33550	187	18	lado	lado	X
bracis-33550	187	19	é	é	PROPN
bracis-33550	187	20	o	o	X
bracis-33550	187	21	chefe	chefe	X
bracis-33550	187	22	do	do	AUX
bracis-33550	187	23	e	e	NOUN
bracis-33550	187	24	o	o	X
bracis-33550	187	25	outro	outro	NOUN
bracis-33550	187	26	é	é	PROPN
bracis-33550	187	27	de	de	PROPN
bracis-33550	187	28	junho	junho	PROPN
bracis-33550	187	29	.	.	PUNCT
bracis-33550	188	1	5	5	NUM
bracis-33550	188	2	discussion	discussion	NOUN
bracis-33550	188	3	and	and	CCONJ
bracis-33550	188	4	 	 	SPACE
bracis-33550	188	5	conclusions	conclusion	NOUN
bracis-33550	188	6	in	in	ADP
bracis-33550	188	7	this	this	DET
bracis-33550	188	8	paper	paper	NOUN
bracis-33550	188	9	,	,	PUNCT
bracis-33550	188	10	we	we	PRON
bracis-33550	188	11	present	present	VERB
bracis-33550	188	12	a	a	DET
bracis-33550	188	13	freely	freely	ADV
bracis-33550	188	14	available	available	ADJ
bracis-33550	188	15	spontaneous	spontaneous	ADJ
bracis-33550	188	16	speech	speech	NOUN
bracis-33550	188	17	corpus	corpus	NOUN
bracis-33550	188	18	for	for	ADP
bracis-33550	188	19	the	the	DET
bracis-33550	188	20	brazilian	brazilian	ADJ
bracis-33550	188	21	portuguese	portuguese	ADJ
bracis-33550	188	22	language	language	NOUN
bracis-33550	188	23	,	,	PUNCT
bracis-33550	188	24	totaling	total	VERB
bracis-33550	188	25	239.30	239.30	NUM
bracis-33550	188	26	 	 	SPACE
bracis-33550	188	27	h	h	NOUN
bracis-33550	188	28	,	,	PUNCT
bracis-33550	188	29	and	and	CCONJ
bracis-33550	188	30	report	report	VERB
bracis-33550	188	31	preliminary	preliminary	ADJ
bracis-33550	188	32	asr	asr	PROPN
bracis-33550	188	33	results	result	NOUN
bracis-33550	188	34	,	,	PUNCT
bracis-33550	188	35	using	use	VERB
bracis-33550	188	36	both	both	CCONJ
bracis-33550	188	37	the	the	DET
bracis-33550	188	38	wav2vec2	wav2vec2	PROPN
bracis-33550	188	39	-	-	PUNCT
bracis-33550	188	40	xlsr-53	xlsr-53	PROPN
bracis-33550	188	41	and	and	CCONJ
bracis-33550	188	42	distil	distil	NOUN
bracis-33550	188	43	-	-	PUNCT
bracis-33550	188	44	whisper	whisper	NOUN
bracis-33550	188	45	models	model	NOUN
bracis-33550	188	46	fine	fine	ADV
bracis-33550	188	47	-	-	PUNCT
bracis-33550	188	48	tuned	tune	VERB
bracis-33550	188	49	and	and	CCONJ
bracis-33550	188	50	trained	train	VERB
bracis-33550	188	51	on	on	ADP
bracis-33550	188	52	our	our	PRON
bracis-33550	188	53	corpus	corpus	NOUN
bracis-33550	188	54	.	.	PUNCT
bracis-33550	189	1	to	to	ADP
bracis-33550	189	2	the	the	DET
bracis-33550	189	3	best	good	ADJ
bracis-33550	189	4	of	of	ADP
bracis-33550	189	5	our	our	PRON
bracis-33550	189	6	knowledge	knowledge	NOUN
bracis-33550	189	7	,	,	PUNCT
bracis-33550	189	8	distil	distil	NOUN
bracis-33550	189	9	-	-	PUNCT
bracis-33550	189	10	whisper	whisper	NOUN
bracis-33550	189	11	has	have	AUX
bracis-33550	189	12	never	never	ADV
bracis-33550	189	13	been	be	AUX
bracis-33550	189	14	evaluated	evaluate	VERB
bracis-33550	189	15	for	for	ADP
bracis-33550	189	16	bp	bp	PROPN
bracis-33550	189	17	datasets	dataset	NOUN
bracis-33550	189	18	.	.	PUNCT
bracis-33550	190	1	for	for	ADP
bracis-33550	190	2	comparison	comparison	NOUN
bracis-33550	190	3	purposes	purpose	NOUN
bracis-33550	190	4	,	,	PUNCT
bracis-33550	190	5	we	we	PRON
bracis-33550	190	6	also	also	ADV
bracis-33550	190	7	bring	bring	VERB
bracis-33550	190	8	the	the	DET
bracis-33550	190	9	tests	test	NOUN
bracis-33550	190	10	on	on	ADP
bracis-33550	190	11	the	the	DET
bracis-33550	190	12	datasets	dataset	NOUN
bracis-33550	190	13	used	use	VERB
bracis-33550	190	14	in	in	ADP
bracis-33550	190	15	previous	previous	ADJ
bracis-33550	190	16	works	work	NOUN
bracis-33550	190	17	.	.	PUNCT
bracis-33550	191	1	the	the	DET
bracis-33550	191	2	wav2vec	wav2vec	PROPN
bracis-33550	191	3	2.0	2.0	NUM
bracis-33550	191	4	based	base	VERB
bracis-33550	191	5	model	model	NOUN
bracis-33550	191	6	from	from	ADP
bracis-33550	191	7	[	[	X
bracis-33550	191	8	14	14	NUM
bracis-33550	191	9	]	]	PUNCT
bracis-33550	191	10	remains	remain	VERB
bracis-33550	191	11	the	the	DET
bracis-33550	191	12	best	good	ADJ
bracis-33550	191	13	for	for	SCONJ
bracis-33550	191	14	the	the	DET
bracis-33550	191	15	common	common	ADJ
bracis-33550	191	16	voice	voice	NOUN
bracis-33550	191	17	dataset	dataset	VERB
bracis-33550	191	18	with	with	ADP
bracis-33550	191	19	a	a	DET
bracis-33550	191	20	wer	wer	NOUN
bracis-33550	191	21	of	of	ADP
bracis-33550	191	22	16.32	16.32	NUM
bracis-33550	191	23	%	%	NOUN
bracis-33550	191	24	,	,	PUNCT
bracis-33550	191	25	even	even	ADV
bracis-33550	191	26	in	in	ADP
bracis-33550	191	27	the	the	DET
bracis-33550	191	28	most	most	ADV
bracis-33550	191	29	recent	recent	ADJ
bracis-33550	191	30	version	version	NOUN
bracis-33550	191	31	(	(	PUNCT
bracis-33550	191	32	cv	cv	PROPN
bracis-33550	191	33	version	version	NOUN
bracis-33550	191	34	17.0	17.0	NUM
bracis-33550	191	35	)	)	PUNCT
bracis-33550	191	36	,	,	PUNCT
bracis-33550	191	37	as	as	SCONJ
bracis-33550	191	38	it	it	PRON
bracis-33550	191	39	is	be	AUX
bracis-33550	191	40	trained	train	VERB
bracis-33550	191	41	in	in	ADP
bracis-33550	191	42	prepared	prepared	ADJ
bracis-33550	191	43	/	/	SYM
bracis-33550	191	44	read	read	NOUN
bracis-33550	191	45	speech	speech	NOUN
bracis-33550	191	46	.	.	PUNCT
bracis-33550	192	1	in	in	ADP
bracis-33550	192	2	the	the	DET
bracis-33550	192	3	coraa	coraa	NOUN
bracis-33550	192	4	v1	v1	NOUN
bracis-33550	192	5	dataset	dataset	NOUN
bracis-33550	192	6	,	,	PUNCT
bracis-33550	192	7	[	[	X
bracis-33550	192	8	6	6	NUM
bracis-33550	192	9	]	]	SYM
bracis-33550	192	10	’s	’s	PART
bracis-33550	192	11	models	model	NOUN
bracis-33550	192	12	,	,	PUNCT
bracis-33550	192	13	which	which	PRON
bracis-33550	192	14	is	be	AUX
bracis-33550	192	15	also	also	ADV
bracis-33550	192	16	trained	train	VERB
bracis-33550	192	17	an	an	DET
bracis-33550	192	18	wav2vec	wav2vec	PROPN
bracis-33550	192	19	2.0	2.0	NUM
bracis-33550	192	20	model	model	NOUN
bracis-33550	192	21	,	,	PUNCT
bracis-33550	192	22	but	but	CCONJ
bracis-33550	192	23	mostly	mostly	ADV
bracis-33550	192	24	based	base	VERB
bracis-33550	192	25	on	on	ADP
bracis-33550	192	26	spontaneous	spontaneous	ADJ
bracis-33550	192	27	speech	speech	NOUN
bracis-33550	192	28	,	,	PUNCT
bracis-33550	192	29	continues	continue	VERB
bracis-33550	192	30	presenting	present	VERB
bracis-33550	192	31	the	the	DET
bracis-33550	192	32	best	good	ADJ
bracis-33550	192	33	values	value	NOUN
bracis-33550	192	34	for	for	ADP
bracis-33550	192	35	this	this	DET
bracis-33550	192	36	dataset	dataset	NOUN
bracis-33550	192	37	with	with	ADP
bracis-33550	192	38	24.18	24.18	NUM
bracis-33550	192	39	%	%	NOUN
bracis-33550	192	40	wer	wer	NOUN
bracis-33550	192	41	,	,	PUNCT
bracis-33550	192	42	but	but	CCONJ
bracis-33550	192	43	our	our	PRON
bracis-33550	192	44	wav2vec2	wav2vec2	PROPN
bracis-33550	192	45	-	-	PUNCT
bracis-33550	192	46	nurc	nurc	PROPN
bracis-33550	192	47	-	-	PUNCT
bracis-33550	192	48	sp-2	sp-2	PROPN
bracis-33550	192	49	model	model	NOUN
bracis-33550	192	50	was	be	AUX
bracis-33550	192	51	in	in	ADP
bracis-33550	192	52	second	second	ADJ
bracis-33550	192	53	position	position	NOUN
bracis-33550	192	54	(	(	PUNCT
bracis-33550	192	55	since	since	SCONJ
bracis-33550	192	56	it	it	PRON
bracis-33550	192	57	is	be	AUX
bracis-33550	192	58	also	also	ADV
bracis-33550	192	59	trained	train	VERB
bracis-33550	192	60	for	for	ADP
bracis-33550	192	61	spontaneous	spontaneous	ADJ
bracis-33550	192	62	speech	speech	NOUN
bracis-33550	192	63	)	)	PUNCT
bracis-33550	192	64	.	.	PUNCT
bracis-33550	193	1	for	for	ADP
bracis-33550	193	2	the	the	DET
bracis-33550	193	3	new	new	ADJ
bracis-33550	193	4	dataset	dataset	NOUN
bracis-33550	193	5	,	,	PUNCT
bracis-33550	193	6	focus	focus	NOUN
bracis-33550	193	7	of	of	ADP
bracis-33550	193	8	this	this	DET
bracis-33550	193	9	paper	paper	NOUN
bracis-33550	193	10	,	,	PUNCT
bracis-33550	193	11	the	the	DET
bracis-33550	193	12	best	good	ADJ
bracis-33550	193	13	results	result	NOUN
bracis-33550	193	14	were	be	AUX
bracis-33550	193	15	the	the	DET
bracis-33550	193	16	distil	distil	NOUN
bracis-33550	193	17	-	-	PUNCT
bracis-33550	193	18	whisper	whisper	NOUN
bracis-33550	193	19	fine	fine	ADV
bracis-33550	193	20	-	-	PUNCT
bracis-33550	193	21	tuned	tune	VERB
bracis-33550	193	22	over	over	ADP
bracis-33550	193	23	nurc	nurc	ADV
bracis-33550	193	24	-	-	PUNCT
bracis-33550	193	25	sp	sp	NOUN
bracis-33550	193	26	audio	audio	NOUN
bracis-33550	193	27	corpus	corpus	NOUN
bracis-33550	193	28	with	with	ADP
bracis-33550	193	29	a	a	DET
bracis-33550	193	30	wer	wer	NOUN
bracis-33550	193	31	of	of	ADP
bracis-33550	193	32	24.22	24.22	NUM
bracis-33550	193	33	%	%	NOUN
bracis-33550	193	34	followed	follow	VERB
bracis-33550	193	35	by	by	ADP
bracis-33550	193	36	a	a	DET
bracis-33550	193	37	fine	fine	ADV
bracis-33550	193	38	-	-	PUNCT
bracis-33550	193	39	tuned	tune	VERB
bracis-33550	193	40	versions	version	NOUN
bracis-33550	193	41	of	of	ADP
bracis-33550	193	42	wav2vec2	wav2vec2	PROPN
bracis-33550	193	43	-	-	PUNCT
bracis-33550	193	44	xlsr-53	xlsr-53	PROPN
bracis-33550	193	45	model	model	NOUN
bracis-33550	193	46	with	with	ADP
bracis-33550	193	47	a	a	DET
bracis-33550	193	48	wer	wer	NOUN
bracis-33550	193	49	of	of	ADP
bracis-33550	193	50	33.73	33.73	NUM
bracis-33550	193	51	%	%	NOUN
bracis-33550	193	52	,	,	PUNCT
bracis-33550	193	53	that	that	PRON
bracis-33550	193	54	is	be	AUX
bracis-33550	193	55	almost	almost	ADV
bracis-33550	193	56	10	10	NUM
bracis-33550	193	57	%	%	NOUN
bracis-33550	193	58	point	point	NOUN
bracis-33550	193	59	worse	bad	ADJ
bracis-33550	193	60	than	than	ADP
bracis-33550	193	61	distil	distil	NOUN
bracis-33550	193	62	-	-	PUNCT
bracis-33550	193	63	whisper	whisper	NOUN
bracis-33550	193	64	’s	’s	NOUN
bracis-33550	193	65	.	.	PUNCT
bracis-33550	194	1	these	these	DET
bracis-33550	194	2	results	result	NOUN
bracis-33550	194	3	indicates	indicate	VERB
bracis-33550	194	4	that	that	SCONJ
bracis-33550	194	5	distil	distil	NOUN
bracis-33550	194	6	-	-	PUNCT
bracis-33550	194	7	whisper	whisper	NOUN
bracis-33550	194	8	is	be	AUX
bracis-33550	194	9	promising	promise	VERB
bracis-33550	194	10	for	for	ADP
bracis-33550	194	11	low	low	ADJ
bracis-33550	194	12	and	and	CCONJ
bracis-33550	194	13	medium	medium	ADJ
bracis-33550	194	14	resource	resource	NOUN
bracis-33550	194	15	languages	language	NOUN
bracis-33550	194	16	and	and	CCONJ
bracis-33550	194	17	should	should	AUX
bracis-33550	194	18	be	be	AUX
bracis-33550	194	19	evaluated	evaluate	VERB
bracis-33550	194	20	with	with	ADP
bracis-33550	194	21	more	more	ADJ
bracis-33550	194	22	bp	bp	PROPN
bracis-33550	194	23	datasets	dataset	NOUN
bracis-33550	194	24	in	in	ADP
bracis-33550	194	25	the	the	DET
bracis-33550	194	26	future	future	NOUN
bracis-33550	194	27	.	.	PUNCT
bracis-33550	195	1	notes	note	NOUN
bracis-33550	195	2	1.https://www.ted.com/talks	1.https://www.ted.com/talks	NUM
bracis-33550	195	3	.	.	PUNCT
bracis-33550	196	1	2.for	2.for	ADP
bracis-33550	196	2	example	example	NOUN
bracis-33550	196	3	,	,	PUNCT
bracis-33550	196	4	[	[	X
bracis-33550	196	5	5	5	NUM
bracis-33550	196	6	]	]	PUNCT
bracis-33550	196	7	comments	comment	NOUN
bracis-33550	196	8	on	on	ADP
bracis-33550	196	9	vowel	vowel	NOUN
bracis-33550	196	10	elision	elision	NOUN
bracis-33550	196	11	between	between	ADP
bracis-33550	196	12	words	word	NOUN
bracis-33550	196	13	that	that	PRON
bracis-33550	196	14	,	,	PUNCT
bracis-33550	196	15	in	in	ADP
bracis-33550	196	16	the	the	DET
bracis-33550	196	17	são	são	PROPN
bracis-33550	196	18	paulo	paulo	PROPN
bracis-33550	196	19	dialect	dialect	PROPN
bracis-33550	196	20	,	,	PUNCT
bracis-33550	196	21	affects	affect	VERB
bracis-33550	196	22	the	the	DET
bracis-33550	196	23	final	final	ADJ
bracis-33550	196	24	posttonic	posttonic	ADJ
bracis-33550	196	25	vowels	vowel	NOUN
bracis-33550	196	26	/a/	/a/	PRON
bracis-33550	196	27	,	,	PUNCT
bracis-33550	196	28	/o/	/o/	NOUN
bracis-33550	196	29	and	and	CCONJ
bracis-33550	196	30	/u/.	/u/.	PUNCT
bracis-33550	196	31	for	for	ADP
bracis-33550	196	32	instance	instance	NOUN
bracis-33550	196	33	,	,	PUNCT
bracis-33550	196	34	in	in	ADP
bracis-33550	196	35	the	the	DET
bracis-33550	196	36	example	example	NOUN
bracis-33550	196	37	(	(	PUNCT
bracis-33550	196	38	a	a	X
bracis-33550	196	39	)	)	PUNCT
bracis-33550	196	40	me’ren[da	me’ren[da	NOUN
bracis-33550	196	41	es]co’lar	es]co’lar	PRON
bracis-33550	196	42	(	(	PUNCT
bracis-33550	196	43	school	school	NOUN
bracis-33550	196	44	lunch	lunch	NOUN
bracis-33550	196	45	)	)	PUNCT
bracis-33550	196	46	–	–	PUNCT
bracis-33550	196	47	>	>	X
bracis-33550	196	48	me’ren[des]co’lar	me’ren[des]co’lar	PROPN
bracis-33550	196	49	,	,	PUNCT
bracis-33550	196	50	the	the	DET
bracis-33550	196	51	vowel	vowel	NOUN
bracis-33550	196	52	/a/	/a/	X
bracis-33550	196	53	is	be	AUX
bracis-33550	196	54	deleted	delete	VERB
bracis-33550	196	55	and	and	CCONJ
bracis-33550	196	56	a	a	DET
bracis-33550	196	57	new	new	ADJ
bracis-33550	196	58	syllable	syllable	NOUN
bracis-33550	196	59	is	be	AUX
bracis-33550	196	60	created	create	VERB
bracis-33550	196	61	(	(	PUNCT
bracis-33550	196	62	[	[	X
bracis-33550	196	63	des	des	X
bracis-33550	196	64	]	]	X
bracis-33550	196	65	)	)	PUNCT
bracis-33550	196	66	.	.	PUNCT
bracis-33550	197	1	3.https://openslr.org/150/.	3.https://openslr.org/150/.	NUM
bracis-33550	197	2	4.https://groups.inf.ed.ac.uk/ami/corpus/overview.shtml	4.https://groups.inf.ed.ac.uk/ami/corpus/overview.shtml	NUM
bracis-33550	197	3	.	.	PUNCT
bracis-33550	198	1	5.https://commonvoice.mozilla.org/en/datasets	5.https://commonvoice.mozilla.org/en/dataset	NOUN
bracis-33550	198	2	.	.	PUNCT
bracis-33550	199	1	6.https://www.openslr.org/51/.	6.https://www.openslr.org/51/.	NUM
bracis-33550	199	2	7.https://github.com/nilc-nlp/coraa	7.https://github.com/nilc-nlp/coraa	NUM
bracis-33550	199	3	.	.	PUNCT
bracis-33550	200	1	8.github.com/nilc-nlp/nurc-sp-audio-corpus	8.github.com/nilc-nlp/nurc-sp-audio-corpus	NUM
bracis-33550	200	2	.	.	PUNCT
bracis-33550	201	1	9.https://librivox.org/pages/about-librivox/.	9.https://librivox.org/pages/about-librivox/.	NUM
bracis-33550	201	2	10.https://www.openslr.org/12	10.https://www.openslr.org/12	NUM
bracis-33550	201	3	.	.	PUNCT
bracis-33550	202	1	11.https://www.openslr.org/60/.	11.https://www.openslr.org/60/.	NUM
bracis-33550	202	2	12.https://arxiv.org/abs/2209.11871	12.https://arxiv.org/abs/2209.11871	NUM
bracis-33550	202	3	.	.	PUNCT
bracis-33550	203	1	13.https://sites.google.com/view/tarsila-c4ai/home	13.https://sites.google.com/view/tarsila-c4ai/home	NUM
bracis-33550	203	2	.	.	PUNCT
bracis-33550	203	3	14.https://github.com/edresson/wav2vec-wrapper	14.https://github.com/edresson/wav2vec-wrapper	NOUN
bracis-33550	203	4	.	.	PUNCT
bracis-33550	204	1	15.https://huggingface.co/edresson/wav2vec2-large-xlsr-coraa-portuguese	15.https://huggingface.co/edresson/wav2vec2-large-xlsr-coraa-portuguese	X
bracis-33550	204	2	.	.	PUNCT
bracis-33550	205	1	16.https://github.com/huggingface/distil-whisper/tree/main/training	16.https://github.com/huggingface/distil-whisper/tree/main/training	NUM
bracis-33550	205	2	.	.	PUNCT
bracis-33550	206	1	17.https://huggingface.co/blog/fine-tune-whisper	17.https://huggingface.co/blog/fine-tune-whisper	NUM
bracis-33550	206	2	.	.	PUNCT
bracis-33550	207	1	18.here	18.here	X
bracis-33550	207	2	,	,	PUNCT
bracis-33550	207	3	we	we	PRON
bracis-33550	207	4	also	also	ADV
bracis-33550	207	5	focus	focus	VERB
bracis-33550	207	6	our	our	PRON
bracis-33550	207	7	analysis	analysis	NOUN
bracis-33550	207	8	on	on	ADP
bracis-33550	207	9	the	the	DET
bracis-33550	207	10	metric	metric	ADJ
bracis-33550	207	11	cer	cer	NOUN
bracis-33550	207	12	,	,	PUNCT
bracis-33550	207	13	because	because	SCONJ
bracis-33550	207	14	for	for	ADP
bracis-33550	207	15	smaller	small	ADJ
bracis-33550	207	16	audios	audio	NOUN
bracis-33550	207	17	,	,	PUNCT
bracis-33550	207	18	with	with	ADP
bracis-33550	207	19	just	just	ADV
bracis-33550	207	20	a	a	DET
bracis-33550	207	21	few	few	ADJ
bracis-33550	207	22	words	word	NOUN
bracis-33550	207	23	,	,	PUNCT
bracis-33550	207	24	this	this	DET
bracis-33550	207	25	metric	metric	NOUN
bracis-33550	207	26	tends	tend	VERB
bracis-33550	207	27	to	to	PART
bracis-33550	207	28	be	be	AUX
bracis-33550	207	29	more	more	ADV
bracis-33550	207	30	reliable	reliable	ADJ
bracis-33550	207	31	.	.	PUNCT
bracis-33550	208	1	references	reference	NOUN
bracis-33550	208	2	ardila	ardila	PROPN
bracis-33550	208	3	,	,	PUNCT
bracis-33550	208	4	r.	r.	PROPN
bracis-33550	208	5	,	,	PUNCT
bracis-33550	208	6	et	et	PROPN
bracis-33550	208	7	al	al	PROPN
bracis-33550	208	8	.	.	PUNCT
bracis-33550	208	9	:	:	PUNCT
bracis-33550	209	1	common	common	ADJ
bracis-33550	209	2	voice	voice	NOUN
bracis-33550	209	3	:	:	PUNCT
bracis-33550	209	4	a	a	DET
bracis-33550	209	5	massively	massively	ADV
bracis-33550	209	6	-	-	PUNCT
bracis-33550	209	7	multilingual	multilingual	ADJ
bracis-33550	209	8	speech	speech	NOUN
bracis-33550	209	9	corpus	corpus	NOUN
bracis-33550	209	10	.	.	PUNCT
bracis-33550	210	1	in	in	ADP
bracis-33550	210	2	:	:	PUNCT
bracis-33550	210	3	calzolari	calzolari	PROPN
bracis-33550	210	4	,	,	PUNCT
bracis-33550	210	5	n.	n.	PROPN
bracis-33550	210	6	,	,	PUNCT
bracis-33550	210	7	et	et	PROPN
bracis-33550	210	8	al	al	PROPN
bracis-33550	210	9	.	.	PUNCT
bracis-33550	211	1	(	(	PUNCT
bracis-33550	211	2	eds	ed	NOUN
bracis-33550	211	3	.	.	PUNCT
bracis-33550	211	4	)	)	PUNCT
bracis-33550	211	5	proceedings	proceeding	NOUN
bracis-33550	211	6	of	of	ADP
bracis-33550	211	7	the	the	DET
bracis-33550	211	8	twelfth	twelfth	ADJ
bracis-33550	211	9	language	language	NOUN
bracis-33550	211	10	resources	resource	NOUN
bracis-33550	211	11	and	and	CCONJ
bracis-33550	211	12	evaluation	evaluation	NOUN
bracis-33550	211	13	conference	conference	NOUN
bracis-33550	211	14	,	,	PUNCT
bracis-33550	211	15	may	may	PROPN
bracis-33550	211	16	2020	2020	NUM
bracis-33550	211	17	,	,	PUNCT
bracis-33550	211	18	marseille	marseille	NOUN
bracis-33550	211	19	,	,	PUNCT
bracis-33550	211	20	france	france	PROPN
bracis-33550	211	21	,	,	PUNCT
bracis-33550	211	22	pp	pp	X
bracis-33550	211	23	.	.	PUNCT
bracis-33550	212	1	4218–4222	4218–4222	NOUN
bracis-33550	212	2	.	.	PUNCT
bracis-33550	213	1	european	european	ADJ
bracis-33550	213	2	language	language	PROPN
bracis-33550	213	3	resources	resource	NOUN
bracis-33550	213	4	association	association	NOUN
bracis-33550	213	5	(	(	PUNCT
bracis-33550	213	6	2020	2020	NUM
bracis-33550	213	7	)	)	PUNCT
bracis-33550	213	8	.	.	PUNCT
bracis-33550	214	1	https://aclanthology.org/2020.lrec-1.520	https://aclanthology.org/2020.lrec-1.520	PROPN
bracis-33550	214	2	baevski	baevski	PROPN
bracis-33550	214	3	,	,	PUNCT
bracis-33550	214	4	a.	a.	PROPN
bracis-33550	214	5	,	,	PUNCT
bracis-33550	214	6	zhou	zhou	PROPN
bracis-33550	214	7	,	,	PUNCT
bracis-33550	214	8	h.	h.	PROPN
bracis-33550	214	9	,	,	PUNCT
bracis-33550	214	10	mohamed	mohamed	PROPN
bracis-33550	214	11	,	,	PUNCT
bracis-33550	214	12	a.	a.	PROPN
bracis-33550	214	13	,	,	PUNCT
bracis-33550	214	14	auli	auli	PROPN
bracis-33550	214	15	,	,	PUNCT
bracis-33550	214	16	m.	m.	NOUN
bracis-33550	214	17	:	:	PUNCT
bracis-33550	214	18	wav2vec	wav2vec	NOUN
bracis-33550	214	19	2.0	2.0	NUM
bracis-33550	214	20	:	:	PUNCT
bracis-33550	214	21	a	a	DET
bracis-33550	214	22	framework	framework	NOUN
bracis-33550	214	23	for	for	ADP
bracis-33550	214	24	self	self	NOUN
bracis-33550	214	25	-	-	PUNCT
bracis-33550	214	26	supervised	supervise	VERB
bracis-33550	214	27	learning	learning	NOUN
bracis-33550	214	28	of	of	ADP
bracis-33550	214	29	speech	speech	NOUN
bracis-33550	214	30	representations	representation	NOUN
bracis-33550	214	31	.	.	PUNCT
bracis-33550	215	1	adv	adv	PROPN
bracis-33550	215	2	.	.	PUNCT
bracis-33550	215	3	neural	neural	PROPN
bracis-33550	215	4	inf	inf	PROPN
bracis-33550	215	5	.	.	PUNCT
bracis-33550	215	6	process	process	NOUN
bracis-33550	215	7	.	.	PUNCT
bracis-33550	216	1	syst	syst	PROPN
bracis-33550	216	2	.	.	PUNCT
bracis-33550	217	1	33	33	NUM
bracis-33550	217	2	,	,	PUNCT
bracis-33550	217	3	12449–12460	12449–12460	NUM
bracis-33550	217	4	(	(	PUNCT
bracis-33550	217	5	2020	2020	NUM
bracis-33550	217	6	)	)	PUNCT
bracis-33550	217	7	.	.	PUNCT
bracis-33550	218	1	https://arxiv.org/abs/2006.11477	https://arxiv.org/abs/2006.11477	NOUN
bracis-33550	218	2	bain	bain	PROPN
bracis-33550	218	3	,	,	PUNCT
bracis-33550	218	4	m.	m.	NOUN
bracis-33550	218	5	,	,	PUNCT
bracis-33550	218	6	huh	huh	INTJ
bracis-33550	218	7	,	,	PUNCT
bracis-33550	218	8	j.	j.	PROPN
bracis-33550	218	9	,	,	PUNCT
bracis-33550	218	10	han	han	PROPN
bracis-33550	218	11	,	,	PUNCT
bracis-33550	218	12	t.	t.	PROPN
bracis-33550	218	13	,	,	PUNCT
bracis-33550	218	14	zisserman	zisserman	NOUN
bracis-33550	218	15	,	,	PUNCT
bracis-33550	218	16	a.	a.	NOUN
bracis-33550	218	17	:	:	PUNCT
bracis-33550	218	18	whisperx	whisperx	NOUN
bracis-33550	218	19	:	:	PUNCT
bracis-33550	218	20	time	time	NOUN
bracis-33550	218	21	-	-	PUNCT
bracis-33550	218	22	accurate	accurate	ADJ
bracis-33550	218	23	speech	speech	NOUN
bracis-33550	218	24	transcription	transcription	NOUN
bracis-33550	218	25	of	of	ADP
bracis-33550	218	26	long	long	ADJ
bracis-33550	218	27	-	-	PUNCT
bracis-33550	218	28	form	form	NOUN
bracis-33550	218	29	audio	audio	NOUN
bracis-33550	218	30	.	.	PUNCT
bracis-33550	219	1	in	in	ADP
bracis-33550	219	2	:	:	PUNCT
bracis-33550	219	3	interspeech	interspeech	NOUN
bracis-33550	219	4	2023	2023	NUM
bracis-33550	219	5	,	,	PUNCT
bracis-33550	219	6	pp	pp	ADJ
bracis-33550	219	7	.	.	PUNCT
bracis-33550	220	1	4489–4493	4489–4493	NUM
bracis-33550	220	2	(	(	PUNCT
bracis-33550	220	3	2023	2023	NUM
bracis-33550	220	4	)	)	PUNCT
bracis-33550	220	5	.	.	PUNCT
bracis-33550	221	1	https://doi.org/10.21437/interspeech.2023-78	https://doi.org/10.21437/interspeech.2023-78	PROPN
bracis-33550	221	2	beckman	beckman	PROPN
bracis-33550	221	3	,	,	PUNCT
bracis-33550	221	4	m.e	m.e	PROPN
bracis-33550	221	5	.	.	PROPN
bracis-33550	221	6	:	:	PUNCT
bracis-33550	222	1	a	a	DET
bracis-33550	222	2	typology	typology	NOUN
bracis-33550	222	3	of	of	ADP
bracis-33550	222	4	spontaneous	spontaneous	ADJ
bracis-33550	222	5	speech	speech	NOUN
bracis-33550	222	6	.	.	PUNCT
bracis-33550	223	1	in	in	ADP
bracis-33550	223	2	:	:	PUNCT
bracis-33550	223	3	computing	compute	VERB
bracis-33550	223	4	prosody	prosody	NOUN
bracis-33550	223	5	:	:	PUNCT
bracis-33550	223	6	computational	computational	ADJ
bracis-33550	223	7	models	model	NOUN
bracis-33550	223	8	for	for	ADP
bracis-33550	223	9	processing	process	VERB
bracis-33550	223	10	spontaneous	spontaneous	ADJ
bracis-33550	223	11	speech	speech	NOUN
bracis-33550	223	12	,	,	PUNCT
bracis-33550	223	13	pp	pp	ADP
bracis-33550	223	14	.	.	PUNCT
bracis-33550	224	1	7–26	7–26	NOUN
bracis-33550	224	2	.	.	PUNCT
bracis-33550	225	1	springer	springer	NOUN
bracis-33550	225	2	,	,	PUNCT
bracis-33550	225	3	new	new	PROPN
bracis-33550	225	4	york	york	PROPN
bracis-33550	225	5	(	(	PUNCT
bracis-33550	225	6	1997	1997	NUM
bracis-33550	225	7	)	)	PUNCT
bracis-33550	225	8	.	.	PUNCT
bracis-33550	226	1	https://doi.org/10.1007/978-1-4612-2258-3_2	https://doi.org/10.1007/978-1-4612-2258-3_2	PROPN
bracis-33550	226	2	bohn	bohn	PROPN
bracis-33550	226	3	,	,	PUNCT
bracis-33550	226	4	g.p	g.p	PROPN
bracis-33550	226	5	.	.	PROPN
bracis-33550	226	6	:	:	PUNCT
bracis-33550	226	7	processos	processos	PROPN
bracis-33550	226	8	e	e	PROPN
bracis-33550	226	9	representações	representações	PROPN
bracis-33550	226	10	lexicais	lexicais	PROPN
bracis-33550	226	11	:	:	PUNCT
bracis-33550	226	12	o	o	PROPN
bracis-33550	226	13	caso	caso	PROPN
bracis-33550	226	14	das	das	PROPN
bracis-33550	226	15	vogais	vogais	PROPN
bracis-33550	226	16	posteriores	posteriores	PROPN
bracis-33550	226	17	do	do	AUX
bracis-33550	226	18	dialeto	dialeto	VERB
bracis-33550	226	19	paulista	paulista	PROPN
bracis-33550	226	20	.	.	PUNCT
bracis-33550	227	1	delta	delta	NOUN
bracis-33550	227	2	:	:	PUNCT
bracis-33550	227	3	documentação	documentação	PROPN
bracis-33550	227	4	e	e	PROPN
bracis-33550	227	5	estudos	estudos	PROPN
bracis-33550	227	6	em	em	PRON
bracis-33550	227	7	linguística	linguística	PROPN
bracis-33550	227	8	teórica	teórica	PROPN
bracis-33550	227	9	e	e	PROPN
bracis-33550	227	10	aplicada	aplicada	PROPN
bracis-33550	227	11	33(2	33(2	NUM
bracis-33550	227	12	)	)	PUNCT
bracis-33550	227	13	,	,	PUNCT
bracis-33550	227	14	september	september	PROPN
bracis-33550	227	15	2017	2017	NUM
bracis-33550	227	16	.	.	PUNCT
bracis-33550	228	1	https://revistas.pucsp.br/index.php/delta/article/view/34370	https://revistas.pucsp.br/index.php/delta/article/view/34370	PROPN
bracis-33550	228	2	candido_junior	candido_junior	PROPN
bracis-33550	228	3	,	,	PUNCT
bracis-33550	228	4	a.	a.	NOUN
bracis-33550	228	5	,	,	PUNCT
bracis-33550	228	6	et	et	PROPN
bracis-33550	228	7	al	al	PROPN
bracis-33550	228	8	.	.	PUNCT
bracis-33550	228	9	:	:	PUNCT
bracis-33550	229	1	coraa	coraa	PROPN
bracis-33550	229	2	asr	asr	PROPN
bracis-33550	229	3	:	:	PUNCT
bracis-33550	229	4	a	a	DET
bracis-33550	229	5	large	large	ADJ
bracis-33550	229	6	corpus	corpus	NOUN
bracis-33550	229	7	of	of	ADP
bracis-33550	229	8	spontaneous	spontaneous	ADJ
bracis-33550	229	9	and	and	CCONJ
bracis-33550	229	10	prepared	prepared	ADJ
bracis-33550	229	11	speech	speech	NOUN
bracis-33550	229	12	manually	manually	ADV
bracis-33550	229	13	validated	validate	VERB
bracis-33550	229	14	for	for	ADP
bracis-33550	229	15	speech	speech	NOUN
bracis-33550	229	16	recognition	recognition	NOUN
bracis-33550	229	17	in	in	ADP
bracis-33550	229	18	brazilian	brazilian	ADJ
bracis-33550	229	19	portuguese	portuguese	PROPN
bracis-33550	229	20	.	.	PUNCT
bracis-33550	230	1	lang	lang	PROPN
bracis-33550	230	2	.	.	PUNCT
bracis-33550	231	1	resour	resour	PROPN
bracis-33550	231	2	.	.	PUNCT
bracis-33550	232	1	eval	eval	PROPN
bracis-33550	232	2	.	.	PUNCT
bracis-33550	233	1	57	57	NUM
bracis-33550	233	2	,	,	PUNCT
bracis-33550	233	3	1139–1171	1139–1171	NUM
bracis-33550	233	4	(	(	PUNCT
bracis-33550	233	5	2023	2023	NUM
bracis-33550	233	6	)	)	PUNCT
bracis-33550	233	7	.	.	PUNCT
bracis-33550	234	1	https://doi.org/10.1007/s10579-022-09621-4	https://doi.org/10.1007/s10579-022-09621-4	PROPN
bracis-33550	234	2	.	.	PUNCT
bracis-33550	235	1	https://link.springer.com/article/10.1007/s10579-022-09621-4	https://link.springer.com/article/10.1007/s10579-022-09621-4	PROPN
bracis-33550	235	2	clifton	clifton	PROPN
bracis-33550	235	3	,	,	PUNCT
bracis-33550	235	4	a.	a.	PROPN
bracis-33550	235	5	,	,	PUNCT
bracis-33550	235	6	et	et	PROPN
bracis-33550	235	7	al	al	PROPN
bracis-33550	235	8	.	.	PUNCT
bracis-33550	235	9	:	:	PUNCT
bracis-33550	235	10	100,000	100,000	NUM
bracis-33550	235	11	podcasts	podcast	NOUN
bracis-33550	235	12	:	:	PUNCT
bracis-33550	235	13	a	a	DET
bracis-33550	235	14	spoken	spoken	ADJ
bracis-33550	235	15	english	english	ADJ
bracis-33550	235	16	document	document	NOUN
bracis-33550	235	17	corpus	corpus	NOUN
bracis-33550	235	18	.	.	PUNCT
bracis-33550	236	1	in	in	ADP
bracis-33550	236	2	:	:	PUNCT
bracis-33550	236	3	scott	scott	PROPN
bracis-33550	236	4	,	,	PUNCT
bracis-33550	236	5	d.	d.	PROPN
bracis-33550	236	6	,	,	PUNCT
bracis-33550	236	7	bel	bel	PROPN
bracis-33550	236	8	,	,	PUNCT
bracis-33550	236	9	n.	n.	PROPN
bracis-33550	236	10	,	,	PUNCT
bracis-33550	236	11	zong	zong	PROPN
bracis-33550	236	12	,	,	PUNCT
bracis-33550	236	13	c.	c.	PROPN
bracis-33550	236	14	(	(	PUNCT
bracis-33550	236	15	eds	eds	PROPN
bracis-33550	236	16	.	.	PUNCT
bracis-33550	236	17	)	)	PUNCT
bracis-33550	236	18	proceedings	proceeding	NOUN
bracis-33550	236	19	of	of	ADP
bracis-33550	236	20	the	the	DET
bracis-33550	236	21	28th	28th	ADJ
bracis-33550	236	22	international	international	ADJ
bracis-33550	236	23	conference	conference	NOUN
bracis-33550	236	24	on	on	ADP
bracis-33550	236	25	computational	computational	ADJ
bracis-33550	236	26	linguistics	linguistic	NOUN
bracis-33550	236	27	,	,	PUNCT
bracis-33550	236	28	barcelona	barcelona	PROPN
bracis-33550	236	29	,	,	PUNCT
bracis-33550	236	30	spain	spain	PROPN
bracis-33550	236	31	(	(	PUNCT
bracis-33550	236	32	online	online	ADJ
bracis-33550	236	33	)	)	PUNCT
bracis-33550	236	34	,	,	PUNCT
bracis-33550	236	35	december	december	PROPN
bracis-33550	236	36	2020	2020	NUM
bracis-33550	236	37	,	,	PUNCT
bracis-33550	236	38	pp	pp	ADV
bracis-33550	236	39	.	.	PUNCT
bracis-33550	236	40	5903–5917	5903–5917	NUM
bracis-33550	236	41	.	.	PUNCT
bracis-33550	237	1	international	international	ADJ
bracis-33550	237	2	committee	committee	PROPN
bracis-33550	237	3	on	on	ADP
bracis-33550	237	4	computational	computational	ADJ
bracis-33550	237	5	linguistics	linguistic	NOUN
bracis-33550	237	6	(	(	PUNCT
bracis-33550	237	7	2020	2020	NUM
bracis-33550	237	8	)	)	PUNCT
bracis-33550	237	9	.	.	PUNCT
bracis-33550	238	1	https://aclanthology.org/2020.coling-main.519	https://aclanthology.org/2020.coling-main.519	ADJ
bracis-33550	238	2	conneau	conneau	NOUN
bracis-33550	238	3	,	,	PUNCT
bracis-33550	238	4	a.	a.	NOUN
bracis-33550	238	5	,	,	PUNCT
bracis-33550	238	6	baevski	baevski	NOUN
bracis-33550	238	7	,	,	PUNCT
bracis-33550	238	8	a.	a.	PROPN
bracis-33550	238	9	,	,	PUNCT
bracis-33550	238	10	collobert	collobert	PROPN
bracis-33550	238	11	,	,	PUNCT
bracis-33550	238	12	r.	r.	PROPN
bracis-33550	238	13	,	,	PUNCT
bracis-33550	238	14	mohamed	mohamed	PROPN
bracis-33550	238	15	,	,	PUNCT
bracis-33550	238	16	a.	a.	PROPN
bracis-33550	238	17	,	,	PUNCT
bracis-33550	238	18	auli	auli	PROPN
bracis-33550	238	19	,	,	PUNCT
bracis-33550	238	20	m.	m.	NOUN
bracis-33550	238	21	:	:	PUNCT
bracis-33550	239	1	unsupervised	unsupervised	ADJ
bracis-33550	239	2	cross	cross	ADJ
bracis-33550	239	3	-	-	ADJ
bracis-33550	239	4	lingual	lingual	ADJ
bracis-33550	239	5	representation	representation	NOUN
bracis-33550	239	6	learning	learn	VERB
bracis-33550	239	7	for	for	ADP
bracis-33550	239	8	speech	speech	NOUN
bracis-33550	239	9	recognition	recognition	NOUN
bracis-33550	239	10	.	.	PUNCT
bracis-33550	240	1	in	in	ADP
bracis-33550	240	2	:	:	PUNCT
bracis-33550	240	3	proceedings	proceeding	NOUN
bracis-33550	240	4	of	of	ADP
bracis-33550	240	5	the	the	DET
bracis-33550	240	6	interspeech	interspeech	NOUN
bracis-33550	240	7	2021	2021	NUM
bracis-33550	240	8	,	,	PUNCT
bracis-33550	240	9	pp	pp	ADP
bracis-33550	240	10	.	.	PUNCT
bracis-33550	241	1	2426–2430	2426–2430	NUM
bracis-33550	241	2	(	(	PUNCT
bracis-33550	241	3	2021	2021	NUM
bracis-33550	241	4	)	)	PUNCT
bracis-33550	241	5	.	.	PUNCT
bracis-33550	242	1	https://doi.org/10.21437/interspeech.2021-329	https://doi.org/10.21437/interspeech.2021-329	NOUN
bracis-33550	242	2	conneau	conneau	NOUN
bracis-33550	242	3	,	,	PUNCT
bracis-33550	242	4	a.	a.	PROPN
bracis-33550	242	5	,	,	PUNCT
bracis-33550	242	6	et	et	PROPN
bracis-33550	242	7	al	al	PROPN
bracis-33550	242	8	.	.	PROPN
bracis-33550	242	9	:	:	PUNCT
bracis-33550	243	1	unsupervised	unsupervised	ADJ
bracis-33550	243	2	cross	cross	ADJ
bracis-33550	243	3	-	-	ADJ
bracis-33550	243	4	lingual	lingual	ADJ
bracis-33550	243	5	representation	representation	NOUN
bracis-33550	243	6	learning	learn	VERB
bracis-33550	243	7	at	at	ADP
bracis-33550	243	8	scale	scale	NOUN
bracis-33550	243	9	.	.	PUNCT
bracis-33550	244	1	in	in	ADP
bracis-33550	244	2	:	:	PUNCT
bracis-33550	244	3	proceedings	proceeding	NOUN
bracis-33550	244	4	of	of	ADP
bracis-33550	244	5	the	the	DET
bracis-33550	244	6	58th	58th	ADJ
bracis-33550	244	7	annual	annual	ADJ
bracis-33550	244	8	meeting	meeting	NOUN
bracis-33550	244	9	of	of	ADP
bracis-33550	244	10	the	the	DET
bracis-33550	244	11	association	association	NOUN
bracis-33550	244	12	for	for	ADP
bracis-33550	244	13	computational	computational	ADJ
bracis-33550	244	14	linguistics	linguistic	NOUN
bracis-33550	244	15	,	,	PUNCT
bracis-33550	244	16	pp	pp	X
bracis-33550	244	17	.	.	PUNCT
bracis-33550	244	18	8440–8451	8440–8451	NUM
bracis-33550	244	19	(	(	PUNCT
bracis-33550	244	20	2020	2020	NUM
bracis-33550	244	21	)	)	PUNCT
bracis-33550	244	22	google	google	PROPN
bracis-33550	244	23	scholar	scholar	NOUN
bracis-33550	244	24	  	  	SPACE
bracis-33550	244	25	gabler	gabler	NOUN
bracis-33550	244	26	,	,	PUNCT
bracis-33550	244	27	p.	p.	NOUN
bracis-33550	244	28	,	,	PUNCT
bracis-33550	244	29	geiger	geiger	PROPN
bracis-33550	244	30	,	,	PUNCT
bracis-33550	244	31	b.c	b.c	PROPN
bracis-33550	244	32	.	.	PROPN
bracis-33550	244	33	,	,	PUNCT
bracis-33550	244	34	schuppler	schuppler	NOUN
bracis-33550	244	35	,	,	PUNCT
bracis-33550	244	36	b.	b.	PROPN
bracis-33550	244	37	,	,	PUNCT
bracis-33550	244	38	kern	kern	PROPN
bracis-33550	244	39	,	,	PUNCT
bracis-33550	244	40	r.	r.	PROPN
bracis-33550	244	41	:	:	PUNCT
bracis-33550	244	42	reconsidering	reconsider	VERB
bracis-33550	244	43	read	read	NOUN
bracis-33550	244	44	and	and	CCONJ
bracis-33550	244	45	spontaneous	spontaneous	ADJ
bracis-33550	244	46	speech	speech	NOUN
bracis-33550	244	47	:	:	PUNCT
bracis-33550	244	48	causal	causal	ADJ
bracis-33550	244	49	perspectives	perspective	NOUN
bracis-33550	244	50	on	on	ADP
bracis-33550	244	51	the	the	DET
bracis-33550	244	52	generation	generation	NOUN
bracis-33550	244	53	of	of	ADP
bracis-33550	244	54	training	training	NOUN
bracis-33550	244	55	data	datum	NOUN
bracis-33550	244	56	for	for	ADP
bracis-33550	244	57	automatic	automatic	ADJ
bracis-33550	244	58	speech	speech	NOUN
bracis-33550	244	59	recognition	recognition	NOUN
bracis-33550	244	60	.	.	PUNCT
bracis-33550	245	1	information	information	NOUN
bracis-33550	245	2	14(2	14(2	NUM
bracis-33550	245	3	)	)	PUNCT
bracis-33550	245	4	(	(	PUNCT
bracis-33550	245	5	2023	2023	NUM
bracis-33550	245	6	)	)	PUNCT
bracis-33550	245	7	google	google	NOUN
bracis-33550	245	8	scholar	scholar	NOUN
bracis-33550	245	9	  	  	SPACE
bracis-33550	245	10	gandhi	gandhi	PROPN
bracis-33550	245	11	,	,	PUNCT
bracis-33550	245	12	s.	s.	PROPN
bracis-33550	245	13	,	,	PUNCT
bracis-33550	245	14	von	von	PROPN
bracis-33550	245	15	platen	platen	PROPN
bracis-33550	245	16	,	,	PUNCT
bracis-33550	245	17	p.	p.	NOUN
bracis-33550	245	18	,	,	PUNCT
bracis-33550	245	19	rush	rush	NOUN
bracis-33550	245	20	,	,	PUNCT
bracis-33550	245	21	a.m.	a.m.	ADV
bracis-33550	245	22	:	:	PUNCT
bracis-33550	245	23	distil	distil	NOUN
bracis-33550	245	24	-	-	PUNCT
bracis-33550	245	25	whisper	whisper	NOUN
bracis-33550	245	26	:	:	PUNCT
bracis-33550	245	27	robust	robust	ADJ
bracis-33550	245	28	knowledge	knowledge	NOUN
bracis-33550	245	29	distillation	distillation	NOUN
bracis-33550	245	30	via	via	ADP
bracis-33550	245	31	large	large	ADJ
bracis-33550	245	32	-	-	PUNCT
bracis-33550	245	33	scale	scale	NOUN
bracis-33550	245	34	pseudo	pseudo	NOUN
bracis-33550	245	35	labelling	labelling	NOUN
bracis-33550	245	36	(	(	PUNCT
bracis-33550	245	37	2023	2023	NUM
bracis-33550	245	38	)	)	PUNCT
bracis-33550	245	39	google	google	NOUN
bracis-33550	245	40	scholar	scholar	NOUN
bracis-33550	245	41	  	  	SPACE
bracis-33550	245	42	garmash	garmash	NOUN
bracis-33550	245	43	,	,	PUNCT
bracis-33550	245	44	e.	e.	PROPN
bracis-33550	245	45	,	,	PUNCT
bracis-33550	245	46	et	et	PROPN
bracis-33550	245	47	al	al	PROPN
bracis-33550	245	48	.	.	PUNCT
bracis-33550	245	49	:	:	PUNCT
bracis-33550	246	1	cem	cem	NOUN
bracis-33550	246	2	mil	mil	NOUN
bracis-33550	246	3	podcasts	podcast	NOUN
bracis-33550	246	4	:	:	PUNCT
bracis-33550	246	5	a	a	DET
bracis-33550	246	6	spoken	spoken	ADJ
bracis-33550	246	7	portuguese	portuguese	ADJ
bracis-33550	246	8	document	document	NOUN
bracis-33550	246	9	corpus	corpus	NOUN
bracis-33550	246	10	for	for	ADP
bracis-33550	246	11	multi	multi	ADJ
bracis-33550	246	12	-	-	ADJ
bracis-33550	246	13	modal	modal	ADJ
bracis-33550	246	14	,	,	PUNCT
bracis-33550	246	15	multi	multi	ADJ
bracis-33550	246	16	-	-	ADJ
bracis-33550	246	17	lingual	lingual	ADJ
bracis-33550	246	18	and	and	CCONJ
bracis-33550	246	19	multi	multi	ADJ
bracis-33550	246	20	-	-	ADJ
bracis-33550	246	21	dialect	dialect	ADJ
bracis-33550	246	22	information	information	NOUN
bracis-33550	246	23	access	access	NOUN
bracis-33550	246	24	research	research	NOUN
bracis-33550	246	25	.	.	PUNCT
bracis-33550	247	1	in	in	ADP
bracis-33550	247	2	:	:	PUNCT
bracis-33550	247	3	arampatzis	arampatzi	NOUN
bracis-33550	247	4	,	,	PUNCT
bracis-33550	247	5	a.	a.	NOUN
bracis-33550	247	6	,	,	PUNCT
bracis-33550	247	7	et	et	PROPN
bracis-33550	247	8	al	al	PROPN
bracis-33550	247	9	.	.	PUNCT
bracis-33550	247	10	(	(	PUNCT
bracis-33550	247	11	eds	ed	NOUN
bracis-33550	247	12	.	.	PUNCT
bracis-33550	247	13	)	)	PUNCT
bracis-33550	248	1	experimental	experimental	ADJ
bracis-33550	248	2	ir	ir	PROPN
bracis-33550	248	3	meets	meet	VERB
bracis-33550	248	4	multilinguality	multilinguality	NOUN
bracis-33550	248	5	.	.	PUNCT
bracis-33550	249	1	multimodality	multimodality	NOUN
bracis-33550	249	2	,	,	PUNCT
bracis-33550	249	3	and	and	CCONJ
bracis-33550	249	4	interaction	interaction	NOUN
bracis-33550	249	5	:	:	PUNCT
bracis-33550	249	6	14th	14th	ADJ
bracis-33550	249	7	international	international	ADJ
bracis-33550	249	8	conference	conference	NOUN
bracis-33550	249	9	of	of	ADP
bracis-33550	249	10	the	the	DET
bracis-33550	249	11	clef	clef	PROPN
bracis-33550	249	12	association	association	PROPN
bracis-33550	249	13	,	,	PUNCT
bracis-33550	249	14	clef	clef	NOUN
bracis-33550	249	15	2023	2023	NUM
bracis-33550	249	16	,	,	PUNCT
bracis-33550	249	17	thessaloniki	thessaloniki	PROPN
bracis-33550	249	18	,	,	PUNCT
bracis-33550	249	19	greece	greece	PROPN
bracis-33550	249	20	,	,	PUNCT
bracis-33550	249	21	18–21	18–21	NUM
bracis-33550	249	22	september	september	PROPN
bracis-33550	249	23	2023	2023	NUM
bracis-33550	249	24	,	,	PUNCT
bracis-33550	249	25	proceedings	proceeding	NOUN
bracis-33550	249	26	,	,	PUNCT
bracis-33550	249	27	pp	pp	ADP
bracis-33550	249	28	.	.	PUNCT
bracis-33550	250	1	48–59	48–59	NUM
bracis-33550	250	2	.	.	PUNCT
bracis-33550	251	1	springer	springer	PROPN
bracis-33550	251	2	,	,	PUNCT
bracis-33550	251	3	heidelberg	heidelberg	PROPN
bracis-33550	251	4	(	(	PUNCT
bracis-33550	251	5	2023	2023	NUM
bracis-33550	251	6	)	)	PUNCT
bracis-33550	251	7	.	.	PUNCT
bracis-33550	252	1	https://doi.org/10.1007/978-3-031-42448-9_5	https://doi.org/10.1007/978-3-031-42448-9_5	VERB
bracis-33550	252	2	graves	grave	NOUN
bracis-33550	252	3	,	,	PUNCT
bracis-33550	252	4	a.	a.	NOUN
bracis-33550	252	5	,	,	PUNCT
bracis-33550	252	6	fernández	fernández	PROPN
bracis-33550	252	7	,	,	PUNCT
bracis-33550	252	8	s.	s.	PROPN
bracis-33550	252	9	,	,	PUNCT
bracis-33550	252	10	gomez	gomez	PROPN
bracis-33550	252	11	,	,	PUNCT
bracis-33550	252	12	f.	f.	PROPN
bracis-33550	252	13	,	,	PUNCT
bracis-33550	252	14	schmidhuber	schmidhuber	PROPN
bracis-33550	252	15	,	,	PUNCT
bracis-33550	252	16	j.	j.	PROPN
bracis-33550	252	17	:	:	PUNCT
bracis-33550	252	18	connectionist	connectionist	ADJ
bracis-33550	252	19	temporal	temporal	ADJ
bracis-33550	252	20	classification	classification	NOUN
bracis-33550	252	21	:	:	PUNCT
bracis-33550	252	22	labelling	label	VERB
bracis-33550	252	23	unsegmented	unsegmente	VERB
bracis-33550	252	24	sequence	sequence	NOUN
bracis-33550	252	25	data	datum	NOUN
bracis-33550	252	26	with	with	ADP
bracis-33550	252	27	recurrent	recurrent	ADJ
bracis-33550	252	28	neural	neural	ADJ
bracis-33550	252	29	networks	network	NOUN
bracis-33550	252	30	.	.	PUNCT
bracis-33550	253	1	in	in	ADP
bracis-33550	253	2	:	:	PUNCT
bracis-33550	253	3	proceedings	proceeding	NOUN
bracis-33550	253	4	of	of	ADP
bracis-33550	253	5	the	the	DET
bracis-33550	253	6	23rd	23rd	ADJ
bracis-33550	253	7	international	international	ADJ
bracis-33550	253	8	conference	conference	NOUN
bracis-33550	253	9	on	on	ADP
bracis-33550	253	10	machine	machine	NOUN
bracis-33550	253	11	learning	learning	NOUN
bracis-33550	253	12	,	,	PUNCT
bracis-33550	253	13	pp	pp	ADJ
bracis-33550	253	14	.	.	PUNCT
bracis-33550	254	1	369–376	369–376	NUM
bracis-33550	254	2	(	(	PUNCT
bracis-33550	254	3	2006	2006	NUM
bracis-33550	254	4	)	)	PUNCT
bracis-33550	254	5	google	google	PROPN
bracis-33550	254	6	scholar	scholar	NOUN
bracis-33550	254	7	  	  	SPACE
bracis-33550	254	8	gris	gris	NOUN
bracis-33550	254	9	,	,	PUNCT
bracis-33550	254	10	l.r.s	l.r.s	PROPN
bracis-33550	254	11	.	.	PUNCT
bracis-33550	254	12	,	,	PUNCT
bracis-33550	254	13	casanova	casanova	PROPN
bracis-33550	254	14	,	,	PUNCT
bracis-33550	254	15	e.	e.	PROPN
bracis-33550	254	16	,	,	PUNCT
bracis-33550	254	17	de	de	PROPN
bracis-33550	254	18	 	 	SPACE
bracis-33550	254	19	oliveira	oliveira	PROPN
bracis-33550	254	20	,	,	PUNCT
bracis-33550	254	21	f.s	f.s	PROPN
bracis-33550	254	22	.	.	PROPN
bracis-33550	254	23	,	,	PUNCT
bracis-33550	254	24	da	da	PROPN
bracis-33550	254	25	 	 	SPACE
bracis-33550	254	26	silva	silva	PROPN
bracis-33550	254	27	 	 	SPACE
bracis-33550	254	28	soares	soares	PROPN
bracis-33550	254	29	,	,	PUNCT
bracis-33550	254	30	a.	a.	PROPN
bracis-33550	254	31	,	,	PUNCT
bracis-33550	254	32	candido	candido	PROPN
bracis-33550	254	33	 	 	SPACE
bracis-33550	254	34	junior	junior	PROPN
bracis-33550	254	35	,	,	PUNCT
bracis-33550	254	36	a.	a.	NOUN
bracis-33550	254	37	:	:	PUNCT
bracis-33550	254	38	brazilian	brazilian	ADJ
bracis-33550	254	39	portuguese	portuguese	ADJ
bracis-33550	254	40	speech	speech	NOUN
bracis-33550	254	41	recognition	recognition	NOUN
bracis-33550	254	42	using	use	VERB
bracis-33550	254	43	wav2vec	wav2vec	PROPN
bracis-33550	254	44	2.0	2.0	NUM
bracis-33550	254	45	.	.	PUNCT
bracis-33550	255	1	in	in	ADP
bracis-33550	255	2	:	:	PUNCT
bracis-33550	255	3	pinheiro	pinheiro	PROPN
bracis-33550	255	4	,	,	PUNCT
bracis-33550	255	5	v.	v.	ADV
bracis-33550	255	6	,	,	PUNCT
bracis-33550	255	7	et	et	PROPN
bracis-33550	255	8	al	al	PROPN
bracis-33550	255	9	.	.	PUNCT
bracis-33550	255	10	(	(	PUNCT
bracis-33550	255	11	eds	ed	NOUN
bracis-33550	255	12	.	.	PUNCT
bracis-33550	255	13	)	)	PUNCT
bracis-33550	255	14	computational	computational	ADJ
bracis-33550	255	15	processing	processing	NOUN
bracis-33550	255	16	of	of	ADP
bracis-33550	255	17	the	the	DET
bracis-33550	255	18	portuguese	portuguese	ADJ
bracis-33550	255	19	language	language	NOUN
bracis-33550	255	20	,	,	PUNCT
bracis-33550	255	21	pp	pp	ADJ
bracis-33550	255	22	.	.	PUNCT
bracis-33550	256	1	333–343	333–343	NUM
bracis-33550	256	2	.	.	PUNCT
bracis-33550	257	1	springer	springer	NOUN
bracis-33550	257	2	,	,	PUNCT
bracis-33550	257	3	cham	cham	PROPN
bracis-33550	257	4	(	(	PUNCT
bracis-33550	257	5	2022	2022	NUM
bracis-33550	257	6	)	)	PUNCT
bracis-33550	257	7	.	.	PUNCT
bracis-33550	258	1	https://doi.org/10.1007/978-3-030-98305-5_31	https://doi.org/10.1007/978-3-030-98305-5_31	PROPN
bracis-33550	258	2	gris	gris	PROPN
bracis-33550	258	3	,	,	PUNCT
bracis-33550	258	4	l.r.s	l.r.s	PROPN
bracis-33550	258	5	.	.	PUNCT
bracis-33550	258	6	,	,	PUNCT
bracis-33550	258	7	marcacini	marcacini	PROPN
bracis-33550	258	8	,	,	PUNCT
bracis-33550	258	9	r.	r.	PROPN
bracis-33550	258	10	,	,	PUNCT
bracis-33550	258	11	junior	junior	PROPN
bracis-33550	258	12	,	,	PUNCT
bracis-33550	258	13	a.c	a.c	PROPN
bracis-33550	258	14	.	.	PROPN
bracis-33550	258	15	,	,	PUNCT
bracis-33550	258	16	casanova	casanova	PROPN
bracis-33550	258	17	,	,	PUNCT
bracis-33550	258	18	e.	e.	PROPN
bracis-33550	258	19	,	,	PUNCT
bracis-33550	258	20	soares	soares	PROPN
bracis-33550	258	21	,	,	PUNCT
bracis-33550	258	22	a.	a.	NOUN
bracis-33550	258	23	,	,	PUNCT
bracis-33550	258	24	aluísio	aluísio	NOUN
bracis-33550	258	25	,	,	PUNCT
bracis-33550	258	26	s.m	s.m	PROPN
bracis-33550	258	27	.	.	PROPN
bracis-33550	258	28	:	:	PUNCT
bracis-33550	258	29	evaluating	evaluate	VERB
bracis-33550	258	30	openai	openai	PROPN
bracis-33550	258	31	’s	’s	PART
bracis-33550	258	32	whisper	whisper	NOUN
bracis-33550	258	33	asr	asr	NOUN
bracis-33550	258	34	for	for	ADP
bracis-33550	258	35	punctuation	punctuation	NOUN
bracis-33550	258	36	prediction	prediction	NOUN
bracis-33550	258	37	and	and	CCONJ
bracis-33550	258	38	topic	topic	NOUN
bracis-33550	258	39	modeling	modeling	NOUN
bracis-33550	258	40	of	of	ADP
bracis-33550	258	41	life	life	NOUN
bracis-33550	258	42	histories	history	NOUN
bracis-33550	258	43	of	of	ADP
bracis-33550	258	44	the	the	DET
bracis-33550	258	45	museum	museum	NOUN
bracis-33550	258	46	of	of	ADP
bracis-33550	258	47	the	the	DET
bracis-33550	258	48	person	person	NOUN
bracis-33550	258	49	(	(	PUNCT
bracis-33550	258	50	2023	2023	NUM
bracis-33550	258	51	)	)	PUNCT
bracis-33550	258	52	google	google	NOUN
bracis-33550	258	53	scholar	scholar	NOUN
bracis-33550	258	54	  	  	SPACE
bracis-33550	258	55	liu	liu	PROPN
bracis-33550	258	56	,	,	PUNCT
bracis-33550	258	57	y.	y.	PROPN
bracis-33550	258	58	,	,	PUNCT
bracis-33550	258	59	shriberg	shriberg	PROPN
bracis-33550	258	60	,	,	PUNCT
bracis-33550	258	61	e.	e.	PROPN
bracis-33550	258	62	,	,	PUNCT
bracis-33550	258	63	stolcke	stolcke	PROPN
bracis-33550	258	64	,	,	PUNCT
bracis-33550	258	65	a.	a.	NOUN
bracis-33550	258	66	,	,	PUNCT
bracis-33550	258	67	hillard	hillard	PROPN
bracis-33550	258	68	,	,	PUNCT
bracis-33550	258	69	d.	d.	PROPN
bracis-33550	258	70	,	,	PUNCT
bracis-33550	258	71	ostendorf	ostendorf	PROPN
bracis-33550	258	72	,	,	PUNCT
bracis-33550	258	73	m.	m.	NOUN
bracis-33550	258	74	,	,	PUNCT
bracis-33550	258	75	harper	harper	NOUN
bracis-33550	258	76	,	,	PUNCT
bracis-33550	258	77	m.	m.	NOUN
bracis-33550	258	78	:	:	PUNCT
bracis-33550	258	79	enriching	enrich	VERB
bracis-33550	258	80	speech	speech	NOUN
bracis-33550	258	81	recognition	recognition	NOUN
bracis-33550	258	82	with	with	ADP
bracis-33550	258	83	automatic	automatic	ADJ
bracis-33550	258	84	detection	detection	NOUN
bracis-33550	258	85	of	of	ADP
bracis-33550	258	86	sentence	sentence	NOUN
bracis-33550	258	87	boundaries	boundary	NOUN
bracis-33550	258	88	and	and	CCONJ
bracis-33550	258	89	disfluencies	disfluency	NOUN
bracis-33550	258	90	.	.	PUNCT
bracis-33550	259	1	ieee	ieee	PROPN
bracis-33550	259	2	trans	trans	PROPN
bracis-33550	259	3	.	.	PUNCT
bracis-33550	259	4	audio	audio	PROPN
bracis-33550	259	5	speech	speech	NOUN
bracis-33550	259	6	lang	lang	PROPN
bracis-33550	259	7	.	.	PUNCT
bracis-33550	259	8	process	process	NOUN
bracis-33550	259	9	.	.	PUNCT
bracis-33550	259	10	,	,	PUNCT
bracis-33550	259	11	1526–1540	1526–1540	NUM
bracis-33550	259	12	(	(	PUNCT
bracis-33550	259	13	2006	2006	NUM
bracis-33550	259	14	)	)	PUNCT
bracis-33550	259	15	google	google	PROPN
bracis-33550	259	16	scholar	scholar	NOUN
bracis-33550	259	17	  	  	SPACE
bracis-33550	259	18	panayotov	panayotov	PROPN
bracis-33550	259	19	,	,	PUNCT
bracis-33550	259	20	v.	v.	PROPN
bracis-33550	259	21	,	,	PUNCT
bracis-33550	259	22	chen	chen	PROPN
bracis-33550	259	23	,	,	PUNCT
bracis-33550	259	24	g.	g.	PROPN
bracis-33550	259	25	,	,	PUNCT
bracis-33550	259	26	povey	povey	NOUN
bracis-33550	259	27	,	,	PUNCT
bracis-33550	259	28	d.	d.	PROPN
bracis-33550	259	29	,	,	PUNCT
bracis-33550	259	30	khudanpur	khudanpur	PROPN
bracis-33550	259	31	,	,	PUNCT
bracis-33550	259	32	s.	s.	PROPN
bracis-33550	259	33	:	:	PUNCT
bracis-33550	259	34	librispeech	librispeech	PROPN
bracis-33550	259	35	:	:	PUNCT
bracis-33550	259	36	an	an	DET
bracis-33550	259	37	asr	asr	NOUN
bracis-33550	259	38	corpus	corpus	NOUN
bracis-33550	259	39	based	base	VERB
bracis-33550	259	40	on	on	ADP
bracis-33550	259	41	public	public	ADJ
bracis-33550	259	42	domain	domain	NOUN
bracis-33550	259	43	audio	audio	NOUN
bracis-33550	259	44	books	book	NOUN
bracis-33550	259	45	.	.	PUNCT
bracis-33550	260	1	in	in	ADP
bracis-33550	260	2	:	:	PUNCT
bracis-33550	260	3	2015	2015	NUM
bracis-33550	260	4	ieee	ieee	NOUN
bracis-33550	260	5	international	international	ADJ
bracis-33550	260	6	conference	conference	NOUN
bracis-33550	260	7	on	on	ADP
bracis-33550	260	8	acoustics	acoustic	NOUN
bracis-33550	260	9	,	,	PUNCT
bracis-33550	260	10	speech	speech	NOUN
bracis-33550	260	11	and	and	CCONJ
bracis-33550	260	12	signal	signal	NOUN
bracis-33550	260	13	processing	processing	NOUN
bracis-33550	260	14	(	(	PUNCT
bracis-33550	260	15	icassp	icassp	PROPN
bracis-33550	260	16	)	)	PUNCT
bracis-33550	260	17	,	,	PUNCT
bracis-33550	260	18	pp	pp	ADP
bracis-33550	260	19	.	.	PUNCT
bracis-33550	261	1	5206–5210	5206–5210	NUM
bracis-33550	261	2	(	(	PUNCT
bracis-33550	261	3	2015	2015	NUM
bracis-33550	261	4	)	)	PUNCT
bracis-33550	261	5	google	google	NOUN
bracis-33550	261	6	scholar	scholar	NOUN
bracis-33550	261	7	  	  	SPACE
bracis-33550	261	8	pratap	pratap	NOUN
bracis-33550	261	9	,	,	PUNCT
bracis-33550	261	10	v.	v.	PROPN
bracis-33550	261	11	,	,	PUNCT
bracis-33550	261	12	xu	xu	PROPN
bracis-33550	261	13	,	,	PUNCT
bracis-33550	261	14	q.	q.	PROPN
bracis-33550	261	15	,	,	PUNCT
bracis-33550	261	16	sriram	sriram	PROPN
bracis-33550	261	17	,	,	PUNCT
bracis-33550	261	18	a.	a.	NOUN
bracis-33550	261	19	,	,	PUNCT
bracis-33550	261	20	synnaeve	synnaeve	NOUN
bracis-33550	261	21	,	,	PUNCT
bracis-33550	261	22	g.	g.	PROPN
bracis-33550	261	23	,	,	PUNCT
bracis-33550	261	24	collobert	collobert	PROPN
bracis-33550	261	25	,	,	PUNCT
bracis-33550	261	26	r.	r.	PROPN
bracis-33550	261	27	:	:	PUNCT
bracis-33550	262	1	mls	mls	PROPN
bracis-33550	262	2	:	:	PUNCT
bracis-33550	262	3	a	a	DET
bracis-33550	262	4	large	large	ADJ
bracis-33550	262	5	-	-	PUNCT
bracis-33550	262	6	scale	scale	NOUN
bracis-33550	262	7	multilingual	multilingual	ADJ
bracis-33550	262	8	dataset	dataset	NOUN
bracis-33550	262	9	for	for	ADP
bracis-33550	262	10	speech	speech	NOUN
bracis-33550	262	11	research	research	NOUN
bracis-33550	262	12	.	.	PUNCT
bracis-33550	263	1	in	in	ADP
bracis-33550	263	2	:	:	PUNCT
bracis-33550	263	3	proceedings	proceeding	NOUN
bracis-33550	263	4	of	of	ADP
bracis-33550	263	5	the	the	DET
bracis-33550	263	6	interspeech	interspeech	NOUN
bracis-33550	263	7	2020	2020	NUM
bracis-33550	263	8	,	,	PUNCT
bracis-33550	263	9	pp	pp	ADV
bracis-33550	263	10	.	.	PUNCT
bracis-33550	264	1	2757–2761	2757–2761	NUM
bracis-33550	264	2	(	(	PUNCT
bracis-33550	264	3	2020	2020	NUM
bracis-33550	264	4	)	)	PUNCT
bracis-33550	264	5	.	.	PUNCT
bracis-33550	265	1	https://doi.org/10.21437/interspeech.2020-2826	https://doi.org/10.21437/interspeech.2020-2826	PROPN
bracis-33550	265	2	radford	radford	PROPN
bracis-33550	265	3	,	,	PUNCT
bracis-33550	265	4	a.	a.	PROPN
bracis-33550	265	5	,	,	PUNCT
bracis-33550	265	6	kim	kim	PROPN
bracis-33550	265	7	,	,	PUNCT
bracis-33550	265	8	j.w	j.w	PROPN
bracis-33550	265	9	.	.	PROPN
bracis-33550	265	10	,	,	PUNCT
bracis-33550	265	11	xu	xu	PROPN
bracis-33550	265	12	,	,	PUNCT
bracis-33550	265	13	t.	t.	PROPN
bracis-33550	265	14	,	,	PUNCT
bracis-33550	265	15	brockman	brockman	PROPN
bracis-33550	265	16	,	,	PUNCT
bracis-33550	265	17	g.	g.	PROPN
bracis-33550	265	18	,	,	PUNCT
bracis-33550	265	19	mcleavey	mcleavey	NOUN
bracis-33550	265	20	,	,	PUNCT
bracis-33550	265	21	c.	c.	NOUN
bracis-33550	265	22	,	,	PUNCT
bracis-33550	265	23	sutskever	sutskever	PROPN
bracis-33550	265	24	,	,	PUNCT
bracis-33550	265	25	i.	i.	NOUN
bracis-33550	265	26	:	:	PUNCT
bracis-33550	265	27	robust	robust	ADJ
bracis-33550	265	28	speech	speech	NOUN
bracis-33550	265	29	recognition	recognition	NOUN
bracis-33550	265	30	via	via	ADP
bracis-33550	265	31	large	large	ADJ
bracis-33550	265	32	-	-	PUNCT
bracis-33550	265	33	scale	scale	NOUN
bracis-33550	265	34	weak	weak	ADJ
bracis-33550	265	35	supervision	supervision	NOUN
bracis-33550	265	36	.	.	PUNCT
bracis-33550	266	1	in	in	ADP
bracis-33550	266	2	:	:	PUNCT
bracis-33550	266	3	krause	krause	PROPN
bracis-33550	266	4	,	,	PUNCT
bracis-33550	266	5	a.	a.	NOUN
bracis-33550	266	6	,	,	PUNCT
bracis-33550	266	7	brunskill	brunskill	NOUN
bracis-33550	266	8	,	,	PUNCT
bracis-33550	266	9	e.	e.	PROPN
bracis-33550	266	10	,	,	PUNCT
bracis-33550	266	11	cho	cho	PROPN
bracis-33550	266	12	,	,	PUNCT
bracis-33550	266	13	k.	k.	PROPN
bracis-33550	266	14	,	,	PUNCT
bracis-33550	266	15	engelhardt	engelhardt	PROPN
bracis-33550	266	16	,	,	PUNCT
bracis-33550	266	17	b.	b.	PROPN
bracis-33550	266	18	,	,	PUNCT
bracis-33550	266	19	sabato	sabato	PROPN
bracis-33550	266	20	,	,	PUNCT
bracis-33550	266	21	s.	s.	PROPN
bracis-33550	266	22	,	,	PUNCT
bracis-33550	266	23	scarlett	scarlett	PROPN
bracis-33550	266	24	,	,	PUNCT
bracis-33550	266	25	j.	j.	PROPN
bracis-33550	266	26	(	(	PUNCT
bracis-33550	266	27	eds	eds	PROPN
bracis-33550	266	28	.	.	PUNCT
bracis-33550	266	29	)	)	PUNCT
bracis-33550	266	30	proceedings	proceeding	NOUN
bracis-33550	266	31	of	of	ADP
bracis-33550	266	32	the	the	DET
bracis-33550	266	33	40th	40th	ADJ
bracis-33550	266	34	international	international	ADJ
bracis-33550	266	35	conference	conference	NOUN
bracis-33550	266	36	on	on	ADP
bracis-33550	266	37	machine	machine	NOUN
bracis-33550	266	38	learning	learning	NOUN
bracis-33550	266	39	,	,	PUNCT
bracis-33550	266	40	proceedings	proceeding	NOUN
bracis-33550	266	41	of	of	ADP
bracis-33550	266	42	machine	machine	NOUN
bracis-33550	266	43	learning	learn	VERB
bracis-33550	266	44	research	research	NOUN
bracis-33550	266	45	,	,	PUNCT
bracis-33550	266	46	23–29	23–29	NUM
bracis-33550	266	47	july	july	PROPN
bracis-33550	266	48	2023	2023	NUM
bracis-33550	266	49	,	,	PUNCT
bracis-33550	266	50	vol	vol	NOUN
bracis-33550	266	51	.	.	PUNCT
bracis-33550	266	52	 	 	SPACE
bracis-33550	267	1	202	202	NUM
bracis-33550	267	2	,	,	PUNCT
bracis-33550	267	3	pp	pp	ADJ
bracis-33550	267	4	.	.	PUNCT
bracis-33550	268	1	28492–28518	28492–28518	NUM
bracis-33550	268	2	.	.	PUNCT
bracis-33550	268	3	pmlr	pmlr	NOUN
bracis-33550	268	4	(	(	PUNCT
bracis-33550	268	5	2023	2023	NUM
bracis-33550	268	6	)	)	PUNCT
bracis-33550	268	7	google	google	NOUN
bracis-33550	268	8	scholar	scholar	NOUN
bracis-33550	268	9	  	  	SPACE
bracis-33550	268	10	rodrigues	rodrigue	NOUN
bracis-33550	268	11	,	,	PUNCT
bracis-33550	268	12	a.c	a.c	PROPN
bracis-33550	268	13	.	.	PROPN
bracis-33550	268	14	,	,	PUNCT
bracis-33550	268	15	et	et	PROPN
bracis-33550	268	16	al	al	PROPN
bracis-33550	268	17	:	:	PUNCT
bracis-33550	268	18	portal	portal	ADJ
bracis-33550	268	19	nurc	nurc	ADJ
bracis-33550	268	20	-	-	PUNCT
bracis-33550	268	21	sp	sp	NOUN
bracis-33550	268	22	:	:	PUNCT
bracis-33550	268	23	design	design	NOUN
bracis-33550	268	24	,	,	PUNCT
bracis-33550	268	25	development	development	NOUN
bracis-33550	268	26	,	,	PUNCT
bracis-33550	268	27	and	and	CCONJ
bracis-33550	268	28	speech	speech	NOUN
bracis-33550	268	29	processing	processing	NOUN
bracis-33550	268	30	corpora	corpora	NOUN
bracis-33550	268	31	resources	resource	NOUN
bracis-33550	268	32	to	to	PART
bracis-33550	268	33	support	support	VERB
bracis-33550	268	34	the	the	DET
bracis-33550	268	35	public	public	ADJ
bracis-33550	268	36	dissemination	dissemination	NOUN
bracis-33550	268	37	of	of	ADP
bracis-33550	268	38	portuguese	portuguese	ADJ
bracis-33550	268	39	spoken	spoken	ADJ
bracis-33550	268	40	language	language	NOUN
bracis-33550	268	41	.	.	PUNCT
bracis-33550	269	1	in	in	ADP
bracis-33550	269	2	:	:	PUNCT
bracis-33550	269	3	gamallo	gamallo	ADJ
bracis-33550	269	4	,	,	PUNCT
bracis-33550	269	5	p.	p.	PROPN
bracis-33550	269	6	,	,	PUNCT
bracis-33550	269	7	et	et	PROPN
bracis-33550	269	8	al	al	PROPN
bracis-33550	269	9	(	(	PUNCT
bracis-33550	269	10	eds	eds	PROPN
bracis-33550	269	11	.	.	PUNCT
bracis-33550	269	12	)	)	PUNCT
bracis-33550	270	1	proceedings	proceeding	NOUN
bracis-33550	270	2	of	of	ADP
bracis-33550	270	3	the	the	DET
bracis-33550	270	4	16th	16th	ADJ
bracis-33550	270	5	international	international	ADJ
bracis-33550	270	6	conference	conference	NOUN
bracis-33550	270	7	on	on	ADP
bracis-33550	270	8	computational	computational	ADJ
bracis-33550	270	9	processing	processing	NOUN
bracis-33550	270	10	of	of	ADP
bracis-33550	270	11	portuguese	portuguese	PROPN
bracis-33550	270	12	,	,	PUNCT
bracis-33550	270	13	pp	pp	ADJ
bracis-33550	270	14	.	.	PUNCT
bracis-33550	271	1	187–195	187–195	NUM
bracis-33550	271	2	.	.	PUNCT
bracis-33550	271	3	association	association	NOUN
bracis-33550	271	4	for	for	ADP
bracis-33550	271	5	computational	computational	ADJ
bracis-33550	271	6	lingustics	lingustic	NOUN
bracis-33550	271	7	(	(	PUNCT
bracis-33550	271	8	2024	2024	NUM
bracis-33550	271	9	)	)	PUNCT
bracis-33550	271	10	google	google	NOUN
bracis-33550	271	11	scholar	scholar	NOUN
bracis-33550	271	12	  	  	SPACE
bracis-33550	271	13	salesky	salesky	NOUN
bracis-33550	271	14	,	,	PUNCT
bracis-33550	271	15	e.	e.	PROPN
bracis-33550	271	16	,	,	PUNCT
bracis-33550	271	17	et	et	PROPN
bracis-33550	271	18	al	al	PROPN
bracis-33550	271	19	.	.	PROPN
bracis-33550	271	20	:	:	PUNCT
bracis-33550	272	1	the	the	DET
bracis-33550	272	2	multilingual	multilingual	ADJ
bracis-33550	272	3	tedx	tedx	ADJ
bracis-33550	272	4	corpus	corpus	NOUN
bracis-33550	272	5	for	for	ADP
bracis-33550	272	6	speech	speech	NOUN
bracis-33550	272	7	recognition	recognition	NOUN
bracis-33550	272	8	and	and	CCONJ
bracis-33550	272	9	translation	translation	NOUN
bracis-33550	272	10	.	.	PUNCT
bracis-33550	273	1	in	in	ADP
bracis-33550	273	2	:	:	PUNCT
bracis-33550	273	3	proceedings	proceeding	NOUN
bracis-33550	273	4	of	of	ADP
bracis-33550	273	5	the	the	DET
bracis-33550	273	6	interspeech	interspeech	NOUN
bracis-33550	273	7	2021	2021	NUM
bracis-33550	273	8	,	,	PUNCT
bracis-33550	273	9	pp	pp	ADV
bracis-33550	273	10	.	.	PUNCT
bracis-33550	273	11	3655–3659	3655–3659	NUM
bracis-33550	273	12	(	(	PUNCT
bracis-33550	273	13	2021	2021	NUM
bracis-33550	273	14	)	)	PUNCT
bracis-33550	273	15	google	google	NOUN
bracis-33550	273	16	scholar	scholar	NOUN
bracis-33550	273	17	  	  	SPACE
bracis-33550	273	18	éva	éva	NOUN
bracis-33550	273	19	székely	székely	ADV
bracis-33550	273	20	,	,	PUNCT
bracis-33550	273	21	henter	henter	PROPN
bracis-33550	273	22	,	,	PUNCT
bracis-33550	273	23	g.e	g.e	PROPN
bracis-33550	273	24	.	.	PROPN
bracis-33550	273	25	,	,	PUNCT
bracis-33550	273	26	beskow	beskow	PROPN
bracis-33550	273	27	,	,	PUNCT
bracis-33550	273	28	j.	j.	PROPN
bracis-33550	273	29	,	,	PUNCT
bracis-33550	273	30	gustafson	gustafson	PROPN
bracis-33550	273	31	,	,	PUNCT
bracis-33550	273	32	j.	j.	PROPN
bracis-33550	273	33	:	:	PUNCT
bracis-33550	273	34	spontaneous	spontaneous	ADJ
bracis-33550	273	35	conversational	conversational	ADJ
bracis-33550	273	36	speech	speech	NOUN
bracis-33550	273	37	synthesis	synthesis	NOUN
bracis-33550	273	38	from	from	ADP
bracis-33550	273	39	found	find	VERB
bracis-33550	273	40	data	datum	NOUN
bracis-33550	273	41	.	.	PUNCT
bracis-33550	274	1	in	in	ADP
bracis-33550	274	2	:	:	PUNCT
bracis-33550	274	3	proceedings	proceeding	NOUN
bracis-33550	274	4	of	of	ADP
bracis-33550	274	5	the	the	DET
bracis-33550	274	6	interspeech	interspeech	NOUN
bracis-33550	274	7	2019	2019	NUM
bracis-33550	274	8	,	,	PUNCT
bracis-33550	274	9	pp	pp	ADV
bracis-33550	274	10	.	.	PUNCT
bracis-33550	275	1	4435–4439	4435–4439	NUM
bracis-33550	275	2	(	(	PUNCT
bracis-33550	275	3	2019	2019	NUM
bracis-33550	275	4	)	)	PUNCT
bracis-33550	275	5	.	.	PUNCT
bracis-33550	276	1	https://doi.org/10.21437/interspeech.2019-2836	https://doi.org/10.21437/interspeech.2019-2836	NOUN
bracis-33550	276	2	vaswani	vaswani	NOUN
bracis-33550	276	3	,	,	PUNCT
bracis-33550	276	4	a.	a.	NOUN
bracis-33550	276	5	,	,	PUNCT
bracis-33550	276	6	et	et	PROPN
bracis-33550	276	7	al	al	PROPN
bracis-33550	276	8	.	.	PUNCT
bracis-33550	276	9	:	:	PUNCT
bracis-33550	277	1	attention	attention	NOUN
bracis-33550	277	2	is	be	AUX
bracis-33550	277	3	all	all	PRON
bracis-33550	277	4	you	you	PRON
bracis-33550	277	5	need	need	VERB
bracis-33550	277	6	.	.	PUNCT
bracis-33550	278	1	in	in	ADP
bracis-33550	278	2	:	:	PUNCT
bracis-33550	278	3	advances	advance	NOUN
bracis-33550	278	4	in	in	ADP
bracis-33550	278	5	neural	neural	ADJ
bracis-33550	278	6	information	information	NOUN
bracis-33550	278	7	p0rocessing	p0rocesse	VERB
bracis-33550	278	8	systems	system	NOUN
bracis-33550	278	9	,	,	PUNCT
bracis-33550	278	10	vol	vol	NOUN
bracis-33550	278	11	.	.	PROPN
bracis-33550	278	12	30	30	NUM
bracis-33550	278	13	(	(	PUNCT
bracis-33550	278	14	2017	2017	NUM
bracis-33550	278	15	)	)	PUNCT
bracis-33550	278	16	google	google	PROPN
bracis-33550	278	17	scholar	scholar	NOUN
bracis-33550	278	18	  	  	SPACE
bracis-33550	278	19	zen	zen	PROPN
bracis-33550	278	20	,	,	PUNCT
bracis-33550	278	21	h.	h.	PROPN
bracis-33550	278	22	,	,	PUNCT
bracis-33550	278	23	et	et	PROPN
bracis-33550	278	24	al	al	PROPN
bracis-33550	278	25	.	.	PROPN
bracis-33550	278	26	:	:	PUNCT
bracis-33550	279	1	libritts	libritts	PROPN
bracis-33550	279	2	:	:	PUNCT
bracis-33550	279	3	a	a	DET
bracis-33550	279	4	corpus	corpus	NOUN
bracis-33550	279	5	derived	derive	VERB
bracis-33550	279	6	from	from	ADP
bracis-33550	279	7	librispeech	librispeech	NOUN
bracis-33550	279	8	for	for	ADP
bracis-33550	279	9	text	text	NOUN
bracis-33550	279	10	-	-	PUNCT
bracis-33550	279	11	to	to	ADP
bracis-33550	279	12	-	-	PUNCT
bracis-33550	279	13	speech	speech	NOUN
bracis-33550	279	14	(	(	PUNCT
bracis-33550	279	15	2019	2019	NUM
bracis-33550	279	16	)	)	PUNCT
bracis-33550	279	17	google	google	NOUN
bracis-33550	279	18	scholar	scholar	NOUN
bracis-33550	279	19	  	  	SPACE
bracis-33550	279	20	download	download	NOUN
bracis-33550	279	21	references	reference	NOUN
bracis-33550	279	22	acknowledgements	acknowledgement	NOUN
bracis-33550	279	23	this	this	DET
bracis-33550	279	24	work	work	NOUN
bracis-33550	279	25	was	be	AUX
bracis-33550	279	26	carried	carry	VERB
bracis-33550	279	27	out	out	ADP
bracis-33550	279	28	at	at	ADP
bracis-33550	279	29	the	the	DET
bracis-33550	279	30	center	center	NOUN
bracis-33550	279	31	for	for	ADP
bracis-33550	279	32	artificial	artificial	ADJ
bracis-33550	279	33	intelligence	intelligence	NOUN
bracis-33550	279	34	(	(	PUNCT
bracis-33550	279	35	c4ai	c4ai	NOUN
bracis-33550	279	36	-	-	PUNCT
bracis-33550	279	37	usp	usp	NOUN
bracis-33550	279	38	)	)	PUNCT
bracis-33550	279	39	,	,	PUNCT
bracis-33550	279	40	with	with	ADP
bracis-33550	279	41	support	support	NOUN
bracis-33550	279	42	by	by	ADP
bracis-33550	279	43	the	the	DET
bracis-33550	279	44	são	são	PROPN
bracis-33550	279	45	paulo	paulo	PROPN
bracis-33550	279	46	research	research	PROPN
bracis-33550	279	47	foundation	foundation	PROPN
bracis-33550	279	48	(	(	PUNCT
bracis-33550	279	49	fapesp	fapesp	NOUN
bracis-33550	279	50	grant	grant	VERB
bracis-33550	279	51	#	#	SYM
bracis-33550	279	52	2019/07665	2019/07665	NUM
bracis-33550	279	53	-	-	SYM
bracis-33550	279	54	4	4	NUM
bracis-33550	279	55	)	)	PUNCT
bracis-33550	279	56	and	and	CCONJ
bracis-33550	279	57	by	by	ADP
bracis-33550	279	58	the	the	DET
bracis-33550	279	59	ibm	ibm	PROPN
bracis-33550	279	60	corporation	corporation	NOUN
bracis-33550	279	61	.	.	PUNCT
bracis-33550	280	1	this	this	DET
bracis-33550	280	2	project	project	NOUN
bracis-33550	280	3	was	be	AUX
bracis-33550	280	4	also	also	ADV
bracis-33550	280	5	supported	support	VERB
bracis-33550	280	6	by	by	ADP
bracis-33550	280	7	the	the	DET
bracis-33550	280	8	ministry	ministry	PROPN
bracis-33550	280	9	of	of	ADP
bracis-33550	280	10	science	science	PROPN
bracis-33550	280	11	,	,	PUNCT
bracis-33550	280	12	technology	technology	NOUN
bracis-33550	280	13	and	and	CCONJ
bracis-33550	280	14	innovation	innovation	NOUN
bracis-33550	280	15	,	,	PUNCT
bracis-33550	280	16	with	with	ADP
bracis-33550	280	17	resources	resource	NOUN
bracis-33550	280	18	of	of	ADP
bracis-33550	280	19	law	law	NOUN
bracis-33550	280	20	no	no	NOUN
bracis-33550	280	21	.	.	PROPN
bracis-33550	281	1	8.248	8.248	NUM
bracis-33550	281	2	,	,	PUNCT
bracis-33550	281	3	of	of	ADP
bracis-33550	281	4	october	october	PROPN
bracis-33550	281	5	23	23	NUM
bracis-33550	281	6	,	,	PUNCT
bracis-33550	281	7	1991	1991	NUM
bracis-33550	281	8	,	,	PUNCT
bracis-33550	281	9	within	within	ADP
bracis-33550	281	10	the	the	DET
bracis-33550	281	11	scope	scope	NOUN
bracis-33550	281	12	of	of	ADP
bracis-33550	281	13	ppi	ppi	NOUN
bracis-33550	281	14	-	-	PUNCT
bracis-33550	281	15	softex	softex	NOUN
bracis-33550	281	16	,	,	PUNCT
bracis-33550	281	17	coordinated	coordinate	VERB
bracis-33550	281	18	by	by	ADP
bracis-33550	281	19	softex	softex	NOUN
bracis-33550	281	20	and	and	CCONJ
bracis-33550	281	21	published	publish	VERB
bracis-33550	281	22	residence	residence	NOUN
bracis-33550	281	23	in	in	ADP
bracis-33550	281	24	tic	tic	ADJ
bracis-33550	281	25	13	13	NUM
bracis-33550	281	26	,	,	PUNCT
bracis-33550	281	27	dou	dou	ADJ
bracis-33550	281	28	01245.010222/2022	01245.010222/2022	NOUN
bracis-33550	281	29	-	-	PUNCT
bracis-33550	281	30	44	44	NUM
bracis-33550	281	31	.	.	PUNCT
bracis-33550	282	1	author	author	NOUN
bracis-33550	282	2	information	information	NOUN
bracis-33550	282	3	authors	author	NOUN
bracis-33550	282	4	and	and	CCONJ
bracis-33550	282	5	affiliations	affiliation	NOUN
bracis-33550	282	6	university	university	PROPN
bracis-33550	282	7	of	of	ADP
bracis-33550	282	8	são	são	PROPN
bracis-33550	282	9	paulo	paulo	PROPN
bracis-33550	282	10	,	,	PUNCT
bracis-33550	282	11	são	são	PROPN
bracis-33550	282	12	carlos	carlos	PROPN
bracis-33550	282	13	,	,	PUNCT
bracis-33550	282	14	sp	sp	PROPN
bracis-33550	282	15	,	,	PUNCT
bracis-33550	282	16	13566	13566	NUM
bracis-33550	282	17	-	-	SYM
bracis-33550	282	18	590	590	NUM
bracis-33550	282	19	,	,	PUNCT
bracis-33550	282	20	brazil	brazil	PROPN
bracis-33550	282	21	rodrigo	rodrigo	PROPN
bracis-33550	282	22	lima	lima	PROPN
bracis-33550	282	23	,	,	PUNCT
bracis-33550	282	24	 	 	SPACE
bracis-33550	282	25	sidney	sidney	PROPN
bracis-33550	282	26	e.	e.	PROPN
bracis-33550	282	27	leal	leal	PROPN
bracis-33550	282	28	 	 	SPACE
bracis-33550	282	29	&	&	CCONJ
bracis-33550	282	30	 	 	SPACE
bracis-33550	282	31	sandra	sandra	PROPN
bracis-33550	282	32	m.	m.	PROPN
bracis-33550	282	33	aluísio	aluísio	PROPN
bracis-33550	282	34	universidade	universidade	PROPN
bracis-33550	282	35	estadual	estadual	PROPN
bracis-33550	282	36	paulista	paulista	PROPN
bracis-33550	282	37	,	,	PUNCT
bracis-33550	282	38	são	são	PROPN
bracis-33550	282	39	josé	josé	PROPN
bracis-33550	282	40	do	do	PROPN
bracis-33550	282	41	rio	rio	PROPN
bracis-33550	282	42	preto	preto	PROPN
bracis-33550	282	43	,	,	PUNCT
bracis-33550	282	44	sp	sp	NOUN
bracis-33550	282	45	,	,	PUNCT
bracis-33550	282	46	15054	15054	NUM
bracis-33550	282	47	-	-	SYM
bracis-33550	282	48	000	000	NUM
bracis-33550	282	49	,	,	PUNCT
bracis-33550	282	50	brazil	brazil	PROPN
bracis-33550	282	51	arnaldo	arnaldo	PROPN
bracis-33550	282	52	candido	candido	PROPN
bracis-33550	282	53	junior	junior	PROPN
bracis-33550	282	54	authors	authors	PROPN
bracis-33550	282	55	rodrigo	rodrigo	PROPN
bracis-33550	282	56	limaview	limaview	PROPN
bracis-33550	282	57	author	author	NOUN
bracis-33550	282	58	publications	publication	NOUN
bracis-33550	282	59	search	search	NOUN
bracis-33550	282	60	author	author	NOUN
bracis-33550	282	61	on	on	ADP
bracis-33550	282	62	:	:	PUNCT
bracis-33550	282	63	pubmed	pubmed	PROPN
bracis-33550	282	64	 	 	SPACE
bracis-33550	282	65	google	google	PROPN
bracis-33550	282	66	scholar	scholar	PROPN
bracis-33550	282	67	sidney	sidney	PROPN
bracis-33550	282	68	e.	e.	PROPN
bracis-33550	282	69	lealview	lealview	PROPN
bracis-33550	282	70	author	author	NOUN
bracis-33550	282	71	publications	publication	NOUN
bracis-33550	282	72	search	search	NOUN
bracis-33550	282	73	author	author	NOUN
bracis-33550	282	74	on	on	ADP
bracis-33550	282	75	:	:	PUNCT
bracis-33550	282	76	pubmed	pubmed	PROPN
bracis-33550	282	77	 	 	SPACE
bracis-33550	282	78	google	google	PROPN
bracis-33550	282	79	scholar	scholar	PROPN
bracis-33550	282	80	arnaldo	arnaldo	PROPN
bracis-33550	282	81	candido	candido	PROPN
bracis-33550	282	82	juniorview	juniorview	PROPN
bracis-33550	282	83	author	author	NOUN
bracis-33550	282	84	publications	publication	VERB
bracis-33550	282	85	search	search	NOUN
bracis-33550	282	86	author	author	NOUN
bracis-33550	282	87	on	on	ADP
bracis-33550	282	88	:	:	PUNCT
bracis-33550	282	89	pubmed	pubmed	PROPN
bracis-33550	282	90	 	 	SPACE
bracis-33550	282	91	google	google	PROPN
bracis-33550	282	92	scholar	scholar	PROPN
bracis-33550	282	93	sandra	sandra	PROPN
bracis-33550	282	94	m.	m.	PROPN
bracis-33550	282	95	aluísioview	aluísioview	PROPN
bracis-33550	282	96	author	author	NOUN
bracis-33550	282	97	publications	publication	VERB
bracis-33550	282	98	search	search	NOUN
bracis-33550	282	99	author	author	NOUN
bracis-33550	282	100	on	on	ADP
bracis-33550	282	101	:	:	PUNCT
bracis-33550	282	102	pubmed	pubmed	PROPN
bracis-33550	282	103	 	 	SPACE
bracis-33550	282	104	google	google	PROPN
bracis-33550	282	105	scholar	scholar	NOUN
bracis-33550	282	106	corresponding	correspond	VERB
bracis-33550	282	107	author	author	NOUN
bracis-33550	282	108	correspondence	correspondence	NOUN
bracis-33550	282	109	to	to	ADP
bracis-33550	282	110	sidney	sidney	PROPN
bracis-33550	282	111	e.	e.	PROPN
bracis-33550	282	112	leal	leal	PROPN
bracis-33550	282	113	.	.	PUNCT
bracis-33550	283	1	editor	editor	NOUN
bracis-33550	283	2	information	information	NOUN
bracis-33550	283	3	editors	editor	NOUN
bracis-33550	283	4	and	and	CCONJ
bracis-33550	283	5	affiliations	affiliation	NOUN
bracis-33550	283	6	universidade	universidade	VERB
bracis-33550	283	7	federal	federal	PROPN
bracis-33550	283	8	fluminense	fluminense	PROPN
bracis-33550	283	9	,	,	PUNCT
bracis-33550	283	10	niterói	niterói	PROPN
bracis-33550	283	11	,	,	PUNCT
bracis-33550	283	12	brazil	brazil	PROPN
bracis-33550	283	13	aline	aline	PROPN
bracis-33550	283	14	paes	paes	PROPN
bracis-33550	283	15	instituto	instituto	PROPN
bracis-33550	283	16	tecnológico	tecnológico	PROPN
bracis-33550	283	17	de	de	PROPN
bracis-33550	283	18	aeronáutica	aeronáutica	PROPN
bracis-33550	283	19	,	,	PUNCT
bracis-33550	283	20	são	são	PROPN
bracis-33550	283	21	josé	josé	PROPN
bracis-33550	283	22	dos	dos	PROPN
bracis-33550	283	23	campos	campos	PROPN
bracis-33550	283	24	,	,	PUNCT
bracis-33550	283	25	brazil	brazil	PROPN
bracis-33550	283	26	filipe	filipe	PROPN
bracis-33550	283	27	a.	a.	PROPN
bracis-33550	283	28	n.	n.	PROPN
bracis-33550	283	29	verri	verri	PROPN
bracis-33550	283	30	rights	right	NOUN
bracis-33550	283	31	and	and	CCONJ
bracis-33550	283	32	permissions	permission	NOUN
bracis-33550	283	33	reprints	reprint	NOUN
bracis-33550	283	34	and	and	CCONJ
bracis-33550	283	35	permissions	permission	VERB
bracis-33550	283	36	copyright	copyright	NOUN
bracis-33550	283	37	information	information	NOUN
bracis-33550	283	38	©	©	ADP
bracis-33550	283	39	2025	2025	NUM
bracis-33550	283	40	the	the	DET
bracis-33550	283	41	author(s	author(s	NOUN
bracis-33550	283	42	)	)	PUNCT
bracis-33550	283	43	,	,	PUNCT
bracis-33550	283	44	under	under	ADP
bracis-33550	283	45	exclusive	exclusive	ADJ
bracis-33550	283	46	license	license	NOUN
bracis-33550	283	47	to	to	ADP
bracis-33550	283	48	springer	springer	NOUN
bracis-33550	283	49	nature	nature	PROPN
bracis-33550	283	50	switzerland	switzerland	PROPN
bracis-33550	283	51	ag	ag	PROPN
bracis-33550	283	52	about	about	ADP
bracis-33550	283	53	this	this	DET
bracis-33550	283	54	paper	paper	NOUN
bracis-33550	283	55	cite	cite	VERB
bracis-33550	283	56	this	this	DET
bracis-33550	283	57	paper	paper	NOUN
bracis-33550	283	58	lima	lima	PROPN
bracis-33550	283	59	,	,	PUNCT
bracis-33550	283	60	r.	r.	PROPN
bracis-33550	283	61	,	,	PUNCT
bracis-33550	283	62	leal	leal	PROPN
bracis-33550	283	63	,	,	PUNCT
bracis-33550	283	64	s.e	s.e	PROPN
bracis-33550	283	65	.	.	PROPN
bracis-33550	283	66	,	,	PUNCT
bracis-33550	283	67	junior	junior	PROPN
bracis-33550	283	68	,	,	PUNCT
bracis-33550	283	69	a.c	a.c	PROPN
bracis-33550	283	70	.	.	PROPN
bracis-33550	283	71	,	,	PUNCT
bracis-33550	283	72	aluísio	aluísio	PROPN
bracis-33550	283	73	,	,	PUNCT
bracis-33550	283	74	s.m	s.m	PROPN
bracis-33550	283	75	.	.	PROPN
bracis-33550	284	1	(	(	PUNCT
bracis-33550	284	2	2025	2025	NUM
bracis-33550	284	3	)	)	PUNCT
bracis-33550	284	4	.	.	PUNCT
bracis-33550	285	1	a	a	DET
bracis-33550	285	2	large	large	ADJ
bracis-33550	285	3	dataset	dataset	NOUN
bracis-33550	285	4	of	of	ADP
bracis-33550	285	5	 	 	SPACE
bracis-33550	285	6	spontaneous	spontaneous	ADJ
bracis-33550	285	7	speech	speech	NOUN
bracis-33550	285	8	with	with	ADP
bracis-33550	285	9	 	 	SPACE
bracis-33550	285	10	the	the	DET
bracis-33550	285	11	 	 	SPACE
bracis-33550	285	12	accent	accent	NOUN
bracis-33550	285	13	spoken	speak	VERB
bracis-33550	285	14	in	in	ADP
bracis-33550	285	15	 	 	SPACE
bracis-33550	285	16	são	são	PROPN
bracis-33550	285	17	paulo	paulo	PROPN
bracis-33550	285	18	for	for	ADP
bracis-33550	285	19	 	 	SPACE
bracis-33550	285	20	automatic	automatic	ADJ
bracis-33550	285	21	speech	speech	NOUN
bracis-33550	285	22	recognition	recognition	NOUN
bracis-33550	285	23	evaluation	evaluation	NOUN
bracis-33550	285	24	.	.	PUNCT
bracis-33550	286	1	in	in	ADP
bracis-33550	286	2	:	:	PUNCT
bracis-33550	286	3	paes	paes	PROPN
bracis-33550	286	4	,	,	PUNCT
bracis-33550	286	5	a.	a.	PROPN
bracis-33550	286	6	,	,	PUNCT
bracis-33550	286	7	verri	verri	NOUN
bracis-33550	286	8	,	,	PUNCT
bracis-33550	286	9	f.a.n	f.a.n	ADJ
bracis-33550	286	10	.	.	PUNCT
bracis-33550	286	11	(	(	PUNCT
bracis-33550	286	12	eds	ed	NOUN
bracis-33550	286	13	)	)	PUNCT
bracis-33550	286	14	intelligent	intelligent	ADJ
bracis-33550	286	15	systems	system	NOUN
bracis-33550	286	16	.	.	PUNCT
bracis-33550	286	17	bracis	bracis	PROPN
bracis-33550	286	18	2024	2024	NUM
bracis-33550	286	19	.	.	PUNCT
bracis-33550	287	1	lecture	lecture	NOUN
bracis-33550	287	2	notes	note	NOUN
bracis-33550	287	3	in	in	ADP
bracis-33550	287	4	computer	computer	NOUN
bracis-33550	287	5	science	science	NOUN
bracis-33550	287	6	(	(	PUNCT
bracis-33550	287	7	)	)	PUNCT
bracis-33550	287	8	,	,	PUNCT
bracis-33550	287	9	vol	vol	NOUN
bracis-33550	287	10	15412	15412	NUM
bracis-33550	287	11	.	.	PUNCT
bracis-33550	288	1	springer	springer	NOUN
bracis-33550	288	2	,	,	PUNCT
bracis-33550	288	3	cham	cham	PROPN
bracis-33550	288	4	.	.	PUNCT
bracis-33550	289	1	https://doi.org/10.1007/978-3-031-79029-4_3	https://doi.org/10.1007/978-3-031-79029-4_3	PROPN
bracis-33550	289	2	download	download	NOUN
bracis-33550	289	3	citation	citation	NOUN
bracis-33550	289	4	.ris	.ris	PUNCT
bracis-33550	289	5	.enw	.enw	PROPN
bracis-33550	289	6	.bib	.bib	PUNCT
bracis-33550	290	1	doi	doi	PROPN
bracis-33550	290	2	:	:	PUNCT
bracis-33550	290	3	https://doi.org/10.1007/978-3-031-79029-4_3	https://doi.org/10.1007/978-3-031-79029-4_3	NOUN
bracis-33550	290	4	published	publish	VERB
bracis-33550	290	5	:	:	PUNCT
bracis-33550	290	6	30	30	NUM
bracis-33550	290	7	january	january	PROPN
bracis-33550	290	8	2025	2025	NUM
bracis-33550	290	9	publisher	publisher	NOUN
bracis-33550	290	10	name	name	NOUN
bracis-33550	290	11	:	:	PUNCT
bracis-33550	290	12	springer	springer	NOUN
bracis-33550	290	13	,	,	PUNCT
bracis-33550	290	14	cham	cham	PROPN
bracis-33550	290	15	print	print	PROPN
bracis-33550	290	16	isbn	isbn	PROPN
bracis-33550	290	17	:	:	PUNCT
bracis-33550	290	18	978	978	NUM
bracis-33550	290	19	-	-	SYM
bracis-33550	290	20	3	3	NUM
bracis-33550	290	21	-	-	PUNCT
bracis-33550	290	22	031	031	NUM
bracis-33550	290	23	-	-	PUNCT
bracis-33550	290	24	79028	79028	NUM
bracis-33550	290	25	-	-	SYM
bracis-33550	290	26	7	7	NUM
bracis-33550	290	27	online	online	ADJ
bracis-33550	290	28	isbn	isbn	NOUN
bracis-33550	290	29	:	:	PUNCT
bracis-33550	290	30	978	978	NUM
bracis-33550	290	31	-	-	SYM
bracis-33550	290	32	3	3	NUM
bracis-33550	290	33	-	-	PUNCT
bracis-33550	290	34	031	031	NUM
bracis-33550	290	35	-	-	PUNCT
bracis-33550	290	36	79029	79029	NUM
bracis-33550	290	37	-	-	SYM
bracis-33550	290	38	4	4	NUM
bracis-33550	290	39	ebook	ebook	NOUN
bracis-33550	290	40	packages	package	NOUN
bracis-33550	290	41	:	:	PUNCT
bracis-33550	290	42	computer	computer	NOUN
bracis-33550	290	43	sciencecomputer	sciencecomputer	NOUN
bracis-33550	290	44	science	science	NOUN
bracis-33550	290	45	(	(	PUNCT
bracis-33550	290	46	r0	r0	NOUN
bracis-33550	290	47	)	)	PUNCT
bracis-33550	290	48	share	share	VERB
bracis-33550	290	49	this	this	DET
bracis-33550	290	50	paper	paper	NOUN
bracis-33550	290	51	anyone	anyone	PRON
bracis-33550	290	52	you	you	PRON
bracis-33550	290	53	share	share	VERB
bracis-33550	290	54	the	the	DET
bracis-33550	290	55	following	follow	VERB
bracis-33550	290	56	link	link	NOUN
bracis-33550	290	57	with	with	ADP
bracis-33550	290	58	will	will	AUX
bracis-33550	290	59	be	be	AUX
bracis-33550	290	60	able	able	ADJ
bracis-33550	290	61	to	to	PART
bracis-33550	290	62	read	read	VERB
bracis-33550	290	63	this	this	DET
bracis-33550	290	64	content	content	NOUN
bracis-33550	290	65	:	:	PUNCT
bracis-33550	290	66	get	get	VERB
bracis-33550	290	67	shareable	shareable	ADJ
bracis-33550	290	68	linksorry	linksorry	NOUN
bracis-33550	290	69	,	,	PUNCT
bracis-33550	290	70	a	a	DET
bracis-33550	290	71	shareable	shareable	ADJ
bracis-33550	290	72	link	link	NOUN
bracis-33550	290	73	is	be	AUX
bracis-33550	290	74	not	not	PART
bracis-33550	290	75	currently	currently	ADV
bracis-33550	290	76	available	available	ADJ
bracis-33550	290	77	for	for	ADP
bracis-33550	290	78	this	this	DET
bracis-33550	290	79	article	article	NOUN
bracis-33550	290	80	.	.	PUNCT
bracis-33550	291	1	copy	copy	VERB
bracis-33550	291	2	shareable	shareable	ADJ
bracis-33550	291	3	link	link	NOUN
bracis-33550	291	4	to	to	PART
bracis-33550	291	5	clipboard	clipboard	NOUN
bracis-33550	291	6	provided	provide	VERB
bracis-33550	291	7	by	by	ADP
bracis-33550	291	8	the	the	DET
bracis-33550	291	9	springer	springer	NOUN
bracis-33550	291	10	nature	nature	PROPN
bracis-33550	291	11	sharedit	sharedit	PROPN
bracis-33550	291	12	content	content	NOUN
bracis-33550	291	13	-	-	PUNCT
bracis-33550	291	14	sharing	share	VERB
bracis-33550	291	15	initiative	initiative	NOUN
bracis-33550	291	16	keywords	keyword	VERB
bracis-33550	291	17	automatic	automatic	ADJ
bracis-33550	291	18	speech	speech	NOUN
bracis-33550	291	19	recognition	recognition	NOUN
bracis-33550	291	20	evaluation	evaluation	NOUN
bracis-33550	291	21	spontaneous	spontaneous	ADJ
bracis-33550	291	22	speech	speech	NOUN
bracis-33550	291	23	brazilian	brazilian	ADJ
bracis-33550	291	24	portuguese	portuguese	ADJ
bracis-33550	291	25	public	public	ADJ
bracis-33550	291	26	speech	speech	NOUN
bracis-33550	291	27	corpora	corpora	NOUN
bracis-33550	291	28	publish	publish	NOUN
bracis-33550	291	29	with	with	ADP
bracis-33550	291	30	us	us	PROPN
bracis-33550	291	31	policies	policy	NOUN
bracis-33550	291	32	and	and	CCONJ
bracis-33550	291	33	ethics	ethic	NOUN
bracis-33550	291	34	search	search	NOUN
bracis-33550	291	35	search	search	NOUN
bracis-33550	291	36	by	by	ADP
bracis-33550	291	37	keyword	keyword	NOUN
bracis-33550	291	38	or	or	CCONJ
bracis-33550	291	39	author	author	NOUN
bracis-33550	291	40	search	search	NOUN
bracis-33550	291	41	navigation	navigation	NOUN
bracis-33550	291	42	find	find	VERB
bracis-33550	291	43	a	a	DET
bracis-33550	291	44	journal	journal	NOUN
bracis-33550	291	45	publish	publish	VERB
bracis-33550	291	46	with	with	ADP
bracis-33550	291	47	us	we	PRON
bracis-33550	291	48	track	track	VERB
bracis-33550	291	49	your	your	PRON
bracis-33550	291	50	research	research	NOUN
bracis-33550	291	51	discover	discover	VERB
bracis-33550	291	52	content	content	NOUN
bracis-33550	291	53	journals	journal	NOUN
bracis-33550	291	54	a	a	DET
bracis-33550	291	55	-	-	PUNCT
bracis-33550	291	56	z	z	NOUN
bracis-33550	291	57	books	book	NOUN
bracis-33550	291	58	a	a	DET
bracis-33550	291	59	-	-	PUNCT
bracis-33550	291	60	z	z	NOUN
bracis-33550	291	61	publish	publish	NOUN
bracis-33550	291	62	with	with	ADP
bracis-33550	291	63	us	us	PROPN
bracis-33550	291	64	journal	journal	PROPN
bracis-33550	291	65	finder	finder	PROPN
bracis-33550	291	66	publish	publish	VERB
bracis-33550	291	67	your	your	PRON
bracis-33550	291	68	research	research	NOUN
bracis-33550	291	69	language	language	NOUN
bracis-33550	291	70	editing	edit	VERB
bracis-33550	291	71	open	open	ADJ
bracis-33550	291	72	access	access	NOUN
bracis-33550	291	73	publishing	publishing	NOUN
bracis-33550	291	74	products	product	NOUN
bracis-33550	291	75	and	and	CCONJ
bracis-33550	291	76	services	service	NOUN
bracis-33550	291	77	our	our	PRON
bracis-33550	291	78	products	product	NOUN
bracis-33550	291	79	librarians	librarian	VERB
bracis-33550	291	80	societies	society	NOUN
bracis-33550	291	81	partners	partner	NOUN
bracis-33550	291	82	and	and	CCONJ
bracis-33550	291	83	advertisers	advertiser	NOUN
bracis-33550	291	84	our	our	PRON
bracis-33550	291	85	brands	brand	NOUN
bracis-33550	291	86	springer	springer	NOUN
bracis-33550	291	87	nature	nature	PROPN
bracis-33550	291	88	portfolio	portfolio	PROPN
bracis-33550	291	89	bmc	bmc	PROPN
bracis-33550	291	90	palgrave	palgrave	PROPN
bracis-33550	291	91	macmillan	macmillan	PROPN
bracis-33550	291	92	apress	apress	PROPN
bracis-33550	291	93	discover	discover	VERB
bracis-33550	291	94	your	your	PRON
bracis-33550	291	95	privacy	privacy	NOUN
bracis-33550	291	96	choices	choice	NOUN
bracis-33550	291	97	/	/	SYM
bracis-33550	291	98	manage	manage	NOUN
bracis-33550	291	99	cookies	cookie	NOUN
bracis-33550	291	100	your	your	PRON
bracis-33550	291	101	us	us	PROPN
bracis-33550	292	1	state	state	NOUN
bracis-33550	292	2	privacy	privacy	NOUN
bracis-33550	292	3	rights	right	NOUN
bracis-33550	292	4	accessibility	accessibility	NOUN
bracis-33550	292	5	statement	statement	NOUN
bracis-33550	292	6	terms	term	NOUN
bracis-33550	292	7	and	and	CCONJ
bracis-33550	292	8	conditions	condition	NOUN
bracis-33550	292	9	privacy	privacy	NOUN
bracis-33550	292	10	policy	policy	NOUN
bracis-33550	292	11	help	help	NOUN
bracis-33550	292	12	and	and	CCONJ
bracis-33550	292	13	support	support	VERB
bracis-33550	292	14	legal	legal	ADJ
bracis-33550	292	15	notice	notice	NOUN
bracis-33550	292	16	cancel	cancel	VERB
bracis-33550	292	17	contracts	contract	NOUN
bracis-33550	292	18	here	here	ADV
bracis-33550	292	19	129.74.145.123	129.74.145.123	NUM
bracis-33550	292	20	hesburgh	hesburgh	PROPN
bracis-33550	292	21	library	library	PROPN
bracis-33550	292	22	er	er	INTJ
bracis-33550	292	23	unit	unit	NOUN
bracis-33550	292	24	(	(	PUNCT
bracis-33550	292	25	3005732405	3005732405	NUM
bracis-33550	292	26	)	)	PUNCT
bracis-33550	292	27	northeast	northeast	ADJ
bracis-33550	292	28	research	research	NOUN
bracis-33550	292	29	libraries	library	NOUN
bracis-33550	292	30	(	(	PUNCT
bracis-33550	292	31	nerl	nerl	PROPN
bracis-33550	292	32	)	)	PUNCT
bracis-33550	292	33	(	(	PUNCT
bracis-33550	292	34	8200828607	8200828607	NUM
bracis-33550	292	35	)	)	PUNCT
bracis-33550	292	36	nerl	nerl	VERB
bracis-33550	292	37	ta	ta	X
bracis-33550	292	38	account	account	NOUN
bracis-33550	292	39	(	(	PUNCT
bracis-33550	292	40	3006206169	3006206169	NUM
bracis-33550	292	41	)	)	PUNCT
bracis-33550	292	42	university	university	NOUN
bracis-33550	292	43	of	of	ADP
bracis-33550	292	44	notre	notre	PROPN
bracis-33550	292	45	dame	dame	PROPN
bracis-33550	292	46	hesburgh	hesburgh	PROPN
bracis-33550	292	47	library	library	NOUN
bracis-33550	292	48	(	(	PUNCT
bracis-33550	292	49	3000184373	3000184373	NUM
bracis-33550	292	50	)	)	PUNCT
bracis-33550	293	1	©	©	ADP
bracis-33550	293	2	2025	2025	NUM
bracis-33550	293	3	springer	springer	NOUN
bracis-33550	293	4	nature	nature	NOUN
