-
Notifications
You must be signed in to change notification settings - Fork 0
/
asr.html
443 lines (391 loc) · 20 KB
/
asr.html
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
<!DOCTYPE html>
<html>
<head lang="en">
<meta charset="UTF-8">
<meta http-equiv="x-ua-compatible" content="ie=edge">
<title>Bengali.AI Massively Crowdsourced OOD-Speech Dataset</title>
<meta name="description" content="">
<meta name="viewport" content="width=device-width, initial-scale=1">
<meta property="og:image" content="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/tsne_all_ppx40.png">
<meta property="og:image:type" content="image/png">
<meta property="og:image:width" content="1296">
<meta property="og:image:height" content="840">
<meta property="og:type" content="website" />
<meta property="og:url" content="https://jonbarron.info/zipnerf/"/>
<meta property="og:title" content="Bengali.AI Speech Recognition Project" />
<meta property="og:description" content="We present OOD-Speech, the first out-of-distribution
(OOD) benchmarking dataset for Bengali automatic
speech recognition (ASR). Being one of the most spo-
ken languages globally, Bengali portrays large diversity
in dialects and prosodic features, which demands ASR
frameworks to be robust towards distribution shifts. For
example, islamic religious sermons in Bengali are deliv-
ered with a tonality that is significantly different from
regular speech. Our training dataset is collected via mas-
sively online crowdsourcing campaigns which resulted in
1177.94 hours collected and curated from 22, 645 native
Bengali speakers from South Asia. Our test dataset com-
prises 23.03 hours of speech collected and manually anno-
tated from 17 different sources, e.g., Bengali TV drama,
Audiobook, Talk show, Online class, and Islamic sermons
to name a few. OOD-Speech is jointly the largest pub-
licly available speech dataset, as well as the first out-of-
distribution ASR benchmarking dataset for Bengali." />
<meta name="twitter:card" content="summary_large_image" />
<meta name="twitter:title" content="Bengali.AI Speech Recognition Project" />
<meta name="twitter:description" content="We present OOD-Speech, the first out-of-distribution
(OOD) benchmarking dataset for Bengali automatic
speech recognition (ASR). Being one of the most spo-
ken languages globally, Bengali portrays large diversity
in dialects and prosodic features, which demands ASR
frameworks to be robust towards distribution shifts. For
example, islamic religious sermons in Bengali are deliv-
ered with a tonality that is significantly different from
regular speech. Our training dataset is collected via mas-
sively online crowdsourcing campaigns which resulted in
1177.94 hours collected and curated from 22, 645 native
Bengali speakers from South Asia. Our test dataset com-
prises 23.03 hours of speech collected and manually anno-
tated from 17 different sources, e.g., Bengali TV drama,
Audiobook, Talk show, Online class, and Islamic sermons
to name a few. OOD-Speech is jointly the largest pub-
licly available speech dataset, as well as the first out-of-
distribution ASR benchmarking dataset for Bengali." />
<meta name="twitter:image" content="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/tsne_all_ppx40.png" />
<link rel="icon" href="data:image/svg+xml,<svg xmlns=%22http://www.w3.org/2000/svg%22 viewBox=%220 0 100 100%22><text y=%22.9em%22 font-size=%2290%22>⚡</text></svg>">
<link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.5/css/bootstrap.min.css">
<link rel="stylesheet" href="https://maxcdn.bootstrapcdn.com/font-awesome/4.4.0/css/font-awesome.min.css">
<link rel="stylesheet" href="https://cdnjs.cloudflare.com/ajax/libs/codemirror/5.8.0/codemirror.min.css">
<link rel="stylesheet" href="css/app.css">
<link rel="stylesheet" href="css/bootstrap.min.css">
<script src="https://ajax.googleapis.com/ajax/libs/jquery/1.11.3/jquery.min.js"></script>
<script src="https://maxcdn.bootstrapcdn.com/bootstrap/3.3.5/js/bootstrap.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/codemirror/5.8.0/codemirror.min.js"></script>
<script src="https://cdnjs.cloudflare.com/ajax/libs/clipboard.js/1.5.3/clipboard.min.js"></script>
<script src="js/app.js"></script>
<script src="js/video_comparison.js"></script>
</head>
<body>
<div class="container" id="main">
<div class="row">
<h2 class="col-md-12 text-center">
<b>OOD-Speech</b>: A Large Bengali Speech Recognition Dataset for
Out-of-Distribution Benchmarking</br>
<small>
INTERSPEECH 2023
</small>
</h2>
</div>
<div class="row">
<div class="col-md-12 text-center">
<ul class="list-inline">
<li>
<a href="https://www.linkedin.com/in/souhardya-saha/">
Souhardya Saha
</a>
</li>
<li>
<a href="https://www.linkedin.com/in/fazle-rakib/">
Fazle Rakib
</a>
</li>
<li>
<a href="https://scholar.google.com/citations?user=JmTaGcYAAAAJ&hl=en">
Shihab Istiak
</a>
</li>
<li>
<a href="https://scholar.google.com/citations?user=2Un1c7QAAAAJ&hl=en">
Samiul Alam
</a>
</li>
<li>
<a href="https://mnansary.github.io/">
Nazmuddoha Ansary
</a>
</li>
<li>
<a href="https://www.linkedin.com/in/mobassir-hossen-916a83137/">
Syed Mobasshir
</a>
</li>
<li>
<a href="https://www.sust.edu/institutes/iict/faculty/10">
Sayma Sultana
</a>
</li>
<li>
<a href="https://www.linkedin.com/in/marsia-haque-meghla-321021207/">
Marsia Meghla
</a>
</li>
<li>
<a href="https://people.bengali.ai/farig/">
Farig Sadeq
</a>
</li>
<li>
<a href="https://appledora.github.io/#/">
Nazia Tasnim
</a>
</li>
<li>
<a href="https://www.linkedin.com/in/mamunur-rahaman-mamun/?originalSubdomain=bd">
Mamun Rahman
</a>
</li>
<li>
<a href="https://scholar.google.com/citations?user=W_okWy0AAAAJ&hl=en">
Tahsin Reasat
</a>
</li>
<li>
<a href="https://people.bengali.ai/sushmit/">
Asif Sushmit
</a>
</li>
<li>
<a href="https://imtiazhumayun.github.io/">
Ahmed Imtiaz Humayun
</a>
</li>
</br>Bengali.AI
</ul>
</div>
</div>
<div class="row">
<div class="text-center">
<ul class="nav nav-pills nav-justified">
<li>
<a href="https://bengaliai.github.io/">
<image src="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/logoBengaliai.jpg" height="60px">
<h4><strong>Bengali.AI Home</strong></h4>
</a>
</li>
<li>
<a href="https://arxiv.org/abs/2305.09688">
<image src="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/interpaper.png" height="60px">
<h4><strong>Interspeech Paper & Supplementary Materials</strong></h4>
</a>
</li>
<li>
<a href="https://arxiv.org/pdf/2206.14053.pdf">
<image src="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/commpaper.png" height="60px">
<h4><strong>Commonvoice + Dlsprint Paper</strong></h4>
</a>
</li>
<li>
<a href="https://shobdo.bengali.ai/prompts">
<image src="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/record.png" height="60px">
<h4><strong>Contribute Data</strong></h4>
</a>
</li>
<li>
<a href="https://commonvoice.mozilla.org/bn/datasets">
<image src="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/download.png" height="60px">
<h4><strong>Download Dataset</strong></h4>
</a>
</li>
<li>
<a href="https://www.kaggle.com/c/dlsprint">
<image src="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/kaggle.png" height="60px">
<h4><strong>DLSprint: Kaggle Community Competition</strong></h4>
</a>
</li>
<li>
<a href="https://www.kaggle.com/competitions/bengaliai-speech">
<image src="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/kaggle.png" height="60px">
<h4><strong>Kaggle Featured Competition Ongoing!!!</strong></h4>
</a>
</li>
</ul>
</div>
</div>
<!-- <div class="row">
<div class="col-md-8 col-md-offset-2">
<video id="v0" width="100%" autoplay loop muted controls>
<source src="img/teaser.mp4" type="video/mp4" />
</video>
</div>
</div> -->
<div class="row">
<div class="col-md-8 col-md-offset-2">
<div class="image">
<img src="https://raw.githubusercontent.com/BengaliAI/bengaliai.github.io/main/images/tsne_all_ppx40.png" alt="Your Image", width = 400, height=320>
</div>
<div class="text">
<p>Domain shift in different domains of speech data.</p>
</div>
</div>
</div>
<div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Abstract
</h3>
<p class="text-justify">
We present OOD-Speech, the first out-of-distribution
(OOD) benchmarking dataset for Bengali automatic
speech recognition (ASR). Being one of the most spo-
ken languages globally, Bengali portrays large diversity
in dialects and prosodic features, which demands ASR
frameworks to be robust towards distribution shifts. For
example, islamic religious sermons in Bengali are deliv-
ered with a tonality that is significantly different from
regular speech. Our training dataset is collected via mas-
sively online crowdsourcing campaigns which resulted in
1177.94 hours collected and curated from 22, 645 native
Bengali speakers from South Asia. Our test dataset com-
prises 23.03 hours of speech collected and manually anno-
tated from 17 different sources, e.g., Bengali TV drama,
Audiobook, Talk show, Online class, and Islamic sermons
to name a few. OOD-Speech is jointly the largest pub-
licly available speech dataset, as well as the first out-of-
distribution ASR benchmarking dataset for Bengali.
</p>
</div>
</div>
<div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Project Details
</h3>
<p class="text-justify">
This is the largest publicly available speech recognition corpus for Bengali. The data crowdsourcing started at 21st Feb 2022 and still ongoing.
The data contributors come from all walks life and the dataset has representative samples from different demographic strata and domains (gender, regional accent, age, background noise etc).
The test data comprises of challanging samples from different types of scripted and spontaneous speech recordings.
A Kaggle community research competition based on an earlier version of this dataset won the Kaggle Community Competition Award for the month of July, 2022.
Currently a dedicated team is extending the dataset to include representative samples for the different major regional dialects of Bengali. More big news coming!! Stay tuned.
</p>
</div>
</div>
<div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
(Still Ongoing) Speech Recog. Project Contributors
</h3>
<p class="text-justify">
Data Collection and Curation: Souhardya Saha, Fazle Rakib, Marsia Meghla, Imtiaz Prio, Rezuwan Hassan, Mayank, Siha Haque, Rezwana Sultana, Tahmid Hossain</p>
<p class="text-justify">
Web Platform: Mamunur Rahman, Manash Kumar Mandal, Shakirul Islam </p>
<p class="text-justify">
Modeling and Analysis: Tahsin Reasat, Nazia Tasnim, Istiak Shihab, Nazmuddoha Ansary, Syed Mobasshir Hossen, Shahrin Nakkhatra, Zaowad Rahabin, Arijit Mukherjee
<p class="text-justify">
Linguists: Tanveer Azmal, Sazia Mehnaz, Kanij Fatema</p>
</div>
</div>
<!-- <div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Video
</h3>
<div class="text-center">
<div style="position:relative;padding-top:56.25%;">
<iframe src="https://youtube.com/embed/xrrhynRzC8k" allowfullscreen style="position:absolute;top:0;left:0;width:100%;height:100%;"></iframe>
</div>
</div>
</div>
</div> -->
<br>
<!-- <div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
360° Video Flythroughs
</h3>
<div class="text-center">
<div style="position:relative;padding-top:56.25%;">
<iframe src="https://www.youtube.com/embed/videoseries?list=PLzPoYEE6Aw7Jzjek1uEIPnpcTDL3u8tb4" title="YouTube video player" frameborder="0" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" allowfullscreen style="position:absolute;top:0;left:0;width:100%;height:100%;"></iframe>
<!-- <iframe src="https://youtube.com/embed/jbE2ri8xEZo" allowfullscreen style="position:absolute;top:0;left:0;width:100%;height:100%;"></iframe> -->
</div>
</div>
</div>
</div> -->
<br>
<!-- <div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
XY aliasing
</h3>
<div class="video-compare-container" id="xyaliasDiv">
<video class="video" id="xyalias" loop playsinline autoPlay muted src="img/xy_alias_swipe_crf27.mp4" onplay="resizeAndPlay(this)"></video>
<canvas height=0 class="videoMerge" id="xyaliasMerge"></canvas>
</div>
<p class="text-justify">
A naive baseline (left) combining mip-NeRF 360 and Instant NGP results in aliasing as the camera moves laterally. Our full method (right) produces prefiltered renderings that do not flicker or shimmer.
</p>
</div>
</div>
-->
<!-- <div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Z aliasing
</h3>
<video id="v0" width="100%" autoplay loop muted controls>
<source src="img/z_alias_pdf_labeled.m4v" type="video/mp4" />
</video>
<p class="text-justify">
The proposal network used for resampling points along rays in mip-NeRF 360 results in an artifact we refer to as <em>z-aliasing</em>, where foreground content alternately appears and disappears as the camera moves toward or away from scene content. Z-aliasing occurs when the initial set of samples from the proposal network is not dense enough and misses thin structures, such as the chair above. Missed content can not be recovered by later rounds of sampling, since no future samples will be placed at that location along the ray. Our improvements to proposal network supervision result in a prefiltered proposal output that preserves the foreground object for all frames in this sequence. The plots above depict samples along a ray for three rounds of resampling (blue, orange, and green lines), with the y axis showing rendering weight (how much each interval contributes to the final rendered color), as a normalized probability density.
</p>
</div>
</div> -->
<div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Citation 1
</h3>
<div class="form-group col-md-10 col-md-offset-1">
<textarea id="bibtex" class="form-control" readonly>
@misc{2305.09688,
Author = {Fazle Rabbi Rakib and Souhardya Saha Dip and Samiul Alam and Nazia Tasnim and Md. Istiak Hossain Shihab and Md. Nazmuddoha Ansary and Syed Mobassir Hossen and Marsia Haque Meghla and Mamunur Mamun and Farig Sadeque and Sayma Sultana Chowdhury and Tahsin Reasat and Asif Sushmit and Ahmed Imtiaz Humayun},
Title = {OOD-Speech: A Large Bengali Speech Recognition Dataset for Out-of-Distribution Benchmarking},
Year = {2023},
Eprint = {arXiv:2305.09688},
}</textarea>
</div>
</div>
</div>
<div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Citation 2
</h3>
<div class="form-group col-md-10 col-md-offset-1">
<textarea id="bibtex" class="form-control" readonly>
@article{alam2022bengali,
title={Bengali common voice speech dataset for automatic speech recognition},
author={Alam, Samiul and Sushmit, Asif and Abdullah, Zaowad and Nakkhatra, Shahrin and Ansary, MD and Hossen, Syed Mobassir and Mehnaz, Sazia Morshed and Reasat, Tahsin and Humayun, Ahmed Imtiaz},
journal={arXiv preprint arXiv:2206.14053},
year={2022}
}</textarea>
</div>
</div>
</div>
<div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Acknowledgements
</h3>
<p class="text-justify">
We thank Arijit Mukherjee, Mayank and Mozilla Foundation. Also thanks to BUET, SUST, DIU, UIU, DU RME, A2I Bangladesh, AUST, NUST, Green U., BracU CSE for their massive support. And the real heroes are the 25,000 people who have contributed for the development of this dataset.
<br><br>
sdf
</p>
</div>
</div>
</div>
<div class="row">
<div class="col-md-8 col-md-offset-2">
<h3>
Contact
</h3>
<p class="text-justify">
research.bengaliai@gmail.com, sushmit@ieee.org, ahmed.imtiaz.prio@gmail.com
<br><br>
.
</p>
</div>
</div>
</div>
</body>
</html>