0.25
0.5
0.75
1.25
1.5
1.75
2
Speech Recognition and Deep Learning
Published on Sep 13, 20159334 Views
Related categories
Chapter list
Speech Recognition and Deep Learning00:00
Speech recognition - 101:09
Speech recognition - 202:40
Speech recognition - 303:23
Outline06:58
Traditional speech models08:34
Basic pipeline - 108:38
Basic pipeline - 209:28
Basic pipeline - 310:32
Basic pipeline - 412:36
Basic pipeline - 515:47
Basic pipeline - 616:41
Features18:02
Example: Spectrogram - 118:21
Example: Spectrogram - 219:31
Back to modeling24:54
Acoustic model24:57
Modeling 1 phoneme - 25:52
Modeling 1 phoneme - 227:40
Modeling 1 phoneme - 327:43
Modeling 1 phoneme - 427:53
Modeling 1 phoneme - 527:55
Modeling 1 phoneme - 627:57
Modeling 1 phoneme - 728:00
Modeling 1 phoneme - 828:01
Inference with 1 phoneme - 128:20
Inference with 1 phoneme - 230:10
Modeling a word31:43
Training from sentences - 132:50
Training from sentences - 234:26
Obstacles...36:11
Language modeling36:52
Putting it together39:54
Decoder - 140:34
Decoder - 240:53
Decoder - 341:18
Decoder visualization - 142:57
Decoder - 444:34
Decoder - 546:09
Decoder visualization - 247:59
Decoder visualization - 348:25
Decoder visualization - 448:33
Decoder visualization - 549:21
Decoder visualization - 649:32
Decoder visualization - 750:03
Decoder visualization - 850:10
Decoder visualization - 950:14
Decoder visualization - 1050:17
Decoder visualization - 1150:19
Decoder visualization - 1250:20
Decoder visualization - 1350:21
Decoder visualization - 1450:24
Decoder visualization - 1550:36
Decoder visualization - 1651:46
Decoder visualization - 1751:52
Decoder visualization - 1852:20
Decoder visualization - 1952:28
Is that all?53:30
Deep Learning56:04
Where can DL help?56:14
Basic pipeline56:52
DNN acoustic models - 158:25
DNN acoustic models - 259:49
DNN acoustic models - 301:00:36
DNN acoustic models - 401:02:46
Early wins for DNN models01:05:12
More powerful acoustic models01:06:18
Rescoring01:08:51
Rescoring with Neural LM01:09:45
Training from unsegmented data with CTC01:11:30
Complexity01:11:32
Network setup01:12:46
Problem01:15:25
Collapsing operator01:16:11
Likelihood of sequence01:17:38
Training - 101:18:50
Training - 201:18:57
Training - 301:19:23
Decoding - 101:21:17
Decoding - 201:21:26
Decoding - 301:21:56
End-to-end learning - 101:22:44
End-to-end learning - 101:23:39
Example transcriptions01:24:38
Conclusion01:25:55
Thank you01:27:01