“ASR:2015-04-08”版本间的差异
来自cslt Wiki
（以“目录 [隐藏] 1 Speech Processing 1.1 AM development 1.1.1 Environment 1.1.2 RNN AM 1.1.3 Mic-Array 1.1.4 Convolutive network 1.1.5 RNN-DAE(Deep based Auto-Encode...”为内容创建页面） |
（→Text Processing） |
||
| （1位用户的2个中间修订版本未显示） | |||
| 第1行： | 第1行： | ||
| − | + | ==Speech Processing == | |
| − | + | === AM development === | |
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ||
| − | + | ==== Environment ==== | |
| − | + | * grid-11 often shut down automatically, too slow computation speed. | |
| − | + | ||
| − | + | ||
| − | |||
| − | |||
| − | |||
| − | + | ==== RNN AM==== | |
| − | + | * details at http://liuc.cslt.org/pages/rnnam.html | |
| − | + | * tuning parameters on monophone NN | |
| − | RNN | + | * run using wsj,MPE |
| − | + | ||
| − | http://cslt. | + | |
| − | Speaker ID | + | |
| − | DNN-based sid --Yiye | + | ==== Mic-Array ==== |
| − | Decode --Yiye | + | * investigate alpha parameter in time domian and frquency domain |
| − | http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=327 | + | * ALPHA>=0 |
| − | Ivector based ASR | + | |
| − | http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?step=view_request&cvssid=340 | + | |
| − | Ivector dimention is smaller, performance is better | + | ====Convolutive network==== |
| − | Augument to hidden layer is better than input layer | + | * HOLD |
| − | train on wsj(testbase dev93+evl92) | + | :* CNN + DNN feature fusion |
| − | Text Processing | + | |
| − | tag LM | + | ====RNN-DAE(Deep based Auto-Encode-RNN)==== |
| − | similar word extension in FST | + | * HOLD -Zhiyong |
| − | check the formula using Bayes and experiment | + | * http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=261 |
| − | RNN LM | + | |
| − | rnn | + | |
| − | code the character-lm using Theano | + | ===Speaker ID=== |
| − | lstm+rnn | + | :* DNN-based sid --Yiye |
| − | check the lstm-rnnlm code about how to Initialize and update learning rate.(hold) | + | :* Decode --Yiye |
| − | W2V based doc classification | + | :* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=327 |
| − | + | ||
| − | + | ===Ivector based ASR=== | |
| − | Translation | + | :* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?step=view_request&cvssid=340 |
| − | v5.0 demo released | + | :* Ivector dimention is smaller, performance is better |
| − | cut the dict and use new segment-tool | + | :* Augument to hidden layer is better than input layer |
| − | Sparse NN in NLP | + | :* train on wsj(testbase dev93+evl92) |
| − | prepare the ACL | + | |
| − | + | ==Text Processing== | |
| − | + | ===tag LM=== | |
| − | + | * similar word extension in FST | |
| − | online learning | + | :* check the formula using Bayes and experiment |
| − | data is ready.prepare the ACL paper | + | :* add more test data |
| − | + | :* test the baseline(no weight) and different weight method | |
| − | + | ||
| + | ====RNN LM==== | ||
| + | *rnn | ||
| + | :* code the character-lm using Theano | ||
| + | *lstm+rnn | ||
| + | :* check the lstm-rnnlm code about how to Initialize and update learning rate.(hold) | ||
| + | |||
| + | ====W2V based doc classification==== | ||
| + | * reproducible test using English data | ||
| + | * Code new version spherical word vector. | ||
| + | * Accomplish movMF model | ||
| + | |||
| + | ===Translation=== | ||
| + | * v5.0 demo released | ||
| + | :* cut the dict and use new segment-tool | ||
| + | |||
| + | ===Sparse NN in NLP=== | ||
| + | * prepare the ACL | ||
| + | :* test result is ok now[http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=lr&step=view_request&cvssid=344]. | ||
| + | :* find the new direction. | ||
| + | |||
| + | ===online learning=== | ||
| + | * data is ready.prepare the ACL paper | ||
| + | :* finish some test. | ||
| + | :* test the result on different time. | ||
| + | |||
| + | ===relation classifier=== | ||
| + | * check code and find the problem that result is different on sigmoid and tanh | ||
2015年4月8日 (三) 10:50的最后版本
Speech Processing
AM development
Environment
- grid-11 often shut down automatically, too slow computation speed.
RNN AM
- details at http://liuc.cslt.org/pages/rnnam.html
- tuning parameters on monophone NN
- run using wsj,MPE
Mic-Array
- investigate alpha parameter in time domian and frquency domain
- ALPHA>=0
Convolutive network
- HOLD
- CNN + DNN feature fusion
RNN-DAE(Deep based Auto-Encode-RNN)
- HOLD -Zhiyong
- http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=261
Speaker ID
- DNN-based sid --Yiye
- Decode --Yiye
- http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=327
Ivector based ASR
- http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?step=view_request&cvssid=340
- Ivector dimention is smaller, performance is better
- Augument to hidden layer is better than input layer
- train on wsj(testbase dev93+evl92)
Text Processing
tag LM
- similar word extension in FST
- check the formula using Bayes and experiment
- add more test data
- test the baseline(no weight) and different weight method
RNN LM
- rnn
- code the character-lm using Theano
- lstm+rnn
- check the lstm-rnnlm code about how to Initialize and update learning rate.(hold)
W2V based doc classification
- reproducible test using English data
- Code new version spherical word vector.
- Accomplish movMF model
Translation
- v5.0 demo released
- cut the dict and use new segment-tool
Sparse NN in NLP
- prepare the ACL
- test result is ok now[1].
- find the new direction.
online learning
- data is ready.prepare the ACL paper
- finish some test.
- test the result on different time.
relation classifier
- check code and find the problem that result is different on sigmoid and tanh
