“ASR:2015-06-29”版本间的差异
来自cslt Wiki
（→Speech Processing） |
（→Text Processing） |
||
| 第47行： | 第47行： | ||
:* check the lstm-rnnlm code about how to Initialize and update learning rate.(hold) | :* check the lstm-rnnlm code about how to Initialize and update learning rate.(hold) | ||
| − | ==== | + | ====Neural Based Document Classification==== |
| − | * | + | * (hold) |
| − | + | ||
| − | + | ||
| − | + | ||
===Order representation === | ===Order representation === | ||
| + | * Nested Dropout | ||
* modify the objective function(hold) | * modify the objective function(hold) | ||
| − | + | ===Balance Representation=== | |
| − | * | + | * Find error signal |
| − | === | + | ===Recommendation=== |
| − | * | + | * Reproduce baseline. |
| − | + | ||
| − | + | ||
| − | === | + | ===DSSM based QA=== |
| − | * | + | * Reproduce baseline. |
| − | + | ||
| − | + | ||
| − | + | ||
2015年7月2日 (四) 12:44的最后版本
Speech Processing
AM development
Environment
RNN AM
- morpheme RNN --zhiyuan
Mic-Array
- hold
- compute EER with kaldi
====Data selection unsupervised learning
- train using aurora4 --zhiyong
- train using wsj --xuewei
RNN-DAE(Deep based Auto-Encode-RNN)
- hold
- deliver to mengyuan
Speaker ID
- DNN-based sid --Lantian
Ivector&Dvector based ASR
- hold --Tian Lan
- Cluster the speakers to speaker-classes, then using the distance or the posterior-probability as the metric
- dark-konowlege using i-vector
- train on wsj(testbase dev93+evl92)
- --hold
Dark knowledge
- test random last output layer when train MPE --zhiyuan
language vector
- hold
Text Processing
RNN LM
- character-lm rnn(hold)
- lstm+rnn
- check the lstm-rnnlm code about how to Initialize and update learning rate.(hold)
Neural Based Document Classification
- (hold)
Order representation
- Nested Dropout
- modify the objective function(hold)
Balance Representation
- Find error signal
Recommendation
- Reproduce baseline.
DSSM based QA
- Reproduce baseline.
