“ASR:2015-04-20”版本间的差异

来自cslt Wiki
跳转至： 导航、 搜索
（→‎Sparse NN in NLP）
（→‎Dark knowledge）
 
（2位用户的3个中间修订版本未显示）
第4行： 第4行：
 
==== Environment ====
 
==== Environment ====
 
* grid-11 often shut down automatically, too slow computation speed.
 
* grid-11 often shut down automatically, too slow computation speed.
−
* add a server(760)
+
* New grid-13 added, using gpu970
 +
* To update the wiki enviroment infomation
  
 
==== RNN AM====
 
==== RNN AM====
 
* details at http://liuc.cslt.org/pages/rnnam.html
 
* details at http://liuc.cslt.org/pages/rnnam.html
−
* tuning parameters on monophone NN
+
* Test monophone on RNN using dark-knowledge
 
* run using wsj,MPE   
 
* run using wsj,MPE   
−
 
  
 
==== Mic-Array ====
 
==== Mic-Array ====
 +
* Change the prediction from  fbank to spectrum features
 
* investigate alpha parameter in time domian and frquency domain  
 
* investigate alpha parameter in time domian and frquency domain  
 
* ALPHA>=0, using data generated by reverber toolkit
 
* ALPHA>=0, using data generated by reverber toolkit
 
* consider theta
 
* consider theta
−
 
−
 
−
====Convolutive network====
 
−
* HOLD
 
−
* CNN + DNN feature fusion
 
  
 
====RNN-DAE(Deep based Auto-Encode-RNN)====
 
====RNN-DAE(Deep based Auto-Encode-RNN)====
−
* HOLD -Zhiyong
+
* HOLD --Zhiyong Zhang
 
* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=261
 
* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=261
−
 
  
 
===Speaker ID===   
 
===Speaker ID===   
−
:* DNN-based sid --Yiye
+
:* DNN-based sid --Yiye Lin
 
:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=327
 
:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=327
  
−
===Ivector based ASR===
+
===Ivector&Dvector based ASR===
−
*hold
+
:* Cluster the speakers to speaker-classes, then using the distance or the posterior-probability as the metric
 +
:* Direct using the dark-knowledge strategy to do the ivector training.
 
:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?step=view_request&cvssid=340
 
:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?step=view_request&cvssid=340
 
:* Ivector dimention is smaller, performance is better
 
:* Ivector dimention is smaller, performance is better
第39行： 第35行：
  
 
===Dark knowledge===
 
===Dark knowledge===
−
:*http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zxw&step=view_request&cvssid=264 --zhiyong
+
:* Ensemble
−
:* trial on logit matching faild --mengyuan
+
::*http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zxw&step=view_request&cvssid=264 --Zhiyong Zhang
−
:* adaptation for chinglish under investigation-mengyuan
+
:* adaptation for chinglish under investigation --Mengyuan Zhao
−
:* unsupervised training with wsj contributes to aurora4 model--xiangyu
+
::* Try to improve the chinglish performance extremly
−
:* test large database with amida--xiangyu
+
:* unsupervised training with wsj contributes to aurora4 model --Xiangyu Zeng
 +
::* test large database with AMIDA
  
 
===bilingual recognition===
 
===bilingual recognition===
−
:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zxw&step=view_request&cvssid=359--zhiyuan
+
:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zxw&step=view_request&cvssid=359 --Zhiyuan Tang
  
 
==Text Processing==
 
==Text Processing==
第72行： 第69行：
 
* test the order feature ,need some result:
 
* test the order feature ,need some result:
 
* large dimension result:http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=lr&step=view_request&cvssid=344
 
* large dimension result:http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=lr&step=view_request&cvssid=344
−
:* sparse-nn on 1000 dimension(le-6,0.71) is better than 200 dimension(le-12,0.69).
+
:* sparse-nn on 1000 dimension(le-6,0.705236) is better than 200 dimension(le-12,0.694678).
  
 
===online learning===
 
===online learning===

2015年4月22日 (三) 08:49的最后版本

Speech Processing

AM development

Environment

  • grid-11 often shut down automatically, too slow computation speed.
  • New grid-13 added, using gpu970
  • To update the wiki enviroment infomation

RNN AM

Mic-Array

  • Change the prediction from fbank to spectrum features
  • investigate alpha parameter in time domian and frquency domain
  • ALPHA>=0, using data generated by reverber toolkit
  • consider theta

RNN-DAE(Deep based Auto-Encode-RNN)

Speaker ID

Ivector&Dvector based ASR

Dark knowledge

  • Ensemble
  • adaptation for chinglish under investigation --Mengyuan Zhao
  • Try to improve the chinglish performance extremly
  • unsupervised training with wsj contributes to aurora4 model --Xiangyu Zeng
  • test large database with AMIDA

bilingual recognition

Text Processing

tag LM

  • similar word extension in FST
  • will check the formula using Bayes and experiment
  • add similarity weight

RNN LM

  • rnn
  • test the ppl and code the character-lm
  • lstm+rnn
  • check the lstm-rnnlm code about how to Initialize and update learning rate.(hold)

W2V based document classification

  • result about norm model [1]
  • try CNN model

Translation

  • v5.0 demo released
  • cut the dict and use new segment-tool

Sparse NN in NLP

  • sparse-nn on 1000 dimension(le-6,0.705236) is better than 200 dimension(le-12,0.694678).

online learning

  • modified the listNet SGD

relation classifier

  • check the CNN code and contact the author of paper
取自“http://cslt.org/mediawiki/index.php?title=ASR:2015-04-20&oldid=14826”