ASR:2015-06-15 - 版本历史

Zxw：/* Speech Processing */

2015-06-25T02:46:28Z

‎Speech Processing

Zxw：/* Speech Processing */

2015-06-17T06:33:09Z

‎Speech Processing

Zxw：/* RNN AM */

2015-06-17T01:02:54Z

‎RNN AM

Zxw：以“==Speech Processing == === AM development === ==== Environment ==== * ==== RNN AM==== morpheme RNN-zhiyuan ==== Mic-Array ==== * hold * Change the prediction f...”为内容创建页面

2015-06-17T01:02:13Z

以“==Speech Processing == === AM development === ==== Environment ==== * ==== RNN AM==== *morpheme RNN-zhiyuan * ==== Mic-Array ==== * hold * Change the prediction f...”为内容创建页面

新页面

==Speech Processing ==
=== AM development ===

==== Environment ====
*
==== RNN AM====
*morpheme RNN-zhiyuan

* ==== Mic-Array ====
* hold
* 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
* compute EER with kaldi

====RNN-DAE(Deep based Auto-Encode-RNN)====
* hold
* deliver to mengyuan
:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=261

===Speaker ID===
* DNN-based sid --Tian Lan
:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=327

===Ivector&Dvector based ASR===
* hold --Tian Lan
* 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
* Ivector dimention is smaller, performance is better
* Augument to hidden layer is better than input layer
* train on wsj(testbase dev93+evl92)

===Dark knowledge===
* Ensemble using 100h dataset to construct diffrernt structures -- Mengyuan
:*http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zxw&step=view_request&cvssid=264 --Zhiyong Zhang
* adaptation English and Chinglish
:* Try to improve the chinglish performance extremly
* unsupervised training with wsj contributes to aurora4 model --Xiangyu Zeng
* test large database with AMIDA
* test hidden layer knowledge transfer--xuewei
* test random last output layer when train MPE--zhiyuan

===bilingual recognition===
* hold
:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zxw&step=view_request&cvssid=359 --Zhiyuan Tang and Mengyuan

===language vector===
* train DNN with language vector--xuewei

==Text Processing==
====RNN LM====
*character-lm rnn(hold)
*lstm+rnn
:* check the lstm-rnnlm code about how to Initialize and update learning rate.(hold)

====W2V based document classification====
* APSIPA paper
* CNN adapt to resolve the low resource problem
===Pair-wise LM===
* draft paper of journal

===Order representation ===
* modify the objective function(hold)
* sup-sampling method to solve the low frequence word(hold)
* journal paper

===binary vector===
* nips paper
===Stochastic ListNet===
*done

===relation classifier===
*done

===plan to do===
* combine LDA with neural network

@@ 第3行： / 第3行： @@
 ==== Environment ====
+*grid-14 does not work --mengyuan
 ==== RNN AM====
-*morpheme RNN-zhiyuan
+*morpheme RNN --zhiyuan
 *RNN MPE --zhiyuan and xuewei
 ==== Mic-Array ====
 * hold
 * compute EER with kaldi
@@ 第24行： / 第22行： @@
 *  DNN-based sid --Lantian
 :* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=327
 ===Ivector&Dvector based ASR===
@@ 第35行： / 第34行： @@
 * test random last output layer when train MPE--zhiyuan
 ===language vector===
 * hold --xuewei
+* train using chinese and chiglish
-===DNN MPE===
-* random the last layer, train MPE--zhiyuan
 ==Text Processing==

@@ 第6行： / 第6行： @@
 ==== RNN AM====
 *morpheme RNN-zhiyuan
 ==== Mic-Array ====
@@ 第21行： / 第22行： @@
 ===Speaker ID===
-*  DNN-based sid --Tian Lan
+*  DNN-based sid --Lantian
 :* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zhangzy&step=view_request&cvssid=327
 ===Ivector&Dvector based ASR===
-*  hold --Tian Lan
+* hold --Tian Lan
 * 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.
+* dark-konowlege using i-vector
-:* 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)
 ===Dark knowledge===
 * test random last output layer when train MPE--zhiyuan
 ===bilingual recognition===
-* hold
+* imbalance dataset(10h,100h and 1400h) to train without share--Zhiyong and mengyuan
-:* http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=zxw&step=view_request&cvssid=359 --Zhiyuan Tang and Mengyuan
+* record utterances mixed with Chinese and English
 ===language vector===
-* train DNN with language vector--xuewei
+* hold --xuewei
 ==Text Processing==

@@ 第7行： / 第7行： @@
 *morpheme RNN-zhiyuan
-* ==== Mic-Array ====
+==== Mic-Array ====
 * hold
 * Change the prediction from  fbank to spectrum features

ASR:2015-06-15 - 版本历史

Zxw：/* Speech Processing */

Zxw：/* Speech Processing */

Zxw：/* RNN AM */

Zxw：以“==Speech Processing == === AM development === ==== Environment ==== * ==== RNN AM==== *morpheme RNN-zhiyuan * ==== Mic-Array ==== * hold * Change the prediction f...”为内容创建页面

Zxw：以“==Speech Processing == === AM development === ==== Environment ==== * ==== RNN AM==== morpheme RNN-zhiyuan ==== Mic-Array ==== * hold * Change the prediction f...”为内容创建页面