2014-09-29 - 版本历史

2014年9月29日 (一) 05:49 Liuc

2014-09-29T05:49:10Z

Lr：/* Text Processing */

2014-09-29T02:38:20Z

‎Text Processing

2014年9月28日 (日) 12:34 Zhangzy

2014-09-28T12:34:50Z

2014年9月28日 (日) 09:09 Cslt

2014-09-28T09:09:24Z

2014年9月28日 (日) 09:07 Cslt

2014-09-28T09:07:39Z

2014年9月28日 (日) 09:07 Cslt

2014-09-28T09:07:16Z

Cslt：以“==Resoruce Building== == Leftover questions== * Investigating LOUDS FST. * CLG embedded decoder plus online compiler. * DNN-GMM co-training * NN LM ==Speech Proc...”为内容创建页面

2014-09-28T09:06:33Z

以“==Resoruce Building== == Leftover questions== * Investigating LOUDS FST. * CLG embedded decoder plus online compiler. * DNN-GMM co-training * NN LM ==Speech Proc...”为内容创建页面

新页面

==Resoruce Building==

== Leftover questions==

* Investigating LOUDS FST.
* CLG embedded decoder plus online compiler.
* DNN-GMM co-training
* NN LM

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

==== Sparse DNN ====
* Investigating layer-based DNN training

====Noise training====
:* First draft of the noisy training journal paper
:* Check abnormal behavior with large sigma (Yinshi, Liuchao)

===Drop out & Rectification & convolutive network===

* Drop out

:* No performance improvement found yet.
:* [http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=wangd&step=view_request&cvssid=261]

* Rectification
:* Dropout NA problem was caused by large magnitude of weights

* Convolutive network
# Test more configurations

====Denoising & Farfield ASR====

* Lasso-based de-reverberation is done with the REVERBERATION toolkit
:* Start to compose the experiment section for the SL paper.

====VAD====

* Noise model training done. Under testing.
* Need to investigate the performance reduction in babble noise. Call Jia.

====Speech rate training====

* Some interesting results with the simple speech rate change algorithm was obtained on the WSJ db
[http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=wangd&step=view_request&cvssid=268]

* Seems ROS model is superior to the normal one with faster speech
* Need to check distribution of ROS on WSJ
* Suggest to extract speech data of different ROS, construct a new test set
* Suggest to use Tencent training data
* Suggest to remove silence when compute ROS

====Scoring====

* Pitch & rythmn done.
* Harmonics hold

====Confidence====

* Basic confidence by using lattice-based posterior + DNN posterior + ROS done
* 23% detection error achieved by balanced model

===Speaker ID===

* Second model done

===Emotion detection===

* delivered to Sinovoice

==Text Processing==
===LM development===

====Domain specific LM====

h2. domain specific count dumped
h2. ngram generation is on going

h2. NUM tag LM:

* HCLG union seems better than G union, when integrating grammar + LM (25->23)
* Boost specific words like wifi if TAG model does not work for a particular word.

===Word2Vector===

====W2V based doc classification====

* Initial results variable Bayesian GMM obtained. Performance is not as good as the conventional GMM.
* Non-linear inter-language transform: English-Spanish-Czch: wv model training done, transform model on investigation
:* probably over-fitting with the MLP training
:* SSA-based local linear mapping still on running

* Knowledge vector started
:* document obtained from wiki

* Character to word conversion
:* Design the transform model

===RNN LM===

* Prepare WSJ database
* Trained model 10000 x 4 + 320 + 10000
* Better performance obtained (4.16-3.47)
* gigaword sampling for Chinese data

==Translation==

* v3.0 demo released
:* still slow

==QA==

* Huilan framework design done
* Investigate better framework

←上一版本		2014年9月28日 (日) 09:07的版本
第1行：		第1行：
−	~~==Resoruce Building==~~
−
−	~~== Leftover questions==~~
−
−	* Investigating LOUDS FST.
−	* CLG embedded decoder plus online compiler.
−	* DNN-GMM co-training
−	* NN LM
−
	==Speech Processing ==		==Speech Processing ==
	=== AM development ===		=== AM development ===

@@ 第3行： / 第3行： @@
 ==== Sparse DNN ====
-* Investigating layer-based DNN training
+* Performance improvement found when pruned slightly
 ====Noise training====
 :* First draft of the noisy training journal paper
 :* Check abnormal behavior with large sigma (Yinshi, Liuchao)
@@ 第21行： / 第24行： @@
 * Convolutive network
 # Test more configurations
@@ 第26行： / 第33行： @@
 ====Denoising & Farfield ASR====
 * Lasso-based de-reverberation is done with the REVERBERATION toolkit
 :* Start to compose the experiment section for the SL paper.
@@ 第31行： / 第39行： @@
 ====VAD====
 * Noise model training done. Under testing.
 * Need to investigate the performance reduction in babble noise. Call Jia.
@@ 第36行： / 第45行： @@
 ====Speech rate training====
-−
+*
 * Some interesting results with the simple speech rate change algorithm was obtained on the WSJ db
 [http://cslt.riit.tsinghua.edu.cn/cgi-bin/cvss/cvss_request.pl?account=wangd&step=view_request&cvssid=268]
@@ 第45行： / 第54行： @@
 * Suggest to use Tencent training data
 * Suggest to remove silence when compute ROS
 ====Scoring====
-* Pitch & rythmn done.
+* global scoring done.
 * Harmonics hold
@@ 第54行： / 第68行： @@
 ====Confidence====
 * Basic confidence by using lattice-based posterior + DNN posterior + ROS done
 * 23% detection error achieved by balanced model
@@ 第59行： / 第75行： @@
 ===Speaker ID===
 * GMM-based test program delivered
-* Implementing GMM registration program
+* GMM registration program done
 ===Emotion detection===
 * Sinovoice is implementing the server

@@ 第72行： / 第72行： @@
 ====Domain specific LM====
 h2. ngram generation is on going
+h2. look the memory and baidu_hi done
 h2. NUM tag LM:
+* maxi work is released.
-* HCLG union seems better than G union, when integrating grammar + LM (25->23)
+* yuanbin continue the tag lm work.
 * Boost specific words like wifi if TAG model does not work for a particular word.
@@ 第88行： / 第88行： @@
 * Initial results variable Bayesian GMM obtained. Performance is not as good as the conventional GMM.
 * Non-linear inter-language transform: English-Spanish-Czch: wv model training done, transform model on investigation
-:* probably over-fitting with the MLP training
+:* SSA-based local linear mapping still on running.
-:* SSA-based local linear mapping still on running
+:* k-means classes change to 2.
 * Knowledge vector started
 :* document obtained from wiki
 * Character to word conversion
-:* Design the transform model
+:* read more paper .
 ===RNN LM===
@@ 第109行： / 第113行： @@
 * v3.0 demo released
 :* still slow
 ===QA===
-* Huilan framework design done
+* liangshan_v1 performance 74.3%.
-* Investigate better framework
+* New framework and GA method is done

@@ 第9行： / 第9行： @@
 :* Check abnormal behavior with large sigma (Yinshi, Liuchao)
-===Drop out & Rectification & convolutive network===
+====Drop out & Rectification & convolutive network====
 * Drop out