“2020-07-27”版本间的差异

来自cslt Wiki
跳转至： 导航、 搜索
 
（3位用户的3个中间修订版本未显示）
第19行： 第19行：
 
|Yunqi Cai
 
|Yunqi Cai
 
||   
 
||   
−
*
+
*experiments on Gaussian regularization
 
||
 
||
 
*  
 
*  
第30行： 第30行：
 
|Zhiyuan Tang
 
|Zhiyuan Tang
 
||  
 
||  
−
*  
+
* Adaptation with nf-dt asr.
 
||
 
||
−
*  
+
* Continue.
 
||
 
||
 
*   
 
*   
第41行： 第41行：
 
|Lantian Li
 
|Lantian Li
 
||  
 
||  
−
*  
+
* Compare two MLE methods.
 +
* Length norm and whitening in NL scoring.
 
||
 
||
−
*  
+
* Self-learning plus NDA.
 
||
 
||
 
*   
 
*   
第52行： 第53行：
 
|Ying Shi
 
|Ying Shi
 
||  
 
||  
−
* sub space flow with gauss. r code  
+
* sub space with trainable mean of r code
 +
* sub space flow with gauss. mixture r code  
 
||
 
||
−
*
+
* analysis the gaussianlity of r space
 +
* verify the robustness of sub space flow
 
||
 
||
 
*   
 
*   

2020年7月27日 (一) 00:01的最后版本

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
  • More investigation on Maximum Gaussianlity training
  • Reformulate the TASLP paper
  • Complete the MG paper
  • Start working on the condition transfer (CT) paper
Yunqi Cai
  • experiments on Gaussian regularization
Zhiyuan Tang
  • Adaptation with nf-dt asr.
  • Continue.
Lantian Li
  • Compare two MLE methods.
  • Length norm and whitening in NL scoring.
  • Self-learning plus NDA.
Ying Shi
  • sub space with trainable mean of r code
  • sub space flow with gauss. mixture r code
  • analysis the gaussianlity of r space
  • verify the robustness of sub space flow
Haoran Sun
  • Analysed latent space of ML-DNF
  • Classification via logpz
  • Analysises of subspace DNF
Yue Fan
Jiawen Kang
  • Read NDA papers;
  • MAML code;
  • MAML experiments.
  • MAML SSMC&MC experiments.
Ruiqi Liu
  • revise the paper draft.
  • Multi-genre generalization.
  • MC experiments data.
  • MC experiments
Haolin Chen
  • Subspace Flow
    • Noisy r var=1, fixed mean
    • GMM Loss
  • Subspace Flow
    • Other noise
    • Improve performance
取自“http://index.cslt.org/mediawiki/index.php?title=2020-07-27&oldid=35303”