“2020-07-06”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43行: 第43行:
 
|Lantian Li
 
|Lantian Li
 
||  
 
||  
*  
+
* Enroll-test mismatch (data collection)
 +
* PyPLDA
 +
* Deisgn on flow-based scoring.
 
||
 
||
*  
+
* Flow-based scoring.
 
||
 
||
 
*   
 
*   
第54行: 第56行:
 
|Ying Shi
 
|Ying Shi
 
||  
 
||  
*  
+
* BP + length norm enhancement method
 
+
* Sub space flow
 
||
 
||
*
+
* Continue on Sub space flow
 +
** Mult-SRN variance
 +
** Cyclie loss
 +
* Sub space flow plan B
 
||
 
||
 
*   
 
*   
第77行: 第82行:
 
|Yue Fan
 
|Yue Fan
 
||  
 
||  
*
+
*Baseline for sound event detection
 +
*CN 's MP4 files organization
 
||
 
||
*
+
*Experiments with different data and models
 
||
 
||
 
*   
 
*   
第101行: 第107行:
 
|Ruiqi Liu
 
|Ruiqi Liu
 
||  
 
||  
*
+
* Paper draft
 
||
 
||
*
+
* Complete the experiment and modify the paper
 
||
 
||
 
*   
 
*   

2020年7月6日 (一) 00:04的最后版本

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
  • More deisgn on prior-based denoising
  • Investigation on the optimum of NL with respect to liklihood ratio test
  • Complete the deepnorm2 paper
  • Continue on the neural Bayes book
Yunqi Cai
Zhiyuan Tang
  • BetaFlow.
  • Layer-by-layer training for flow.
  • Det-constant flow.
  • Analyse det-constant (mostly 1) flow.
Lantian Li
  • Enroll-test mismatch (data collection)
  • PyPLDA
  • Deisgn on flow-based scoring.
  • Flow-based scoring.
Ying Shi
  • BP + length norm enhancement method
  • Sub space flow
  • Continue on Sub space flow
    • Mult-SRN variance
    • Cyclie loss
  • Sub space flow plan B
Haoran Sun
Yue Fan
  • Baseline for sound event detection
  • CN 's MP4 files organization
  • Experiments with different data and models
Jiawen Kang
  • Write multi-condition paper.
  • Arrange near-far data.
  • Arrange cross-channel data.
  • Modify code for new data.
Ruiqi Liu
  • Paper draft
  • Complete the experiment and modify the paper
Haolin Chen
  • FFJORD/Glow test on BP denoising
  • Review papers on self-supervised learning in speech processing
  • Prepare for the report
  • Subspace method for denoising