“2021-04-19”版本间的差异

来自cslt Wiki
跳转至: 导航搜索
 
(6位用户的8个中间修订版本未显示)
第17行: 第17行:
 
|Yunqi Cai
 
|Yunqi Cai
 
||   
 
||   
*
+
*Intelligent Materials and sensitivity analysis investigation
 
||
 
||
 
*  
 
*  
第28行: 第28行:
 
|Lantian Li
 
|Lantian Li
 
||  
 
||  
*  
+
* Study on robust MAML and its code review.
*  
+
* Test of margin loss.
 
||
 
||
 
*  
 
*  
第40行: 第40行:
 
|Ying Shi
 
|Ying Shi
 
||  
 
||  
*  
+
* HUAWEI project
*  
+
** DAE
 +
** ASR posteriori
 
||
 
||
*  
+
* HUAWEI project
 +
** Asymmetric loss
 +
** Knowledge distillation
 +
* Research
 +
** speech enhancement with deep prior(DAE + clean prior)
 
||
 
||
 
*   
 
*   
第66行: 第71行:
 
|Jiao Han
 
|Jiao Han
 
||  
 
||  
*  
+
* MG normalization experiments and result analysis.
 
||
 
||
 
*  
 
*  
第91行: 第96行:
 
|Di Wang
 
|Di Wang
 
||  
 
||  
*
+
* Find out the "hard trials" of different models through PLDA rating.
 
||
 
||
*
+
* Test the "hard trials" on more powerful SOTA speaker models.
 +
* Human test.
 +
* Complete the SVM experiment.
 
||
 
||
 
*   
 
*   
第102行: 第109行:
 
|Tiankai Zhi
 
|Tiankai Zhi
 
||  
 
||  
*  
+
* MG experiments,observation within class
 
||
 
||
 
*  
 
*  
第112行: 第119行:
 
|Chen Chen
 
|Chen Chen
 
||  
 
||  
*  
+
* APP bug fix & function add
 +
||
 +
* APP bug fix & function add
 
||
 
||
 
*  
 
*  
||
 
 
 
|-
 
|-
  
 
|}
 
|}

2021年4月19日 (一) 03:09的最后版本

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
  • AI 100 writing
  • AI 100 writing
Yunqi Cai
  • Intelligent Materials and sensitivity analysis investigation
Lantian Li
  • Study on robust MAML and its code review.
  • Test of margin loss.
Ying Shi
  • HUAWEI project
    • DAE
    • ASR posteriori
  • HUAWEI project
    • Asymmetric loss
    • Knowledge distillation
  • Research
    • speech enhancement with deep prior(DAE + clean prior)
Haoran Sun
Jiao Han
  • MG normalization experiments and result analysis.
Yang Zhang
  • built a web page for fake audio detection
  • fixed web network issue of huazheng project
  • add recording function for fake audio detection web page
  • recheck the CN-Celeb data
  • add NDA loss for PyTorch based speaker verification
Di Wang
  • Find out the "hard trials" of different models through PLDA rating.
  • Test the "hard trials" on more powerful SOTA speaker models.
  • Human test.
  • Complete the SVM experiment.
Tiankai Zhi
  • MG experiments,observation within class
Chen Chen
  • APP bug fix & function add
  • APP bug fix & function add