Difference between revisions of "2020-02-17"

From cslt Wiki
Jump to: navigation, search
(以“{| class="wikitable" !People !! This Week !! Next Week !! Task Tracking (<font color="red">DeadLine</font>) |- |- |Dong Wang || * || * || * |- |- |Yunqi Cai...”为内容创建页面)
 
 
(12 intermediate revisions by 11 users not shown)
Line 6: Line 6:
 
|Dong Wang
 
|Dong Wang
 
||  
 
||  
*  
+
* Code cleaning for DNF
 +
* More review on DNF theory
 +
* More review on speech enhancement with deep generative models
 
||
 
||
*  
+
* More investigation on subspace DNF
 +
* Draft paper for DNF properties
 
||
 
||
 
*   
 
*   
Line 17: Line 20:
 
|Yunqi Cai
 
|Yunqi Cai
 
||  
 
||  
*  
+
* retest the plda score(lenth norm false)
 +
* train and test sota xvector
 
||
 
||
*  
+
* more investigation on Between class var of DNF
 
||
 
||
 
*   
 
*   
Line 28: Line 32:
 
|Zhiyuan Tang
 
|Zhiyuan Tang
 
||  
 
||  
*  
+
* Confirmed that Glow can disentangle audio (content/non-content, silence/non-silence, to be confirmed)
 +
* Noise audio cannot be disentangled.
 
||
 
||
*  
+
* Try to disentangel noise audio.
 
||
 
||
 
*   
 
*   
Line 39: Line 44:
 
|Lantian Li
 
|Lantian Li
 
||  
 
||  
*  
+
* Update SOTA.
 +
* Complete NF + LDA/PLDA.
 
||
 
||
*  
+
* DNF + LDA/PLDA
 
||
 
||
 
*   
 
*   
Line 50: Line 56:
 
|Ying Shi
 
|Ying Shi
 
||  
 
||  
*
+
* double flow denoise experiment and visualization
 +
* DAE baseline
 
||
 
||
*
+
* DAE baseline
 +
* timit ASR baseline
 
||
 
||
 
*   
 
*   
Line 83: Line 91:
 
|Yue Fan
 
|Yue Fan
 
||  
 
||  
*
+
*Arrange the work of the CN2
 
||
 
||
*
+
*baseline on svm&gmm
 
||
 
||
 
*   
 
*   
Line 94: Line 102:
 
|Jiawen Kang
 
|Jiawen Kang
 
||  
 
||  
*  
+
* Prepare SOTA vectors
 +
* Prepare CNceleb as test set
 +
* Continute experiments
 
||
 
||
*
+
* Continute experiments
 +
* evaluatte models on diffierent geners
 +
* Prepare paper writing
 
||
 
||
 
*   
 
*   
Line 105: Line 117:
 
|Ruiqi Liu
 
|Ruiqi Liu
 
||  
 
||  
*  
+
* Experiments about ResNet and prepare vectors.
 +
 
 
||
 
||
*  
+
* Supplement the last experimental data  and prepare paper writing.
 +
 
 
||
 
||
 
*   
 
*   
Line 116: Line 130:
 
|Sitong Cheng
 
|Sitong Cheng
 
||  
 
||  
*  
+
* modify DNF code
 
||
 
||
*  
+
* prepare paper writing
 +
* step 2&3 with DNF
 
||
 
||
 
*   
 
*   
Line 127: Line 142:
 
|Zhixin Liu
 
|Zhixin Liu
 
||  
 
||  
*  
+
* Get initial conclusion of task 1
 
||
 
||
*  
+
* Analyze results
 +
* Prepare DNF
 +
||
 +
 +
|-
 +
 
 +
|-
 +
|Haolin Chen
 +
||
 +
* Run experiments with white noise, consider adding discriminative structure
 +
||
 +
* Run experiments with realnoise, Evaluation on result.
 
||
 
||
 
*   
 
*   

Latest revision as of 00:14, 17 February 2020

People This Week Next Week Task Tracking (DeadLine)
Dong Wang
  • Code cleaning for DNF
  • More review on DNF theory
  • More review on speech enhancement with deep generative models
  • More investigation on subspace DNF
  • Draft paper for DNF properties
Yunqi Cai
  • retest the plda score(lenth norm false)
  • train and test sota xvector
  • more investigation on Between class var of DNF
Zhiyuan Tang
  • Confirmed that Glow can disentangle audio (content/non-content, silence/non-silence, to be confirmed)
  • Noise audio cannot be disentangled.
  • Try to disentangel noise audio.
Lantian Li
  • Update SOTA.
  • Complete NF + LDA/PLDA.
  • DNF + LDA/PLDA
Ying Shi
  • double flow denoise experiment and visualization
  • DAE baseline
  • DAE baseline
  • timit ASR baseline
Wenqiang Du
Haoran Sun
Yue Fan
  • Arrange the work of the CN2
  • baseline on svm&gmm
Jiawen Kang
  • Prepare SOTA vectors
  • Prepare CNceleb as test set
  • Continute experiments
  • Continute experiments
  • evaluatte models on diffierent geners
  • Prepare paper writing
Ruiqi Liu
  • Experiments about ResNet and prepare vectors.
  • Supplement the last experimental data and prepare paper writing.
Sitong Cheng
  • modify DNF code
  • prepare paper writing
  • step 2&3 with DNF
Zhixin Liu
  • Get initial conclusion of task 1
  • Analyze results
  • Prepare DNF
Haolin Chen
  • Run experiments with white noise, consider adding discriminative structure
  • Run experiments with realnoise, Evaluation on result.