Deep Generative Factorization For Speech Signal(ICASSP21)

来自cslt Wiki
2020年10月23日 (五) 08:10Sunhaoran（讨论 | 贡献）的版本

跳转至： 导航、 搜索

Introduction

  • This paper presented a speech information factorization method based on a novel deep generative model that we called factorial discriminative normalization flow. Qualitative and quantitative experimental results show that compared to all other models, the proposed factorial DNF can retain the class structure corresponding to multiple information factors, and changing one factor will cause little distortion on other factors. This demonstrates that factorial DNF can well factorize speech signal into different information factors.

Members

  • Haoran Sun, Lantian Li, Yunqi Cai, Yang Zhang, Thomas Fang Zheng, Dong Wang

Publications

  • Haoran Sun, Lantian Li, Yunqi Cai, Yang Zhang, Thomas Fang Zheng, Dong Wang, "Deep Generative Factorization For Speech Signal", 2020. pdf

Source Code

  • xxx

Factorial DNF

  • xxx

Experiments

Data

  • xx

Encoding

  • The latent codes generated by VAE and NF almost lose the class structure; DNF can retain the class structure of the information factor corresponding to the class labels in the model training; Factorial DNF can retain the class structure corresponding to all the information factors.
  • The latent codes generated by various models are as below, plotted by t-SNE. In the first row (a) to (e), each color represents a phone; in the second row (f) to (j), each color represents a speaker. ‘Phone DNF’ denotes DNF trained with phone labels; ‘Speaker DNF’ denotes DNF trained with speaker labels.

Fdnf tsne.png

Factor manipulation

x' = f(f−1(x) + µA,c2 − µA,c1)
  • MLP posteriors on the target class before and after phone/speaker manipulation are as below. ‘f-DNF’ denotes factorial DNF. δ(·) denotes the difference on posteriors p(·|x') and p(·|x) 
                        Phone Manipulation
 Model |  p(q2|x) | p(q2|x') |  δ(q2)  ||  p(s|x) |  p(s|x') |   δ(s)
  VAE  |   0.013  |  0.312   |  0.299   ||  0.612  |  0.454   |  -0.158 
  NF   |   0.013  |  0.410   |  0.397   ||  0.612  |  0.489   |  -0.123 
  DNF  |   0.013  |  0.619   |  0.606   ||  0.612  |  0.335   |  -0.277  
 f-DNF |   0.013  |  0.636  |  0.623   ||  0.612  |  0.536  |  -0.076 

                        Speaker Manipulation
 Model |  p(s2|x) | p(s2|x') |  δ(s2)  ||  p(q|x) |  p(q|x') |   δ(q)
  VAE  |   0.010  |  0.303   |  0.293   ||  0.520  |  0.509  |  -0.011
  NF   |   0.010  |  0.435   |  0.425   ||  0.520  |  0.484   |  -0.036 
  DNF  |   0.010  |  0.700   |  0.690   ||  0.520  |  0.349   |  -0.171  
 f-DNF |   0.010  |  0.710  |  0.700   ||  0.520  |  0.503   |  -0.017

Future Work

  • Test factorial DNF on larger datasets.
  • Establish general theories for deep generative factorization.
取自“http://index.cslt.org/mediawiki/index.php?title=Deep_Generative_Factorization_For_Speech_Signal(ICASSP21)&oldid=35583”