A Factorial Deep Markov Model for Unsupervised Disentangled Representation Learning from Speech

Sameer Khurana, Shafiq Rayhan Joty, Ahmed Ali, James Glass

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

We present the Factorial Deep Markov Model (FDMM) for representation learning of speech. The FDMM learns disentangled, interpretable and lower dimensional latent representations from speech without supervision. We use a static and dynamic latent variable to exploit the fact that information in a speech signal evolves at different time scales. Latent representations learned by the FDMM outperform a baseline i-vector system on speaker verification and dialect identification while also reducing the error rate of a phone recognition system in a domain mismatch scenario.

Original languageEnglish
Title of host publication2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
Pages6540-6544
Number of pages5
ISBN (Electronic)9781479981311
DOIs
Publication statusPublished - 1 May 2019
Event44th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Brighton, United Kingdom
Duration: 12 May 201917 May 2019

Publication series

NameICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
Volume2019-May
ISSN (Print)1520-6149

Conference

Conference44th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019
CountryUnited Kingdom
CityBrighton
Period12/5/1917/5/19

Keywords

  • Disentangled Representation Learning
  • Factorial Deep Markov Model
  • Variational Inference

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Khurana, S., Rayhan Joty, S., Ali, A., & Glass, J. (2019). A Factorial Deep Markov Model for Unsupervised Disentangled Representation Learning from Speech. In 2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings (pp. 6540-6544). [8683131] (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings; Vol. 2019-May). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ICASSP.2019.8683131

A Factorial Deep Markov Model for Unsupervised Disentangled Representation Learning from Speech. / Khurana, Sameer; Rayhan Joty, Shafiq; Ali, Ahmed; Glass, James.

2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. p. 6540-6544 8683131 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings; Vol. 2019-May).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Khurana, S, Rayhan Joty, S, Ali, A & Glass, J 2019, A Factorial Deep Markov Model for Unsupervised Disentangled Representation Learning from Speech. in 2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings., 8683131, ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings, vol. 2019-May, Institute of Electrical and Electronics Engineers Inc., pp. 6540-6544, 44th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019, Brighton, United Kingdom, 12/5/19. https://doi.org/10.1109/ICASSP.2019.8683131
Khurana S, Rayhan Joty S, Ali A, Glass J. A Factorial Deep Markov Model for Unsupervised Disentangled Representation Learning from Speech. In 2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2019. p. 6540-6544. 8683131. (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). https://doi.org/10.1109/ICASSP.2019.8683131
Khurana, Sameer ; Rayhan Joty, Shafiq ; Ali, Ahmed ; Glass, James. / A Factorial Deep Markov Model for Unsupervised Disentangled Representation Learning from Speech. 2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 6540-6544 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).
@inproceedings{656a8d1ee2484a92992303eb84e2d515,
title = "A Factorial Deep Markov Model for Unsupervised Disentangled Representation Learning from Speech",
abstract = "We present the Factorial Deep Markov Model (FDMM) for representation learning of speech. The FDMM learns disentangled, interpretable and lower dimensional latent representations from speech without supervision. We use a static and dynamic latent variable to exploit the fact that information in a speech signal evolves at different time scales. Latent representations learned by the FDMM outperform a baseline i-vector system on speaker verification and dialect identification while also reducing the error rate of a phone recognition system in a domain mismatch scenario.",
keywords = "Disentangled Representation Learning, Factorial Deep Markov Model, Variational Inference",
author = "Sameer Khurana and {Rayhan Joty}, Shafiq and Ahmed Ali and James Glass",
year = "2019",
month = "5",
day = "1",
doi = "10.1109/ICASSP.2019.8683131",
language = "English",
series = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
pages = "6540--6544",
booktitle = "2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings",

}

TY - GEN

T1 - A Factorial Deep Markov Model for Unsupervised Disentangled Representation Learning from Speech

AU - Khurana, Sameer

AU - Rayhan Joty, Shafiq

AU - Ali, Ahmed

AU - Glass, James

PY - 2019/5/1

Y1 - 2019/5/1

N2 - We present the Factorial Deep Markov Model (FDMM) for representation learning of speech. The FDMM learns disentangled, interpretable and lower dimensional latent representations from speech without supervision. We use a static and dynamic latent variable to exploit the fact that information in a speech signal evolves at different time scales. Latent representations learned by the FDMM outperform a baseline i-vector system on speaker verification and dialect identification while also reducing the error rate of a phone recognition system in a domain mismatch scenario.

AB - We present the Factorial Deep Markov Model (FDMM) for representation learning of speech. The FDMM learns disentangled, interpretable and lower dimensional latent representations from speech without supervision. We use a static and dynamic latent variable to exploit the fact that information in a speech signal evolves at different time scales. Latent representations learned by the FDMM outperform a baseline i-vector system on speaker verification and dialect identification while also reducing the error rate of a phone recognition system in a domain mismatch scenario.

KW - Disentangled Representation Learning

KW - Factorial Deep Markov Model

KW - Variational Inference

UR - http://www.scopus.com/inward/record.url?scp=85068965658&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85068965658&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2019.8683131

DO - 10.1109/ICASSP.2019.8683131

M3 - Conference contribution

AN - SCOPUS:85068965658

T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

SP - 6540

EP - 6544

BT - 2019 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2019 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

ER -