Multiple Choice Learning for Efficient Speech Separation with Many Speakers

  • David Perera
  • , François Derrida
  • , Théo Mariotte
  • , Gaël Richard
  • , Slim Essid

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Training speech separation models in the supervised setting raises a permutation problem: finding the best assignation between the model predictions and the ground truth separated signals. This inherently ambiguous task is customarily solved using Permutation Invariant Training (PIT). In this article, we instead consider using the Multiple Choice Learning (MCL) framework, which was originally introduced to tackle ambiguous tasks. We demonstrate experimentally on the popular WSJ0-mix and LibriMix benchmarks that MCL matches the performances of PIT, while being computationally advantageous. This opens the door to a promising research direction, as MCL can be naturally extended to handle a variable number of speakers, or to tackle speech separation in the unsupervised setting.

Original languageEnglish
Title of host publication2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025 - Proceedings
EditorsBhaskar D Rao, Isabel Trancoso, Gaurav Sharma, Neelesh B. Mehta
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9798350368741
DOIs
Publication statusPublished - 1 Jan 2025
Event2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025 - Hyderabad, India
Duration: 6 Apr 202511 Apr 2025

Publication series

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

Conference

Conference2025 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2025
Country/TerritoryIndia
CityHyderabad
Period6/04/2511/04/25

Keywords

  • Librimix
  • PIT
  • WSJ0-mix
  • cocktail party
  • multiple choice learning
  • speech separation

Fingerprint

Dive into the research topics of 'Multiple Choice Learning for Efficient Speech Separation with Many Speakers'. Together they form a unique fingerprint.

Cite this