English

Universal Paralinguistic Speech Representations Using Self-Supervised Conformers

Sound 2022-12-14 v4 Machine Learning Audio and Speech Processing

Abstract

Many speech applications require understanding aspects beyond the words being spoken, such as recognizing emotion, detecting whether the speaker is wearing a mask, or distinguishing real from synthetic speech. In this work, we introduce a new state-of-the-art paralinguistic representation derived from large-scale, fully self-supervised training of a 600M+ parameter Conformer-based architecture. We benchmark on a diverse set of speech tasks and demonstrate that simple linear classifiers trained on top of our time-averaged representation outperform nearly all previous results, in some cases by large margins. Our analyses of context-window size demonstrate that, surprisingly, 2 second context-windows achieve 96\% the performance of the Conformers that use the full long-term context on 7 out of 9 tasks. Furthermore, while the best per-task representations are extracted internally in the network, stable performance across several layers allows a single universal representation to reach near optimal performance on all tasks.

Keywords

Cite

@article{arxiv.2110.04621,
  title  = {Universal Paralinguistic Speech Representations Using Self-Supervised Conformers},
  author = {Joel Shor and Aren Jansen and Wei Han and Daniel Park and Yu Zhang},
  journal= {arXiv preprint arXiv:2110.04621},
  year   = {2022}
}
R2 v1 2026-06-24T06:45:49.510Z