English

LanSER: Language-Model Supported Speech Emotion Recognition

Computation and Language 2023-09-11 v1 Machine Learning Sound Audio and Speech Processing

Abstract

Speech emotion recognition (SER) models typically rely on costly human-labeled data for training, making scaling methods to large speech datasets and nuanced emotion taxonomies difficult. We present LanSER, a method that enables the use of unlabeled data by inferring weak emotion labels via pre-trained large language models through weakly-supervised learning. For inferring weak labels constrained to a taxonomy, we use a textual entailment approach that selects an emotion label with the highest entailment score for a speech transcript extracted via automatic speech recognition. Our experimental results show that models pre-trained on large datasets with this weak supervision outperform other baseline models on standard SER datasets when fine-tuned, and show improved label efficiency. Despite being pre-trained on labels derived only from text, we show that the resulting representations appear to model the prosodic content of speech.

Keywords

Cite

@article{arxiv.2309.03978,
  title  = {LanSER: Language-Model Supported Speech Emotion Recognition},
  author = {Taesik Gong and Josh Belanich and Krishna Somandepalli and Arsha Nagrani and Brian Eoff and Brendan Jou},
  journal= {arXiv preprint arXiv:2309.03978},
  year   = {2023}
}

Comments

Presented at INTERSPEECH 2023

R2 v1 2026-06-28T12:15:41.419Z