Exploring Local Interpretable Model-Agnostic Explanations for Speech Emotion Recognition with Distribution-Shift

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Hjuler, Maja J., Clemmensen, Line H., Das, Sneha
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866912314240794624
author Hjuler, Maja J.
Clemmensen, Line H.
Das, Sneha
author_facet Hjuler, Maja J.
Clemmensen, Line H.
Das, Sneha
contents We introduce EmoLIME, a version of local interpretable model-agnostic explanations (LIME) for black-box Speech Emotion Recognition (SER) models. To the best of our knowledge, this is the first attempt to apply LIME in SER. EmoLIME generates high-level interpretable explanations and identifies which specific frequency ranges are most influential in determining emotional states. The approach aids in interpreting complex, high-dimensional embeddings such as those generated by end-to-end speech models. We evaluate EmoLIME, qualitatively, quantitatively, and statistically, across three emotional speech datasets, using classifiers trained on both hand-crafted acoustic features and Wav2Vec 2.0 embeddings. We find that EmoLIME exhibits stronger robustness across different models than across datasets with distribution shifts, highlighting its potential for more consistent explanations in SER tasks within a dataset.
format Preprint
id arxiv_https___arxiv_org_abs_2504_05368
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Exploring Local Interpretable Model-Agnostic Explanations for Speech Emotion Recognition with Distribution-Shift
Hjuler, Maja J.
Clemmensen, Line H.
Das, Sneha
Sound
Audio and Speech Processing
We introduce EmoLIME, a version of local interpretable model-agnostic explanations (LIME) for black-box Speech Emotion Recognition (SER) models. To the best of our knowledge, this is the first attempt to apply LIME in SER. EmoLIME generates high-level interpretable explanations and identifies which specific frequency ranges are most influential in determining emotional states. The approach aids in interpreting complex, high-dimensional embeddings such as those generated by end-to-end speech models. We evaluate EmoLIME, qualitatively, quantitatively, and statistically, across three emotional speech datasets, using classifiers trained on both hand-crafted acoustic features and Wav2Vec 2.0 embeddings. We find that EmoLIME exhibits stronger robustness across different models than across datasets with distribution shifts, highlighting its potential for more consistent explanations in SER tasks within a dataset.
title Exploring Local Interpretable Model-Agnostic Explanations for Speech Emotion Recognition with Distribution-Shift
topic Sound
Audio and Speech Processing
url https://arxiv.org/abs/2504.05368