Metraipolis › archive › dc_6bee17a26413
From Speech to Editable Concepts: Probing Emotion Recognition with Concept Bottleneck Models — arXiv:2609.39453
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.39453
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-07T12:45:25.079Z
- SHA-256 of the archived text
68c4b9d8acd1b6b5abb214c3ef115708a265b98b1b8d317fddcc3e60435cd4db- Stable URL of this page
https://metraipolis.com/documento/dc_6bee17a26413
How to cite
From Speech to Editable Concepts: Probing Emotion Recognition with Concept Bottleneck Models — arXiv:2609.39453 — arXiv (en), https://arxiv.org/abs/2609.39453 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_6bee17a26413 (Metraipolis archive dc_6bee17a26413, sha256 68c4b9d8acd1b6b5abb214c3ef115708a265b98b1b8d317fddcc3e60435cd4db)
BibTeX
@misc{dc6bee17a26413,
title = {From Speech to Editable Concepts: Probing Emotion Recognition with Concept Bottleneck Models — arXiv:2609.39453},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.39453},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_6bee17a26413; sha256 68c4b9d8acd1b6b5abb214c3ef115708a265b98b1b8d317fddcc3e60435cd4db},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.39453. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Speech emotion recognition (SER) is the task of assigning emotion labels to utterances. Early systems relied on acoustic features, whereas recent approaches combine multiple modalities, most commonly speech and text. Still, performance remains poor on many datasets. Large language models (LLMs) have therefore attracted interest for SER, as they can process diverse inputs jointly with instructions. However, direct audio input raises questions of explainability. To address similar questions in image classification, concept bottleneck models were introduced. This work adapts concept bottlenecks to SER to examine how individual predictions depend on transcripts, acoustic descriptions and speaker attributes. Experiments test three LLMs on CREMA-D, IEMOCAP and MELD, with concepts extracted by separate tools. On scripted corpora, LLMs are strongly biased towards the transcript in the zero-shot setting, which lowers Macro-F1 from 27.8 to 5.8 on CREMA-D. Fine-tuning removes this bias, and the transcript raises Macro-F1 from 41.8 to 45.1. Removing speech rate changes 48% of Neutral predictions to Disgust on CREMA-D; removing intensity level on MELD changes predictions despite little change in Macro-F1. These findings show that aggregate performance changes alone do not capture the effects of concept removal on individual predictions. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-10-01 in cs.AI). Categorie arXiv: Sound (cs.SD) ; Artificial Intelligence (cs.AI); Computation and Language (cs.CL); Audio and Speech Processing (eess.AS). Pagina del lavoro: https://arxiv.org/abs/2609.39453 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito). Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina. Letto il 2026-10-07T12:14:14.088Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-10-01?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).