Metraipolis › archive › dc_5afca18e8063

Measuring Fairness in Large Audio Language Models via Semantic-Aware Bias Estimation — arXiv:2608.13624

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.13624
License (as archived)
CC0-1.0
Archived
2026-10-19T02:21:01.314Z
SHA-256 of the archived text
d431949f9b6460ca23b773bdd361feb27f333fc433e1aa4942379ecfe8cf735b
Stable URL of this page
https://metraipolis.com/documento/dc_5afca18e8063

How to cite

Measuring Fairness in Large Audio Language Models via Semantic-Aware Bias Estimation — arXiv:2608.13624 — arXiv (en), https://arxiv.org/abs/2608.13624 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_5afca18e8063 (Metraipolis archive dc_5afca18e8063, sha256 d431949f9b6460ca23b773bdd361feb27f333fc433e1aa4942379ecfe8cf735b)
BibTeX
@misc{dc5afca18e8063,
  title = {Measuring Fairness in Large Audio Language Models via Semantic-Aware Bias Estimation — arXiv:2608.13624},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.13624},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_5afca18e8063; sha256 d431949f9b6460ca23b773bdd361feb27f333fc433e1aa4942379ecfe8cf735b},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.13624. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large Audio Language Models (LALMs) have seen increasing use for audio understanding tasks such as speech recognition and audio question answering, raising concerns about fairness across demographic subgroups. Fairness evaluation in spoken-input settings is challenging due to confounding factors, including semantic variation in spoken content and speaker-specific characteristics. Ignoring these factors can result in misleading conclusions about model bias. We propose a semantic-aware mixed-effects regression framework for fairness evaluation in LALMs that explicitly accounts for these confounders. Our approach incorporates sentence-level semantic embeddings of reference text as covariates and models speaker identity as a random effect. Notably, semantic representations are extracted from the same LALM under evaluation, enabling semantic control over variation as perceived by the model itself. Experiments on simulated data and real-world benchmarks demonstrate that the proposed approach substantially reduces spurious fairness findings and yields more robust and interpretable estimates of subgroup performance differences.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-17 in cs.AI).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI); Sound (cs.SD).
Pagina del lavoro: https://arxiv.org/abs/2608.13624 — 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 2026-10-19T02:02:02.853Z dalla pagina https://arxiv.org/catchup/cs.AI/2026-08-17?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).