Metraipolis › archive › dc_2be340eea5c8
Relevant Evidence Decoding for Audio-Visual Hallucination Mitigation — arXiv:2610.02976
- Source
- arXiv
- Original
- https://arxiv.org/abs/2610.02976
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-05T22:49:24.957Z
- SHA-256 of the archived text
9097cf70fdead426cc9865bde7da2367ad661884075b10af51c15395bdc203cd- Stable URL of this page
https://metraipolis.com/documento/dc_2be340eea5c8
How to cite
Relevant Evidence Decoding for Audio-Visual Hallucination Mitigation — arXiv:2610.02976 — arXiv, https://arxiv.org/abs/2610.02976 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_2be340eea5c8 (Metraipolis archive dc_2be340eea5c8, sha256 9097cf70fdead426cc9865bde7da2367ad661884075b10af51c15395bdc203cd)
BibTeX
@misc{dc2be340eea5c8,
title = {Relevant Evidence Decoding for Audio-Visual Hallucination Mitigation — arXiv:2610.02976},
howpublished = {arXiv},
url = {https://arxiv.org/abs/2610.02976},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_2be340eea5c8; sha256 9097cf70fdead426cc9865bde7da2367ad661884075b10af51c15395bdc203cd},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv, https://arxiv.org/abs/2610.02976. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Audio-Visual Large Language Models (AV-LLMs) remain prone to cross-modal hallucinations, where one modality incorrectly affects predictions about another. Although contrastive decoding reduces hallucinations in vision-language models, its direct extension to AV-LLMs overlooks a key challenge: different questions require different perceptual evidence, including audio, video, or their interaction. Notably, we observe that joint audio-visual inference can weaken the prediction even when a model can recover the correct answer from a single informative modality. For example, when asked which instrument is heard, a model may correctly predict violin from the audio alone. Once a video showing a guitar is added, its confidence in violin may drop. In this paper, we introduce Relevant Evidence Decoding (RED), a training-free method that identifies question-relevant evidence and selectively strengthens its contribution. RED uses pointwise mutual information to quantify the predictive support provided by audio and video beyond the question alone. It decomposes their joint contribution into audio, video, and residual interaction components. A question-only inference pass determines the required evidence type, after which the model augments the original audio-visual prediction with the corresponding PMI contribution. Across three audio-visual hallucination benchmarks and three AV-LLMs, RED improves accuracy over standard decoding by up to 7% on CMM, 6.3% on AVHBench, and 3.8% on SVHalluc, with an average relative time to first token of 1.5x standard decoding. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno). Categorie arXiv: Artificial Intelligence (cs.AI). Pagina del lavoro: https://arxiv.org/abs/2610.02976 — 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 2026-10-05T22:45:55.659Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).