Metraipolis › archive › dc_d415f27cd2e2

Enhancing Audio Reasoning via Semantic Summary Prediction — arXiv:2609.20849

Source: arXiv (en)
Original: https://arxiv.org/abs/2609.20849
License (as archived): CC0-1.0
Archived: 2026-10-10T01:27:04.809Z
SHA-256 of the archived text: 4d0b4f4cf76871431f7f2207f3672e63581c686f24d0a8fbe40911834bf0eb9a
Stable URL of this page: https://metraipolis.com/documento/dc_d415f27cd2e2

How to cite

Enhancing Audio Reasoning via Semantic Summary Prediction — arXiv:2609.20849 — arXiv (en), https://arxiv.org/abs/2609.20849 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_d415f27cd2e2 (Metraipolis archive dc_d415f27cd2e2, sha256 4d0b4f4cf76871431f7f2207f3672e63581c686f24d0a8fbe40911834bf0eb9a)

BibTeX

@misc{dcd415f27cd2e2,
  title = {Enhancing Audio Reasoning via Semantic Summary Prediction — arXiv:2609.20849},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.20849},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_d415f27cd2e2; sha256 4d0b4f4cf76871431f7f2207f3672e63581c686f24d0a8fbe40911834bf0eb9a},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.20849. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large Audio Language Models (LALMs) perform well on complex question answering but often show a reasoning gap, where explicit Chain-of-Thought (CoT) reduces accuracy compared to direct answers. We hypothesize that long reasoning sequences shift attention away from the audio input. To address this, we propose SPARE (Semantic Prediction for Audio REasoning), which introduces a register token aligned with the final conclusion using a cosine similarity loss with a Sentence-BERT embedding. This conditions the model's latent space with the target semantic goal before reasoning begins. Experiments on MMAU and MMAR with SALMONN show improved zero-shot reasoning and stronger early attention to audio without additional inference cost.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-21 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Sound (cs.SD); Audio and Speech Processing (eess.AS).
Pagina del lavoro: https://arxiv.org/abs/2609.20849 — 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 2026-10-10T01:00:30.259Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-21?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).