Metraipolis › archive › dc_b2c203ff2302

Hierarchical Dual-Graph Retrieval for Long-form Audio Reasoning — arXiv:2610.41390

Source
arXiv (en)
Original
https://arxiv.org/abs/2610.41390
License (as archived)
CC0-1.0
Archived
2026-11-03T08:20:08.176Z
SHA-256 of the archived text
e6cd0b12ef8ef328aebc24df2121b2d49eb611cd54770399c0c064e4342bffdc
Stable URL of this page
https://metraipolis.com/documento/dc_b2c203ff2302

How to cite

Hierarchical Dual-Graph Retrieval for Long-form Audio Reasoning — arXiv:2610.41390 — arXiv (en), https://arxiv.org/abs/2610.41390 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_b2c203ff2302 (Metraipolis archive dc_b2c203ff2302, sha256 e6cd0b12ef8ef328aebc24df2121b2d49eb611cd54770399c0c064e4342bffdc)
BibTeX
@misc{dcb2c203ff2302,
  title = {Hierarchical Dual-Graph Retrieval for Long-form Audio Reasoning — arXiv:2610.41390},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.41390},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_b2c203ff2302; sha256 e6cd0b12ef8ef328aebc24df2121b2d49eb611cd54770399c0c064e4342bffdc},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.41390. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Large audio language models (LALMs) struggle with long-form audio reasoning, as processing entire recordings often limits to capture query-relevant content distributed across distant segments while incurring substantial computational cost. We introduce AudioHelix, a hierarchical DualGraph framework tailored for LALMs' retrieval-augmented generation (RAG) systems. Unlike existing text-only or non-graph RAG approaches, AudioHelix connects related textual and acoustic evidence across temporally distant segments for multi-hop reasoning. Extensive experiments on TraceAV-Bench, MMAU-Pro, and AudioMarathon demonstrate that AudioHelix consistently outperforms full-audio processing and existing RAG methods. The project site is this https URL .

Tema degli esploratori: statistica e scienza dei dati (elenco arXiv cs.IR, nuovi lavori del giorno).
Categorie arXiv: Sound (cs.SD) ; Information Retrieval (cs.IR); Audio and Speech Processing (eess.AS).
Pagina del lavoro: https://arxiv.org/abs/2610.41390 — 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 2026-11-03T08:06:23.756Z dalla pagina https://arxiv.org/list/cs.IR/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).