Metraipolis › archive › dc_edc867f9aaed
Training-Free Contextual ASR via SpeechLLM-Based Error-Aware Selective Retrieval — arXiv:2609.30694
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2609.30694
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-08T04:19:22.477Z
- SHA-256 of the archived text
4340d37ba0f697939e1695a51e7f863edfb6f9e7ca5143881f529ee47d1a2d24- Stable URL of this page
https://metraipolis.com/documento/dc_edc867f9aaed
How to cite
Training-Free Contextual ASR via SpeechLLM-Based Error-Aware Selective Retrieval — arXiv:2609.30694 — arXiv (en), https://arxiv.org/abs/2609.30694 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_edc867f9aaed (Metraipolis archive dc_edc867f9aaed, sha256 4340d37ba0f697939e1695a51e7f863edfb6f9e7ca5143881f529ee47d1a2d24)
BibTeX
@misc{dcedc867f9aaed,
title = {Training-Free Contextual ASR via SpeechLLM-Based Error-Aware Selective Retrieval — arXiv:2609.30694},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2609.30694},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_edc867f9aaed; sha256 4340d37ba0f697939e1695a51e7f863edfb6f9e7ca5143881f529ee47d1a2d24},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.30694. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Recognition of domain-specific and low-frequency terms remains challenging for automatic speech recognition (ASR). Although contextual biasing can improve their recognition, directly providing a large terminology dictionary introduces many irrelevant biasing terms. Retrieval-based contextual biasing addresses this issue by selecting candidate terms from an external dictionary, but querying many recognized words requires numerous dictionary lookups and may yield poorly targeted candidates. We propose a training-free contextual ASR framework in which a pretrained speech large language model (SpeechLLM) jointly generates an ASR hypothesis and localizes error spans likely to involve domain-specific terms. Only the localized spans are used to retrieve phonologically similar terms from an external terminology dictionary. The same SpeechLLM then re-recognizes the audio conditioned on the first-pass hypothesis and the retrieved terms, without task-specific model training. To assess applicability across domains, we evaluate the framework on medical, air traffic control, and financial speech. The proposed method substantially reduces dictionary queries while improving the recall and ranking of relevant terminology candidates and second-pass ASR performance across all three domains. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-28 in eess.AS). Categorie arXiv: Sound (cs.SD) ; Audio and Speech Processing (eess.AS). Pagina del lavoro: https://arxiv.org/abs/2609.30694 — 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 2026-10-08T03:50:06.037Z dalla pagina https://arxiv.org/catchup/eess.AS/2026-09-28?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).