Metraipolis › archive › dc_4af0d89f13fe

How to Recognize New Words: A Comparison Between Context Biasing Methods and Speech LLMs — arXiv:2608.05759

Source
arXiv (en)
Original
https://arxiv.org/abs/2608.05759
License (as archived)
CC0-1.0
Archived
2026-10-25T21:27:13.723Z
SHA-256 of the archived text
919e177e684c7c68022ded55a7c138376947226bb5c288d177c63a7652e966c9
Stable URL of this page
https://metraipolis.com/documento/dc_4af0d89f13fe

How to cite

How to Recognize New Words: A Comparison Between Context Biasing Methods and Speech LLMs — arXiv:2608.05759 — arXiv (en), https://arxiv.org/abs/2608.05759 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_4af0d89f13fe (Metraipolis archive dc_4af0d89f13fe, sha256 919e177e684c7c68022ded55a7c138376947226bb5c288d177c63a7652e966c9)
BibTeX
@misc{dc4af0d89f13fe,
  title = {How to Recognize New Words: A Comparison Between Context Biasing Methods and Speech LLMs — arXiv:2608.05759},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2608.05759},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_4af0d89f13fe; sha256 919e177e684c7c68022ded55a7c138376947226bb5c288d177c63a7652e966c9},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.05759. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Recognizing new and rare words - named entities, acronyms, domain specific special words, and other items scarce in training data - remains a key challenge for automatic speech recognition (ASR). We compare two strategies for this: context biasing methods, where an ASR model is extended such that during inference a word list can be supplied, and speech large language models (LLMs) prompted with context directly. We evaluate two context biasing methods based on Whisper against three speech LLMs across read and non-read speech, reporting biased, unbiased, and overall word error rate (WER). The context biasing methods cut biased WER by up to 88% relative while leaving other words largely unaffected. Speech LLMs excel on read speech but generalize less well to non-read speech, and prove sensitive to distractor count and prompt word order. We characterize the resulting trade-offs to guide method selection.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-07 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2608.05759 — 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 2026-10-25T21:24:38.204Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-07?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).