Metraipolis › archive › dc_f778d566e375

KeyASRBench: Measuring Whole-Phrase Keyword Preservation in Real-World ASR — arXiv:2611.10852

Source
arXiv (en)
Original
https://arxiv.org/abs/2611.10852
License (as archived)
CC0-1.0
Archived
2026-11-10T18:42:52.427Z
SHA-256 of the archived text
1913005f0b5001d7ad1297475702d998e2e75fc1321fd3dd9c25c33aa0cad2d6
Stable URL of this page
https://metraipolis.com/documento/dc_f778d566e375

How to cite

KeyASRBench: Measuring Whole-Phrase Keyword Preservation in Real-World ASR — arXiv:2611.10852 — arXiv (en), https://arxiv.org/abs/2611.10852 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f778d566e375 (Metraipolis archive dc_f778d566e375, sha256 1913005f0b5001d7ad1297475702d998e2e75fc1321fd3dd9c25c33aa0cad2d6)
BibTeX
@misc{dcf778d566e375,
  title = {KeyASRBench: Measuring Whole-Phrase Keyword Preservation in Real-World ASR — arXiv:2611.10852},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2611.10852},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f778d566e375; sha256 1913005f0b5001d7ad1297475702d998e2e75fc1321fd3dd9c25c33aa0cad2d6},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2611.10852. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Public automatic speech recognition (ASR) benchmarks increasingly report low sequence error rates, yet production systems still use hotword lists, contextual biasing, retrieval, and post-correction. This motivates a complementary question: does a transcript preserve selected names, products, technical terms, versions, and relational phrases? We present KeyASRBench, a bilingual pilot that operationalizes occurrence-based Keyword Error Rate (KER) over frozen whole-phrase annotations. The 10-hour model-conditioned challenge set contains 1,501 Chinese and English clips with 4,723 keyword occurrences. Across five locally inferred checkpoints and two official API systems, language-specific MER/WER ranges from 5.31% to 12.91%, while KER ranges from 13.61% to 33.81%. A disjoint 30-minute model-blind control yields KER of 10.65%-27.31%. The control confirms that selected-phrase errors extend beyond challenge-set selection.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI).
Pagina del lavoro: https://arxiv.org/abs/2611.10852 — 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 2026-11-10T18:37:37.612Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).