Metraipolis › archive › dc_46ef58e40bf7
Towards Multilingual End-of-Turn Detection using Frozen Speech Encoders — arXiv:2610.39755
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2610.39755
- License (as archived)
- CC0-1.0
- Archived
- 2026-11-01T03:01:10.700Z
- SHA-256 of the archived text
04b5a9e4e41421095ade9e93f60e18ca502d0c2d85bdd4bb4535c83a0705c54c- Stable URL of this page
https://metraipolis.com/documento/dc_46ef58e40bf7
How to cite
Towards Multilingual End-of-Turn Detection using Frozen Speech Encoders — arXiv:2610.39755 — arXiv (en), https://arxiv.org/abs/2610.39755 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_46ef58e40bf7 (Metraipolis archive dc_46ef58e40bf7, sha256 04b5a9e4e41421095ade9e93f60e18ca502d0c2d85bdd4bb4535c83a0705c54c)
BibTeX
@misc{dc46ef58e40bf7,
title = {Towards Multilingual End-of-Turn Detection using Frozen Speech Encoders — arXiv:2610.39755},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2610.39755},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_46ef58e40bf7; sha256 04b5a9e4e41421095ade9e93f60e18ca502d0c2d85bdd4bb4535c83a0705c54c},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.39755. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
A spoken dialogue system must judge when the user has yielded the floor. This end-of-turn (EoT) detector is normally trained and tuned afresh for each language, which is costly to scale to new languages. We instead build one on a frozen multilingual encoder, training only a small fixed-size head. Using five CALLHOME languages, we investigate three questions. First, how well a head trained on one language transfers zero-shot to the others. Second, how much target-language adaptation is required to recover full task performance. Third, whether a single fixed decision threshold holds across languages without per-language tuning. We find that the head transfers well, even between tonal and pitch-accent languages, and that a few hours of audio recover nearly all the loss, matching in-language training. A frozen multilingual encoder with one lightweight head is a practical, language-agnostic baseline for EoT detection. Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv eess.AS, nuovi lavori del giorno). Categorie arXiv: Audio and Speech Processing (eess.AS). Pagina del lavoro: https://arxiv.org/abs/2610.39755 — 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 2026-11-01T02:26:00.570Z dalla pagina https://arxiv.org/list/eess.AS/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).