Metraipolis › archive › dc_f4abdfc88bd2

Full-Duplex Speech Models Take the Floor When Asked, Not When Needed — arXiv:2609.19596

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.19596
License (as archived)
CC0-1.0
Archived
2026-10-10T03:48:02.134Z
SHA-256 of the archived text
64aca25452d7879a08c726e5eda612a2fb38ae6b01565572bf40a62c07d87ebb
Stable URL of this page
https://metraipolis.com/documento/dc_f4abdfc88bd2

How to cite

Full-Duplex Speech Models Take the Floor When Asked, Not When Needed — arXiv:2609.19596 — arXiv (en), https://arxiv.org/abs/2609.19596 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_f4abdfc88bd2 (Metraipolis archive dc_f4abdfc88bd2, sha256 64aca25452d7879a08c726e5eda612a2fb38ae6b01565572bf40a62c07d87ebb)
BibTeX
@misc{dcf4abdfc88bd2,
  title = {Full-Duplex Speech Models Take the Floor When Asked, Not When Needed — arXiv:2609.19596},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.19596},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_f4abdfc88bd2; sha256 64aca25452d7879a08c726e5eda612a2fb38ae6b01565572bf40a62c07d87ebb},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.19596. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Full-duplex speech models listen and speak at once, promising always-on assistants. Yet they must also decide when they should speak. Human listeners speak when addressed or when the speaker stops, but also self-select to correct a false claim, supply a missing word, or warn of danger. We ask whether full-duplex models do the same. To separate the reason to speak from the opportunity, we construct context-matched English monologues in which only the trigger utterance varies within a topic, define 10 conditions from turn-allocation rules, and compress inter-word pauses to limit opportunities created by silence. Across five model families, being addressed and silence are far more reliable triggers than false facts or hazards. Frame-level text-token probabilities in Moshi and PersonaPlex are lower for false facts than for Neutral when averaged over the first 2\,s after trigger end. Pauses or permission to interrupt do not close this gap either. Given the floor, Moshi and PersonaPlex answer most direct questions, yet the proportion of non-empty false-fact replies that challenge the claim is only .14--.15, and the proportion of hazard replies that warn of danger is .04--.07. This paper thus identifies a gap in both speech initiation and response content. Closing it requires genuine content understanding and intervention decisions grounded in it.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-18 in cs.CL).
Categorie arXiv: Computation and Language (cs.CL) ; Human-Computer Interaction (cs.HC).
Pagina del lavoro: https://arxiv.org/abs/2609.19596 — 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 2026-10-10T03:24:08.125Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-09-18?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).