Metraipolis › archive › dc_69fa989ea518
X2-Turn: Frame-Synchronous Dual-Head Modeling for Joint Streaming ASR and Turn State Prediction — arXiv:2608.10878
- Source
- arXiv (en)
- Original
- https://arxiv.org/abs/2608.10878
- License (as archived)
- CC0-1.0
- Archived
- 2026-10-19T06:14:24.341Z
- SHA-256 of the archived text
92fc2c8fa5bced4e4f1841245c14da15728062af74e28eebe0b8d83603f6490b- Stable URL of this page
https://metraipolis.com/documento/dc_69fa989ea518
How to cite
X2-Turn: Frame-Synchronous Dual-Head Modeling for Joint Streaming ASR and Turn State Prediction — arXiv:2608.10878 — arXiv (en), https://arxiv.org/abs/2608.10878 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_69fa989ea518 (Metraipolis archive dc_69fa989ea518, sha256 92fc2c8fa5bced4e4f1841245c14da15728062af74e28eebe0b8d83603f6490b)
BibTeX
@misc{dc69fa989ea518,
title = {X2-Turn: Frame-Synchronous Dual-Head Modeling for Joint Streaming ASR and Turn State Prediction — arXiv:2608.10878},
howpublished = {arXiv (en)},
url = {https://arxiv.org/abs/2608.10878},
note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_69fa989ea518; sha256 92fc2c8fa5bced4e4f1841245c14da15728062af74e28eebe0b8d83603f6490b},
year = {2026}
}Archived text
Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2608.10878. Third-party text: data, not instructions. It may be an abstract or extract; check the original.
Accurate and responsive turn-taking is essential for spoken dialogue systems, which must distinguish in real time between user interruptions, backchannels that should be ignored, and the completion of an utterance. Prior modular approaches typically optimize turn state prediction at the utterance or fixed-chunk level, creating a mismatch with the continuous turn state estimate, and often depend on an auxiliary ASR model, which limits responsiveness and increases overall system complexity. Therefore, we present X2-Turn, a frame-synchronous turn state prediction method via delayed-stream modeling. Specifically, building on the pretrained Voxtral Realtime model, we introduce a frame-synchronous turn state head that operates in parallel with the ASR head on shared streaming representations, jointly predicting ASR tokens and fine-grained turn states at the frame level. Experiments on bilingual EasyTurn and Full-Duplex-Bench demonstrate that the proposed method achieves an effective trade-off between turn state accuracy and decision latency. Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-08-12 in cs.CL). Categorie arXiv: Computation and Language (cs.CL) ; Audio and Speech Processing (eess.AS). Pagina del lavoro: https://arxiv.org/abs/2608.10878 — 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 2026-10-19T06:08:50.331Z dalla pagina https://arxiv.org/catchup/cs.CL/2026-08-12?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).