Metraipolis › archive › dc_57db48fb24ae

CharDuplex: Building Character-Consistent Full-Duplex Spoken Dialogue Models — arXiv:2609.34461

Source
arXiv (en)
Original
https://arxiv.org/abs/2609.34461
License (as archived)
CC0-1.0
Archived
2026-10-08T01:37:19.710Z
SHA-256 of the archived text
e8d0913fa894c3cf50213976576fac787b9b2b9e03d4584b2d57af1e8f914181
Stable URL of this page
https://metraipolis.com/documento/dc_57db48fb24ae

How to cite

CharDuplex: Building Character-Consistent Full-Duplex Spoken Dialogue Models — arXiv:2609.34461 — arXiv (en), https://arxiv.org/abs/2609.34461 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_57db48fb24ae (Metraipolis archive dc_57db48fb24ae, sha256 e8d0913fa894c3cf50213976576fac787b9b2b9e03d4584b2d57af1e8f914181)
BibTeX
@misc{dc57db48fb24ae,
  title = {CharDuplex: Building Character-Consistent Full-Duplex Spoken Dialogue Models — arXiv:2609.34461},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2609.34461},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_57db48fb24ae; sha256 e8d0913fa894c3cf50213976576fac787b9b2b9e03d4584b2d57af1e8f914181},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2609.34461. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

Full-duplex speech models are moving voice interaction beyond conventional turn-taking, yet natural conversation is shaped not only by when an agent speaks, but also by how it behaves as a conversational character. We present CharDuplex, a character-driven full-duplex speech model that combines real-time spoken interaction with persona-conditioned behavior. We first adapt GLM-4-Voice to an always-on dual-stream architecture and train the model for full-duplex conversation. Then a fully automated pipeline constructs character-conditioned dialogue data from open-source character descriptions for character-conditioned supervised fine-tuning. The model is further refined with the proposed FDGym, where an LLM-simulated user dynamically interacts with the model, enabling reinforcement learning over evolving multi-turn interactions. On SpeechRole-Eval, CharDuplex achieves the highest average score among the evaluated open-source models, while remaining competitive with closed-source systems. It also demonstrates competitive general speech intelligence and strong full-duplex interaction capabilities. CharDuplex demonstrates a practical training recipe for building full-duplex voice assistants that are not only interactive, but also character-consistent.

Tema degli esploratori: agenti e intelligenza artificiale (retro-catalogo arXiv: annunci del 2026-09-29 in eess.AS).
Categorie arXiv: Audio and Speech Processing (eess.AS).
Pagina del lavoro: https://arxiv.org/abs/2609.34461 — 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 2026-10-08T01:11:22.073Z dalla pagina https://arxiv.org/catchup/eess.AS/2026-09-29?abs=True (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).