{"ID":6138150,"CreatedAt":"2026-07-09T01:07:32.349475501Z","UpdatedAt":"2026-07-11T08:29:44.454854982Z","DeletedAt":null,"paper_url":"https://arxiv.org/abs/2607.07148","arxiv_id":"2607.07148","title":"Decoupling Conversational Dynamics in Full-Duplex Spoken Models through Reinforcement Learning","abstract":"Recent full-duplex spoken dialogue models have demonstrated compelling progress toward human-like interaction, enabling agents to respond with low latency, produce backchannels, and handle user barge-ins. Yet these improvements in conversational dynamics often come with weaker reasoning and instruction-following abilities, revealing a potential tension between interactive dynamics and intelligence capability. In this paper, we argue that such an intelligence--dynamics trade-off is not fundamental: conversational dynamics can instead be learned as a separate real-time decision policy from human dialogue data. To this end, we propose DuplexPO, a reinforcement learning (RL) framework that decouples when to speak from what to say. It preserves the semantic response capability of an instruction-tuned assistant, while optimizing its temporal interaction behavior over selected high-impact windows from long human conversations. To quantitatively optimize these dynamics, we formulate the Factorized Conversational Dynamics Reward (FCDR) to enable fine-grained temporal credit assignment for turn initiation, backchanneling, yielding, and regularized participation. The policy is then optimized with a GRPO-style objective. Experiments show that DuplexPO substantially improves full-duplex behaviors, including timely backchannels, smooth turn-taking, and barge-in handling, while maintaining strong reasoning and instruction-following performance. Moreover, improvements in dynamics-oriented metrics are reflected in better user experience, suggesting that optimizing conversational timing as a standalone objective can promote more natural full-duplex interaction.","short_abstract":"Recent full-duplex spoken dialogue models have demonstrated compelling progress toward human-like interaction, enabling agents to respond with low latency, produce backchannels, and handle user barge-ins. Yet these improvements in conversational dynamics often come with weaker reasoning and instruction-following abilit...","url_abs":"https://arxiv.org/abs/2607.07148","url_pdf":"https://arxiv.org/pdf/2607.07148v1","authors":"[\"Yuxin Li\",\"Donghang Wu\",\"Guan-Ting Lin\",\"Hung-yi Lee\",\"Chengwei Qin\",\"Zhehuai Chen\",\"Chen Chen\"]","published":"2026-07-08T08:39:34Z","proceeding":"eess.AS","tasks":"[\"eess.AS\"]","methods":"[\"Reinforcement Learning\"]","has_code":false}