作者位次 #3/9ACL 2026Poster· Natural Language Processing
Dual-Axis Generative Reward Model Toward Semantic and Turn-taking Robustness in Interactive Spoken Dialogue Models
Yifu Chen · Shengpeng Ji · Zhengqing Liu · Qian Chen · Wen Wang · Ziqing Wang · Yangzhuo Li · Tianle Liang · Zhou Zhao
摘要
Achieving seamless, human-like interaction remains a key challenge for full-duplex spoken dialogue models (SDMs). Reinforcement learning (RL) has substantially enhanced text- and vision-language models, while well-designed reward signals are crucial for the performance of RL. We consider RL a pro...