MEGA Hub

Cocktail-Talker: Multi-Speaker Dialog Modeling in Noisy Social Environments with Turn Action GRPO

Authors

Do you know Xilin Jiang?You can claim authorship or link another user.Do you know Riki Shimizu?You can claim authorship or link another user.Do you know Sukru Samet Dindar?You can claim authorship or link another user.Do you know Junkai Wu?You can claim authorship or link another user.Do you know Zhongweiyang Xu?You can claim authorship or link another user.Do you know Nima Mesgarani?You can claim authorship or link another user.

Abstract

Spoken dialog systems are typically designed for clean, dyadic interactions in which a single user and an assistant take turns speaking. Real-world social conversations, however, are often more ambiguous: multiple speakers may participate in the same conversation amid irrelevant speech and background noise. Each utterance may be directed to the assistant, addressed to another speaker, or completely irrelevant. In such settings, the assistant must decide not only what to say, but also whether to speak at all. In this paper, we introduce Cocktail-Talker, a speech LLM framework for multi-speaker spoken dialog modeling in noisy social environments. We model the assistant's behavior with three action tokens: <|respond|>, <|listen|>, and <|ignore|>, placed before a response or silence. Cocktail-Talker is trained via supervised finetuning and reinforcement learning to generate the appropriate action token and, only in <|respond|> mode, a speech response. To prepare the training data, we develop Cocktail-DialogGen, an LLM-based data pipeline that simulates realistic multi-speaker dialogs with speaker roles across diverse social settings. Together, these components take a step toward spoken dialog systems that interact more naturally and selectively in complex social environments.

Community

00