Yuqin Dai

Large Language Models (LLMs) have demonstrated remarkable capabilities in open-domain dialogues. However, their performance in service dialogues remains suboptimal, as these require agents to guide users toward specific business objectives while dynamically tracking states and adapting strategies...