TPE: Towards Better Compositional Reasoning over Cognitive Tools via Multi-persona Collaboration

Hongru Wang, Huimin Wang, Lingzhi Wang, Minda Hu, Rui Wang, Boyang Xue, Yongfeng Huang, Kam-Fai Wong

Abstract: Previous works in tool learning mainly focus on the external function tools, such as models and APIs, while overlooking the existence of cognitive tools inside the cognitive/thinking processing of human mind. In this paper, we expand the scope of tools, centering on cognitive tools within the context of dialogue systems, such as multiple sources and psychological or tutoring strategies being dynamically applied in a single turn to compose helpful responses. To further enhance the reasoning and planning capability of LLMs with these cognitive tools, we introduce a multi-persona collaboration framework: Think-Plan-Execute (TPE). This framework decouples the response generation process into three distinct roles: Thinker, Planner, and Executor. Specifically, the Thinker analyzes the internal status exhibited in the dialogue context, such as user emotions and preferences, to formulate a global guideline. The Planner then generates executable plans to call different cognitive tools (e.g., sources or strategies), while the Executor compiles all intermediate results into a coherent response. Experimental results on three public datasets (FoCus, CIMA, and PsyQA) demonstrate the effectiveness and robustness of TPE, revealing the promising potential of cognitive tools as a substantial expansion of current toolset.