Think Socially via Cognitive Reasoning

2025-09-26

AI Summary

While large language models (LLMs) excel at formal logical reasoning, they struggle with ambiguous, polysemous, and underdetermined social cues prevalent in real-world interpersonal contexts. Method: This paper introduces “cognitive reasoning” as a novel paradigm, formalizing human social cognition through a tree-structured cognitive flow framework that explicitly captures its associative, incremental, and explanatory nature. We train LLMs via supervised fine-tuning and multi-objective reinforcement learning on a self-constructed cognitive flow dataset, enabling dynamic interpretation and adaptive response to social ambiguity. Contribution/Results: Experiments demonstrate substantial improvements in LLM performance on complex social decision-making tasks—outperforming conventional logic-based baselines—and yield computationally grounded insights into the mechanisms of human social cognition.

Technology Category

• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: (Large) Language Models

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• User Modeling, Personalization and Recommendation: Large Language Models (LLM) for user modeling and recommendation
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems

Abstract

LLMs trained for logical reasoning excel at step-by-step deduction to reach verifiable answers. However, this paradigm is ill-suited for navigating social situations, which induce an interpretive process of analyzing ambiguous cues that rarely yield a definitive outcome. To bridge this gap, we introduce Cognitive Reasoning, a paradigm modeled on human social cognition. It formulates the interpretive process into a structured cognitive flow of interconnected cognitive units (e.g., observation or attribution), which combine adaptively to enable effective social thinking and responses. We then propose CogFlow, a complete framework that instills this capability in LLMs. CogFlow first curates a dataset of cognitive flows by simulating the associative and progressive nature of human thought via tree-structured planning. After instilling the basic cognitive reasoning capability via supervised fine-tuning, CogFlow adopts reinforcement learning to enable the model to improve itself via trial and error, guided by a multi-objective reward that optimizes both cognitive flow and response quality. Extensive experiments show that CogFlow effectively enhances the social cognitive capabilities of LLMs, and even humans, leading to more effective social decision-making.

Problem

• Enhancing social reasoning in LLMs
• Modeling human social cognition processes
• Improving ambiguous social cue interpretation

Innovation

• Cognitive Reasoning models human social cognition
• CogFlow framework uses tree-structured planning
• Combines supervised fine-tuning with reinforcement learning