Enhancing Commentary Strategies for Imperfect Information Card Games: A Study of Large Language Models in Guandan Commentary

2024-06-23

arXiv.org

AI Summary

To address the challenge of generating high-quality, real-time Chinese commentary for imperfect-information card games (e.g., Guandan), where large language models (LLMs) struggle with strategic depth and linguistic fluency, this paper proposes an intelligent commentary system integrating reinforcement learning (RL) with LLMs. Methodologically, it introduces Theory of Mind (ToM) modeling—novel in LLM-based strategy understanding—to jointly enable state-guided reasoning, ToM-informed policy analysis, and stylistic retrieval, augmented by retrieval-augmented generation (RAG) and multi-stage information filtering. The system is trained end-to-end on open-source LLMs. Evaluation shows professional-level performance in strategic accuracy, narrative coherence, and naturalness of Chinese expression—outperforming GPT-4 across multiple metrics. This work establishes a new paradigm for explainable AI commentary in complex, adversarial game settings.

Technology Category

• Multiagent Systems: Adversarial Agents
• Cognitive Modeling & Cognitive Systems: Social Cognition And Interaction
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

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

Abstract

Recent advancements in large language models (LLMs) have unlocked the potential for generating high-quality game commentary. However, producing insightful and engaging commentary for complex games with incomplete information remains a significant challenge. In this paper, we introduce a novel commentary method that combine Reinforcement Learning (RL) and LLMs, tailored specifically for the Chinese card game extit{Guandan}. Our system leverages RL to generate intricate card-playing scenarios and employs LLMs to generate corresponding commentary text, effectively emulating the strategic analysis and narrative prowess of professional commentators. The framework comprises a state commentary guide, a Theory of Mind (ToM)-based strategy analyzer, and a style retrieval module, which seamlessly collaborate to deliver detailed and context-relevant game commentary in the Chinese language environment. We empower LLMs with ToM capabilities and refine both retrieval and information filtering mechanisms. This facilitates the generation of personalized commentary content. Our experimental results showcase the substantial enhancement in performance achieved by the proposed commentary framework when applied to open-source LLMs, surpassing the performance of GPT-4 across multiple evaluation metrics.

Problem

• Enhancing commentary for imperfect information card games using LLMs
• Combining RL and LLMs for Guandan game commentary
• Improving Chinese language commentary with ToM and style retrieval

Innovation

• Combines Reinforcement Learning and LLMs
• Uses Theory of Mind for strategy analysis
• Enhances retrieval and filtering mechanisms