Cognitive Diversity Harnessing: Complementary Theory-of-Mind Coalitions

Shao et al.'s surprising finding that higher ToM doesn't guarantee better cooperation opens a fascinating avenue. We propose actively exploiting cognitive diversity rather than seeking homogeneity. Inspired by their stable matching approach but diverging fundamentally, this research would develop algorithms that identify complementary cognitive pairings (e.g., high-ToM strategists with low-ToM executors) to outperform uniform teams. This extends Guzman Piedrahita et al.'s work on LLM behavioral patterns by creating intentional heterogeneity. For instance, in disaster response scenarios like Wang et al.'s MITL tasks, high-ToM agents could handle complex temporal planning while low-ToM agents execute reliably under stress. The novelty lies in treating cognitive variation as a resource rather than a limitation, potentially revolutionizing how we design agent teams for real-world unpredictability.

References:

1. Cognitive Insights and Stable Coalition Matching for Fostering Multi-Agent Cooperation. Jiaqi Shao, Tianjun Yuan, Tao Lin, Bing Luo (2024).
2. Cognitive Insights and Stable Coalition Matching for Fostering Multi-Agent Cooperation. Jiaqi Shao, Tianjun Yuan, Tao Lin, Xuanyu Cao, Bing Luo (2024). arXiv.org.
3. Corrupted by Reasoning: Reasoning Language Models Become Free-Riders in Public Goods Games. David Guzman Piedrahita, Yongjin Yang, Mrinmaya Sachan, Giorgia Ramponi, Bernhard Schölkopf, Zhijing Jin (2025). arXiv.org.
4. Decentralized Multi-agent Coordination under MITL Tasks and Communication Constraints. W. Wang, Georg Friedrich Schuppe, Jana Tumova (2022). International Conference on Cyber-Physical Systems.
5. Decentralized Multi-agent Coordination under MITL Specifications and Communication Constraints. W. Wang, Georg Friedrich Schuppe, Jana Tumova (2023). 2023 31st Mediterranean Conference on Control and Automation (MED).