# Concept Learning for Cooperative Multi-Agent Reinforcement Learning

arXiv: https://arxiv.org/abs/2507.20143 · HTML: https://mlsanity.com/paper/2507.20143/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 21 of its 28 references that are in the map.
- First posted on arXiv 27 Jul 2025.

## Builds on (most cited first, 21 of 21)

- [Concept Bottleneck Models](https://mlsanity.com/paper/2007.04612/) (2020)
- [QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/1803.11485/) (2018)
- [The StarCraft Multi-Agent Challenge](https://mlsanity.com/paper/1902.04043/) (2019)
- [Interpretation of Neural Networks is Fragile](https://mlsanity.com/paper/1710.10547/) (2017)
- [QTRAN: Learning to Factorize with Transformation for Cooperative Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/1905.05408/) (2019)
- [QPLEX: Duplex Dueling Multi-Agent Q-Learning](https://mlsanity.com/paper/2008.01062/) (2020)
- [Weighted QMIX: Expanding Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2006.10800/) (2020)
- [Verifiable Reinforcement Learning via Policy Extraction](https://mlsanity.com/paper/1805.08328/) (2018)
- [Reliable Post hoc Explanations: Modeling Uncertainty in Explainability](https://mlsanity.com/paper/2008.05030/) (2020)
- [Shared Experience Actor-Critic for Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2006.07169/) (2020)
- [Celebrating Diversity in Shared Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2106.02195/) (2021)
- [Qatten: A General Framework for Cooperative Multiagent Reinforcement Learning](https://mlsanity.com/paper/2002.03939/) (2020)
- [Shapley Q-value: A Local Reward Approach to Solve Global Reward Games](https://mlsanity.com/paper/1907.05707/) (2019)
- [Graying the black box: Understanding DQNs](https://mlsanity.com/paper/1602.02658/) (2016)
- [Guided Deep Reinforcement Learning for Swarm Systems](https://mlsanity.com/paper/1709.06011/) (2017)
- [SHAQ: Incorporating Shapley Value Theory into Multi-Agent Q-Learning](https://mlsanity.com/paper/2105.15013/) (2021)
- [Q-value Path Decomposition for Deep Multiagent Reinforcement Learning](https://mlsanity.com/paper/2002.03950/) (2020)
- [Interactive Disentanglement: Learning Concepts by Interacting with their Prototype Representations](https://mlsanity.com/paper/2112.02290/) (2021)
- [Concept Gradient: Concept-based Interpretation Without Linear Assumption](https://mlsanity.com/paper/2208.14966/) (2022)
- [NA2Q: Neural Attention Additive Model for Interpretable Multi-Agent Q-Learning](https://mlsanity.com/paper/2304.13383/) (2023)
- [MIXRTs: Toward Interpretable Multi-Agent Reinforcement Learning via Mixing Recurrent Soft Decision Trees](https://mlsanity.com/paper/2209.07225/) (2022)

## Abstract (arXiv)

Despite substantial progress in applying neural networks (NN) to multi-agent reinforcement learning (MARL) areas, they still largely suffer from a lack of transparency and interoperability. However, its implicit cooperative mechanism is not yet fully understood due to black-box networks. In this work, we study an interpretable value decomposition framework via concept bottleneck models, which promote trustworthiness by conditioning credit assignment on an intermediate level of human-like cooperation concepts. To address this problem, we propose a novel value-based method, named Concepts learning for Multi-agent Q-learning (CMQ), that goes beyond the current performance-vs-interpretability trade-off by learning interpretable cooperation concepts. CMQ represents each cooperation concept as a supervised vector, as opposed to existing models where the information flowing through their end-to-end mechanism is concept-agnostic. Intuitively, using individual action value conditioning on global state embeddings to represent each concept allows for extra cooperation representation capacity. Empirical evaluations on the StarCraft II micromanagement challenge and level-based foraging (LBF) show that CMQ achieves superior performance compared with the state-of-the-art counterparts. The results also demonstrate that CMQ provides more cooperation concept representation capturing meaningful cooperation modes, and supports test-time concept interventions for detecting potential biases of cooperation mode and identifying spurious artifacts that impact cooperation.