Min Chen
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Min Chen has a total of 16 co-authors. Here are the top co-authors by number of shared publications:
- Zhiqiang Pu
Institute of Automation, Chinese Academy of Sciences, Beijing, China
5 shared publications - Shijie Wang
Shanghai AI Laboratory
3 shared publications - Yi Pan
Institute of Automation, Chinese Academy of Sciences, Beijing, China
3 shared publications - Boyin Liu
School of Artificial Intelligence University of Chinese Academy of Sciences Beijing China
2 shared publications - Hao Ma
Tencent
2 shared publications - Jianqiang Yi
Institute of Automation, Chinese Academy of Sciences, Beijing, China
2 shared publications
Latest Publications
- 18 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. PP (8) , 1-14
MA2RL: Masked Autoencoders for Generalizable Multi-Agent Reinforcement Learning
To develop generalizable models in multi-agent reinforcement learning, recent approaches have been devoted to discovering task-independent skills for each agent, which generalize across tasks and facilitate agents’ cooperation. However, particularly in partially observed settings, such approaches struggle with sample efficiency and generalization capabilities due to two primary challenges: (a) How to incorporate global states into coordinating the skills of different agents? (b) How to learn generalizable and consistent skill semantics when each agent only receives partial observations? To address these challenges, we propose a framework called Masked Autoencoders for Multi-Agent Reinforcement Learning (MA2RL), which encourages agents to infer unobserved entities by reconstructing entity-states from the entity perspective. The entity perspective helps MA2RL generalize to diverse tasks with varying agent numbers and action spaces. Specifically, we treat local entity-observations as masked contexts of the global entity-states, and MA2RL can infer the latent representation of dynamically masked entities, facilitating the assignment of task-independent skills and the learning of skill semantics. Extensive experiments demonstrate that MA2RL achieves significant improvements relative to state-of-the-art approaches, demonstrating extraordinary performance, remarkable zero-shot generalization capabilities and advantageous transfer-ability.
- 18 February 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Artificial Intelligence
- Vol. PP (8) , 1-14
MA2RL: Masked Autoencoders for Generalizable Multi-Agent Reinforcement Learning
To develop generalizable models in multi-agent reinforcement learning, recent approaches have been devoted to discovering task-independent skills for each agent, which generalize across tasks and facilitate agents’ cooperation. However, particularly in partially observed settings, such approaches struggle with sample efficiency and generalization capabilities due to two primary challenges: (a) How to incorporate global states into coordinating the skills of different agents? (b) How to learn generalizable and consistent skill semantics when each agent only receives partial observations? To address these challenges, we propose a framework called Masked Autoencoders for Multi-Agent Reinforcement Learning (MA2RL), which encourages agents to infer unobserved entities by reconstructing entity-states from the entity perspective. The entity perspective helps MA2RL generalize to diverse tasks with varying agent numbers and action spaces. Specifically, we treat local entity-observations as masked contexts of the global entity-states, and MA2RL can infer the latent representation of dynamically masked entities, facilitating the assignment of task-independent skills and the learning of skill semantics. Extensive experiments demonstrate that MA2RL achieves significant improvements relative to state-of-the-art approaches, demonstrating extraordinary performance, remarkable zero-shot generalization capabilities and advantageous transfer-ability.
- 19 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive and Developmental Systems
- Vol. 17 (5) , 1110-1124
Efficient Multitask Reinforcement Learning via Task-Specific Action Correction
Multi-task reinforcement learning (MTRL) holds potential for building general-purpose agents, enabling them to generalize across a variety of tasks. However, MTRL may still be susceptible to conflicts between tasks. A primary reason for this problem is that a universal policy struggles to balance short-term and dense learning signals across various tasks, e.g. , distinct reward functions in reinforcement learning. In social cognitive theory, internalized future goals, as a form of cognitive representations, can effectively mitigate potential short-term conflicts in multitask settings. Considering the benefits of future goals, we propose a novel and general framework called Task-Specific Action Correction (TSAC) from the goal perspective as an orthogonal research to previous MTRL methods. Specifically, to avoid myopia, TSAC introduces goal-oriented sparse rewards and decomposes policy learning into two separate policies: a shared policy (SP) and an action correction policy (ACP). The SP outputs a short-term perspective action based on guiding dense rewards. To alleviate conflicts resulting from excessive focus on specific tasks’ details in SP, the ACP incorporates goal-oriented sparse rewards, enabling an agent to adopt a long-term perspective to output a correction action and achieve generalization across tasks. Finally, the actions output by SP and ACP are combined based on the action correction function to form a final action that interact with the environment. Extensive experiments conducted on Meta-World and multi-task StarCraft II multi-agent scenarios demonstrate that TSAC outperforms existing state-of-the-art methods, achieving significant improvements in sample efficiency, generalization and effective action execution across tasks.
- 19 February 2025
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive and Developmental Systems
- Vol. 17 (5) , 1110-1124
Efficient Multitask Reinforcement Learning via Task-Specific Action Correction
Multi-task reinforcement learning (MTRL) holds potential for building general-purpose agents, enabling them to generalize across a variety of tasks. However, MTRL may still be susceptible to conflicts between tasks. A primary reason for this problem is that a universal policy struggles to balance short-term and dense learning signals across various tasks, e.g. , distinct reward functions in reinforcement learning. In social cognitive theory, internalized future goals, as a form of cognitive representations, can effectively mitigate potential short-term conflicts in multitask settings. Considering the benefits of future goals, we propose a novel and general framework called Task-Specific Action Correction (TSAC) from the goal perspective as an orthogonal research to previous MTRL methods. Specifically, to avoid myopia, TSAC introduces goal-oriented sparse rewards and decomposes policy learning into two separate policies: a shared policy (SP) and an action correction policy (ACP). The SP outputs a short-term perspective action based on guiding dense rewards. To alleviate conflicts resulting from excessive focus on specific tasks’ details in SP, the ACP incorporates goal-oriented sparse rewards, enabling an agent to adopt a long-term perspective to output a correction action and achieve generalization across tasks. Finally, the actions output by SP and ACP are combined based on the action correction function to form a final action that interact with the environment. Extensive experiments conducted on Meta-World and multi-task StarCraft II multi-agent scenarios demonstrate that TSAC outperforms existing state-of-the-art methods, achieving significant improvements in sample efficiency, generalization and effective action execution across tasks.
- Top Cited Papers
- 12 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/CAA Journal of Automatica Sinica
- Vol. 11 (1) , 37-57
Orientation and Decision-Making for Soccer Based on Sports Analytics and AI: A Systematic Review
Due to ever-growing soccer data collection approaches and progressing artificial intelligence (AI) methods, soccer analysis, evaluation, and decision-making have received increasing interest from not only the professional sports analytics realm but also the academic AI research community. AI brings game-changing approaches for soccer analytics where soccer has been a typical benchmark for AI research. The combination has been an emerging topic. In this paper, soccer match analytics are taken as a complete observation-orientation-decision-action (OODA) loop. In addition, as in AI frameworks such as that for reinforcement learning, interacting with a virtual environment enables an evolving model. Therefore, both soccer analytics in the real world and virtual domains are discussed. With the intersection of the OODA loop and the real-virtual domains, available soccer data, including event and tracking data, and diverse orientation and decision-making models for both real-world and virtual soccer matches are comprehensively reviewed. Finally, some promising directions in this interdisciplinary area are pointed out. It is claimed that paradigms for both professional sports analytics and AI research could be combined. Moreover, it is quite promising to bridge the gap between the real and virtual domains for soccer match analysis and decision-making.
- 12 January 2024
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE/CAA Journal of Automatica Sinica
- Vol. 11 (1) , 37-57
Top Cited PapersOrientation and Decision-Making for Soccer Based on Sports Analytics and AI: A Systematic Review
Due to ever-growing soccer data collection approaches and progressing artificial intelligence (AI) methods, soccer analysis, evaluation, and decision-making have received increasing interest from not only the professional sports analytics realm but also the academic AI research community. AI brings game-changing approaches for soccer analytics where soccer has been a typical benchmark for AI research. The combination has been an emerging topic. In this paper, soccer match analytics are taken as a complete observation-orientation-decision-action (OODA) loop. In addition, as in AI frameworks such as that for reinforcement learning, interacting with a virtual environment enables an evolving model. Therefore, both soccer analytics in the real world and virtual domains are discussed. With the intersection of the OODA loop and the real-virtual domains, available soccer data, including event and tracking data, and diverse orientation and decision-making models for both real-world and virtual soccer matches are comprehensively reviewed. Finally, some promising directions in this interdisciplinary area are pointed out. It is claimed that paradigms for both professional sports analytics and AI research could be combined. Moreover, it is quite promising to bridge the gap between the real and virtual domains for soccer match analysis and decision-making.
- 22 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive and Developmental Systems
- Vol. 16 (4) , 1302-1314
QFuture: Learning Future Expectation Cognition in Multiagent Reinforcement Learning
In multi-agent reinforcement learning (MARL), agents must learn to cooperate by observing the environment and selecting actions that maximize their rewards. However, this learning process can be hampered by myopia, wherein agents’ strategies fail to consider the long-term consequences of their actions. A primary reason for this problem is the inaccurate estimation of the long-term value of each action. Socially, humans derive future expectation cognition from available information to anticipate potential future outcomes and adjust their actions accordingly to avoid myopia. Motivated by these insights, this paper proposes a novel framework called QFuture to address the myopia problem. Specifically, we first design a future expectation cognition module (FECM) in this framework to build future expectation cognition in the calculation of individual actionvalue (IAV) and joint action-value (JAV). We model future expectation cognition as random variables in FECM, which learn representation by maximizing mutual information with the future trajectory based on current information. Furthermore, a return-based regularizer is designed to reflect ”expectation” and ensure informativeness in the future expectation representation module (FERM) which encodes the future trajectory. Experiments on StarCraft II micromanagement tasks and Google Research Football show that QFuture achieves significant state-of-the-art performance. Demonstrative videos are available at https://sites.google.com/view/qfuture .
- 22 December 2023
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Transactions on Cognitive and Developmental Systems
- Vol. 16 (4) , 1302-1314
QFuture: Learning Future Expectation Cognition in Multiagent Reinforcement Learning
In multi-agent reinforcement learning (MARL), agents must learn to cooperate by observing the environment and selecting actions that maximize their rewards. However, this learning process can be hampered by myopia, wherein agents’ strategies fail to consider the long-term consequences of their actions. A primary reason for this problem is the inaccurate estimation of the long-term value of each action. Socially, humans derive future expectation cognition from available information to anticipate potential future outcomes and adjust their actions accordingly to avoid myopia. Motivated by these insights, this paper proposes a novel framework called QFuture to address the myopia problem. Specifically, we first design a future expectation cognition module (FECM) in this framework to build future expectation cognition in the calculation of individual actionvalue (IAV) and joint action-value (JAV). We model future expectation cognition as random variables in FECM, which learn representation by maximizing mutual information with the future trajectory based on current information. Furthermore, a return-based regularizer is designed to reflect ”expectation” and ensure informativeness in the future expectation representation module (FERM) which encodes the future trajectory. Experiments on StarCraft II micromanagement tasks and Google Research Football show that QFuture achieves significant state-of-the-art performance. Demonstrative videos are available at https://sites.google.com/view/qfuture .
- 15 April 2023
- Published by Springer Nature
- p. 3-14
Knowledge Transfer from Situation Evaluation to Multi-agent Reinforcement Learning
- 15 April 2023
- Published by Springer Nature
- p. 3-14
Knowledge Transfer from Situation Evaluation to Multi-agent Reinforcement Learning