Author name cluster

Lihe Li

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

11 papers

2 author rows

AAMAS Conference 2026 Conference Paper

Boosting Offline MARL under Imbalanced Datasets via Compositional Diffusion Models

Lihe Li
Shenghe Hu
Bingxuan Lan
Yuqi Bian
Huan ZHang
Zhao Ming
Chongjie Zhang
Lei Yuan

Offline multi-agent reinforcement learning (MARL) is hampered by agent-quality imbalance in datasets, where the entanglement of expert and suboptimal behaviors from heterogeneous behavior policies inhibits effective policy learning. Conventional offline MARL methods overfit to these suboptimal behaviors, leading to significant performance degradation. A promising solution is data augmentationusinggenerativemodelslikediffusionmodel, which cangeneratebalanced, high-qualitytrajectoriestoenrichthedataset. However, existing methods usually adopt a standard diffusion process, conditioning generation solely on team-level signals such as globalreturn. Thiscoarseguidancelacksactive, fine-grained, agentlevelcontrol, limitingthediffusionmodel’sabilitytoproducehighquality cooperative behaviors that generalize beyond the dataset. Toaddressthis, weproposeCompositionalDiffusionforImbalanced Datasets (CODI), a novel framework that leverages large language models (LLMs) and diffusion models to generate balanced, highqualitytrajectories. CODIfirstdistillsanagentqualitylabelerfrom anLLMtoannotatethedataset. Itthenemploysaconditionaldiffusion model that generates trajectory segments based on not only return-to-go but also fine-grained agent quality labels. Crucially, This work is licensed under a Creative Commons Attribution International 4. 0 License. Proc. of the 25th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2026), C. Amato, L. Dennis, V. Mascardi, J. Thangarajah (eds.), May 25 – 29, 2026, Paphos, Cyprus. © 2026 International Foundation for Autonomous Agents and Multiagent Systems (www. ifaamas. org). https: //doi. org/10. 65109/WOLI7576 to effective compose scattered high-quality behaviors and enable generalization, CODI decomposes the target team quality into indistribution agent-level labels for compositional diffusion generation. Thesegeneratedsegmentsaresubsequentlystitchedintocomplete trajectories, augmenting the dataset. Extensiveevaluationon challenging imbalanced datasets, where only a single agent is an expert, showsthatCODIsuccessfullymitigatesdataimbalanceand facilitates the learning of strong cooperative policies, recovering 63% of the performance achieved with a balanced expert dataset and substantially outperforming baseline methods.

PDF

AAAI Conference 2026 Conference Paper

Multi-agent In-context Coordination via Decentralized Memory Retrieval

Tao Jiang
Zichuan Lin
Lihe Li
Yi-Chen Li
Cong Guan
Lei Yuan
Zongzhang Zhang
Yang Yu

Large transformer models, trained on diverse datasets, have demonstrated impressive few-shot performance on previously unseen tasks without requiring parameter updates. This capability has also been explored in Reinforcement Learning (RL), where agents interact with the environment to retrieve context and maximize cumulative rewards, showcasing strong adaptability in complex settings. However, in cooperative Multi-Agent Reinforcement Learning (MARL), where agents must coordinate toward a shared goal, decentralized policy deployment can lead to mismatches in task alignment and reward assignment, limiting the efficiency of policy adaptation. To address this challenge, we introduce Multi-agent In-context Coordination via Decentralized Memory Retrieval (MAICC), a novel approach designed to enhance coordination by fast adaptation. Our method involves training a centralized embedding model to capture fine-grained trajectory representations, followed by decentralized models that approximate the centralized one to obtain team-level task information. Based on the learned embeddings, relevant trajectories are retrieved as context, which, combined with the agents' current sub-trajectories, inform decision-making. During decentralized execution, we introduce a novel memory mechanism that effectively balances test-time online data with offline memory. Based on the constructed memory, we propose a hybrid utility score that incorporates both individual- and team-level returns, ensuring credit assignment across agents. Extensive experiments on cooperative MARL benchmarks, including Level-Based Foraging (LBF) and SMAC (v1/v2), show that MAICC enables faster adaptation to unseen tasks compared to existing methods.

PDF Details DOI

NeurIPS Conference 2025 Conference Paper

Adaptable Safe Policy Learning from Multi-task Data with Constraint Prioritized Decision Transformer

Ruiqi Xue
Ziqian Zhang
Lihe Li
Cong Guan
Lei Yuan
Yang Yu

Learning safe reinforcement learning (RL) policies from offline multi-task datasets without direct environmental interaction is crucial for efficient and reliable deployment of RL agents. Benefiting from their scalability and strong in-context learning capabilities, recent approaches attempt to utilize Decision Transformer (DT) architectures for offline safe RL, demonstrating promising adaptability across varying safety budgets. However, these methods primarily focus on single-constraint scenarios and struggle with diverse constraint configurations across multiple tasks. Additionally, their reliance on heuristically defined Return-To-Go (RTG) inputs limits flexibility and reduces learning efficiency, particularly in complex multi-task environments. To address these limitations, we propose CoPDT, a novel DT-based framework designed to enhance adaptability to diverse constraints and varying safety budgets. Specifically, CoPDT introduces a constraint prioritized prompt encoder, which leverages sparse binary cost signals to accurately identify constraints, and a constraint prioritized Return-To-Go (CPRTG) token mechanism, which dynamically generates RTGs based on identified constraints and corresponding safety budgets. Extensive experiments on the OSRL benchmark demonstrate that CoPDT achieves superior efficiency and significantly enhanced safety compliance across diverse multi-task scenarios, surpassing state-of-the-art DT-based methods by satisfying safety constraints in more than twice as many tasks.

PDF Details

ICLR Conference 2025 Conference Paper

Efficient Multi-agent Offline Coordination via Diffusion-based Trajectory Stitching

Lei Yuan 0005
Yuqi Bian
Lihe Li
Ziqian Zhang
Cong Guan
Yang Yu 0001

Learning from offline data without interacting with the environment is a promising way to fully leverage the intelligent decision-making capabilities of multi-agent reinforcement learning (MARL). Previous approaches have primarily focused on developing learning techniques, such as conservative methods tailored to MARL using limited offline data. However, these methods often overlook the temporal relationships across different timesteps and spatial relationships between teammates, resulting in low learning efficiency in imbalanced data scenarios. To comprehensively explore the data structure of MARL and enhance learning efficiency, we propose Multi-Agent offline coordination via Diffusion-based Trajectory Stitching (MADiTS), a novel diffusion-based data augmentation pipeline that systematically generates trajectories by stitching high-quality coordination segments together. MADiTS first generates trajectory segments using a trained diffusion model, followed by applying a bidirectional dynamics constraint to ensure that the trajectories align with environmental dynamics. Additionally, we develop an offline credit assignment technique to identify and optimize the behavior of underperforming agents in the generated segments. This iterative procedure continues until a satisfactory augmented episode trajectory is generated within the predefined limit or is discarded otherwise. Empirical results on imbalanced datasets of multiple benchmarks demonstrate that MADiTS significantly improves MARL performance.

Details

ICML Conference 2025 Conference Paper

Learning to Reuse Policies in State Evolvable Environments

Ziqian Zhang
Bohan Yang 0018
Lihe Li
Yuqi Bian
Ruiqi Xue
Feng Chen 0042
Yi-Chen Li 0001
Lei Yuan 0005

The policy trained via reinforcement learning (RL) makes decisions based on sensor-derived state features. It is common for state features to evolve for reasons such as periodic sensor maintenance or the addition of new sensors for performance improvement. The deployed policy fails in new state space when state features are unseen during training. Previous work tackles this challenge by training a sensor-invariant policy or generating multiple policies and selecting the appropriate one with limited samples. However, both directions struggle to guarantee the performance when faced with unpredictable evolutions. In this paper, we formalize this problem as state evolvable reinforcement learning (SERL), where the agent is required to mitigate policy degradation after state evolutions without costly exploration. We propose Lapse by reusing policies learned from the old state space in two distinct aspects. On one hand, Lapse directly reuses the robust old policy by composing it with a learned state reconstruction model to handle vanishing sensors. On the other hand, the behavioral experience from the old policy is reused by Lapse to train a newly adaptive policy through offline learning, better utilizing new sensors. To leverage advantages of both policies in different scenarios, we further propose automatic ensemble weight adjustment to effectively aggregate them. Theoretically, we justify that robust policy reuse helps mitigate uncertainty and error from both evolution and reconstruction. Empirically, Lapse achieves a significant performance improvement, outperforming the strongest baseline by about $2\times$ in benchmark environments.

Details

ICML Conference 2025 Conference Paper

LLM-Assisted Semantically Diverse Teammate Generation for Efficient Multi-agent Coordination

Lihe Li
Lei Yuan 0005
Pengsen Liu
Tao Jiang
Yang Yu 0001

Training with diverse teammates is the key for learning generalizable agents. Typical approaches aim to generate diverse teammates by utilizing techniques like randomization, designing regularization terms, or reducing policy compatibility, etc. However, such teammates lack semantic information, resulting in inefficient teammate generation and poor adaptability of the agents. To tackle these challenges, we propose Semantically Diverse Teammate Generation (SemDiv), a novel framework leveraging the capabilities of large language models (LLMs) to discover and learn diverse coordination behaviors at the semantic level. In each iteration, SemDiv first generates a novel coordination behavior described in natural language, then translates it into a reward function to train a teammate policy. Once the policy is verified to be meaningful, novel, and aligned with the behavior, the agents train a policy for coordination. Through this iterative process, SemDiv efficiently generates a diverse set of semantically grounded teammates, enabling agents to develop specialized policies, and select the most suitable ones through language-based reasoning to adapt to unseen teammates. Experiments show that SemDiv generates teammates covering a wide range of coordination behaviors, including those unreachable by baseline methods. Evaluation across four MARL environments, each with five unseen representative teammates, demonstrates SemDiv’s superior coordination and adaptability. Our code is available at https: //github. com/lilh76/SemDiv.

Details

IJCAI Conference 2024 Conference Paper

Continual Multi-Objective Reinforcement Learning via Reward Model Rehearsal

Lihe Li
Ruotong Chen
Ziqian Zhang
Zhichao Wu
Yi-Chen Li
Cong Guan
Yang Yu
Lei Yuan

Multi-objective reinforcement learning (MORL) approaches address real-world problems with multiple objectives by learning policies maximizing returns weighted by different user preferences. Typical methods assume the objectives remain unchanged throughout the agent's lifetime. However, in some real-world situations, the agent may encounter dynamically changing learning objectives, i. e. , different vector-valued reward functions at different learning stages. This issue has not been considered in problem formulation or algorithm design. To address this issue, we formalize the setting as a continual MORL (CMORL) problem for the first time, accounting for the evolution of objectives throughout the learning process. Subsequently, we propose Continual Multi-Objective Reinforcement Learning via Reward Model Rehearsal (CORe3), incorporating a dynamic agent network for rapid adaptation to new objectives. Moreover, we develop a reward model rehearsal technique to recover the reward signals for previous objectives, thus alleviating catastrophic forgetting. Experiments on four CMORL benchmarks showcase that CORe3 effectively learns policies satisfying different preferences on all encountered objectives, and outperforms the best baseline by 171%, highlighting the capability of CORe3 to handle situations with evolving objectives.

PDF Details DOI

AAMAS Conference 2024 Conference Paper

Cost-aware Offline Safe Meta Reinforcement Learning with Robust In-Distribution Online Task Adaptation

Cong Guan
Ruiqi Xue
Ziqian Zhang
Lihe Li
Yi-Chen Li
Lei Yuan
Yang Yu

Despite the gained prominence made by reinforcement learning (RL) in various domains, ensuring safety in real-world applications remains a significant challenge. Offline safe RL, which learns safe policies from pre-collected data, has emerged to address these concerns. However, existing approaches assume a single constraint mode and lack adaptability to diverse safety constraints. In realworld scenarios, we often find ourselves working with datasets gathered from various tasks, with the aim of developing a generalized policy capable of handling unknown tasks during testing. To deal with such offline safe meta RL problem, we introduce a novel framework called COSTA, which is designed to facilitate the learning of a safe generalized policy that can adapt and be transferred ∗Yang Yu is the corresponding author. This work is licensed under a Creative Commons Attribution International 4. 0 License. Proc. of the 23rd International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2024), N. Alechina, V. Dignum, M. Dastani, J. S. Sichman (eds.), May 6 – 10, 2024, Auckland, New Zealand. © 2024 International Foundation for Autonomous Agents and Multiagent Systems (www. ifaamas. org). to unknown tasks during testing. COSTA addresses two key challenges in offline safe meta RL: First, it develops a cost-aware task inference module using contrastive learning to distinguish tasks based on safety constraints, mitigating the MDP ambiguity problem. Second, COSTA introduces a novel metric, Safe In-Distribution Score (SIDS), to assess the in-distribution degree of trajectories, out of the consideration of both reward maximization and cost constraint satisfaction. Trajectories collected with a safe exploration policy are filtered using SIDS for robust online task adaptation. Experimental results in a tailored benchmark suite within the Mujoco environments demonstrate that COSTA consistently balances safety and reward maximization, outperforming multiple baselines.

PDF

NeurIPS Conference 2024 Conference Paper

Multi-Agent Domain Calibration with a Handful of Offline Data

Tao Jiang
Lei Yuan
Lihe Li
Cong Guan
Zongzhang Zhang
Yang Yu

The shift in dynamics results in significant performance degradation of policies trained in the source domain when deployed in a different target domain, posing a challenge for the practical application of reinforcement learning (RL) in real-world scenarios. Domain transfer methods aim to bridge this dynamics gap through techniques such as domain adaptation or domain calibration. While domain adaptation involves refining the policy through extensive interactions in the target domain, it may not be feasible for sensitive fields like healthcare and autonomous driving. On the other hand, offline domain calibration utilizes only static data from the target domain to adjust the physics parameters of the source domain (e. g. , a simulator) to align with the target dynamics, enabling the direct deployment of the trained policy without sacrificing performance, which emerges as the most promising for policy deployment. However, existing techniques primarily rely on evolution algorithms for calibration, resulting in low sample efficiency. To tackle this issue, we propose a novel framework Madoc (\textbf{M}ulti-\textbf{a}gent \textbf{do}main \textbf{c}alibration). Firstly, we formulate a bandit RL objective to match the target trajectory distribution by learning a couple of classifiers. We then address the challenge of a large domain parameter space by modeling domain calibration as a cooperative multi-agent reinforcement learning (MARL) problem. Specifically, we utilize a Variational Autoencoder (VAE) to automatically cluster physics parameters with similar effects on the dynamics, grouping them into distinct agents. These grouped agents train calibration policies coordinately to adjust multiple parameters using MARL. Our empirical evaluation on 21 offline locomotion tasks in D4RL and NeoRL benchmarks showcases the superior performance of our method compared to strong existing offline model-based RL, offline domain calibration, and hybrid offline-and-online RL baselines.

PDF Details DOI

UAI Conference 2023 Conference Paper

Fast Teammate Adaptation in the Presence of Sudden Policy Change

Ziqian Zhang
Lei Yuan 0005
Lihe Li
Ke Xue 0001
Chengxing Jia
Cong Guan
Chao Qian 0001
Yang Yu 0001

Cooperative multi-agent reinforcement learning (MARL), where agents coordinates with teammate(s) for a shared goal, may sustain non-stationary caused by the policy change of teammates. Prior works mainly concentrate on the policy change cross episodes, ignoring the fact that teammates may suffer from sudden policy change within an episode, which might lead to miscoordination and poor performance. We formulate the problem as an open Dec-POMDP, where we control some agents to coordinate with uncontrolled teammates, whose policies could be changed within one episode. Then we develop a new framework \textit{\textbf{Fas}t \textbf{t}eammates \textbf{a}da\textbf{p}tation (\textbf{Fastap})} to address the problem. Concretely, we first train versatile teammates’ policies and assign them to different clusters via the Chinese Restaurant Process (CRP). Then, we train the controlled agent(s) to coordinate with the sampled uncontrolled teammates by capturing their identifications as context for fast adaptation. Finally, each agent applies its local information to anticipate the teammates’ context for decision-making accordingly. This process proceeds alternately, leading to a robust policy that can adapt to any teammates during the decentralized execution phase. We show in multiple multi-agent benchmarks that Fastap can achieve superior performance than multiple baselines in stationary and non-stationary scenarios.

Details

AAAI Conference 2023 Conference Paper

Robust Multi-Agent Coordination via Evolutionary Generation of Auxiliary Adversarial Attackers

Lei Yuan
Ziqian Zhang
Ke Xue
Hao Yin
Feng Chen
Cong Guan
Lihe Li
Chao Qian

Cooperative Multi-agent Reinforcement Learning (CMARL) has shown to be promising for many real-world applications. Previous works mainly focus on improving coordination ability via solving MARL-specific challenges (e.g., non-stationarity, credit assignment, scalability), but ignore the policy perturbation issue when testing in a different environment. This issue hasn't been considered in problem formulation or efficient algorithm design. To address this issue, we firstly model the problem as a Limited Policy Adversary Dec-POMDP (LPA-Dec-POMDP), where some coordinators from a team might accidentally and unpredictably encounter a limited number of malicious action attacks, but the regular coordinators still strive for the intended goal. Then, we propose Robust Multi-Agent Coordination via Evolutionary Generation of Auxiliary Adversarial Attackers (ROMANCE), which enables the trained policy to encounter diversified and strong auxiliary adversarial attacks during training, thus achieving high robustness under various policy perturbations. Concretely, to avoid the ego-system overfitting to a specific attacker, we maintain a set of attackers, which is optimized to guarantee the attackers high attacking quality and behavior diversity. The goal of quality is to minimize the ego-system coordination effect, and a novel diversity regularizer based on sparse action is applied to diversify the behaviors among attackers. The ego-system is then paired with a population of attackers selected from the maintained attacker set, and alternately trained against the constantly evolving attackers. Extensive experiments on multiple scenarios from SMAC indicate our ROMANCE provides comparable or better robustness and generalization ability than other baselines.

PDF Details DOI