第39届国际机器学习会议（International Conference on Machine Learning, ICML 2022）于北京时间7月17日至7月23日，在美国马里兰州巴尔的摩市以线上线下结合的方式举办。

本文列举了会议主题与强化学习（Reinforcement Learning, RL）有关的论文：

• [1]. EAT-C: Environment-Adversarial sub-Task Curriculum for Efficient Reinforcement Learning.
• [2]. Optimizing Sequential Experimental Design with Deep Reinforcement Learning.
• [3]. Interactive Inverse Reinforcement Learning for Cooperative Games.
• [4]. Reinforcement Learning from Partial Observation: Linear Function Approximation with Provable Sample Efficiency.
• [5]. Stabilizing Off-Policy Deep Reinforcement Learning from Pixels.
• [6]. Human-in-the-loop: Provably Efficient Preference-based Reinforcement Learning with General Function Approximation.
• [7]. Adversarially Trained Actor Critic for Offline Reinforcement Learning.
• [8]. Balancing Sample Efficiency and Suboptimality in Inverse Reinforcement Learning.
• [9]. Guarantees for Epsilon-Greedy Reinforcement Learning with Function Approximation.
• [10]. DreamerPro: Reconstruction-Free Model-Based Reinforcement Learning with Prototypical Representations.
• [11]. Branching Reinforcement Learning.
• [12]. Provable Reinforcement Learning with a Short-Term Memory.
• [13]. DRIBO: Robust Deep Reinforcement Learning via Multi-View Information Bottleneck.
• [14]. Cascaded Gaps: Towards Logarithmic Regret for Risk-Sensitive Reinforcement Learning.
• [15]. Fast Population-Based Reinforcement Learning on a Single Machine.
• [16]. Revisiting Some Common Practices in Cooperative Multi-Agent Reinforcement Learning.
• [17]. Blocks Assemble! Learning to Assemble with Large-Scale Structured Reinforcement Learning.
• [18]. Retrieval-Augmented Reinforcement Learning.
• [19]. The State of Sparse Training in Deep Reinforcement Learning.
• [20]. Learning Pseudometric-based Action Representations for Offline Reinforcement Learning.
• [21]. Leveraging Approximate Symbolic Models for Reinforcement Learning via Skill Diversity.
• [22]. Provably Efficient Offline Reinforcement Learning for Partially Observable Markov Decision Processes.
• [23]. Off-Policy Reinforcement Learning with Delayed Rewards.
• [24]. Bisimulation Makes Analogies in Goal-Conditioned Reinforcement Learning.
• [25]. Nearly Minimax Optimal Reinforcement Learning with Linear Function Approximation.
• [26]. On the Role of Discount Factor in Offline Reinforcement Learning.
• [27]. MASER: Multi-Agent Reinforcement Learning with Subgoals Generated from Experience Replay Buffer.
• [28]. Federated Reinforcement Learning: Linear Speedup Under Markovian Sampling.
• [29]. Curriculum Reinforcement Learning via Constrained Optimal Transport.
• [30]. Showing Your Offline Reinforcement Learning Work: Online Evaluation Budget Matters.
• [31]. Goal Misgeneralization in Deep Reinforcement Learning.
• [32]. Scalable Deep Reinforcement Learning Algorithms for Mean Field Games.
• [33]. Phasic Self-Imitative Reduction for Sparse-Reward Goal-Conditioned Reinforcement Learning.
• [34]. Deconfounded Value Decomposition for Multi-Agent Reinforcement Learning.
• [35]. PMIC: Improving Multi-Agent Reinforcement Learning with Progressive Mutual Information Collaboration.
• [36]. Delayed Reinforcement Learning by Imitation.
• [37]. Constrained Variational Policy Optimization for Safe Reinforcement Learning.
• [38]. Welfare Maximization in Competitive Equilibrium: Reinforcement Learning for Markov Exchange Economy.
• [39]. Learning Dynamics and Generalization in Deep Reinforcement Learning.
• [40]. Pessimism meets VCG: Learning Dynamic Mechanism Design via Offline Reinforcement Learning.
• [41]. On Improving Model-Free Algorithms for Decentralized Multi-Agent Reinforcement Learning.
• [42]. Optimizing Tensor Network Contraction Using Reinforcement Learning.
• [43]. A Simple Reward-free Approach to Constrained Reinforcement Learning.
• [44]. EqR: Equivariant Representations for Data-Efficient Reinforcement Learning.
• [45]. The Primacy Bias in Deep Reinforcement Learning.
• [46]. History Compression via Language Models in Reinforcement Learning.
• [47]. Plan Better Amid Conservatism: Offline Multi-Agent Reinforcement Learning with Actor Rectification.