强化学习（Reinforcement Learning, RL），又称再励学习、评价学习或增强学习，是机器学习的范式和方法论之一，用于描述和解决智能体（agent）在与环境的交互过程中通过学习策略以达成回报最大化或实现特定目标的问题。
本专栏整理了近几年国际顶级会议中，涉及强化学习（Reinforcement Learning, RL）领域的论文。顶级会议包括但不限于：ICML、AAAI、IJCAI、NIPS、ICLR、AAMAS、CVPR、ICRA等。

今天给大家分享的是2022年人工智能AAAI会议（AAAI Conference on Artificial Intelligence, AAAI）中涉及“强化学习”主题的论文。AAAI旨在促进人工智能的研究和负责任的使用，AAAI还旨在增加公众对人工智能的了解，改善人工智能从业者的教学和培训，并为研究计划者和资助方提供关于当前人工智能发展的重要性和潜力以及未来方向的指导。

• [1]. Backprop-Free Reinforcement Learning with Active Neural Generative Coding.
• [2]. Multi-Sacle Dynamic Coding Improved Spiking Actor Network for Reinforcement Learning.
• [3]. CADRE: A Cascade Deep Reinforcement Learning Framework for Vision-Based Autonomous Urban Driving.
• [4]. Achieving Zero Constraint Violation for Constrained Reinforcement Learning via Primal-Dual Approach.
• [5]. OAM: An Option-Action Reinforcement Learning Framework for Universal Multi-Intersection Control.
• [6]. EMVLight: A Decentralized Reinforcement Learning Framework for Efficient Passage of Emergency Vehicles.
• [7]. DeepThermal: Combustion Optimization for Thermal Power Generating Units Using Offline Reinforcement Learning.
• [8]. AlphaHoldem: High-Performance Artificial Intelligence for Heads-Up No-Limit Poker via End-to-End Reinforcement Learning.
• [9]. Conservative and Adaptive Penalty for Model-Based Safe Reinforcement Learning.
• [10]. Robust Adversarial Reinforcement Learning with Dissipation Inequation Constraint.
• [11]. Enforcement Heuristics for Argumentation with Deep Reinforcement Learning.
• [12]. Programmatic Modeling and Generation of Real-Time Strategic Soccer Environments for Reinforcement Learning.
• [13]. Learning by Competition of Self-Interested Reinforcement Learning Agents.
• [14]. Reinforcement Learning with Stochastic Reward Machines.
• [15]. Reinforcement Learning Based Dynamic Model Combination for Time Series Forecasting.
• [16]. Theoretical Guarantees of Fictitious Discount Algorithms for Episodic Reinforcement Learning and Global Convergence of Policy Gradient Methods.
• [17]. Learning Action Translator for Meta Reinforcement Learning on Sparse-Reward Tasks.
• [18]. Wasserstein Unsupervised Reinforcement Learning.
• [19]. Reinforcement Learning of Causal Variables Using Mediation Analysis.
• [20]. Globally Optimal Hierarchical Reinforcement Learning for Linearly-Solvable Markov Decision Processes.
• [21]. Creativity of AI: Automatic Symbolic Option Discovery for Facilitating Deep Reinforcement Learning.
• [22]. Same State, Different Task: Continual Reinforcement Learning without Interference.
• [23]. Introducing Symmetries to Black Box Meta Reinforcement Learning.
• [24]. Deep Reinforcement Learning Policies Learn Shared Adversarial Features across MDPs.
• [25]. Conjugated Discrete Distributions for Distributional Reinforcement Learning.
• [26]. Learn Goal-Conditioned Policy with Intrinsic Motivation for Deep Reinforcement Learning.
• [27]. Fast and Data Efficient Reinforcement Learning from Pixels via Non-parametric Value Approximation.
• [28]. Recursive Reasoning Graph for Multi-Agent Reinforcement Learning.
• [29]. Exploring Safer Behaviors for Deep Reinforcement Learning.
• [30]. Constraint Sampling Reinforcement Learning: Incorporating Expertise for Faster Learning.
• [31]. Unsupervised Reinforcement Learning in Multiple Environments.
• [32]. Control-Oriented Model-Based Reinforcement Learning with Implicit Differentiation.
• [33]. Blockwise Sequential Model Learning for Partially Observable Reinforcement Learning.
• [34]. Offline Reinforcement Learning as Anti-exploration.
• [35]. Regularization Guarantees Generalization in Bayesian Reinforcement Learning through Algorithmic Stability.
• [36]. Sample-Efficient Reinforcement Learning via Conservative Model-Based Actor-Critic.
• [37]. Controlling Underestimation Bias in Reinforcement Learning via Quasi-median Operation.
• [38]. Structure Learning-Based Task Decomposition for Reinforcement Learning in Non-stationary Environments.
• [39]. Generalizing Reinforcement Learning through Fusing Self-Supervised Learning into Intrinsic Motivation.
• [40]. Reinforcement Learning Augmented Asymptotically Optimal Index Policy for Finite-Horizon Restless Bandits.
• [41]. Constraints Penalized Q-learning for Safe Offline Reinforcement Learning.
• [42]. Q-Ball: Modeling Basketball Games Using Deep Reinforcement Learning.
• [43]. Natural Black-Box Adversarial Examples against Deep Reinforcement Learning.
• [44]. SimSR: Simple Distance-Based State Representations for Deep Reinforcement Learning.
• [45]. State Deviation Correction for Offline Reinforcement Learning.
• [46]. Multi-Agent Reinforcement Learning with General Utilities via Decentralized Shadow Reward Actor-Critic.
• [47]. A Multi-Agent Reinforcement Learning Approach for Efficient Client Selection in Federated Learning.
• [48]. Batch Active Learning with Graph Neural Networks via Multi-Agent Deep Reinforcement Learning.
• [49]. Stackelberg Actor-Critic: Game-Theoretic Reinforcement Learning Algorithms.
• [50]. Invariant Action Effect Model for Reinforcement Learning.
• [51]. Locality Matters: A Scalable Value Decomposition Approach for Cooperative Multi-Agent Reinforcement Learning.
• [52]. Concentration Network for Reinforcement Learning of Large-Scale Multi-Agent Systems.
• [53]. A Deeper Understanding of State-Based Critics in Multi-Agent Reinforcement Learning.
• [54]. Goal Recognition as Reinforcement Learning.
• [55]. NICE: Robust Scheduling through Reinforcement Learning-Guided Integer Programming.
• [56]. MAPDP: Cooperative Multi-Agent Reinforcement Learning to Solve Pickup and Delivery Problems.
• [57]. Eye of the Beholder: Improved Relation Generalization for Text-Based Reinforcement Learning Agents.
• [58]. Text-Based Interactive Recommendation via Offline Reinforcement Learning.
• [59]. Multi-Agent Reinforcement Learning Controller to Maximize Energy Efficiency for Multi-Generator Industrial Wave Energy Converter.
• [60]. Bayesian Model-Based Offline Reinforcement Learning for Product Allocation.
• [61]. Reinforcement Learning for Datacenter Congestion Control.
• [62]. Creating Interactive Crowds with Reinforcement Learning.
• [63]. Using Graph-Aware Reinforcement Learning to Identify Winning Strategies in Diplomacy Games (Student Abstract).
• [64]. Reinforcement Learning Explainability via Model Transforms (Student Abstract).
• [65]. Using Reinforcement Learning for Operating Educational Campuses Safely during a Pandemic (Student Abstract).
• [66]. Criticality-Based Advice in Reinforcement Learning (Student Abstract).
• [67]. VeNAS: Versatile Negotiating Agent Strategy via Deep Reinforcement Learning (Student Abstract).