Multi Agent Reinforcement Learning
Multi-agent reinforcement learning (MARL) focuses on developing algorithms that enable multiple independent agents to learn optimal strategies within a shared environment, often to achieve a common goal. Current research emphasizes improving sample efficiency and generalization, exploring novel architectures like equivariant graph neural networks and specialized network structures (e.g., Bottom-Up Networks), and addressing challenges posed by non-stationarity and partial observability through techniques such as auxiliary prioritization and global state inference with diffusion models. MARL's significance lies in its potential to solve complex real-world problems across diverse domains, including robotics, traffic control, and healthcare, by enabling effective coordination and collaboration among multiple agents.
Papers - Page 17
Learning to Model Diverse Driving Behaviors in Highly Interactive Autonomous Driving Scenarios with Multi-Agent Reinforcement Learning
Liu Weiwei, Hu Wenxuan, Jing Wei, Lei Lanxin, Gao Lingping, Liu YongA Neuro-Symbolic Approach to Multi-Agent RL for Interpretability and Probabilistic Decision Making
Chitra Subramanian, Miao Liu, Naweed Khan, Jonathan Lenchner, Aporva Amarnath, Sarathkrishna Swaminathan, Ryan Riegel, Alexander Gray