Maximum Correntropy Value Decomposition for Multi-agent Deep Reinforcemen Learning
원문 초록
We explore value decomposition solutions for multi-agent deep reinforcement learning in the popular paradigm of centralized training with decentralized execution(CTDE). As the recognized best solution to CTDE, Weighted QMIX is cutting-edge on StarCraft Multi-agent Challenge (SMAC), with a weighting scheme implemented on QMIX to place more emphasis on the optimal joint actions. However, the fixed weight requires manual tuning according to the application scenarios, which painfully prevents Weighted QMIX from being used in broader engineering applications. In this paper, we first demonstrate the flaw of Weighted QMIX using an ordinary One-Step Matrix Game (OMG), that no matter how the weight is chosen, Weighted QMIX struggles to deal with non-monotonic value decomposition problems with a large variance of reward distributions. Then we characterize the problem of value decomposition as an Underfitting One-edged Robust Regression problem and make the first attempt to give a solution to the value decomposition problem from the perspective of information-theoretical learning. We introduce the Maximum Correntropy Criterion (MCC) as a cost function to dynamically adapt the weight to eliminate the effects of minimum in reward distributions. We simplify the implementation and propose a new algorithm called MCVD. A preliminary experiment conducted on OMG shows that MCVD could deal with non-monotonic value decomposition problems with a large tolerance of kernel bandwidth selection. Further experiments are carried out on Cooperative-Navigation and multiple SMAC scenarios, where MCVD exhibits unprecedented ease of implementation, broad applicability, and stability.
함께 살펴볼 논문
공통 태그를 근거로 연결했습니다. 실제 인용 관계나 논문의 우수성을 뜻하지 않습니다.
Hierarchical Reinforcement Learning in StarCraft Micromanagement with Influence Maps and Cluster-based Scripts
Adaptive Reparametrized Time for Score-Based Diffusion Sampling
심리학 기반 추론과 역할 인식 정책 최적화를 통한 일반 역할 연기 에이전트 개선
Improving General Role-Playing Agents via Psychology-Grounded Reasoning and Role-Aware Policy Optimization
When Can Learned Uncertainty Be Trusted? An Audit of Distributional Reinforcement Learning
보상 인플레이션: 강화학습을 위한 건강한 자극
Reward Inflation: A Healthy Stimulus for Reinforcement Learning
Learning Dexterous Manipulation Using Contact Wrench Guidance From Human Demonstration
토론
이 논문과 연결된 포럼 글이 아직 없습니다.