Credit Assignment with Meta-Policy Gradient for Multi-Agent Reinforcement Learning
arXiv:2102.12957
Abstract
Reward decomposition is a critical problem in centralized training with decentralized execution~(CTDE) paradigm for multi-agent reinforcement learning. To take full advantage of global information, which exploits the states from all agents and the related environment for decomposing Q values into individual credits, we propose a general meta-learning-based Mixing Network with Meta Policy Gradient~(MNMPG) framework to distill the global hierarchy for delicate reward decomposition. The excitation signal for learning global hierarchy is deduced from the episode reward difference between before and after "exercise updates" through the utility network. Our method is generally applicable to the CTDE method using a monotonic mixing network. Experiments on the StarCraft II micromanagement benchmark demonstrate that our method just with a simple utility network is able to outperform the current state-of-the-art MARL algorithms on 4 of 5 super hard scenarios. Better performance can be further achieved when combined with a role-based utility network.
References in corpus (17)
- Dota 2 with Large Scale Deep Reinforcement Learning
- Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments
- Human-level performance in first-person multiplayer games with population-based deep reinforcement learning
- Counterfactual Multi-Agent Policy Gradients
- Emergent Tool Use From Multi-Agent Autocurricula
- Social Influence as Intrinsic Motivation for Multi-Agent Deep Reinforcement Learning
- QTRAN: Learning to Factorize with Transformation for Cooperative Multi-Agent Reinforcement Learning
- QPLEX: Duplex Dueling Multi-Agent Q-Learning
- Qatten: A General Framework for Cooperative Multiagent Reinforcement Learning
- MAVEN: Multi-Agent Variational Exploration
- RODE: Learning Roles to Decompose Multi-Agent Tasks
- Learning Implicit Credit Assignment for Cooperative Multi-Agent Reinforcement Learning
- Discovery of Useful Questions as Auxiliary Tasks
- Learning to Explore with Meta-Policy Gradient
- Learning Individually Inferred Communication for Multi-Agent Cooperation
- Exploration with Unreliable Intrinsic Reward in Multi-Agent Reinforcement Learning
- Joint Policy Search for Multi-agent Collaboration with Imperfect Information