Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Deep Coordination Graphs

About

This paper introduces the deep coordination graph (DCG) for collaborative multi-agent reinforcement learning. DCG strikes a flexible trade-off between representational capacity and generalization by factoring the joint value function of all agents according to a coordination graph into payoffs between pairs of agents. The value can be maximized by local message passing along the graph, which allows training of the value function end-to-end with Q-learning. Payoff functions are approximated with deep neural networks that employ parameter sharing and low-rank approximations to significantly improve sample efficiency. We show that DCG can solve predator-prey tasks that highlight the relative overgeneralization pathology, as well as challenging StarCraft II micromanagement tasks.

Wendelin B\"ohmer, Vitaly Kurin, Shimon Whiteson• 2019

Related benchmarks

TaskDatasetResultRank
Multi-Agent Reinforcement LearningSimple Spread N=6
Collisions0.1425
23
Multi-Agent Reinforcement LearningSimple Spread N=3
Collisions0.0322
23
Multi-Agent Reinforcement LearningSimple Spread N=4
Collisions7.56
23
Multi-Agent Reinforcement LearningMPE Speaker-Listener
Return21.2
17
Cooperative Multi-Agent Reinforcement LearningCrypto (last 2% of train)
Mean Episodic Reward50
13
Cooperative Multi-Agent Reinforcement LearningReference (last 2% of train)
Mean Episodic Reward-27.34
13
Cooperative Multi-Agent Reinforcement LearningDisperse (last 2% of train)
Mean Episodic Reward-1.16
13
Cooperative Multi-Agent Reinforcement LearningSpeaker-Listener (last 2% of train)
Mean Episodic Reward-21.19
13
Cooperative Multi-Agent Reinforcement LearningAdversary (last 2% of train)
Mean Episodic Reward40.35
13
Multi-Agent Reinforcement LearningMPE Adversary (test)
Final Test Return38.8
6
Showing 10 of 12 rows

Other info

Follow for update