Joint Policy Search for Multi-agent Collaboration with Imperfect Information

Abstract

To learn good joint policies for multi-agent collaboration with imperfect information remains a fundamental challenge. While for two-player zero-sum games, coordinate-ascent approaches (optimizing one agent's policy at a time, e.g., self-play) work with guarantees, in multi-agent cooperative setting they often converge to sub-optimal Nash equilibrium. On the other hand, directly modeling joint policy changes in imperfect information game is nontrivial due to complicated interplay of policies (e.g., upstream updates affect downstream state reachability). In this paper, we show global changes of game values can be decomposed to policy changes localized at each information set, with a novel term named policy-change density. Based on this, we propose Joint Policy Search(JPS) that iteratively improves joint policies of collaborative agents in imperfect information games, without re-evaluating the entire game. On multi-agent collaborative tabular games, JPS is proven to never worsen performance and can improve solutions provided by unilateral approaches (e.g, CFR), outperforming algorithms designed for collaborative policy learning (e.g. BAD). Furthermore, for real-world games, JPS has an online form that naturally links with gradient updates. We test it to Contract Bridge, a 4-player imperfect-information game where a team of $2$ collaborates to compete against the other. In its bidding phase, players bid in turn to find a good contract through a limited information channel. Based on a strong baseline agent that bids competitive bridge purely through domain-agnostic self-play, JPS improves collaboration of team players and outperforms WBridge5, a championship-winning software, by $+0.63$ IMPs (International Matching Points) per board over 1k games, substantially better than previous SoTA ($+0.41$ IMPs/b) under Double-Dummy evaluation.

arXiv commentMinor fix of the algorithm block

References in corpus (12)

• Playing Atari with Deep Reinforcement Learning
• Learning to Communicate with Deep Multi-Agent Reinforcement Learning
• DeepStack: Expert-Level Artificial Intelligence in No-Limit Poker
• Emergent Tool Use From Multi-Agent Autocurricula
• Anytime Point-Based Approximations for Large POMDPs
• The Hanabi Challenge: A New Frontier for AI Research
• Deep Reinforcement Learning from Self-Play in Imperfect-Information Games
• Safe and Nested Subgame Solving for Imperfect-Information Games
• Suphx: Mastering Mahjong with Deep Reinforcement Learning
• Combining Deep Reinforcement Learning and Search for Imperfect-Information Games
• ELF OpenGo: An Analysis and Open Reimplementation of AlphaZero
• Finding Friend and Foe in Multi-Agent Games

Cited by in corpus (3)

• Credit Assignment with Meta-Policy Gradient for Multi-Agent Reinforcement Learning
• Learned Belief Search: Efficiently Improving Policies in Partially Observable Settings
• BridgeHand2Vec Bridge Hand Representation