# Joint Policy Search for Multi-agent Collaboration with Imperfect Information

arXiv: https://arxiv.org/abs/2008.06495 · HTML: https://mlsanity.com/paper/2008.06495/

## Key facts

- Cited by 13 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 14 Aug 2020.

## Built on by (most cited first, 13 of 13)

- [Pgx: Hardware-Accelerated Parallel Game Simulators for Reinforcement Learning](https://mlsanity.com/paper/2303.17503/) (2023)
- [Near-Optimal Learning of Extensive-Form Games with Imperfect Information](https://mlsanity.com/paper/2202.01752/) (2022)
- [The Power of Regularization in Solving Extensive-Form Games](https://mlsanity.com/paper/2206.09495/) (2022)
- [Position: Towards Bidirectional Human-AI Alignment](https://mlsanity.com/paper/2406.09264/) (2024)
- [Efficient Phi-Regret Minimization in Extensive-Form Games via Online Mirror Descent](https://mlsanity.com/paper/2205.15294/) (2022)
- [Learned Belief Search: Efficiently Improving Policies in Partially Observable Settings](https://mlsanity.com/paper/2106.09086/) (2021)
- [Sample-Efficient Learning of Correlated Equilibria in Extensive-Form Games](https://mlsanity.com/paper/2205.07223/) (2022)
- [A Survey of Decision Making in Adversarial Games](https://mlsanity.com/paper/2207.07971/) (2022)
- [The Update-Equivalence Framework for Decision-Time Planning](https://mlsanity.com/paper/2304.13138/) (2023)
- [Credit Assignment with Meta-Policy Gradient for Multi-Agent Reinforcement Learning](https://mlsanity.com/paper/2102.12957/) (2021)
- [BridgeHand2Vec Bridge Hand Representation](https://mlsanity.com/paper/2310.06624/) (2023)
- [A Simple, Solid, and Reproducible Baseline for Bridge Bidding AI](https://mlsanity.com/paper/2406.10306/) (2024)
- [On-line Learning in Tree MDPs by Treating Policies as Bandit Arms](https://mlsanity.com/paper/2605.04979/) (2026)

## Abstract (arXiv)

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.