AI Paper Atlas

Research map / Multi-agent systems and games / Cooperative multi-agent RL

Correlated Policy Optimization in Multi-Agent Subteams

Dingyang Chen, Jianing Ye, Zhenyu Zhang, Xiaolong Kuang, Xinyang Shen, Ozalp Ozer, Chongjie Zhang, Qi Zhang

ICLR 2026 · Multi-agent systems and games › Cooperative multi-agent RL · approach: Coordination graphs

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use fewer, larger subteams (full DAG instead of product DAG) (ablation)tabular softmax BN policy gradient · Coordination Game N=5normalized suboptimalityImproved
use full DAG instead of product DAG (ablation)tabular softmax BN policy gradient · Coordination Game N=3normalized suboptimalityImproved
use 1+2 subteams instead of product (ablation)tabular softmax BN policy gradient · Coordination Game N=3final performance: 2.94e-01 -> 3.96e-01No clear effect
use full DAG instead of product DAG (fitted transition error) (ablation)BN policy · Coordination Game N=5decomposition error |eP|: 6.03e-01 -> 7.13e-08Improved
use dynamic heuristic DAG instead of random DAG under same edge budget (ablation)MAPPO/MADDPG · Coordination Game, Aloha, Predator-PreyrewardImproved
use heuristic DAG instead of full DAG (ablation)MADDPG · Predator-Prey N=15rewardImproved
use heuristic subgroups instead of meta-learned subgroupsVAST · Warehouse N=16, Battle N=40rewardImproved
use full correlated DAG instead of product DAG (ablation)MAPPO · SMAC bane vs. banewin rateNo clear effect

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy