AI Paper Atlas

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

Interaction Pattern Disentangling for Multi-Agent Reinforcement Learning

Shunyu Liu, Jie Song, Yihe Zhou, Na Yu, Kaixuan Chen, Zunlei Feng, Mingli Song

TPAMI 2024 · Multi-agent systems and games › Cooperative multi-agent RL · approach: MARL benchmarks and grouping · 19 citations (OpenAlex)

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
add contrastive disagreement (CD) loss (ablation)OPT · SMAC and multi-task SMACperformanceImproved
add sparsemax to prototype generation (ablation)OPT · SMAC and multi-task SMACperformanceImproved
add conditional mutual information (CMI) loss (ablation)OPT · multi-task SMACstability and performanceImproved
increase interaction prototype number from 2 to 4 (ablation)OPT · SMACperformance gainImproved
increase prototype number beyond 4 (ablation)OPT · SMACperformance gainHurt
use OPT instead of QMIXQMIX · SMAC corridortest win rate: 0.20 -> 0.88Improved
use OPT instead of QMIXQMIX · SMAC 5m_vs_6mtest win rate: 0.57 -> 0.86Improved
use OPT instead of VDNVDN · SMAC 10m_vs_11mtest win rate: 0.94 -> 0.97No clear effect
train OPT for 20M steps instead of 2M (ablation)OPT · SMAC corridortest win rate: 0.88 -> 0.92Improved
use OPT instead of MAPPO (20M)MAPPO · SMAC corridortest win rate: 0.99 -> 0.88Hurt
use OPT instead of MAPPOMAPPO · GRF counterattack_hardtest win rate: 0.38 -> 0.46Improved
use OPT instead of Attn (QMIX)Attn (QMIX) · Multi-task SMAC 5-11sztest win rate: 0.01 -> 0.27Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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