Research map / Multi-agent systems and games / Cooperative multi-agent RL
Learning Disentangled Multi-Agent World Model for Decentralized Control
Di Xue, Jing Jiang, Shaowei Zhang, Wenhao Guo, Lei Yuan, Zongzhang Zhang, Yang Yu
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add transformer-based centralized critic (vs MLP critic, DMAWM-TC) (ablation) | DMAWM · SMACv2 protoss 5 vs 5 | win rate | Improved |
| add transformer-based interaction predictor (vs per-agent MLP, DMAWM-IP) (ablation) | DMAWM · SMACv2 protoss 5 vs 5 | win rate | Improved |
| add off-policy critic loss grounded on real trajectories (vs beta_off = 0, DMAWM-OC) (ablation) | DMAWM · SMACv2 protoss 5 vs 5 | win rate | Improved |
| lengthen the imagination horizon (4 to 20) (ablation) | DMAWM · SMACv2 (3 maps) | win rate | No clear effect |
| align joint prior with factorized posterior to disentangle agent latents (vs communication-based MAMBA) | DMAWM · SMACv2 protoss 5 vs 5 | unobservable-teammate action prediction accuracy: 85.0% -> 42.3% (lower means less leakage) | Improved |
| use disentangled latent world model instead of MAMBA | DMAWM · SMAC corridor | win rate: 18.6 -> 71.5 | Improved |
| use disentangled latent world model instead of MAPPO on a trivial map | DMAWM · SMAC 2s vs 1sc | win rate: 100.0 -> 98.3 | No clear effect |
| use DMAWM instead of MAPPO trained with 10x more environment steps | DMAWM · SMACv2 | win rate | Improved |
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.
- QMIX — QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning (ICML 2018) · better on 1 of 1 shared result
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
- Simplifying Latent Dynamics with Softly State-Invariant World Models (NeurIPS 2024 · similarity 0.88)
- Revisiting Multi-Agent World Modeling from a Diffusion-Inspired Perspective (NeurIPS 2025 · similarity 0.88)
- COMBO: Compositional World Models for Embodied Multi-Agent Cooperation (ICLR 2025 · similarity 0.88)
- Agent2World: Learning to Generate Symbolic World Models via Adaptive Multi-Agent Feedback (NeurIPS 2026 · similarity 0.86)
- Acting Beyond Learning: Imagination-Assisted Decision-Making in the Visual-based Multi-Agent Cooperative Scenarios (AAAI 2025 · similarity 0.86)
- Latent Collaboration in Multi-Agent Systems (ICML 2026 · similarity 0.86)
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.
