⌁ Awesome World Model Hub Papers · Datasets · Projects
中文 GitHub
← Back to papers

GAWM: Global-Aware World Model for Multi-Agent Reinforcement Learning

arXiv 25.1 2025 59.6 method

TLDR

GAWM enhances multi-agent world models with a Transformer to fuse local observations, improving global state representation and convergence.

Reasoning

The paper clearly identifies a key limitation in model-based MARL (global consistency) and proposes a novel Transformer-based fusion mechanism. Strengths include addressing a practical issue and showing improved stability and convergence. Weaknesses: no explicit real-world experiments or benchmarks mentioned; the abstract is cut off, but likely only simulated environments.

Read-first score

Read-first score 59.6, weighted from topical fit, citation, graph, method, reproducibility, and recency signals. Original total remains 46.

Recency 8%
86.7

Uses a gentle age decay so recent papers surface without erasing older foundations. 2025

Methodology quality 25%
70

Screens visible abstract and analysis fields for experiment, dataset, baseline, metric, and limitation evidence. markers=benchmark,experiment,result

Topical relevance 42%
65.7

Uses existing LLM keyword relevance scores normalized to 0-100. world model,world simulator,generative world model,interactive world model,video world model,world dynamics prediction,model-based reinforcement learning world model

Reproducibility 25%
30

Screens links and visible text for paper, code, dataset, artifact, and repository signals. pdf=True; code=False; dataset=False; markers=none

Field roles

FrontierMethodology anchor

Rank sensitivity

Stability: volatile; rank range: 267.

Keyword Scores

world model
10
model-based reinforcement learning world model
10
world dynamics prediction
8
generative world model
7
world simulator
6
interactive world model
5
video world model
0

Deep Analysis

Innovations

  • Global-aware world model for multi-agent reinforcement learning that enhances centralized world model's ability to achieve globally unified and accurate state representation under CTDE paradigm
  • Transformer architecture to fuse local observation information from different agents for improved global state extraction and representation

Methodology

GAWM is a model-based MARL method that uses a centralized world model with an additional Transformer architecture to fuse local observations from multiple agents, enabling globally unified and accurate state representation while adhering to the CTDE paradigm. The world model is used for data sample augmentation to improve sample efficiency, and the method is evaluated on the SMAC benchmark.

Key Results

GAWM outperforms various model-free and model-based approaches, achieving exceptional performance in the challenging domains of SMAC.

Resources

Paper

Tags