{
  "annotation_id": "2c8f985b-36db-4420-a4bf-db7c4aa23212",
  "date_created": "2026-02-17T05:53:11.557000Z",
  "date_modified": "2026-02-17T05:53:11.557000Z",
  "file_hash": "62b2cbff265661648f910237e76dc340b783357f695e594ed051447c3fda3526",
  "private": false,
  "record": {
    "abstract": "Offline multi-agent reinforcement learning (MARL) aims to solve cooperative decision-making problems in multi-agent systems using pre-collected datasets. Existing offline MARL methods primarily constrain training within the dataset distribution, resulting in overly conservative policies that struggle to generalize beyond the support of the data. While model-based approaches offer a promising solution by expanding the original dataset with synthetic data generated from a learned world model, the high dimensionality, non-stationarity, and complexity of multi-agent systems make it challenging to accurately estimate the transitions and reward functions in offline MARL. Given the difficulty of directly modeling joint dynamics, we propose a local-to-global (LOGO) world model, a novel framework that leverages local predictions-which are easier to estimate-to infer global state dynamics, thus improving prediction accuracy while implicitly capturing agent-wise dependencies. Using the trained world model, we generate synthetic data to augment the original dataset, expanding the effective state-action space. To ensure reliable policy learning, we further introduce an uncertainty-aware sampling mechanism that adaptively weights synthetic data by prediction uncertainty, reducing approximation error propagation to policies. In contrast to conventional ensemble-based methods, our approach requires only an additional encoder for uncertainty estimation, significantly reducing computational overhead while maintaining accuracy. Extensive experiments across 8 scenarios against 8 baselines demonstrate that our method surpasses state-of-the-art baselines on standard offline MARL benchmarks, establishing a new model-based baseline for generalizable offline multi-agent learning.",
    "arxiv_id": "2601.07463",
    "authors": [
      "Sijia li",
      "Xinran Li",
      "Shibo Chen",
      "Jun Zhang"
    ],
    "categories": [
      "cs.AI",
      "cs.LG"
    ],
    "license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
    "title": "Puzzle it Out: Local-to-Global World Model for Offline Multi-Agent Reinforcement Learning",
    "url": "https://arxiv.org/abs/2601.07463",
    "version": "v1"
  },
  "schema_id": "dorsal/arxiv",
  "source": {
    "execution_id": "1edc3294-9925-43ee-8e58-c309ae2eecda",
    "id": "arXiv Dataset",
    "type": "Model",
    "variant": "snapshot-2026-01-17",
    "version": "0.1.0"
  },
  "user_id": 1000002
}