English

Agent2World: Learning to Generate Symbolic World Models via Adaptive Multi-Agent Feedback

Artificial Intelligence 2025-12-30 v1 Computation and Language

Abstract

Symbolic world models (e.g., PDDL domains or executable simulators) are central to model-based planning, but training LLMs to generate such world models is limited by the lack of large-scale verifiable supervision. Current approaches rely primarily on static validation methods that fail to catch behavior-level errors arising from interactive execution. In this paper, we propose Agent2World, a tool-augmented multi-agent framework that achieves strong inference-time world-model generation and also serves as a data engine for supervised fine-tuning, by grounding generation in multi-agent feedback. Agent2World follows a three-stage pipeline: (i) A Deep Researcher agent performs knowledge synthesis by web searching to address specification gaps; (ii) A Model Developer agent implements executable world models; And (iii) a specialized Testing Team conducts adaptive unit testing and simulation-based validation. Agent2World demonstrates superior inference-time performance across three benchmarks spanning both Planning Domain Definition Language (PDDL) and executable code representations, achieving consistent state-of-the-art results. Beyond inference, Testing Team serves as an interactive environment for the Model Developer, providing behavior-aware adaptive feedback that yields multi-turn training trajectories. The model fine-tuned on these trajectories substantially improves world-model generation, yielding an average relative gain of 30.95% over the same model before training. Project page: https://agent2world.github.io.

Keywords

Cite

@article{arxiv.2512.22336,
  title  = {Agent2World: Learning to Generate Symbolic World Models via Adaptive Multi-Agent Feedback},
  author = {Mengkang Hu and Bowei Xia and Yuran Wu and Ailing Yu and Yude Zou and Qiguang Chen and Shijian Wang and Jiarui Jin and Kexin Li and Wenxiang Jiao and Yuan Lu and Ping Luo},
  journal= {arXiv preprint arXiv:2512.22336},
  year   = {2025}
}

Comments

48 pages, 15 tables, 7 figures, Project page: https://agent2world.github.io

R2 v1 2026-07-01T08:42:08.178Z