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

TL;DR

Agent2World introduces a multi-agent framework that generates and refines symbolic world models through interactive feedback, significantly improving inference accuracy and training effectiveness for model-based planning tasks.

Contribution

It presents a novel multi-agent pipeline that combines knowledge synthesis, model implementation, and adaptive testing to enhance symbolic world model generation and training.

Findings

Achieves state-of-the-art inference performance on PDDL and executable code benchmarks.

Provides behavior-aware adaptive feedback for multi-turn training trajectories.

Improves world-model generation accuracy by an average of 30.95% after fine-tuning.

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.