Digital agents require diverse, large-scale UI trajectories to generalize across real-world tasks, yet collecting such data is prohibitively expensive in both human annotation, infra and engineering perspectives. To this end, we introduce UI-Simulator, a scalable paradigm that generates structured UI states and transitions to synthesize training trajectories at scale. Our paradigm integrates a digital world simulator for diverse UI states, a guided rollout process for coherent exploration, and a trajectory wrapper that produces high-quality and diverse trajectories for agent training. We further propose UI-Simulator-Grow, a targeted scaling strategy that enables more rapid and data-efficient scaling by prioritizing high-impact tasks and synthesizes informative trajectory variants. Experiments on WebArena and AndroidWorld show that UI-Simulator rivals or surpasses open-source agents trained on real UIs with significantly better robustness, despite using weaker teacher models. Moreover, UI-Simulator-Grow matches the performance of Llama-3-70B-Instruct using only Llama-3-8B-Instruct as the base model, highlighting the potential of targeted synthesis scaling paradigm to continuously and efficiently enhance the digital agents.
@article{arxiv.2510.14969,
title = {LLMs as Scalable, General-Purpose Simulators For Evolving Digital Agent Training},
author = {Yiming Wang and Da Yin and Yuedong Cui and Ruichen Zheng and Zhiqian Li and Zongyu Lin and Di Wu and Xueqing Wu and Chenchen Ye and Yu Zhou and Kai-Wei Chang},
journal= {arXiv preprint arXiv:2510.14969},
year = {2025}
}
Comments
Preprint. Project page: https://ui-simulator.notion.site/llms-as-scalable-digital-world-simulator; Code and data: https://github.com/WadeYin9712/UI-Simulator