English

PAVXploreRL: Physical-Action-Visual World Model Reinforcement Learning with Action Exploration

Computer Vision and Pattern Recognition 2026-07-18 v1

Abstract

Action-conditioned world models are a key component of embodied AI, serving as scalable policy evaluators that reduce reliance on expensive real-world rollouts. To accurately capture diverse action-induced dynamics, such models should satisfy three key objectives-Physical Plausibility (P), Action Adherence (A), and Visual Fidelity (V), collectively referred to as PAV-while remaining robust to both in-distribution (ID) expert demonstrations and out-of-distribution (OOD) actions. However, existing methods primarily rely on ID action-video pairs and pixel-level reconstruction losses, which do not explicitly optimize PAV objectives and generalize poorly beyond expert data. To address this, we propose PAVXploreRL, a reinforcement learning framework built on a pretrained latent world model that explicitly optimizes PAV objectives through reward-driven training. To improve action generalization, our method jointly leverages ID trajectories and noise-driven OOD action exploration, without paired video supervision. Experiments show that PAVXploreRL consistently outperforms pretrained baselines, achieving a 5.6% average gain across benchmarks and producing higher-quality PAV properties. As a policy evaluator, it also yields more reliable performance estimates and reduces the overestimation bias of prior expert-only world models such as Ctrl-World. Code: https://github.com/Social-AI-Studio/PAVXploreRL

Cite

@article{arxiv.2607.16602,
  title  = {PAVXploreRL: Physical-Action-Visual World Model Reinforcement Learning with Action Exploration},
  author = {Han Wang and Zijun Wang and Shuoshuo Xue and Rui Cao and Fenjiao Cheng and Xiaodang Liang and Roy Ka-Wei Lee},
  journal= {arXiv preprint arXiv:2607.16602},
  year   = {2026}
}

Comments

9 pages, 5 figures