[2512.08991] Deterministic World Models for Verification of Closed-loop Vision-based Systems

Computer Science > Computer Vision and Pattern Recognition arXiv:2512.08991 (cs) [Submitted on 8 Dec 2025 (v1), last revised 1 Apr 2026 (this version, v2)] Title:Deterministic World Models for Verification of Closed-loop Vision-based Systems Authors:Yuang Geng, Zhuoyang Zhou, Zhongzheng Zhang, Siyuan Pan, Hoang-Dung Tran, Ivan Ruchkin View a PDF of the paper titled Deterministic World Models for Verification of Closed-loop Vision-based Systems, by Yuang Geng and 4 other authors View PDF HTML (experimental) Abstract:Verifying closed-loop vision-based control systems remains a fundamental challenge due to the high dimensionality of images and the difficulty of modeling visual environments. While generative models are increasingly used as camera surrogates in verification, their reliance on stochastic latent variables introduces unnecessary overapproximation error. To address this bottleneck, we propose a Deterministic World Model (DWM) that maps system states directly to generative images, effectively eliminating uninterpretable latent variables to ensure precise input bounds. The DWM is trained with a dual-objective loss function that combines pixel-level reconstruction accuracy with a control difference loss to maintain behavioral consistency with the real system. We integrate DWM into a verification pipeline utilizing Star-based reachability analysis (StarV) and employ conformal prediction to derive rigorous statistical bounds on the trajectory deviation between the world...