Designing Latent Safety Filters using Pre-Trained Vision Models
This work addresses safety concerns for vision-based control systems in critical settings like robotics, though it appears incremental as it builds on existing safety filter and PVR methods.
The paper tackles the challenge of ensuring safety in vision-based control systems by investigating the effectiveness of pre-trained vision models (PVRs) for designing safety filters, finding that they can serve as effective backbones for classifiers, Hamilton-Jacobi reachability-based filters, and latent world models with trade-offs in training approaches.
Ensuring safety of vision-based control systems remains a major challenge hindering their deployment in critical settings. Safety filters have gained increased interest as effective tools for ensuring the safety of classical control systems, but their applications in vision-based control settings have so far been limited. Pre-trained vision models (PVRs) have been shown to be effective perception backbones for control in various robotics domains. In this paper, we are interested in examining their effectiveness when used for designing vision-based safety filters. We use them as backbones for classifiers defining failure sets, for Hamilton-Jacobi (HJ) reachability-based safety filters, and for latent world models. We discuss the trade-offs between training from scratch, fine-tuning, and freezing the PVRs when training the models they are backbones for. We also evaluate whether one of the PVRs is superior across all tasks, evaluate whether learned world models or Q-functions are better for switching decisions to safe policies, and discuss practical considerations for deploying these PVRs on resource-constrained devices.