V-OCBF: Learning Safety Filters from Offline Data via Value-Guided Offline Control Barrier Functions

This paper introduces Value-Guided Offline Control Barrier Functions (V-OCBF), a framework for learning neural Control Barrier Functions (CBFs) entirely from offline demonstrations. It provides rigorous safety guarantees for autonomous systems without relying on online interaction or requiring full knowledge of system dynamics, making it crucial for the safe and scalable deployment of AI in robotics and control.