Online Safety Verification

Online safety verification focuses on ensuring the safe operation of autonomous systems and digital technologies, primarily by proactively identifying and mitigating potential risks. Current research emphasizes developing efficient algorithms, such as those leveraging neural networks and Hamilton-Jacobi reachability analysis, to verify safety properties in real-time, even with dynamic environments and limited information. This work is crucial for deploying trustworthy AI systems in safety-critical applications like autonomous driving and robotics, as well as for enhancing the security and privacy of online services and data. The development of robust and computationally efficient verification methods is a key challenge driving ongoing research.