LCSR Seminar Series: Jaime Fernández Fisac

Jaime Fernández Fisac, an assistant professor of Electrical and Computer Engineering at Princeton University, presents "Neural Safety Filters: Scaling Safety-Critical Control Without Surrendering Guarantees". The expanding reach of robotics and artificial intelligence brings acute reliability challenges in the face of limited information, rapidly changing conditions, and close human interaction. Recent years have seen a surge of interest in safety filters, runtime schemes that monitor a system's candidate actions and intervene to preempt catastrophic failures. Yet the restrictive assumptions of model-based approaches and the weakened guarantees of data-driven alternatives have reinforced the perception of a fundamental safety-performance tradeoff. This talk presents theor etical and experimental evidence against the view that strong safety guarantees must come at the expense of performance. Building on a unified understanding of safety filter mechanics, it introduces a general class of neural safety filters that extend the rigorous guarantees of safety-critical control to regimes beyond the reach of classical synthesis. For more about this event, visit the LCSR website.