AI Agents in the Physical World: Balancing Safety & Performance

Balancing safety and performance is crucial to deploying autonomous agents in physical environments. In particular, autonomous racing is a domain that penalizes safe but conservative policies, highlighting the need for robust, adaptive strategies. Current approaches either make simplifying assumptions about other agents or lack robust mechanisms for online adaptation. In this talk we explore the following research themes on learning-based perception, planning and control at the limits of performance.