When Sensors Lie: From Noise to Trustworthy Decisions in Cyber-Physical Systems

This seminar by Christiana Chamon Garcia presents an uncertainty-aware, physics-informed perspective on sensor trust: combining stochastic modeling, system constraints, multimodal fusion, and machine learning to determine not only whether a measurement is unusual, but what explanations are plausible, how confident the system should be, and whether the discrepancy can lead to a security- or safety-relevant outcome. The discussion will highlight open challenges in attribution, generalizable evaluation, hardware-rooted trust, and decision-making under uncertainty for resilient cyber-physical systems. Online via WebEx.