2022/09/20 by Craig Innes, Subramanian Ramamoorthy, Innes, Craig +1 · 3 citations
Computer Science · Engineering · #Adversarial Robustness in Machine Learning #Control Systems and Identification #FOS: Computer and information sciences #Machine Learning (cs.LG) #Robotics (cs.RO) #Target Tracking and Data Fusion in Sensor Networks
paper · pdf · doi:10.48550/arxiv.2209.09674
openalex publication_date 2022/09/20 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
Testing black-box perceptual-control systems in simulation faces two difficulties. Firstly, perceptual inputs in simulation lack the fidelity of real-world sensor inputs. Secondly, for a reasonably accurate perception system, encountering a rare failure trajectory may require running infeasibly many simulations. This paper combines perception error models -- surrogates for a sensor-based detection system -- with state-dependent adaptive importance sampling. This allows us to efficiently assess the rare failure probabilities for real-world perceptual control systems within simulation. Our experiments with an autonomous braking system equipped with an RGB obstacle-detector show that our method can calculate accurate failure probabilities with an inexpensive number of simulations. Further, we show how choice of safety metric can influence the process of learning proposal distributions capable of reliably sampling high-probability failures.