Certifiable Deep Importance Sampling for Rare-Event Simulation of Black-Box Systems
This work addresses the challenge of certifiable rare-event simulation for safety-critical AI-driven physical systems, such as autonomous vehicles, offering a method to prevent dangerous underestimation in black-box scenarios.
The authors tackled the problem of rare-event simulation for black-box systems, where traditional importance sampling methods lack efficiency guarantees, by proposing Deep-Prabilistic Accelerated Evaluation (Deep-PrAE), a framework that converts versatile black-box samplers into ones with relaxed efficiency certificates, enabling accurate estimation of rare-event probability bounds and demonstrating effectiveness in safety-testing intelligent driving algorithms.
Rare-event simulation techniques, such as importance sampling (IS), constitute powerful tools to speed up challenging estimation of rare catastrophic events. These techniques often leverage the knowledge and analysis on underlying system structures to endow desirable efficiency guarantees. However, black-box problems, especially those arising from recent safety-critical applications of AI-driven physical systems, can fundamentally undermine their efficiency guarantees and lead to dangerous under-estimation without diagnostically detected. We propose a framework called Deep Probabilistic Accelerated Evaluation (Deep-PrAE) to design statistically guaranteed IS, by converting black-box samplers that are versatile but could lack guarantees, into one with what we call a relaxed efficiency certificate that allows accurate estimation of bounds on the rare-event probability. We present the theory of Deep-PrAE that combines the dominating point concept with rare-event set learning via deep neural network classifiers, and demonstrate its effectiveness in numerical examples including the safety-testing of intelligent driving algorithms.