Certifiably Adversarially Robust Detection of Out-of-Distribution Data
This addresses a critical safety issue in applications like autonomous systems by enabling reliable uncertainty assessment, though it is incremental as it builds on existing certified robustness techniques.
The paper tackles the problem of deep neural networks being overconfident on out-of-distribution (OOD) inputs by proposing a method to provide certifiable worst-case guarantees for OOD detection, achieving this without significant loss in accuracy.
Deep neural networks are known to be overconfident when applied to out-of-distribution (OOD) inputs which clearly do not belong to any class. This is a problem in safety-critical applications since a reliable assessment of the uncertainty of a classifier is a key property, allowing the system to trigger human intervention or to transfer into a safe state. In this paper, we aim for certifiable worst case guarantees for OOD detection by enforcing not only low confidence at the OOD point but also in an $l_\infty$-ball around it. For this purpose, we use interval bound propagation (IBP) to upper bound the maximal confidence in the $l_\infty$-ball and minimize this upper bound during training time. We show that non-trivial bounds on the confidence for OOD data generalizing beyond the OOD dataset seen at training time are possible. Moreover, in contrast to certified adversarial robustness which typically comes with significant loss in prediction performance, certified guarantees for worst case OOD detection are possible without much loss in accuracy.