Detecting Out-of-Distribution Inputs to Deep Generative Models Using Typicality

Eric Nalisnick, Akihiro Matsukawa, Yee Whye Teh, Balaji Lakshminarayanan

Keywords: Deep generative models, out-of-distribution detection, safety

TL;DR: We propose detecting out-of-distribution inputs to deep generative models via a goodness-of-fit test based on the model entropy.

Abstract: Recent work has shown that deep generative models can assign higher likelihood to out-of-distribution data sets than to their training data [Nalisnick et al., 2019; Choi et al., 2019]. We posit that this phenomenon is caused by a mismatch between the model's typical set and its areas of high probability density. In-distribution inputs should reside in the former but not necessarily in the latter, as previous work has presumed [Bishop, 1994]. To determine whether or not inputs reside in the typical set, we propose a statistically principled, easy-to-implement test using the empirical distribution of model likelihoods. The test is model agnostic and widely applicable, only requiring that the likelihood can be computed or closely approximated. We report experiments showing that our procedure can successfully detect the out-of-distribution sets in several of the challenging cases reported by Nalisnick et al. [2019].