Entropic Issues in Likelihood-Based OOD DetectionDownload PDF

27 Sept 2021, 17:50 (modified: 10 Dec 2021, 13:41)DGMs and Applications @ NeurIPS 2021 PosterReaders: Everyone
Keywords: OOD detection, anomaly detection, deep generative models
TL;DR: We show that the entropy of the data can cause issues in likelihood-based out-of-distribution (OOD) detection
Abstract: Deep generative models trained by maximum likelihood remain very popular methods for reasoning about data probabilistically. However, it has been observed that they can assign higher likelihoods to out-of-distribution (OOD) data than in-distribution data, thus calling into question the meaning of these likelihood values. In this work we provide a novel perspective on this phenomenon, decomposing the average likelihood into a KL divergence term and an entropy term. We argue that the latter can explain the curious OOD behaviour mentioned above, suppressing likelihood values on datasets with higher entropy. Although our idea is simple, we have not seen it explored yet in the literature. This analysis provides further explanation for the success of OOD detection methods based on likelihood ratios, as the problematic entropy term cancels out in expectation. Finally, we discuss how this observation relates to recent success in OOD detection with manifold-supported models, for which the above decomposition does not hold.
1 Reply

Loading