Out-of-Distribution Detection with Class Ratio Estimation
This work addresses the problem of detecting OOD images for machine learning practitioners, offering a simpler and effective approach, though it appears incremental as it builds on existing density ratio methods.
The paper tackles the unreliability of density-based out-of-distribution (OOD) detection by proposing a framework that unifies density ratio methods and directly estimates density ratios through class ratio estimation, achieving competitive results on OOD image problems without requiring deep generative models.
Density-based Out-of-distribution (OOD) detection has recently been shown unreliable for the task of detecting OOD images. Various density ratio based approaches achieve good empirical performance, however methods typically lack a principled probabilistic modelling explanation. In this work, we propose to unify density ratio based methods under a novel framework that builds energy-based models and employs differing base distributions. Under our framework, the density ratio can be viewed as the unnormalized density of an implicit semantic distribution. Further, we propose to directly estimate the density ratio of a data sample through class ratio estimation. We report competitive results on OOD image problems in comparison with recent work that alternatively requires training of deep generative models for the task. Our approach enables a simple and yet effective path towards solving the OOD detection problem.