A Variational Information Theoretic Approach to Out-of-Distribution Detection
This work addresses the challenge of detecting out-of-distribution data for neural network applications, offering a general and explainable framework, though it appears incremental as it builds on existing OOD feature methods.
The paper tackles the problem of out-of-distribution (OOD) detection in neural networks by proposing a novel information-theoretic loss functional that separates in-distribution and OOD features, resulting in a new shaping function that outperforms existing ones on OOD benchmarks.
We present a theory for the construction of out-of-distribution (OOD) detection features for neural networks. We introduce random features for OOD through a novel information-theoretic loss functional consisting of two terms, the first based on the KL divergence separates resulting in-distribution (ID) and OOD feature distributions and the second term is the Information Bottleneck, which favors compressed features that retain the OOD information. We formulate a variational procedure to optimize the loss and obtain OOD features. Based on assumptions on OOD distributions, one can recover properties of existing OOD features, i.e., shaping functions. Furthermore, we show that our theory can predict a new shaping function that out-performs existing ones on OOD benchmarks. Our theory provides a general framework for constructing a variety of new features with clear explainability.