Are Bayesian neural networks intrinsically good at out-of-distribution detection?
arXiv:2107.12248
Abstract
The need to avoid confident predictions on unfamiliar data has sparked interest in out-of-distribution (OOD) detection. It is widely assumed that Bayesian neural networks (BNN) are well suited for this task, as the endowed epistemic uncertainty should lead to disagreement in predictions on outliers. In this paper, we question this assumption and provide empirical evidence that proper Bayesian inference with common neural network architectures does not necessarily lead to good OOD detection. To circumvent the use of approximate inference, we start by studying the infinite-width case, where Bayesian inference can be exact considering the corresponding Gaussian process. Strikingly, the kernels induced under common architectural choices lead to uncertainties that do not reflect the underlying data generating process and are therefore unsuited for OOD detection. Finally, we study finite-width networks using HMC, and observe OOD behavior that is consistent with the infinite-width case. Overall, our study discloses fundamental problems when naively using BNNs for OOD detection and opens interesting avenues for future research.
arXiv commentPublished at UDL Workshop, ICML 2021
References in corpus (8)
- Can You Trust Your Model's Uncertainty? Evaluating Predictive Uncertainty Under Dataset Shift
- Normalizing Flows for Probabilistic Modeling and Inference
- Multiplicative Normalizing Flows for Variational Bayesian Neural Networks
- A Simple Baseline for Bayesian Uncertainty in Deep Learning
- Implicit Weight Uncertainty in Neural Networks
- Detecting Out-of-Distribution Inputs to Deep Generative Models Using Typicality
- Bayesian Hypernetworks
- Posterior Meta-Replay for Continual Learning