English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

The need to avoid confident predictions on unfamiliar data has sparkedinterest in out-of-distribution (OOD) detection. It is widely assumed thatBayesian neural networks (BNN) are well suited for this task, as the endowedepistemic uncertainty should lead to disagreement in predictions on outliers.In this paper, we question this assumption and provide empirical evidence thatproper Bayesian inference with common neural network architectures does notnecessarily lead to good OOD detection. To circumvent the use of approximateinference, we start by studying the infinite-width case, where Bayesianinference can be exact considering the corresponding Gaussian process.Strikingly, the kernels induced under common architectural choices lead touncertainties that do not reflect the underlying data generating process andare therefore unsuited for OOD detection. Finally, we study finite-widthnetworks using HMC, and observe OOD behavior that is consistent with theinfinite-width case. Overall, our study discloses fundamental problems whennaively using BNNs for OOD detection and opens interesting avenues for futureresearch.

Are Bayesian neural networks intrinsically good at out-of-distribution detection?