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 Tools annual

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 Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The distribution of the weights of modern deep neural networks (DNNs) -crucial for uncertainty quantification and robustness - is an eminently complexobject due to its extremely high dimensionality. This paper proposes one of thefirst large-scale explorations of the posterior distribution of deep BayesianNeural Networks (BNNs), expanding its study to real-world vision tasks andarchitectures. Specifically, we investigate the optimal approach forapproximating the posterior, analyze the connection between posterior qualityand uncertainty quantification, delve into the impact of modes on theposterior, and explore methods for visualizing the posterior. Moreover, weuncover weight-space symmetries as a critical aspect for understanding theposterior. To this extent, we develop an in-depth assessment of the impact ofboth permutation and scaling symmetries that tend to obfuscate the Bayesianposterior. While the first type of transformation is known for duplicatingmodes, we explore the relationship between the latter and L2 regularization,challenging previous misconceptions. Finally, to help the community improve ourunderstanding of the Bayesian posterior, we will shortly release the firstlarge-scale checkpoint dataset, including thousands of real-world models andour codes.

A Symmetry-Aware Exploration of Bayesian Neural Network Posteriors