Encoding the latent posterior of Bayesian Neural Networks for uncertainty quantification - Archive ouverte HAL Access content directly
Conference Papers Year :

Encoding the latent posterior of Bayesian Neural Networks for uncertainty quantification

Abstract

Bayesian Neural Networks (BNNs) have been long considered an ideal, yet unscalable solution for improving the robustness and the predictive uncertainty of deep neural networks. While they could capture more accurately the posterior distribution of the network parameters, most BNN approaches are either limited to small networks or rely on constraining assumptions, e.g., parameter independence. These drawbacks have enabled prominence of simple, but computationally heavy approaches such as Deep Ensembles, whose training and testing costs increase linearly with the number of networks. In this work we aim for efficient deep BNNs amenable to complex computer vision architectures, e.g., ResNet50 DeepLabV3+, and tasks, e.g., semantic segmentation, with fewer assumptions on the parameters. We achieve this by leveraging variational autoencoders (VAEs) to learn the interaction and the latent distribution of the parameters at each network layer. Our approach, Latent-Posterior BNN (LP-BNN), is compatible with the recent BatchEnsemble method, leading to highly efficient (in terms of computation and memory during both training and testing) ensembles. LP-BNNs attain competitive results across multiple metrics in several challenging benchmarks for image classification, semantic segmentation and out-of-distribution detection.
Fichier principal
Vignette du fichier
LPBNN_Proceedings.pdf (3.32 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-03097035 , version 1 (05-01-2021)
hal-03097035 , version 2 (25-03-2021)

Licence

Attribution - CC BY 4.0

Identifiers

  • HAL Id : hal-03097035 , version 2

Cite

Gianni Franchi, Andrei Bursuc, Emanuel Aldea, Séverine Dubuisson, Isabelle Bloch. Encoding the latent posterior of Bayesian Neural Networks for uncertainty quantification. NeurIPS workshop on Bayesian Deep Learning, Dec 2020, Vancouver, Canada. ⟨hal-03097035v2⟩
161 View
123 Download

Share

Gmail Facebook Twitter LinkedIn More