Showing 1–1 of 1 results for author: Pantalos, G

A Bayesian Approach to Invariant Deep Neural Networks

Authors: Nikolaos Mourdoukoutas, Marco Federici, Georges Pantalos, Mark van der Wilk, Vincent Fortuin

Abstract: We propose a novel Bayesian neural network architecture that can learn invariances from data alone by inferring a posterior distribution over different weight-sharing schemes. We show that our model outperforms other non-invariant architectures, when trained on datasets that contain specific invariances. The same holds true when no data augmentation is performed. We propose a novel Bayesian neural network architecture that can learn invariances from data alone by inferring a posterior distribution over different weight-sharing schemes. We show that our model outperforms other non-invariant architectures, when trained on datasets that contain specific invariances. The same holds true when no data augmentation is performed. △ Less

Submitted 2 November, 2021; v1 submitted 20 July, 2021; originally announced July 2021.

Comments: 8 pages, 3 figures, To be published in ICML UDL 2021