Imposing Functional Priors on Bayesian Neural Networks

Bogdan Kozyrskiy, Dimitrios Milios, Maurizio Filippone

2023

Abstract

Specifying sensible priors for Bayesian neural networks (BNNs) is key to obtain state-of-the-art predictive performance while obtaining sound predictive uncertainties. However, this is generally difficult because of the complex way prior distributions induce distributions over the functions that BNNs can represent. Switching the focus from the prior over the weights to such functional priors allows for the reasoning on what meaningful prior information should be incorporated. We propose to enforce such meaningful functional priors through Gaussian processes (GPs), which we view as a form of implicit prior over the weights, and we employ scalable Markov chain Monte Carlo (MCMC) to obtain samples from an approximation to the posterior distribution over BNN weights. Unlike previous approaches, our proposal does not require the modification of the original BNN model, it does not require any expensive preliminary optimization, and it can use any inference techniques and any functional prior that can be expressed in closed form. We illustrate the effectiveness of our approach with an extensive experimental campaign.

Paper Citation

in Harvard Style

Kozyrskiy B., Milios D. and Filippone M. (2023). Imposing Functional Priors on Bayesian Neural Networks. In Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-626-2, pages 450-457. DOI: 10.5220/0011742900003411

in Bibtex Style

@conference{icpram23,
author={Bogdan Kozyrskiy and Dimitrios Milios and Maurizio Filippone},
title={Imposing Functional Priors on Bayesian Neural Networks},
booktitle={Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2023},
pages={450-457},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011742900003411},
isbn={978-989-758-626-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 12th International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - Imposing Functional Priors on Bayesian Neural Networks
SN - 978-989-758-626-2
AU - Kozyrskiy B.
AU - Milios D.
AU - Filippone M.
PY - 2023
SP - 450
EP - 457
DO - 10.5220/0011742900003411