Adversarial Robustness Certification for Bayesian Neural Networks
Open Access
- 11 September 2024
- book chapter
- Published by Springer Nature
- p. 3-28
Abstract
We study the problem of certifying the robustness of Bayesian neural networks (BNNs) to adversarial input perturbations. Specifically, we define two notions of robustness for BNNs in an adversarial setting: probabilistic robustness and decision robustness. The former deals with the probabilistic behaviour of the network, that is, it ensures robustness across different stochastic realisations of the network, while the latter provides guarantees for the overall (output) decision of the BNN. Although these robustness properties cannot be computed analytically, we present a unified computational framework for efficiently and formally bounding them. Our approach is based on weight interval sampling, integration and bound propagation techniques, and can be applied to BNNs with a large number of parameters independently of the (approximate) inference method employed to train the BNN. We evaluate the effectiveness of our method on tasks including airborne collision avoidance, medical imaging and autonomous driving, demonstrating that it can compute non-trivial guarantees on medium size images (i.e., over 16 thousand input parameters).Keywords
All Related Versions
This publication has 25 references indexed in Scilit:
- Probabilistic reach-avoid for Bayesian neural networksArtificial Intelligence, 2024
- Provable Preimage Under-Approximation for Neural Networks2024
- Wild patterns: Ten years after the rise of adversarial machine learningPattern Recognition, 2018
- Feature-Guided Black-Box Safety Testing of Deep Neural Networks2018
- Reluplex: An Efficient SMT Solver for Verifying Deep Neural Networks2017
- Safety Verification of Deep Neural Networks2017
- Man vs. computer: Benchmarking machine learning algorithms for traffic sign recognitionNeural Networks, 2012
- Efficient Implementation of Interval Matrix MultiplicationLecture Notes in Computer Science, 2012
- Bayesian Learning for Neural Networks1996
- Statistical Decision Theory and Bayesian Analysis1985