Publication Type
Conference Proceeding Article
Version
acceptedVersion
Publication Date
5-2024
Abstract
Adversarial examples are manipulated samples used to deceive machine learning models, posing a serious threat in safety-critical applications. Existing safety certificates for machine learning models are limited to individual input examples, failing to capture generalization to unseen data. To address this limitation, we propose novel generalization bounds based on the PAC-Bayesian and randomized smoothing frameworks, providing certificates that predict the model’s performance and robustness on unseen test samples based solely on the training data. We present an effective procedure to train and compute the first non-vacuous generalization bounds for neural networks in adversarial settings. Experimental results on the widely recognized MNIST and CIFAR-10 datasets demonstrate the efficacy of our approach, yielding the first robust risk certificates for stochastic convolutional neural networks under the $L_2$ threat model. Our method offers valuable tools for evaluating model susceptibility to real-world adversarial risks.
Keywords
Bayesian, Generalisation, Generalization bound, Machine learning models, Neural-networks, Performance, Safety critical applications, Stochastic neural network, Test samples, Training data
Discipline
Databases and Information Systems | Data Storage Systems
Research Areas
Data Science and Engineering; Information Systems and Management
Publication
Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, Valencia, Spain, 2024 May 2-4
Volume
238
First Page
4528
Last Page
4536
Identifier
https://proceedings.mlr.press/v238/mustafa24a.html
Publisher
ML Research Press
City or Country
New York
Citation
WALEED, Mustafa; PHILIPP, Liznerski; LEDENT, Antoine; DENNIS, Wagner; PUYU, Wang; and MARIUS, Kloft.
Non-vacuous generalization bounds for adversarial risk in stochastic neural networks. (2024). Proceedings of The 27th International Conference on Artificial Intelligence and Statistics, Valencia, Spain, 2024 May 2-4. 238, 4528-4536.
Available at: https://ink.library.smu.edu.sg/sis_research/9306
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.