Deterministic Gaussian Averaged Neural Networks
Machine Learning
2020-06-12 v1 Cryptography and Security
Computer Vision and Pattern Recognition
Machine Learning
Abstract
We present a deterministic method to compute the Gaussian average of neural networks used in regression and classification. Our method is based on an equivalence between training with a particular regularized loss, and the expected values of Gaussian averages. We use this equivalence to certify models which perform well on clean data but are not robust to adversarial perturbations. In terms of certified accuracy and adversarial robustness, our method is comparable to known stochastic methods such as randomized smoothing, but requires only a single model evaluation during inference.
Cite
@article{arxiv.2006.06061,
title = {Deterministic Gaussian Averaged Neural Networks},
author = {Ryan Campbell and Chris Finlay and Adam M Oberman},
journal= {arXiv preprint arXiv:2006.06061},
year = {2020}
}