Name
Abstract | ||
|---|---|---|
We study the robustness of classifiers to various kinds of random noise models. In particular, we consider noise drawn uniformly from the $ell_p$ ball for $p [1, infty]$ and Gaussian noise with an arbitrary covariance matrix. We characterize this robustness to random noise in terms of the distance to the decision boundary of the classifier. This analysis applies to linear classifiers as well as classifiers with locally approximately flat decision boundaries, a condition which is satisfied by state-of-the-art deep neural networks. The predicted robustness is verified experimentally. |
Year | Venue | Field |
|---|---|---|
2018 | AISTATS | Mathematical optimization,Random noise,Algorithm,Robustness (computer science),Covariance matrix,Classifier (linguistics),Decision boundary,Gaussian noise,Deep neural networks,Mathematics |
DocType | ISSN | Citations |
Conference | 21st International Conference on Artificial Intelligence and
Statistics (AISTATS) 2018, Apr 2018, Playa Blanca, Spain. 2018,
http://www.aistats.org/ | 1 |
PageRank | References | Authors |
0.36 | 12 | 3 |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Jean-Yves Franceschi | 1 | 1 | 0.36 |
| Alhussein Fawzi | 2 | 766 | 36.80 |
| Omar Fawzi | 3 | 71 | 10.23 |