O
Omar Fawzi
Researcher at École normale supérieure de Lyon
Publications - 106
Citations - 6146
Omar Fawzi is an academic researcher from École normale supérieure de Lyon. The author has contributed to research in topics: Quantum & Quantum information. The author has an hindex of 28, co-authored 106 publications receiving 4836 citations. Previous affiliations of Omar Fawzi include National Technical University & McGill University.
Papers
More filters
Proceedings ArticleDOI
Universal Adversarial Perturbations
TL;DR: The surprising existence of universal perturbations reveals important geometric correlations among the high-dimensional decision boundary of classifiers and outlines potential security breaches with the existence of single directions in the input space that adversaries can possibly exploit to break a classifier on most natural images.
Posted Content
Universal adversarial perturbations
TL;DR: In this paper, the existence of a universal (image-agnostic) and very small perturbation vector that causes natural images to be misclassified with high probability was shown.
Journal ArticleDOI
Analysis of classifiers’ robustness to adversarial perturbations
TL;DR: In this article, the authors provide a theoretical framework for analyzing the robustness of classifiers to adversarial perturbations, and show fundamental upper bounds on the adversarial robustness.
Journal ArticleDOI
Quantum Conditional Mutual Information and Approximate Markov Chains
TL;DR: In this paper, it was shown that the quantum conditional mutual information I(A : C|B) of an arbitrary state is an upper bound on its distance to the closest reconstructed state.
Journal ArticleDOI
Practical device-independent quantum cryptography via entropy accumulation.
TL;DR: A property of entropy, termed “entropy accumulation”, is presented, which asserts that the total amount of entropy of a large system is the sum of its parts, which is used to prove the security of cryptographic protocols, including device-independent quantum key distribution, while achieving essentially optimal parameters.