Youssef Allouah

Latest

• On the Privacy-Robustness-Utility Trilemma in Distributed Learning
• Fixing by Mixing: A Recipe for Optimal Byzantine ML under Heterogeneity