Publications
See the bottom of the page for details regarding my contributions.
2024
1. Tackling Byzantine Clients in Federated Learning
Youssef Allouah, Sadegh Farhadkhani, Rachid Guerraoui, Nirupam Gupta, Rafael Pinot, Geovani Rizk, Sasha Voitovych
ICML , July 2024
Paper
2. Generalized Bradley-Terry Models for Score Estimation from Paired Comparisons
Julien Fageot, Sadegh Farhadkhani, Lê Nguyên Hoang, Oscar Villemaud
AAAI , February 2024
Paper
2023
3. Epidemic Learning: Boosting Decentralized Learning with Randomized Communication
Martijn De Vos, Sadegh Farhadkhani, Rachid Guerraoui, Anne-Marie Kermarrec, Rafael Pires, Rishi Sharma
NeurIPS , December 2023
Paper Video (by me)
4. Robust Collaborative Learning with Linear Gradient Overhead
Sadegh Farhadkhani, Rachid Guerraoui, Nirupam Gupta, Lê-Nguyên Hoang, Rafael Pinot, John Stephan
ICML , July 2023
Paper Video (by me)
5. Fixing by Mixing: A Recipe for Optimal Byzantine ML under Heterogeneity
Youssef Allouah, Sadegh Farhadkhani, Rachid Guerraoui, Nirupam Gupta, Rafael Pinot, John Stephan
AISTATS , April 2023
Paper
6. On the Strategyproofness of the Geometric Median
El-Mahdi El-Mhamdi, Sadegh Farhadkhani, Rachid Guerraoui, Lê-Nguyên Hoang
AISTATS , April 2023
Paper Video (by me)
2022
7. Byzantine Machine Learning Made Easy By Resilient Averaging of Momentums
Sadegh Farhadkhani, Rachid Guerraoui, Nirupam Gupta, Rafael Pinot, John Stephan
ICML , July 2022
Paper Video (by John, in person talk in Baltimore)
8. An Equivalence Between Data Poisoning and Byzantine Gradient Attacks
Sadegh Farhadkhani, Rachid Guerraoui, Lê Nguyên Hoang, Oscar Villemaud
ICML , July 2022
Paper Video (by Lê, in person talk in Baltimore)
* I did not attend ICML 2022 as I received my visa late.
2021
9. Collaborative Learning in the Jungle (Decentralized, Byzantine, Heterogeneous, Asynchronous and Nonconvex Learning)
El-Mahdi El-Mhamdi, Sadegh Farhadkhani, Rachid Guerraoui, Arsany Guirguis, Lê Nguyên Hoang, Sébastien Rouault
NeurIPS , December 2021
Paper Video (by Lê and me)
* The authors' list is always in alphabetical order in our group. The lead author(s) is usually indicated as the "corresponding author" in the papers. In [1, 3, 4, 6, 8, 9], I was a lead author and I played a primary role (coming up with the idea, algorithms, proofs, etc.). The main idea of [5] was initiated in one of our discussions with YA inspired by [4]. Youssef then did all the heavy lifting for this work. Initially, [7] was NG and RP’s project, studying the impact of momentum in Byzantine ML. In a group meeting, I noticed that their approach can be generalized by changing the definition of robustness to the one used in [9]. I then joined the project and proved that this definition is satisfied by almost all existing works, which led to a unifying framework for Byzantine machine learning.