Byzantine machine learning: A primer
The problem of Byzantine resilience in distributed machine learning, aka Byzantine machine
learning, consists of designing distributed algorithms that can train an accurate model …
learning, consists of designing distributed algorithms that can train an accurate model …
Fixing by mixing: A recipe for optimal byzantine ml under heterogeneity
Y Allouah, S Farhadkhani… - International …, 2023 - proceedings.mlr.press
Byzantine machine learning (ML) aims to ensure the resilience of distributed learning
algorithms to misbehaving (or Byzantine) machines. Although this problem received …
algorithms to misbehaving (or Byzantine) machines. Although this problem received …
Byzantine-robust learning on heterogeneous datasets via bucketing
In Byzantine robust distributed or federated learning, a central server wants to train a
machine learning model over data distributed across multiple workers. However, a fraction …
machine learning model over data distributed across multiple workers. However, a fraction …
Robust distributed learning: tight error bounds and breakdown point under data heterogeneity
The theory underlying robust distributed learning algorithms, designed to resist adversarial
machines, matches empirical observations when data is homogeneous. Under data …
machines, matches empirical observations when data is homogeneous. Under data …
Byzantine-robust learning on heterogeneous data via gradient splitting
Federated learning has exhibited vulnerabilities to Byzantine attacks, where the Byzantine
attackers can send arbitrary gradients to a central server to destroy the convergence and …
attackers can send arbitrary gradients to a central server to destroy the convergence and …
An equivalence between data poisoning and byzantine gradient attacks
S Farhadkhani, R Guerraoui… - … on Machine Learning, 2022 - proceedings.mlr.press
To study the resilience of distributed learning, the “Byzantine" literature considers a strong
threat model where workers can report arbitrary gradients to the parameter server. Whereas …
threat model where workers can report arbitrary gradients to the parameter server. Whereas …
RFVIR: A robust federated algorithm defending against Byzantine attacks
Federated learning (FL) is susceptible to Byzantine attacks due to its inherently distributed
and privacy-preserving nature. Most model parameters-based defense methods become …
and privacy-preserving nature. Most model parameters-based defense methods become …
On the strategyproofness of the geometric median
EM El-Mhamdi, S Farhadkhani… - International …, 2023 - proceedings.mlr.press
The geometric median, an instrumental component of the secure machine learning toolbox,
is known to be effective when robustly aggregating models (or gradients), gathered from …
is known to be effective when robustly aggregating models (or gradients), gathered from …
Renyi differential privacy of propose-test-release and applications to private and robust machine learning
JT Wang, S Mahloujifar, S Wang… - Advances in Neural …, 2022 - proceedings.neurips.cc
Abstract Propose-Test-Release (PTR) is a differential privacy framework that works with
local sensitivity of functions, instead of their global sensitivity. This framework is typically …
local sensitivity of functions, instead of their global sensitivity. This framework is typically …
Byzantine-Resilient Federated Principal Subspace Estimation
This work studies the problem of reliably estimating a subspace in a federated setting, when
some nodes' outputs can be compromised by Byzantine attacks. Typically, the subspace of …
some nodes' outputs can be compromised by Byzantine attacks. Typically, the subspace of …