Timezone: »

Machine Learning with Adversaries: Byzantine Tolerant Gradient Descent
Peva Blanchard · El Mahdi El-Mhamdi · Rachid Guerraoui · Julien Stainer

Tue Dec 05 06:30 PM -- 10:30 PM (PST) @ Pacific Ballroom #22
We study the resilience to Byzantine failures of distributed implementations of Stochastic Gradient Descent (SGD). So far, distributed machine learning frameworks have largely ignored the possibility of failures, especially arbitrary (i.e., Byzantine) ones. Causes of failures include software bugs, network asynchrony, biases in local datasets, as well as attackers trying to compromise the entire system. Assuming a set of $n$ workers, up to $f$ being Byzantine, we ask how resilient can SGD be, without limiting the dimension, nor the size of the parameter space. We first show that no gradient aggregation rule based on a linear combination of the vectors proposed by the workers (i.e, current approaches) tolerates a single Byzantine failure. We then formulate a resilience property of the aggregation rule capturing the basic requirements to guarantee convergence despite $f$ Byzantine workers. We propose \emph{Krum}, an aggregation rule that satisfies our resilience property, which we argue is the first provably Byzantine-resilient algorithm for distributed SGD. We also report on experimental evaluations of Krum.

Author Information

Peva Blanchard
El Mahdi El-Mhamdi (EPFL)
Rachid Guerraoui (EPFL)

EPFL Professor

Julien Stainer

More from the Same Authors