Stochastic, Distributed and Federated Optimization for Machine Learning

Jakub Konečný

We study optimization algorithms for the finite sum problems frequently arising in machine learning applications. First, we propose novel variants of stochastic gradient descent wi…