Publications

(2021). A Field Guide to Federated Optimization. Arxiv.

Samuel Horvath, Stefanos Laskaridis, Mario Almeida, Ilias Leontiadis, Stylianos Venieris, Nicholas Lane (2021). FjORD: Fair and Accurate Federated Learning under heterogeneous targets with Ordered Dropout. Arxiv.

Samuel Horvath, Aaron Klein, Peter Richtarik, Cedric Archambeau (2021). Hyperparameter Transfer Learning with Adaptive Complexity. AISTATS 2021.

Wenlin Chen, Samuel Horvath, Peter Richtarik (2020). Optimal Client Sampling for Federated Learning. Arxiv.

Filip Hanzely, Slavomir Hanzely, Samuel Horvath, Peter Richtarik (2020). Lower Bounds and Optimal Algorithms for Personalized Federated Learning. NeurIPS.

Samuel Horvath, Peter Richtarik (2020). A Better Alternative to Error Feedback for Communication-Efficient Distributed Learning. Arxiv.

Aleksandr Beznosikov, Samuel Horvath, Peter Richtarik, Mher Safaryan (2020). On Biased Compression for Distributed Learning. Arxiv.

Samuel Horvath, Lihua Lei, Peter Richtarik, Michael I. Jordan (2020). Adaptivity of Stochastic Gradient Methods for Nonconvex Optimization. Arxiv.

Samuel Horvath, Chen-Yu Ho, Ludovit Horvath, Atal Narayan Sahu, Marco Canini, Peter Richtarik (2019). Natural Compression for Distributed Deep Learning. Arxiv.

Samuel Horvath, Peter Richtarik (2019). Nonconvex Variance Reduced Optimization with Arbitrary Sampling. In ICML.

PDF Cite Poster Slides

Samuel Horvath, Dmitry Kovalev, Konstantin Mishchenko, Sebastian Stich, Peter Richtarik (2019). Stochastic Distributed Learning with Gradient Quantization and Variance Reduction. Arxiv.

Dmitry Kovalev, Samuel Horvath, Peter Richtarik (2019). Don’t Jump Through Hoops and Remove Those Loops: SVRG and Katyusha are Better Without the Outer Loop. Arxiv.

PDF Cite Slides