Momentum provably improves error feedback!

I Fatkhullin, A Tyurin, P Richtárik - Advances in Neural …, 2023‏ - proceedings.neurips.cc
Due to the high communication overhead when training machine learning models in a
distributed environment, modern algorithms invariably rely on lossy communication …

EF21-P and friends: Improved theoretical communication complexity for distributed optimization with bidirectional compression

K Gruntkowska, A Tyurin… - … Conference on Machine …, 2023‏ - proceedings.mlr.press
In this work we focus our attention on distributed optimization problems in the context where
the communication time between the server and the workers is non-negligible. We obtain …

Lower bounds and nearly optimal algorithms in distributed learning with communication compression

X Huang, Y Chen, W Yin… - Advances in Neural …, 2022‏ - proceedings.neurips.cc
Recent advances in distributed optimization and learning have shown that communication
compression is one of the most effective means of reducing communication. While there …

Communication acceleration of local gradient methods via an accelerated primal-dual algorithm with an inexact prox

A Sadiev, D Kovalev… - Advances in Neural …, 2022‏ - proceedings.neurips.cc
Inspired by a recent breakthrough of Mishchenko et al.[2022], who for the first time showed
that local gradient steps can lead to provable communication acceleration, we propose an …

Queuing dynamics of asynchronous Federated Learning

L Leconte, M Jonckheere… - International …, 2024‏ - proceedings.mlr.press
We study asynchronous federated learning mechanisms with nodes having potentially
different computational speeds. In such an environment, each node is allowed to work on …

Achieving lossless gradient sparsification via map** to alternative space in federated learning

DY Kim, DJ Han, J Seo, J Moon - Forty-first International Conference …, 2024‏ - openreview.net
Handling the substantial communication burden in federated learning (FL) still remains a
significant challenge. Although recent studies have attempted to compress the local …

Anchor sampling for federated learning with partial client participation

F Wu, S Guo, Z Qu, S He, Z Liu… - … Conference on Machine …, 2023‏ - proceedings.mlr.press
Compared with full client participation, partial client participation is a more practical scenario
in federated learning, but it may amplify some challenges in federated learning, such as data …

2Direction: Theoretically faster distributed training with bidirectional communication compression

A Tyurin, P Richtarik - Advances in Neural Information …, 2023‏ - proceedings.neurips.cc
We consider distributed convex optimization problems in the regime when the
communication between the server and the workers is expensive in both uplink and …

On the convergence of fedprox with extrapolation and inexact prox

H Li, P Richtárik - arxiv preprint arxiv:2410.01410, 2024‏ - arxiv.org
Enhancing the FedProx federated learning algorithm (Li et al., 2020) with server-side
extrapolation, Li et al.(2024a) recently introduced the FedExProx method. Their theoretical …

Improving the worst-case bidirectional communication complexity for nonconvex distributed optimization under function similarity

K Gruntkowska, A Tyurin, P Richtárik - arxiv preprint arxiv:2402.06412, 2024‏ - arxiv.org
Effective communication between the server and workers plays a key role in distributed
optimization. In this paper, we focus on optimizing the server-to-worker communication …