English
Related papers

Related papers: Global Momentum Compression for Sparse Communicati…

200 papers

Gradient compression is a recent and increasingly popular technique for reducing the communication cost in distributed training of large-scale machine learning models. In this work we focus on developing efficient distributed methods that…

Optimization and Control · Mathematics 2020-10-02 Xun Qian , Peter Richtárik , Tong Zhang

Modern realities and trends in learning require more and more generalization ability of models, which leads to an increase in both models and training sample size. It is already difficult to solve such tasks in a single device mode. This is…

Optimization and Control · Mathematics 2024-12-03 Dmitry Bylinkin , Kirill Degtyarev , Aleksandr Beznosikov

As datasets and models become increasingly large, distributed training has become a necessary component to allow deep neural networks to train in reasonable amounts of time. However, distributed training can have substantial communication…

Machine Learning · Computer Science 2021-10-18 Jose Javier Gonzalez Ortiz , Jonathan Frankle , Mike Rabbat , Ari Morcos , Nicolas Ballas

Stochastic Gradient Descent (SGD) is the key learning algorithm for many machine learning tasks. Because of its computational costs, there is a growing interest in accelerating SGD on HPC resources like GPU clusters. However, the…

Machine Learning · Computer Science 2021-01-20 Peng Jiang , Gagan Agrawal

Gradient compression is a widely-established remedy to tackle the communication bottleneck in distributed training of large deep neural networks (DNNs). Under the error-feedback framework, Top-$k$ sparsification, sometimes with $k$ as…

Machine Learning · Computer Science 2021-08-03 Atal Narayan Sahu , Aritra Dutta , Ahmed M. Abdelmoniem , Trambak Banerjee , Marco Canini , Panos Kalnis

Pre-training Transformer models is resource-intensive, and recent studies have shown that sign momentum is an efficient technique for training large-scale deep learning models, particularly Transformers. However, its application in…

Machine Learning · Computer Science 2025-03-11 Shuhua Yu , Ding Zhou , Cong Xie , An Xu , Zhi Zhang , Xin Liu , Soummya Kar

Distributed stochastic gradient methods are widely used to preserve data privacy and ensure scalability in large-scale learning tasks. While existing theory on distributed momentum Stochastic Gradient Descent (mSGD) mainly focuses on…

Optimization and Control · Mathematics 2025-05-19 Difei Cheng , Ruinan Jin , Hong Qiao , Bo Zhang

Decentralized learning provides a scalable alternative to parameter-server-based training, yet its performance is often hindered by limited peer-to-peer communication. In this paper, we study how communication should be scheduled over time,…

Machine Learning · Computer Science 2026-04-28 Tongtian Zhu , Tianyu Zhang , Mingze Wang , Zhanpeng Zhou , Can Wang

Distributed learning, particularly Federated Learning (FL), faces a significant bottleneck in the communication cost, particularly the uplink transmission of client-to-server updates, which is often constrained by asymmetric bandwidth…

Machine Learning · Computer Science 2026-02-19 Tomas Ortega , Chun-Yin Huang , Xiaoxiao Li , Hamid Jafarkhani

Modern large scale machine learning applications require stochastic optimization algorithms to be implemented on distributed computational architectures. A key bottleneck is the communication overhead for exchanging information such as…

Machine Learning · Computer Science 2017-10-31 Jianqiao Wangni , Jialei Wang , Ji Liu , Tong Zhang

We propose Adaptive Compressed Gradient Descent (AdaCGD) - a novel optimization algorithm for communication-efficient training of supervised machine learning models with adaptive compression level. Our approach is inspired by the recently…

Machine Learning · Computer Science 2022-11-02 Maksim Makarenko , Elnur Gasanov , Rustem Islamov , Abdurakhmon Sadiev , Peter Richtarik

Graph Neural Networks (GNNs) are a computationally efficient method to learn embeddings and classifications on graph data. However, GNN training has low computational intensity, making communication costs the bottleneck for scalability.…

Machine Learning · Computer Science 2025-04-08 Ujjaini Mukhodopadhyay , Alok Tripathy , Oguz Selvitopi , Katherine Yelick , Aydin Buluc

This paper proposes a prediction-based gradient compression method for distributed learning with event-triggered communication. Our goal is to reduce the amount of information transmitted from the distributed agents to the parameter server…

Information Theory · Computer Science 2024-10-04 Adrian Edin , Zheng Chen , Michel Kieffer , Mikael Johansson

We study the problem of how to distribute the training of large-scale deep learning models in the parallel computing environment. We propose a new distributed stochastic optimization method called Elastic Averaging SGD (EASGD). We analyze…

Machine Learning · Computer Science 2016-05-10 Sixin Zhang

Large-scale deep neural networks (DNN) exhibit excellent performance for various tasks. As DNNs and datasets grow, distributed training becomes extremely time-consuming and demands larger clusters. A main bottleneck is the resulting…

Distributed, Parallel, and Cluster Computing · Computer Science 2025-05-27 Yisu Wang , Ruilong Wu , Xinjiao Li , Dirk Kutscher

Consensus-based decentralized stochastic gradient descent (D-SGD) is a widely adopted algorithm for decentralized training of machine learning models across networked agents. A crucial part of D-SGD is the consensus-based model averaging,…

Information Theory · Computer Science 2025-02-12 Daniel Pérez Herrera , Zheng Chen , Erik G. Larsson

Communication-efficient distributed training algorithms have received considerable interest recently due to their benefits for training Large Language Models (LLMs) in bandwidth-constrained settings, such as across datacenters and over the…

Machine Learning · Computer Science 2025-11-07 Amir Sarfi , Benjamin Thérien , Joel Lidin , Eugene Belilovsky

When the data is distributed across multiple servers, lowering the communication cost between the servers (or workers) while solving the distributed learning problem is an important problem and is the focus of this paper. In particular, we…

Machine Learning · Computer Science 2020-03-25 Anis Elgabli , Jihong Park , Amrit S. Bedi , Mehdi Bennis , Vaneet Aggarwal

Distributed stochastic optimization intertwines (i) stochastic gradient noise, (ii) communication compression, and (iii) adaptive/normalized updates. While each factor has been studied in isolation, their joint effect under realistic…

Applying Differentially Private Stochastic Gradient Descent (DPSGD) to training modern, large-scale neural networks such as transformer-based models is a challenging task, as the magnitude of noise added to the gradients at each iteration…

Machine Learning · Computer Science 2022-07-07 Ryuichi Ito , Seng Pei Liew , Tsubasa Takahashi , Yuya Sasaki , Makoto Onizuka