paper-with-me

홈 › Papers

Accelerating Decentralized Optimization via Overlapping Local Steps

2026-01-04 · Yijie Zhou, Shi Pu arxiv

Decentralized optimization has emerged as a critical paradigm for distributed learning, enabling scalable training while preserving data privacy through peer-to-peer collaboration. However, existing methods often suffer from communication bottlenecks due to frequent synchronization between nodes. We present Overlapping Local Decentralized SGD (OLDSGD), a novel approach to accelerate decentralized training by computation-communication overlapping, significantly reducing network idle time. With a deliberately designed update, OLDSGD preserves the same average update as Local SGD while avoiding communication-induced stalls. Theoretically, we establish non-asymptotic convergence rates for smooth non-convex objectives, showing that OLDSGD retains the same iteration complexity as standard Local Decentralized SGD while improving per-iteration runtime. Empirical results demonstrate OLDSGD's consistent improvements in wall-clock time convergence under different levels of communication delays. With minimal modifications to existing frameworks, OLDSGD offers a practical solution for faster decentralized learning without sacrificing theoretical guarantees.

📄 PDF Abstract BibTeX arXiv:2601.01493

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Adaptive Decentralized Composite Optimization via Three-Operator Splitting

2026-02-19 · Xiaokai Chen, Ilya Kuruzov, Gesualdo Scutari arxiv

The paper studies decentralized optimization over networks, where agents minimize a sum of {\it locally} smooth (strongly) convex losses and plus a nonsmooth convex extended value term. We propose decentralized methods w…

On the Benefits of Multiple Gossip Steps in Communication-Constrained Decentralized Optimization

2020-11-20 · Abolfazl Hashemi, Anish Acharya, Rudrajit Das, Haris Vikalo 외

In decentralized optimization, it is common algorithmic practice to have nodes interleave (local) gradient descent iterations with gossip (i.e. averaging over the network) steps. Motivated by the training of large-scale …

Asynchronous Decentralized SGD with Quantized and Local Updates

2019-10-27 · NeurIPS 2021 12 · Giorgi Nadiradze, Amirmojtaba Sabour, Peter Davies, Shigang Li 외

Decentralized optimization is emerging as a viable alternative for scalable distributed machine learning, but also introduces new challenges in terms of synchronization costs. To this end, several communication-reduction…

BlockingDistributed OptimizationImage ClassificationQuantization

Decentralized Gradient Tracking with Local Steps

2023-01-03 · Yue Liu, Tao Lin, Anastasia Koloskova, Sebastian U. Stich

Gradient tracking (GT) is an algorithm designed for solving decentralized optimization problems over a network (such as training a machine learning model). A key feature of GT is a tracking mechanism that allows to overc…

Quasi-Global Momentum: Accelerating Decentralized Deep Learning on Heterogeneous Data

2021-02-09 · Tao Lin, Sai Praneeth Karimireddy, Sebastian U. Stich, Martin Jaggi

Decentralized training of deep learning models is a key element for enabling data privacy and on-device learning over networks. In realistic learning scenarios, the presence of heterogeneity across different clients' loc…

Deep Learning