paper-with-me

홈 › Papers

Convergence Analysis of Randomized Subspace Normalized SGD under Heavy-Tailed Noise

2026-01-28 · Gaku Omiya, Pierre-Louis Poirion, Akiko Takeda arxiv

Randomized subspace methods reduce per-iteration cost; however, in nonconvex optimization, most analyses are expectation-based, and high-probability bounds remain scarce even under sub-Gaussian noise. We first prove that randomized subspace SGD (RS-SGD) admits a high-probability convergence bound under sub-Gaussian noise, achieving the same order of oracle complexity as prior in-expectation results. Motivated by the prevalence of heavy-tailed gradients in modern machine learning, we then propose randomized subspace normalized SGD (RS-NSGD), which integrates direction normalization into subspace updates. Assuming the noise has bounded $p$-th moments, we establish both in-expectation and high-probability convergence guarantees, and show that RS-NSGD can achieve better oracle complexity than full-dimensional normalized SGD.

📄 PDF Abstract BibTeX arXiv:2601.20399

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Optimal Randomized First-Order Methods for Least-Squares Problems

2020-02-21 · ICML 2020 1 · Jonathan Lacotte, Mert Pilanci

We provide an exact analysis of a class of randomized algorithms for solving overdetermined least-squares problems. We consider first-order methods, where the gradients are pre-conditioned by an approximation of the Hess…

Convergence Analysis of Prediction Markets via Randomized Subspace Descent

2015-12-01 · NeurIPS 2015 12 · Rafael Frongillo, Mark D. Reid

Prediction markets are economic mechanisms for aggregating information about future events through sequential interactions with traders. The pricing mechanisms in these markets are known to be related to optimization al…

Prediction

Convergence Analysis of Inexact Randomized Iterative Methods

2019-03-19 · Nicolas Loizou, Peter Richtárik

In this paper we present a convergence rate analysis of inexact variants of several randomized iterative methods. Among the methods studied are: stochastic gradient descent, stochastic Newton, stochastic proximal point a…

Greedy Low-Rank Gradient Compression for Distributed Learning with Convergence Guarantees

2025-07-11 · Chuyan Chen, Yutong He, Pengrui Li, Weichen Jia 외 arxiv

Distributed optimization is pivotal for large-scale signal processing and machine learning, yet communication overhead remains a major bottleneck. Low-rank gradient compression, in which the transmitted gradients are app…

Distributed Optimization

A Memory Efficient Randomized Subspace Optimization Method for Training Large Language Models

2025-02-11 · Yiming Chen, Yuan Zhang, Yin Liu, Kun Yuan 외

The memory challenges associated with training Large Language Models (LLMs) have become a critical concern, particularly when using the Adam optimizer. To address this issue, numerous memory-efficient techniques have bee…