paper-with-me

Papers

From low probability to high confidence in stochastic convex optimization

2019-07-31 · Damek Davis, Dmitriy Drusvyatskiy, Lin Xiao, Junyu Zhang

Standard results in stochastic convex optimization bound the number of samples that an algorithm needs to generate a point with small function value in expectation. More nuanced high probability guarantees are rare, and typically either rely on "light-tail" noise assumptions or exhibit worse sample complexity. In this work, we show that a wide class of stochastic optimization algorithms for strongly convex problems can be augmented with high confidence bounds at an overhead cost that is only logarithmic in the confidence level and polylogarithmic in the condition number. The procedure we propose, called proxBoost, is elementary and builds on two well-known ingredients: robust distance estimation and the proximal point method. We discuss consequences for both streaming (online) algorithms and offline algorithms based on empirical risk minimization.

📄 PDF Abstract BibTeX arXiv:1907.13307

Code (0)

등록된 구현이 없습니다.

Tasks

Stochastic OptimizationVocal Bursts Intensity Prediction

Similar Papers 제목 키워드 기반

High Probability Complexity Bounds for Non-Smooth Stochastic Optimization with Heavy-Tailed Noise

2021-06-10 · Eduard Gorbunov, Marina Danilova, Innokentiy Shibaev, Pavel Dvurechensky 외

Stochastic first-order methods are standard for training large-scale machine learning models. Random behavior may cause a particular run of an algorithm to result in a highly suboptimal objective value, whereas theoretic…

Stochastic Optimization

Noise-Adaptive High-Probability Regret Bounds for Online Convex Optimization

2026-06-06 · Wentao Zhang, Yutong Zhang, Wentao Mo arxiv

We study high-probability regret bounds for online convex optimization (OCO) with strongly convex losses and establish three results that resolve open questions at the intersection of noise adaptivity, feedback structure…

Nearly Optimal Robust Method for Convex Compositional Problems with Heavy-Tailed Noise

2020-06-17 · Yan Yan, Xin Man, Tianbao Yang

In this paper, we propose robust stochastic algorithms for solving convex compositional problems of the form $f(\E_\xi g(\cdot; \xi)) + r(\cdot)$ by establishing {\bf sub-Gaussian confidence bounds} under weak assumption…

Gradient Clipping Helps in Non-Smooth Stochastic Optimization with Heavy-Tailed Noise

2021-05-21 · NeurIPS 2021 12 · Eduard Gorbunov, Marina Danilova, Innokentiy Andreevich Shibaev, Pavel Dvurechensky 외

Thanks to their practical efficiency and random nature of the data, stochastic first-order methods are standard for training large-scale machine learning models. Random behavior may cause a particular run of an algorithm…

Stochastic Optimization

Stochastic Non-convex Optimization with Strong High Probability Second-order Convergence

2017-10-25 · Mingrui Liu, Tianbao Yang

In this paper, we study stochastic non-convex optimization with non-convex random functions. Recent studies on non-convex optimization revolve around establishing second-order convergence, i.e., converging to a nearly se…

Vocal Bursts Intensity Prediction