paper-with-me

홈 › Papers

Convexity of Optimization Curves: Local Sharp Thresholds, Robustness Impossibility, and New Counterexamples

2025-09-10 · Le Duc Hieu arxiv

We study when the \emph{optimization curve} of first-order methods -- the sequence \${f(x\_n)}*{n\ge0}\$ produced by constant-stepsize iterations -- is convex, equivalently when the forward differences \$f(x\_n)-f(x*{n+1})\$ are nonincreasing. For gradient descent (GD) on convex \$L\$-smooth functions, the curve is convex for all stepsizes \$η\le 1.75/L\$, and this threshold is tight. Moreover, gradient norms are nonincreasing for all \$η\le 2/L\$, and in continuous time (gradient flow) the curve is always convex. These results complement and refine the classical smooth convex optimization toolbox, connecting discrete and continuous dynamics as well as worst-case analyses.

📄 PDF Abstract BibTeX arXiv:2509.08954

Code (0)

등록된 구현이 없습니다.

Similar Papers 제목 키워드 기반

Local and Global Uniform Convexity Conditions

2021-02-09 · Thomas Kerdreux, Alexandre d'Aspremont, Sebastian Pokutta

We review various characterizations of uniform convexity and smoothness on norm balls in finite-dimensional spaces and connect results stemming from the geometry of Banach spaces with \textit{scaling inequalities} used i…

Learning Theory

New Lower Bounds for Stochastic Non-Convex Optimization through Divergence Decomposition

2025-02-19 · El Mehdi Saad, Wei-Cheng Lee, Francesco Orabona

We study fundamental limits of first-order stochastic optimization in a range of nonconvex settings, including L-smooth functions satisfying Quasar-Convexity (QC), Quadratic Growth (QG), and Restricted Secant Inequalitie…

Stochastic Optimization

An Elastica Geodesic Approach With Convexity Shape Prior

2021-01-01 · ICCV 2021 10 · Da Chen, Laurent D. Cohen, Jean-Marie Mirebeau, Xue-Cheng Tai

The minimal geodesic models based on the Eikonal equations are capable of finding suitable solutions in various image segmentation scenarios. Existing geodesic-based segmentation approaches usually exploit the image …

Image SegmentationSegmentationSemantic Segmentation

Online stochastic gradient descent on non-convex losses from high-dimensional inference

2020-03-23 · Gerard Ben Arous, Reza Gheissari, Aukosh Jagannath

Stochastic gradient descent (SGD) is a popular algorithm for optimization problems arising in high-dimensional inference tasks. Here one produces an estimator of an unknown parameter from independent samples of data by i…

General Classificationparameter estimationRetrievalVocal Bursts Intensity Prediction

Characterization of the basin of convexity for multi-snapshot spike deconvolution via variable projection

2026-07-10 · Meghna Kalra, Maxime Ferreira Da Costa, Kiryung Lee arxiv

We study the problem of multi-snapshot spike deconvolution, where the goal is to recover the locations of sparse impulses from their noisy convolution with a known point spread function (PSF) across multiple snapshots. W…