ResearchTrend.AI
  • Papers
  • Communities
  • Events
  • Blog
  • Pricing
Papers
Communities
Social Events
Terms and Conditions
Pricing
Parameter LabParameter LabTwitterGitHubLinkedInBlueskyYoutube

© 2025 ResearchTrend.AI, All rights reserved.

  1. Home
  2. Papers
  3. 2006.14571
120
16

Sparse Convex Optimization via Adaptively Regularized Hard Thresholding

25 June 2020
Kyriakos Axiotis
M. Sviridenko
ArXiv (abs)PDFHTML
Abstract

The goal of Sparse Convex Optimization is to optimize a convex function fff under a sparsity constraint s≤s∗γs\leq s^*\gammas≤s∗γ, where s∗s^*s∗ is the target number of non-zero entries in a feasible solution (sparsity) and γ≥1\gamma\geq 1γ≥1 is an approximation factor. There has been a lot of work to analyze the sparsity guarantees of various algorithms (LASSO, Orthogonal Matching Pursuit (OMP), Iterative Hard Thresholding (IHT)) in terms of the Restricted Condition Number κ\kappaκ. The best known algorithms guarantee to find an approximate solution of value f(x∗)+ϵf(x^*)+\epsilonf(x∗)+ϵ with the sparsity bound of γ=O(κmin⁡{log⁡f(x0)−f(x∗)ϵ,κ})\gamma = O\left(\kappa\min\left\{\log \frac{f(x^0)-f(x^*)}{\epsilon}, \kappa\right\}\right)γ=O(κmin{logϵf(x0)−f(x∗)​,κ}), where x∗x^*x∗ is the target solution. We present a new Adaptively Regularized Hard Thresholding (ARHT) algorithm that makes significant progress on this problem by bringing the bound down to γ=O(κ)\gamma=O(\kappa)γ=O(κ), which has been shown to be tight for a general class of algorithms including LASSO, OMP, and IHT. This is achieved without significant sacrifice in the runtime efficiency compared to the fastest known algorithms. We also provide a new analysis of OMP with Replacement (OMPR) for general fff, under the condition s>s∗κ24s > s^* \frac{\kappa^2}{4}s>s∗4κ2​, which yields Compressed Sensing bounds under the Restricted Isometry Property (RIP). When compared to other Compressed Sensing approaches, it has the advantage of providing a strong tradeoff between the RIP condition and the solution sparsity, while working for any general function fff that meets the RIP condition.

View on arXiv
Comments on this paper