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. 2411.07776
22
2

On theoretical guarantees and a blessing of dimensionality for nonconvex sampling

12 November 2024
Martin Chak
ArXivPDFHTML
Abstract

Existing guarantees for algorithms sampling from nonlogconcave measures on Rd\mathbb{R}^dRd are generally inexplicit or unscalable. Even for the class of measures with logdensities that have bounded Hessians and are strongly concave outside a Euclidean ball of radius RRR, no available theory is comprehensively satisfactory with respect to both RRR and ddd. In this paper, it is shown that complete polynomial complexity can in fact be achieved if R≤cdR\leq c\sqrt{d}R≤cd​, whilst an exponential number of point evaluations is generally necessary for any algorithm as soon as R≥CdR\geq C\sqrt{d}R≥Cd​ for constants C>c>0C>c>0C>c>0. A simple importance sampler with tail-matching proposal achieves the former, owing to a blessing of dimensionality. On the other hand, if strong concavity outside a ball is replaced by a distant dissipativity condition, then sampling guarantees must generally scale exponentially with ddd in all parameter regimes.

View on arXiv
Comments on this paper