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. 2310.05787
14
1

Exact threshold for approximate ellipsoid fitting of random points

9 October 2023
Antoine Maillard
Afonso S. Bandeira
ArXivPDFHTML
Abstract

We consider the problem (P)(\rm P)(P) of exactly fitting an ellipsoid (centered at 000) to nnn standard Gaussian random vectors in Rd\mathbb{R}^dRd, as n,d→∞n, d \to \inftyn,d→∞ with n/d2→α>0n / d^2 \to \alpha > 0n/d2→α>0. This problem is conjectured to undergo a sharp transition: with high probability, (P)(\rm P)(P) has a solution if α<1/4\alpha < 1/4α<1/4, while (P)(\rm P)(P) has no solutions if α>1/4\alpha > 1/4α>1/4. So far, only a trivial bound α>1/2\alpha > 1/2α>1/2 is known to imply the absence of solutions, while the sharpest results on the positive side assume α≤η\alpha \leq \etaα≤η (for η>0\eta > 0η>0 a small constant) to prove that (P)(\rm P)(P) is solvable. In this work we study universality between this problem and a so-called "Gaussian equivalent", for which the same transition can be rigorously analyzed. Our main results are twofold. On the positive side, we prove that if α<1/4\alpha < 1/4α<1/4, there exist an ellipsoid fitting all the points up to a small error, and that the lengths of its principal axes are bounded above and below. On the other hand, for α>1/4\alpha > 1/4α>1/4, we show that achieving small fitting error is not possible if the length of the ellipsoid's shortest axis does not approach 000 as d→∞d \to \inftyd→∞ (and in particular there does not exist any ellipsoid fit whose shortest axis length is bounded away from 000 as d→∞d \to \inftyd→∞). To the best of our knowledge, our work is the first rigorous result characterizing the expected phase transition in ellipsoid fitting at α=1/4\alpha = 1/4α=1/4. In a companion non-rigorous work, the first author and D. Kunisky give a general analysis of ellipsoid fitting using the replica method of statistical physics, which inspired the present work.

View on arXiv
Comments on this paper