Ellipsoid Fitting Up to a Constant

TL;DR

This paper improves the understanding of the maximum number of Gaussian points that can lie on a single ellipsoid in high dimensions, providing a tighter bound that advances the conjecture by Saunderson, Parrilo, and Willsky.

Contribution

It offers a significantly tighter analysis of a random semidefinite program related to ellipsoid fitting, establishing a lower bound proportional to the dimension squared, up to a constant.

Findings

Proves that the maximum number of Gaussian points on an ellipsoid is at least proportional to d^2.

Introduces a refined method for analyzing correlated random matrices.

Resolves one side of the SPW conjecture up to a constant factor.

Abstract

In [Sau11,SPW13], Saunderson, Parrilo and Willsky asked the following elegant geometric question: what is the largest $m= m(d)$ such that there is an ellipsoid in $\mathbb{R}^d$ that passes through $v_1, v_2, \ldots, v_m$ with high probability when the $v_i$s are chosen independently from the standard Gaussian distribution $N(0,I_{d})$. The existence of such an ellipsoid is equivalent to the existence of a positive semidefinite matrix $X$ such that $v_i^{\top}X v_i =1$ for every $1 \leq i \leq m$ - a natural example of a random semidefinite program. SPW conjectured that $m= (1-o(1)) d^2/4$ with high probability. Very recently, Potechin, Turner, Venkat and Wein and Kane and Diakonikolas proved that $m \geq d^2/\log^{O(1)}(d)$ via certain explicit constructions. In this work, we give a substantially tighter analysis of their construction to prove that $m \geq d^2/C$ for an absolute constant $C>0$. This resolves one direction of the SPW conjecture up to a constant. Our analysis proceeds via the method of Graphical Matrix Decomposition that has recently been used to analyze correlated random matrices arising in various areas [BHK+19]. Our key new technical tool is a refined method to prove singular value upper bounds on certain correlated random matrices that are tight up to absolute dimension-independent constants. In contrast, all previous methods that analyze such matrices lose logarithmic factors in the dimension.