Semidefinite Relaxations of Products of Nonnegative Forms on the Sphere
Chenyang Yuan, Pablo A. Parrilo

TL;DR
This paper introduces a polynomial-size semidefinite relaxation for maximizing the geometric mean of non-negative quadratic forms on the sphere, providing constant-factor approximation guarantees and extending to higher degrees.
Contribution
It develops a novel, efficient SDP relaxation for a non-convex polynomial optimization problem, with proven approximation bounds and a hierarchy of relaxations for general degrees.
Findings
The relaxation achieves a constant factor approximation for quadratic forms.
The analysis is tight as the degree increases, approaching the approximation bound.
A hierarchy of relaxations and rounding algorithms are proposed for broader applicability.
Abstract
We study the problem of maximizing the geometric mean of low-degree non-negative forms on the real or complex sphere in variables. We show that this highly non-convex problem is NP-hard even when the forms are quadratic and is equivalent to optimizing a homogeneous polynomial of degree on the sphere. The standard Sum-of-Squares based convex relaxation for this polynomial optimization problem requires solving a semidefinite program (SDP) of size , with multiplicative approximation guarantees of . We exploit the compact representation of this polynomial to introduce a SDP relaxation of size polynomial in and , and prove that it achieves a constant factor multiplicative approximation when maximizing the geometric mean of non-negative quadratic forms. We also show that this analysis is asymptotically tight, with a sequence of instances…
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
Topicsadvanced mathematical theories · Numerical methods in inverse problems · Mathematical Analysis and Transform Methods
