English

Lower Bounds for Convexity Testing

Computational Complexity 2024-10-24 v1 Data Structures and Algorithms

Abstract

We consider the problem of testing whether an unknown and arbitrary set SRnS \subseteq \mathbb{R}^n (given as a black-box membership oracle) is convex, versus ε\varepsilon-far from every convex set, under the standard Gaussian distribution. The current state-of-the-art testing algorithms for this problem make 2O~(n)poly(1/ε)2^{\tilde{O}(\sqrt{n})\cdot \mathrm{poly}(1/\varepsilon)} non-adaptive queries, both for the standard testing problem and for tolerant testing. We give the first lower bounds for convexity testing in the black-box query model: - We show that any one-sided tester (which may be adaptive) must use at least nΩ(1)n^{\Omega(1)} queries in order to test to some constant accuracy ε>0\varepsilon>0. - We show that any non-adaptive tolerant tester (which may make two-sided errors) must use at least 2Ω(n1/4)2^{\Omega(n^{1/4})} queries to distinguish sets that are ε1\varepsilon_1-close to convex versus ε2\varepsilon_2-far from convex, for some absolute constants 0<ε1<ε20<\varepsilon_1<\varepsilon_2. Finally, we also show that for any constant c>0c>0, any non-adaptive tester (which may make two-sided errors) must use at least n1/4cn^{1/4 - c} queries in order to test to some constant accuracy ε>0\varepsilon>0.

Keywords

Cite

@article{arxiv.2410.17958,
  title  = {Lower Bounds for Convexity Testing},
  author = {Xi Chen and Anindya De and Shivam Nadimpalli and Rocco A. Servedio and Erik Waingarten},
  journal= {arXiv preprint arXiv:2410.17958},
  year   = {2024}
}

Comments

52 pages, to appear in SODA 2025

R2 v1 2026-06-28T19:33:01.295Z