English

Sample-based high-dimensional convexity testing

Computational Complexity 2017-06-29 v1

Abstract

In the problem of high-dimensional convexity testing, there is an unknown set SRnS \subseteq \mathbb{R}^n which is promised to be either convex or ε\varepsilon-far from every convex body with respect to the standard multivariate normal distribution N(0,1)n\mathcal{N}(0, 1)^n. The job of a testing algorithm is then to distinguish between these two cases while making as few inspections of the set SS as possible. In this work we consider sample-based testing algorithms, in which the testing algorithm only has access to labeled samples (x,S(x))(\boldsymbol{x},S(\boldsymbol{x})) where each x\boldsymbol{x} is independently drawn from N(0,1)n\mathcal{N}(0, 1)^n. We give nearly matching sample complexity upper and lower bounds for both one-sided and two-sided convexity testing algorithms in this framework. For constant ε\varepsilon, our results show that the sample complexity of one-sided convexity testing is 2Θ~(n)2^{\tilde{\Theta}(n)} samples, while for two-sided convexity testing it is 2Θ~(n)2^{\tilde{\Theta}(\sqrt{n})}.

Keywords

Cite

@article{arxiv.1706.09362,
  title  = {Sample-based high-dimensional convexity testing},
  author = {Xi Chen and Adam Freilich and Rocco A. Servedio and Timothy Sun},
  journal= {arXiv preprint arXiv:1706.09362},
  year   = {2017}
}