English

Parameterized Convexity Testing

Data Structures and Algorithms 2021-10-26 v1

Abstract

In this work, we develop new insights into the fundamental problem of convexity testing of real-valued functions over the domain [n][n]. Specifically, we present a nonadaptive algorithm that, given inputs \eps(0,1),sN\eps \in (0,1), s \in \mathbb{N}, and oracle access to a function, \eps\eps-tests convexity in O(log(s)/\eps)O(\log (s)/\eps), where ss is an upper bound on the number of distinct discrete derivatives of the function. We also show that this bound is tight. Since sns \leq n, our query complexity bound is at least as good as that of the optimal convexity tester (Ben Eliezer; ITCS 2019) with complexity O(log\epsn\eps)O(\frac{\log \eps n}{\eps}); our bound is strictly better when s=o(n)s = o(n). The main contribution of our work is to appropriately parameterize the complexity of convexity testing to circumvent the worst-case lower bound (Belovs et al.; SODA 2020) of Ω(log(\epsn)\eps)\Omega(\frac{\log (\eps n)}{\eps}) expressed in terms of the input size and obtain a more efficient algorithm.

Keywords

Cite

@article{arxiv.2110.13012,
  title  = {Parameterized Convexity Testing},
  author = {Abhiruk Lahiri and Ilan Newman and Nithin Varma},
  journal= {arXiv preprint arXiv:2110.13012},
  year   = {2021}
}

Comments

11 pages, accepted in SOSA 2022