English

Tolerant Junta Testing and the Connection to Submodular Optimization and Function Isomorphism

Data Structures and Algorithms 2016-11-04 v2 Computational Complexity Discrete Mathematics

Abstract

A function f ⁣:{1,1}n{1,1}f\colon \{-1,1\}^n \to \{-1,1\} is a kk-junta if it depends on at most kk of its variables. We consider the problem of tolerant testing of kk-juntas, where the testing algorithm must accept any function that is ϵ\epsilon-close to some kk-junta and reject any function that is ϵ\epsilon'-far from every kk'-junta for some ϵ=O(ϵ)\epsilon'= O(\epsilon) and k=O(k)k' = O(k). Our first result is an algorithm that solves this problem with query complexity polynomial in kk and 1/ϵ1/\epsilon. This result is obtained via a new polynomial-time approximation algorithm for submodular function minimization (SFM) under large cardinality constraints, which holds even when only given an approximate oracle access to the function. Our second result considers the case where k=kk'=k. We show how to obtain a smooth tradeoff between the amount of tolerance and the query complexity in this setting. Specifically, we design an algorithm that given ρ(0,1/2)\rho\in(0,1/2) accepts any function that is ϵρ16\frac{\epsilon\rho}{16}-close to some kk-junta and rejects any function that is ϵ\epsilon-far from every kk-junta. The query complexity of the algorithm is O(klogkϵρ(1ρ)k)O\big( \frac{k\log k}{\epsilon\rho(1-\rho)^k} \big). Finally, we show how to apply the second result to the problem of tolerant isomorphism testing between two unknown Boolean functions ff and gg. We give an algorithm for this problem whose query complexity only depends on the (unknown) smallest kk such that either ff or gg is close to being a kk-junta.

Keywords

Cite

@article{arxiv.1607.03938,
  title  = {Tolerant Junta Testing and the Connection to Submodular Optimization and Function Isomorphism},
  author = {Eric Blais and Clément L. Canonne and Talya Eden and Amit Levi and Dana Ron},
  journal= {arXiv preprint arXiv:1607.03938},
  year   = {2016}
}

Comments

Polished the writing, corrected typos, and fixed an issue in the proof of Theorem 1.2

R2 v1 2026-06-22T14:54:06.162Z