English

Junta Distance Approximation with Sub-Exponential Queries

Data Structures and Algorithms 2021-06-02 v1 Computational Complexity

Abstract

Leveraging tools of De, Mossel, and Neeman [FOCS, 2019], we show two different results pertaining to the \emph{tolerant testing} of juntas. Given black-box access to a Boolean function f:{±1}n{±1}f:\{\pm1\}^{n} \to \{\pm1\}, we give a poly(k,1ε)poly(k, \frac{1}{\varepsilon}) query algorithm that distinguishes between functions that are γ\gamma-close to kk-juntas and (γ+ε)(\gamma+\varepsilon)-far from kk'-juntas, where k=O(kε2)k' = O(\frac{k}{\varepsilon^2}). In the non-relaxed setting, we extend our ideas to give a 2O~(k/ε)2^{\tilde{O}(\sqrt{k/\varepsilon})} (adaptive) query algorithm that distinguishes between functions that are γ\gamma-close to kk-juntas and (γ+ε)(\gamma+\varepsilon)-far from kk-juntas. To the best of our knowledge, this is the first subexponential-in-kk query algorithm for approximating the distance of ff to being a kk-junta (previous results of Blais, Canonne, Eden, Levi, and Ron [SODA, 2018] and De, Mossel, and Neeman [FOCS, 2019] required exponentially many queries in kk). Our techniques are Fourier analytical and make use of the notion of "normalized influences" that was introduced by Talagrand [AoP, 1994].

Cite

@article{arxiv.2106.00287,
  title  = {Junta Distance Approximation with Sub-Exponential Queries},
  author = {Vishnu Iyer and Avishay Tal and Michael Whitmeyer},
  journal= {arXiv preprint arXiv:2106.00287},
  year   = {2021}
}

Comments

To appear in CCC 2021

R2 v1 2026-06-24T02:41:45.771Z