English

Sharper bounds on the Fourier concentration of DNFs

Computational Complexity 2021-10-19 v2

Abstract

In 1992 Mansour proved that every size-ss DNF formula is Fourier-concentrated on sO(loglogs)s^{O(\log\log s)} coefficients. We improve this to sO(loglogk)s^{O(\log\log k)} where kk is the read number of the DNF. Since kk is always at most ss, our bound matches Mansour's for all DNFs and strengthens it for small-read ones. The previous best bound for read-kk DNFs was sO(k3/2)s^{O(k^{3/2})}. For kk up to Θ~(loglogs)\tilde{\Theta}(\log\log s), we further improve our bound to the optimal poly(s)\mathrm{poly}(s); previously no such bound was known for any k=ωs(1)k = \omega_s(1). Our techniques involve new connections between the term structure of a DNF, viewed as a set system, and its Fourier spectrum.

Keywords

Cite

@article{arxiv.2109.04525,
  title  = {Sharper bounds on the Fourier concentration of DNFs},
  author = {Victor Lecomte and Li-Yang Tan},
  journal= {arXiv preprint arXiv:2109.04525},
  year   = {2021}
}

Comments

19 pages; to appear at FOCS 2021