Local Enumeration and Majority Lower Bounds
Abstract
Depth-3 circuit lower bounds and -SAT algorithms are intimately related; the state-of-the-art -circuit lower bound and the -SAT algorithm are based on the same combinatorial theorem. In this paper we define a problem which reveals new interactions between the two. Define Enum(, ) problem as: given an -variable -CNF and an initial assignment , output all satisfying assignments at Hamming distance from , assuming that there are no satisfying assignments of Hamming distance less than from . Observe that: an upper bound on the complexity of Enum(, ) implies: - Depth-3 circuits: Any circuit computing the Majority function has size at least . - -SAT: There exists an algorithm solving -SAT in time . A simple construction shows that . Thus, matching upper bounds would imply a -circuit lower bound of and a -SAT upper bound of . The former yields an unrestricted depth-3 lower bound of solving a long standing open problem, and the latter breaks the Super Strong Exponential Time Hypothesis. In this paper, we propose a randomized algorithm for Enum(, ) and introduce new ideas to analyze it. We demonstrate the power of our ideas by considering the first non-trivial instance of the problem, i.e., Enum(, ). We show that the expected running time of our algorithm is , substantially improving on the trivial bound of . This already improves lower bounds for Majority function to . The previous bound was which follows from the work of H{\aa}stad, Jukna, and Pudl\'ak (Comput. Complex.'95).
Cite
@article{arxiv.2403.09134,
title = {Local Enumeration and Majority Lower Bounds},
author = {Mohit Gurumukhani and Ramamohan Paturi and Pavel Pudlák and Michael Saks and Navid Talebanfard},
journal= {arXiv preprint arXiv:2403.09134},
year = {2024}
}