The Average-Case Complexity of Counting Cliques in Erdos-Renyi Hypergraphs
Abstract
We consider the problem of counting -cliques in -uniform Erdos-Renyi hypergraphs with edge density , and show that its fine-grained average-case complexity can be based on its worst-case complexity. We prove the following: 1. Dense Erdos-Renyi graphs and hypergraphs: Counting -cliques on with and constant matches its worst-case time complexity up to a factor. Assuming randomized ETH, it takes time to count -cliques in if and are constant. 2. Sparse Erdos-Renyi graphs and hypergraphs: When , we give several algorithms exploiting the sparsity of that are faster than the best known worst-case algorithms. Complementing this, based on a fine-grained worst-case assumption, our results imply a different average-case phase diagram for each fixed depicting a tradeoff between a runtime lower bound and . Surprisingly, in the hypergraph case (), these lower bounds are tight against our algorithms exactly when is above the Erd\H{o}s-R\'{e}nyi -clique percolation threshold. This is the first worst-case-to-average-case hardness reduction for a problem on Erd\H{o}s-R\'{e}nyi hypergraphs that we are aware of. We also give a variant of our result for computing the parity of the -clique count that tolerates higher error probability.
Keywords
Cite
@article{arxiv.1903.08247,
title = {The Average-Case Complexity of Counting Cliques in Erdos-Renyi Hypergraphs},
author = {Enric Boix-Adserà and Matthew Brennan and Guy Bresler},
journal= {arXiv preprint arXiv:1903.08247},
year = {2021}
}
Comments
44 pages, 2 figures, appeared in FOCS'19, accepted to SICOMP special edition