The Average-Case Complexity of Counting Cliques in Erdos-Renyi Hypergraphs

We consider the problem of counting $k$-cliques in $s$-uniform Erdos-Renyi hypergraphs $G(n,c,s)$ with edge density $c$, 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 $k$-cliques on $G(n,c,s)$ with $k$ and $c$ constant matches its worst-case time complexity up to a $\mathrm{p}\mathrm{o}\mathrm{l}\mathrm{y}\mathrm{l}\mathrm{o}\mathrm{g}(n)$ factor. Assuming randomized ETH, it takes $n^{\mathrm{\Omega }(k)}$ time to count $k$-cliques in $G(n,c,s)$ if $k$ and $c$ are constant.
2. Sparse Erdos-Renyi graphs and hypergraphs: When $c=\mathrm{\Theta }(n^{-\alpha })$, we give several algorithms exploiting the sparsity of $G(n,c,s)$ 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 $\alpha$ depicting a tradeoff between a runtime lower bound and $k$. Surprisingly, in the hypergraph case ($s \ge 3$),theselowerboundsaretightagainstouralgorithmsexactlywhen$c$ is above the Erd\H{o}s-R'{e}nyi $k$-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 $k$-clique count that tolerates higher error probability.