Almost Optimal Bounds for Sublinear-Time Sampling of $k$-Cliques: Sampling Cliques is Harder Than Counting

TL;DR

This paper establishes nearly tight bounds for sublinear-time algorithms that sample $k$-cliques uniformly in graphs, revealing a fundamental complexity separation between counting and sampling tasks.

Contribution

It provides the first tight bounds for sublinear-time $k$-clique sampling, demonstrating a separation from approximate counting in the same regime.

Findings

Lower bounds are derived using graph constructions with hidden cliques.

Upper bounds are achieved via auxiliary graph constructions and existing edge-sampling algorithms.

The results show sampling is strictly harder than counting in the sublinear regime.

Abstract

In this work, we consider the problem of sampling a $k$-clique in a graph from an almost uniform distribution in sublinear time in the general graph query model. Specifically the algorithm should output each $k$-clique with probability $(1\pm \epsilon)/n_k$, where $n_k$ denotes the number of $k$-cliques in the graph and $\epsilon$ is a given approximation parameter. We prove that the query complexity of this problem is \[ \Theta^*\left(\max\left\{ \left(\frac{(n\alpha)^{k/2}}{ n_k}\right)^{\frac{1}{k-1}} ,\; \min\left\{n\alpha,\frac{n\alpha^{k-1}}{n_k} \right\}\right\}\right). \] where $n$ is the number of vertices in the graph, $\alpha$ is its arboricity, and $\Theta^*$ suppresses the dependence on $(\log n/\epsilon)^{O(k)}$. Interestingly, this establishes a separation between approximate counting and approximate uniform sampling in the sublinear regime. For example, if $k=3$, $\alpha = O(1)$, and $n_3$ (the number of triangles) is $\Theta(n)$, then we get a lower bound of $\Omega(n^{1/4})$ (for constant $\epsilon$), while under these conditions, a $(1\pm \epsilon)$-approximation of $n_3$ can be obtained by performing $\textrm{poly}(\log(n/\epsilon))$ queries (Eden, Ron and Seshadhri, SODA20). Our lower bound follows from a construction of a family of graphs with arboricity $\alpha$ such that in each graph there are $n_k$ cliques (of size $k$), where one of these cliques is "hidden" and hence hard to sample. Our upper bound is based on defining a special auxiliary graph $H_k$, such that sampling edges almost uniformly in $H_k$ translates to sampling $k$-cliques almost uniformly in the original graph $G$. We then build on a known edge-sampling algorithm (Eden, Ron and Rosenbaum, ICALP19) to sample edges in $H_k$, where the challenge is simulate queries to $H_k$ while being given access only to $G$.