Coresets for Clustering in Euclidean Spaces: Importance Sampling is Nearly Optimal

TL;DR

This paper introduces a nearly optimal importance sampling framework for constructing small coresets for Euclidean clustering problems, significantly improving previous bounds and providing new dimensionality reduction techniques.

Contribution

The paper presents a unified two-stage importance sampling method for coresets in Euclidean clustering, reducing coreset size and computational complexity compared to prior work.

Findings

Coreset size for k-median is O(^{-4} k)

New dimensionality reduction connects subspace approximation and clustering

Lower bound matches the upper bound in size dependence

Abstract

Given a collection of $n$ points in $\mathbb{R}^d$, the goal of the $(k,z)$-clustering problem is to find a subset of $k$ "centers" that minimizes the sum of the $z$-th powers of the Euclidean distance of each point to the closest center. Special cases of the $(k,z)$-clustering problem include the $k$-median and $k$-means problems. Our main result is a unified two-stage importance sampling framework that constructs an $\varepsilon$-coreset for the $(k,z)$-clustering problem. Compared to the results for $(k,z)$-clustering in [Feldman and Langberg, STOC 2011], our framework saves a $\varepsilon^2 d$ factor in the coreset size. Compared to the results for $(k,z)$-clustering in [Sohler and Woodruff, FOCS 2018], our framework saves a $\operatorname{poly}(k)$ factor in the coreset size and avoids the $\exp(k/\varepsilon)$ term in the construction time. Specifically, our coreset for $k$-median ($z=1$) has size $\tilde{O}(\varepsilon^{-4} k)$ which, when compared to the result in [Sohler and Woodruff, STOC 2018], saves a $k$ factor in the coreset size. Our algorithmic results rely on a new dimensionality reduction technique that connects two well-known shape fitting problems: subspace approximation and clustering, and may be of independent interest. We also provide a size lower bound of $\Omega\left(k\cdot \min \left\{2^{z/20},d \right\}\right)$ for a $0.01$-coreset for $(k,z)$-clustering, which has a linear dependence of size on $k$ and an exponential dependence on $z$ that matches our algorithmic results.