Compressive K-means

TL;DR

The paper introduces Compressive K-means (CKM), a scalable clustering method that estimates centers from a compressed data sketch, significantly reducing computation time while maintaining accuracy on large datasets.

Contribution

CKM provides a novel, memory-efficient clustering approach that operates on compressed data sketches, reducing dependence on dataset size and initialization sensitivity.

Findings

CKM performs similarly to Lloyd-Max with a sketch size proportional to the number of clusters and ambient dimension.

CKM is two orders of magnitude faster than multiple Lloyd-Max runs on large datasets.

CKM achieves lower classification errors on handwritten digit data.

Abstract

The Lloyd-Max algorithm is a classical approach to perform K-means clustering. Unfortunately, its cost becomes prohibitive as the training dataset grows large. We propose a compressive version of K-means (CKM), that estimates cluster centers from a sketch, i.e. from a drastically compressed representation of the training dataset. We demonstrate empirically that CKM performs similarly to Lloyd-Max, for a sketch size proportional to the number of cen-troids times the ambient dimension, and independent of the size of the original dataset. Given the sketch, the computational complexity of CKM is also independent of the size of the dataset. Unlike Lloyd-Max which requires several replicates, we further demonstrate that CKM is almost insensitive to initialization. For a large dataset of 10^7 data points, we show that CKM can run two orders of magnitude faster than five replicates of Lloyd-Max, with similar clustering performance on artificial data. Finally, CKM achieves lower classification errors on handwritten digits classification.