English

Hilbert Curve Projection Distance for Distribution Comparison

Machine Learning 2024-02-07 v4 Machine Learning

Abstract

Distribution comparison plays a central role in many machine learning tasks like data classification and generative modeling. In this study, we propose a novel metric, called Hilbert curve projection (HCP) distance, to measure the distance between two probability distributions with low complexity. In particular, we first project two high-dimensional probability distributions using Hilbert curve to obtain a coupling between them, and then calculate the transport distance between these two distributions in the original space, according to the coupling. We show that HCP distance is a proper metric and is well-defined for probability measures with bounded supports. Furthermore, we demonstrate that the modified empirical HCP distance with the LpL_p cost in the dd-dimensional space converges to its population counterpart at a rate of no more than O(n1/2max{d,p})O(n^{-1/2\max\{d,p\}}). To suppress the curse-of-dimensionality, we also develop two variants of the HCP distance using (learnable) subspace projections. Experiments on both synthetic and real-world data show that our HCP distance works as an effective surrogate of the Wasserstein distance with low complexity and overcomes the drawbacks of the sliced Wasserstein distance.

Keywords

Cite

@article{arxiv.2205.15059,
  title  = {Hilbert Curve Projection Distance for Distribution Comparison},
  author = {Tao Li and Cheng Meng and Hongteng Xu and Jun Yu},
  journal= {arXiv preprint arXiv:2205.15059},
  year   = {2024}
}

Comments

33 pages, 11 figures

R2 v1 2026-06-24T11:33:03.336Z