English

Hypergraph $p$-Laplacian equations for data interpolation and semi-supervised learning

Numerical Analysis 2025-04-08 v2 Machine Learning Numerical Analysis

Abstract

Hypergraph learning with pp-Laplacian regularization has attracted a lot of attention due to its flexibility in modeling higher-order relationships in data. This paper focuses on its fast numerical implementation, which is challenging due to the non-differentiability of the objective function and the non-uniqueness of the minimizer. We derive a hypergraph pp-Laplacian equation from the subdifferential of the pp-Laplacian regularization. A simplified equation that is mathematically well-posed and computationally efficient is proposed as an alternative. Numerical experiments verify that the simplified pp-Laplacian equation suppresses spiky solutions in data interpolation and improves classification accuracy in semi-supervised learning. The remarkably low computational cost enables further applications.

Keywords

Cite

@article{arxiv.2411.12601,
  title  = {Hypergraph $p$-Laplacian equations for data interpolation and semi-supervised learning},
  author = {Kehan Shi and Martin Burger},
  journal= {arXiv preprint arXiv:2411.12601},
  year   = {2025}
}

Comments

17 pages

R2 v1 2026-06-28T20:05:11.270Z