Research

Paper

TESTING March 19, 2026

Hardness of High-Dimensional Linear Classification

Authors

Alexander Munteanu, Simon Omlor, Jeff M. Phillips

Abstract

We establish new exponential in dimension lower bounds for the Maximum Halfspace Discrepancy problem, which models linear classification. Both are fundamental problems in computational geometry and machine learning in their exact and approximate forms. However, only $O(n^d)$ and respectively $\tilde O(1/\varepsilon^d)$ upper bounds are known and complemented by polynomial lower bounds that do not support the exponential in dimension dependence. We close this gap up to polylogarithmic terms by reduction from widely-believed hardness conjectures for Affine Degeneracy testing and $k$-Sum problems. Our reductions yield matching lower bounds of $\tildeΩ(n^d)$ and respectively $\tildeΩ(1/\varepsilon^d)$ based on Affine Degeneracy testing, and $\tildeΩ(n^{d/2})$ and respectively $\tildeΩ(1/\varepsilon^{d/2})$ conditioned on $k$-Sum. The first bound also holds unconditionally if the computational model is restricted to make sidedness queries, which corresponds to a widely spread setting implemented and optimized in many contemporary algorithms and computing paradigms.

Metadata

arXiv ID: 2603.19061
Provider: ARXIV
Primary Category: cs.CG
Published: 2026-03-19
Fetched: 2026-03-20 06:02

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2603.19061v1</id>\n    <title>Hardness of High-Dimensional Linear Classification</title>\n    <updated>2026-03-19T15:53:41Z</updated>\n    <link href='https://arxiv.org/abs/2603.19061v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2603.19061v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>We establish new exponential in dimension lower bounds for the Maximum Halfspace Discrepancy problem, which models linear classification. Both are fundamental problems in computational geometry and machine learning in their exact and approximate forms. However, only $O(n^d)$ and respectively $\\tilde O(1/\\varepsilon^d)$ upper bounds are known and complemented by polynomial lower bounds that do not support the exponential in dimension dependence. We close this gap up to polylogarithmic terms by reduction from widely-believed hardness conjectures for Affine Degeneracy testing and $k$-Sum problems. Our reductions yield matching lower bounds of $\\tildeΩ(n^d)$ and respectively $\\tildeΩ(1/\\varepsilon^d)$ based on Affine Degeneracy testing, and $\\tildeΩ(n^{d/2})$ and respectively $\\tildeΩ(1/\\varepsilon^{d/2})$ conditioned on $k$-Sum. The first bound also holds unconditionally if the computational model is restricted to make sidedness queries, which corresponds to a widely spread setting implemented and optimized in many contemporary algorithms and computing paradigms.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.CG'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.DS'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.LG'/>\n    <category scheme='http://arxiv.org/schemas/atom' term='stat.ML'/>\n    <published>2026-03-19T15:53:41Z</published>\n    <arxiv:comment>SoCG 2026</arxiv:comment>\n    <arxiv:primary_category term='cs.CG'/>\n    <author>\n      <name>Alexander Munteanu</name>\n    </author>\n    <author>\n      <name>Simon Omlor</name>\n    </author>\n    <author>\n      <name>Jeff M. Phillips</name>\n    </author>\n  </entry>"
}