$\ell_p$-Norm Multiple Kernel One-Class Fisher Null-Space

TL;DR

This paper introduces a novel $$-norm multiple kernel learning algorithm for one-class classification based on Fisher null-space principles, optimized via a min-max saddle point approach, with joint learning extensions and extensive empirical validation.

Contribution

It proposes a new $$-norm based multiple kernel learning method for one-class classification using Fisher null-space, including a joint learning extension and an efficient optimization strategy.

Findings

Outperforms baseline and other algorithms on diverse datasets.

Effective in learning shared kernel weights across related tasks.

Demonstrates robustness and versatility in application domains.

Abstract

The paper addresses the multiple kernel learning (MKL) problem for one-class classification (OCC). For this purpose, based on the Fisher null-space one-class classification principle, we present a multiple kernel learning algorithm where a general $\ell_p$-norm constraint ($p\geq1$) on kernel weights is considered. We cast the proposed one-class MKL task as a min-max saddle point Lagrangian optimisation problem and propose an efficient method to solve it. An extension of the proposed one-class MKL approach is also considered where several related one-class MKL tasks are learned jointly by constraining them to share common kernel weights. An extensive assessment of the proposed method on a range of data sets from different application domains confirms its merits against the baseline and several other algorithms.