Distance Metric Learning for Kernel Machines

TL;DR

This paper evaluates existing Mahalanobis metric learning algorithms as pre-processing for SVMs and introduces SVML, a new method that jointly learns a metric and SVM parameters, improving classification accuracy.

Contribution

The paper presents SVML, a novel algorithm that integrates metric learning with SVM training, outperforming existing methods on benchmark datasets.

Findings

Existing Mahalanobis metric learning algorithms do not significantly improve SVM-RBF performance.

SVML outperforms state-of-the-art metric learning algorithms in accuracy.

SVML is a competitive alternative to Euclidean metrics with effective model selection.

Abstract

Recent work in metric learning has significantly improved the state-of-the-art in k-nearest neighbor classification. Support vector machines (SVM), particularly with RBF kernels, are amongst the most popular classification algorithms that uses distance metrics to compare examples. This paper provides an empirical analysis of the efficacy of three of the most popular Mahalanobis metric learning algorithms as pre-processing for SVM training. We show that none of these algorithms generate metrics that lead to particularly satisfying improvements for SVM-RBF classification. As a remedy we introduce support vector metric learning (SVML), a novel algorithm that seamlessly combines the learning of a Mahalanobis metric with the training of the RBF-SVM parameters. We demonstrate the capabilities of SVML on nine benchmark data sets of varying sizes and difficulties. In our study, SVML outperforms all alternative state-of-the-art metric learning algorithms in terms of accuracy and establishes itself as a serious alternative to the standard Euclidean metric with model selection by cross validation.