Exploring dual information in distance metric learning for clustering

TL;DR

This paper investigates how incorporating dual information from pairwise constraints enhances distance metric learning for clustering, leading to improved alignment with expert knowledge and better clustering outcomes.

Contribution

It introduces a novel approach that exploits dual information in semi-supervised metric learning, improving the effectiveness of clustering with side-information.

Findings

Dual information integration improves clustering performance.

Filtering constraints enhances metric learning effectiveness.

Experiments demonstrate benefits of the proposed method.

Abstract

Distance metric learning algorithms aim to appropriately measure similarities and distances between data points. In the context of clustering, metric learning is typically applied with the assist of side-information provided by experts, most commonly expressed in the form of cannot-link and must-link constraints. In this setting, distance metric learning algorithms move closer pairs of data points involved in must-link constraints, while pairs of points involved in cannot-link constraints are moved away from each other. For these algorithms to be effective, it is important to use a distance metric that matches the expert knowledge, beliefs, and expectations, and the transformations made to stick to the side-information should preserve geometrical properties of the dataset. Also, it is interesting to filter the constraints provided by the experts to keep only the most useful and reject those that can harm the clustering process. To address these issues, we propose to exploit the dual information associated with the pairwise constraints of the semi-supervised clustering problem. Experiments clearly show that distance metric learning algorithms benefit from integrating this dual information.