Kernel-based Distance Metric Learning in the Output Space
This work addresses distance metric learning for classification, but it appears incremental as it builds on existing kernel-based approaches.
The paper tackles the problem of learning a non-linear distance metric by mapping data to a lower-dimensional output space and using a Mahalanobis metric, with experimental results showing advantages over traditional and kernel-based methods in classification tasks.
In this paper we present two related, kernel-based Distance Metric Learning (DML) methods. Their respective models non-linearly map data from their original space to an output space, and subsequent distance measurements are performed in the output space via a Mahalanobis metric. The dimensionality of the output space can be directly controlled to facilitate the learning of a low-rank metric. Both methods allow for simultaneous inference of the associated metric and the mapping to the output space, which can be used to visualize the data, when the output space is 2- or 3-dimensional. Experimental results for a collection of classification tasks illustrate the advantages of the proposed methods over other traditional and kernel-based DML approaches.