Fast Metric Learning For Deep Neural Networks

TL;DR

This paper introduces a fast metric learning method that learns similarity measures directly from data, enabling quicker convergence and improved accuracy in classification tasks.

Contribution

It proposes a novel approach that learns target vectors from similarity constraints without initial features, then maps features to this space for effective similarity measurement.

Findings

Faster convergence in training

Higher accuracy on classification datasets

Effective similarity measurement for retrieval tasks

Abstract

Similarity metrics are a core component of many information retrieval and machine learning systems. In this work we propose a method capable of learning a similarity metric from data equipped with a binary relation. By considering only the similarity constraints, and initially ignoring the features, we are able to learn target vectors for each instance using one of several appropriately designed loss functions. A regression model can then be constructed that maps novel feature vectors to the same target vector space, resulting in a feature extractor that computes vectors for which a predefined metric is a meaningful measure of similarity. We present results on both multiclass and multi-label classification datasets that demonstrate considerably faster convergence, as well as higher accuracy on the majority of the intrinsic evaluation tasks and all extrinsic evaluation tasks.