Metric Learning for Temporal Sequence Alignment

TL;DR

This paper introduces a method to learn a Mahalanobis distance for aligning multivariate time series, improving alignment accuracy by optimizing a learned similarity measure and combining features.

Contribution

It presents a novel metric learning approach for temporal sequence alignment, casting it as a structured prediction problem with tractable optimization.

Findings

Improved alignment performance on real audio data.

Effective feature selection enhances alignment accuracy.

Abstract

In this paper, we propose to learn a Mahalanobis distance to perform alignment of multivariate time series. The learning examples for this task are time series for which the true alignment is known. We cast the alignment problem as a structured prediction task, and propose realistic losses between alignments for which the optimization is tractable. We provide experiments on real data in the audio to audio context, where we show that the learning of a similarity measure leads to improvements in the performance of the alignment task. We also propose to use this metric learning framework to perform feature selection and, from basic audio features, build a combination of these with better performance for the alignment.