Modeling Time Series Similarity with Siamese Recurrent Networks

TL;DR

This paper introduces Siamese Recurrent Networks (SRNs) that learn to measure similarity between time series by training on labeled pairs, improving performance in tasks like signature and voice recognition especially with limited data.

Contribution

The paper proposes SRNs that combine time-series modeling and metric learning, enabling similarity measurement with external supervision and few examples.

Findings

SRNs outperform traditional methods in signature recognition.

SRNs effectively handle out-of-domain classification.

SRNs excel in one-shot learning scenarios.

Abstract

Traditional techniques for measuring similarities between time series are based on handcrafted similarity measures, whereas more recent learning-based approaches cannot exploit external supervision. We combine ideas from time-series modeling and metric learning, and study siamese recurrent networks (SRNs) that minimize a classification loss to learn a good similarity measure between time series. Specifically, our approach learns a vectorial representation for each time series in such a way that similar time series are modeled by similar representations, and dissimilar time series by dissimilar representations. Because it is a similarity prediction models, SRNs are particularly well-suited to challenging scenarios such as signature recognition, in which each person is a separate class and very few examples per class are available. We demonstrate the potential merits of SRNs in within-domain and out-of-domain classification experiments and in one-shot learning experiments on tasks such as signature, voice, and sign language recognition.