Autoencoding Time Series for Visualisation

TL;DR

This paper introduces a novel method for visualizing time series data by converting them into vector representations using echo state networks, then applying an autoencoder to produce meaningful 2D visualizations that capture underlying dynamics.

Contribution

The work presents a new approach combining echo state networks and autoencoders for effective time series visualization with a principled reconstruction error objective.

Findings

Effective visualization of synthetic and real time series data

Captures latent dynamics accurately

Provides clear visual representations

Abstract

We present an algorithm for the visualisation of time series. To that end we employ echo state networks to convert time series into a suitable vector representation which is capable of capturing the latent dynamics of the time series. Subsequently, the obtained vector representations are put through an autoencoder and the visualisation is constructed using the activations of the bottleneck. The crux of the work lies with defining an objective function that quantifies the reconstruction error of these representations in a principled manner. We demonstrate the method on synthetic and real data.