Improving Clinical Predictions through Unsupervised Time Series Representation Learning

TL;DR

This paper demonstrates that unsupervised sequence-to-sequence models, especially forecasting models with attention, improve clinical prediction accuracy from medical time series data compared to supervised methods.

Contribution

Introduces a novel forecasting Seq2Seq model with attention for unsupervised learning on medical time series, enhancing clinical prediction performance.

Findings

Forecasting Seq2Seq with attention outperforms autoencoder models.

Unsupervised learning offers performance benefits over end-to-end supervised models.

Attention mechanism improves the quality of learned representations.

Abstract

In this work, we investigate unsupervised representation learning on medical time series, which bears the promise of leveraging copious amounts of existing unlabeled data in order to eventually assist clinical decision making. By evaluating on the prediction of clinically relevant outcomes, we show that in a practical setting, unsupervised representation learning can offer clear performance benefits over end-to-end supervised architectures. We experiment with using sequence-to-sequence (Seq2Seq) models in two different ways, as an autoencoder and as a forecaster, and show that the best performance is achieved by a forecasting Seq2Seq model with an integrated attention mechanism, proposed here for the first time in the setting of unsupervised learning for medical time series.