Icentia11K: An Unsupervised Representation Learning Dataset for Arrhythmia Subtype Discovery

TL;DR

This paper introduces Icentia11K, the largest public ECG dataset with 11,000 patients, aimed at advancing unsupervised representation learning for arrhythmia subtype discovery and anomalous signal detection.

Contribution

It provides a large-scale ECG dataset and proposes an unsupervised learning framework to discover new arrhythmia subtypes and improve semi-supervised ECG analysis.

Findings

Clustering of known arrhythmia subtypes in PCA embeddings.

Baseline feature extractors demonstrate potential for subtype discovery.

Dataset enables semi-supervised ECG modeling.

Abstract

We release the largest public ECG dataset of continuous raw signals for representation learning containing 11 thousand patients and 2 billion labelled beats. Our goal is to enable semi-supervised ECG models to be made as well as to discover unknown subtypes of arrhythmia and anomalous ECG signal events. To this end, we propose an unsupervised representation learning task, evaluated in a semi-supervised fashion. We provide a set of baselines for different feature extractors that can be built upon. Additionally, we perform qualitative evaluations on results from PCA embeddings, where we identify some clustering of known subtypes indicating the potential for representation learning in arrhythmia sub-type discovery.