Weak Supervision for Affordable Modeling of Electrocardiogram Data

Abstract

Analysing electrocardiograms (ECGs) is an inexpensive and non-invasive, yet powerful way to diagnose heart disease. ECG studies using Machine Learning to automatically detect abnormal heartbeats so far depend on large, manually annotated datasets. While collecting vast amounts of unlabeled data can be straightforward, the point-by-point annotation of abnormal heartbeats is tedious and expensive. We explore the use of multiple weak supervision sources to learn diagnostic models of abnormal heartbeats via human designed heuristics, without using ground truth labels on individual data points. Our work is among the first to define weak supervision sources directly on time series data. Results show that with as few as six intuitive time series heuristics, we are able to infer high quality probabilistic label estimates for over 100,000 heartbeats with little human effort, and use the estimated labels to train competitive classifiers evaluated on held out test data.

Intelligent Clinical Decision Support

“…Answering such questions is often easier, yields more reliable labels, and requires less annotation effort to achieve equivalent performance. We have also demonstrated how to completely avoid the laborious process of pointillistic labeling of reference data for clinical applications of AI using weak supervision [ 49 ]. By harvesting multiple labeling functions that a human expert would use in their mind to assess each case at hand, one can automate the process of data annotation.…”

“…This is particularly appealing when facing large amounts of clinical data needing annotation. In one such exercise, we have shown that a handful of labeling functions derived from basic clinical knowledge can eliminate the need for manual data annotation and yield predictive models of performance comparable to the equivalent models trained on data point-by-point labeled by expert clinicians when evaluated on the task of detecting arrhythmia in ECG signals [ 49 ].…”