The non-invasive and easily accessible characteristics of electrocardiogram (ECG) attract many studies targeting AI-enabled cardiovascular-related disease screening tools based on ECG. However, the high cost of manual labels makes high-performance deep learning models challenging to obtain. Hence, we propose a new self-supervised representation learning framework, contrastive heartbeats (CT-HB), which learns general and robust electrocardiogram representations for efficient training on various downstream tasks. We employ a novel heartbeat sampling method to define positive and negative pairs of heartbeats for contrastive learning by utilizing the periodic and meaningful patterns of electrocardiogram signals. Using the CT-HB framework, the self-supervised learning model learns personalized heartbeat representations representing the specific cardiology context of a patient. Evaluations on public benchmark datasets and a private large-scale real-world dataset with multiple tasks demonstrate that the learned semantic representations result in better performance on downstream tasks and retain high performance while supervised learning suffers performance degradation with fewer supervised labels in downstream tasks.

Multi-task learning has increased in importance due to its superior performance by learning multiple different tasks simultaneously and its ability to perform several different tasks using a single model. In medical phenotyping, task labels are costly to acquire and might contain a certain degree of label noise. This decreases the efficiency of using additional human labels as auxiliary tasks when applying multi-task learning to medical phenotyping. In this work, we proposed an effective multi-task learning framework, CO-TASK, to boost multi-task learning performance by generating auxiliary tasks through COmbination of TASK Labels. The proposed CO-TASK framework generates auxiliary tasks without additional labeling effort, is robust to a certain degree of label noise, and can be applied in parallel with various multi-task learning techniques. We evaluated our performance using the CIFAR-MTL dataset and demonstrated its effectiveness in medical phenotyping using two large-scale ECG phenotyping datasets, an 18 diseases multi-label ECG-P18 dataset and an echocardiogram diagnostic from electrocardiogram dataset ECG-EchoLVH. On the CIFAR-MTL dataset, we doubled the average per-task performance gain of the multi-task learning model from 4.38% to 9.78%. With the proposed task-aware imbalance data sampler, the CO-TASK framework can effectively deal with the different imbalance ratios for the different tasks in electrocardiogram phenotyping datasets. The proposed framework combined with noisy annotations as minor tasks increased the sensitivity by 7.1% compared to the single-task model while maintaining the same specificity as the doctor annotations on the ECG-EchoLVH dataset.