mlsanityData to 7 Oct 2026

Patient Contrastive Learning: a Performant, Expressive, and Practical Approach to ECG Modeling

arXiv 2104.04569 ↗First posted 9 Apr 2021cs.LG eess.SPcited by 24

Not on the map. Why?

Abstract (arXiv)

Supervised machine learning applications in health care are often limited due to a scarcity of labeled training data. To mitigate this effect of small sample size, we introduce a pre-training approach, Patient Contrastive Learning of Representations (PCLR), which creates latent representations of ECGs from a large number of unlabeled examples. The resulting representations are expressive, performant, and practical across a wide spectrum of clinical tasks. We develop PCLR using a large health care system with over 3.2 million 12-lead ECGs, and demonstrate substantial improvements across multiple new tasks when there are fewer than 5,000 labels. We release our model to extract ECG representations at https://github.com/broadinstitute/ml4h/tree/master/model_zoo/PCLR.

Key facts. Cited by 24 papers in the map (data through 7 Oct 2026). Builds on 15 of its 41 references that are in the map. First posted on arXiv 9 Apr 2021.

Builds on

15 of 41 references.

References · 15

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  3. 2016XGBoost: A Scalable Tree Boosting Systemcited by 5,120
  4. 2016SGDR: Stochastic Gradient Descent with Warm Restartscited by 4,993
  5. 2020What Makes for Good Views for Contrastive Learning?cited by 843
  6. 2020Contrastive Learning of Medical Visual Representations from Paired Images and Textcited by 636
  7. 2013Network In Networkcited by 600
  8. 2021Big Self-Supervised Models Advance Medical Image Classificationcited by 320
  9. 2020Contrastive learning of global and local features for medical image segmentation with limited annotationscited by 309
  10. 2019Revisiting Self-Supervised Visual Representation Learningcited by 250
  11. 2020CLOCS: Contrastive Learning of Cardiac Signals Across Space, Time, and Patientscited by 170
  12. 2019Automatic diagnosis of the 12-lead ECG using a deep neural networkcited by 164
3 more
  1. 2020Subject-Aware Contrastive Learning for Biosignalscited by 75
  2. 2017Cardiologist-Level Arrhythmia Detection with Convolutional Neural Networkscited by 52
  3. 2017Self-Supervised Learning for Spinal MRIscited by 24

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2021: 1; 2022: 1; 2023: 7; 2024: 4; 2025: 4; 2026: 7.

24 papers.

Sort:

By area:

  1. 2023Large-scale Training of Foundation Models for Wearable Biosignalscited by 89
  2. 2024RelCon: Relative Contrastive Learning for a Motion Foundation Model for Wearable Datacited by 17
  3. 2023Event-Based Contrastive Learning for Medical Time Seriescited by 13
  4. 2023Clinically Labeled Contrastive Learning for OCT Biomarker Classificationcited by 8
  5. 2021Leveraging Time Irreversibility with Order-Contrastive Pre-trainingcited by 7
  6. 2023Deep Metric Learning for the Hemodynamics Inference with Electrocardiogram Signalscited by 4
  7. 2024Contrastive Learning Is Not Optimal for Quasiperiodic Time Seriescited by 4
  8. 2024Unlocking Telemetry Potential: Self-Supervised Learning for Continuous Clinical Electrocardiogram Monitoringcited by 1
  9. 2024Estimating ECG Intervals from Lead-I Alone: External Validation of Supervised Modelscited by 1
  10. 2025Enhancing Contrastive Learning-based Electrocardiogram Pretrained Model with Patient Memory Queuecited by 1
  11. 2026Beyond Patient Invariance: Learning Cardiac Dynamics via Action-Conditioned JEPAscited by 1
  12. 2022Self-Supervised Mental Disorder Classifiers via Time Reversalcited by 0
12 more
  1. 2023Subject-based Non-contrastive Self-Supervised Learning for ECG Signal Processingcited by 0
  2. 2023A Comparison of Self-Supervised Pretraining Approaches for Predicting Disease Risk from Chest Radiograph Imagescited by 0
  3. 2023Learning Beyond Similarities: Incorporating Dissimilarities between Positive Pairs in Self-Supervised Time Series Learningcited by 0
  4. 2025CuPID: Leveraging Masked Single-Lead ECG Modelling for Enhancing the Representationscited by 0
  5. 2025Parallel-Learning of Invariant and Tempo-variant Attributes of Single-Lead Cardiac Signals: PLITAcited by 0
  6. 2025Leveraging Self-Supervised Learning Methods for Remote Screening of Subjects with Paroxysmal Atrial Fibrillationcited by 0
  7. 2026Position: Evaluation of ECG Representations Must Be Fixedcited by 0
  8. 2026How Do Electrocardiogram Models Scale?cited by 0
  9. 2026Patient-Aware Contrastive Learning Preserves Per-Patient Structure in RR-Interval Representationscited by 0
  10. 2026The Impact of Temporal Context Length and Encoding Strategies on Self-Supervised ECG Representation Learningcited by 0
  11. 2026Beat-Synchronous Tokenization for ECG Transformerscited by 0
  12. 2026Learning Cardiac Features: ECG Biometrics Across Time and~Exercisecited by 0