mlsanityData to 7 Oct 2026

3KG: Contrastive Learning of 12-Lead Electrocardiograms using Physiologically-Inspired Augmentations

arXiv 2106.04452 ↗First posted 21 Apr 2021physics.med-ph cs.LG eess.SPcited by 36

Not on the map. Why?

Abstract (arXiv)

We propose 3KG, a physiologically-inspired contrastive learning approach that generates views using 3D augmentations of the 12-lead electrocardiogram. We evaluate representation quality by fine-tuning a linear layer for the downstream task of 23-class diagnosis on the PhysioNet 2020 challenge training data and find that 3KG achieves a 9.1% increase in mean AUC over the best self-supervised baseline when trained on 1% of labeled data. Our empirical analysis shows that combining spatial and temporal augmentations produces the strongest representations. In addition, we investigate the effect of this physiologically-inspired pretraining on downstream performance on different disease subgroups and find that 3KG makes the greatest gains for conduction and rhythm abnormalities. Our method allows for flexibility in incorporating other self-supervised strategies and highlights the potential for similar modality-specific augmentations for other biomedical signals.

Key facts. Cited by 36 papers in the map (data through 7 Oct 2026). Builds on 21 of its 44 references that are in the map. First posted on arXiv 21 Apr 2021.

Builds on

21 of 44 references.

References · 21

  1. 2014Adam: A Method for Stochastic Optimizationcited by 36,960
  2. 2020A Simple Framework for Contrastive Learning of Visual Representationscited by 11,979
  3. 2018Representation Learning with Contrastive Predictive Codingcited by 7,662
  4. 2019Momentum Contrast for Unsupervised Visual Representation Learningcited by 7,482
  5. 2020Bootstrap your own latent: A new approach to self-supervised Learningcited by 4,712
  6. 2020wav2vec 2.0: A Framework for Self-Supervised Learning of Speech Representationscited by 3,853
  7. 2020Improved Baselines with Momentum Contrastive Learningcited by 2,282
  8. 2020Big Self-Supervised Models are Strong Semi-Supervised Learnerscited by 1,375
  9. 2019Contrastive Multiview Codingcited by 1,297
  10. 2018Learning deep representations by mutual information estimation and maximizationcited by 1,265
  11. 2020What Makes for Good Views for Contrastive Learning?cited by 843
  12. 2015Cyclical Learning Rates for Training Neural Networkscited by 627
9 more
  1. 2020CLOCS: Contrastive Learning of Cardiac Signals Across Space, Time, and Patientscited by 170
  2. 2020Self-supervised ECG Representation Learning for Emotion Recognitioncited by 99
  3. 2021Self-supervised representation learning from 12-lead ECG datacited by 80
  4. 2020Subject-Aware Contrastive Learning for Biosignalscited by 75
  5. 2020MoCo-CXR: MoCo Pretraining Improves Representation and Transferability of Chest X-ray Modelscited by 61
  6. 2021MedAug: Contrastive learning leveraging patient metadata improves representations for chest X-ray interpretationcited by 52
  7. 2019Self-supervised Learning for ECG-based Emotion Recognitioncited by 32
  8. 2018Automated and Interpretable Patient ECG Profiles for Disease Detection, Tracking, and Discoverycited by 14
  9. 2020PCPs: Patient Cardiac Prototypescited by 1

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 5; 2023: 11; 2024: 9; 2025: 6; 2026: 5.

36 papers.

Sort:

  1. 2023Large-scale Training of Foundation Models for Wearable Biosignalscited by 89
  2. 2024Guiding Masked Representation Learning to Capture Spatio-Temporal Relationship of Electrocardiogramcited by 84
  3. 2023ECG-QA: A Comprehensive Question Answering Dataset Combined With Electrocardiogramcited by 52
  4. 2022Lead-agnostic Self-supervised Learning for Local and Global Representations of Electrocardiogramcited by 38
  5. 2024SleepFM: Multi-modal Representation Learning for Sleep Across Brain Activity, ECG and Respiratory Signalscited by 38
  6. 2024ECG Semantic Integrator (ESI): A Foundation ECG Model Pretrained with LLM-Enhanced Cardiological Textcited by 25
  7. 2024RelCon: Relative Contrastive Learning for a Motion Foundation Model for Wearable Datacited by 17
  8. 2023Sequential Multi-Dimensional Self-Supervised Learning for Clinical Time Seriescited by 16
  9. 2024Boosting Masked ECG-Text Auto-Encoders as Discriminative Learnerscited by 16
  10. 2023Towards quantitative precision for ECG analysis: Leveraging state space models, self-supervision and patient metadatacited by 15
  11. 2024Wearable Accelerometer Foundation Models for Health via Knowledge Distillationcited by 15
  12. 2023Event-Based Contrastive Learning for Medical Time Seriescited by 13
24 more
  1. 2022Data Augmentation for Electrocardiogramscited by 12
  2. 2022Analysis of Augmentations for Contrastive ECG Representation Learningcited by 12
  3. 2024ECG-Byte: A Tokenizer for End-to-End Generative Electrocardiogram Language Modelingcited by 9
  4. 2023In-Distribution and Out-of-Distribution Self-supervised ECG Representation Learning for Arrhythmia Detectioncited by 7
  5. 2023Deep Metric Learning for the Hemodynamics Inference with Electrocardiogram Signalscited by 4
  6. 2024Electrocardiogram-Language Model for Few-Shot Question Answering with Meta Learningcited by 4
  7. 2025Self-Alignment Learning to Improve Myocardial Infarction Detection from Single-Lead ECGcited by 4
  8. 2022Pretraining ECG Data with Adversarial Masking Improves Model Generalizability for Data-Scarce Taskscited by 3
  9. 2025Self-DANA: A Resource-Efficient Channel-Adaptive Self-Supervised Approach for ECG Foundation Modelscited by 3
  10. 2022LEAVES: Learning Views for Time-Series Biobehavioral Data in Contrastive Learningcited by 2
  11. 2023Which Augmentation Should I Use? An Empirical Investigation of Augmentations for Self-Supervised Phonocardiogram Representation Learningcited by 2
  12. 2023Riemannian Prediction of Anatomical Diagnoses in Congenital Heart Disease based on 12-lead ECGscited by 2
  13. 2026OC-Distill: Ontology-aware Contrastive Learning with Cross-Modal Distillation for ICU Risk Predictioncited by 2
  14. 2023Automated Cardiovascular Record Retrieval by Multimodal Learning between Electrocardiogram and Clinical Reportcited by 1
  15. 2024Finding "Good Views" of Electrocardiogram Signals for Inferring Abnormalities in Cardiac Conditioncited by 1
  16. 2025Learning ECG Representations via Poly-Window Contrastive Learningcited by 1
  17. 2025ECG-Soup: Harnessing Multi-Layer Synergy for ECG Foundation Modelscited by 1
  18. 2025Domain Knowledge is Power: Leveraging Physiological Priors for Self Supervised Representation Learning in Electrocardiographycited by 1
  19. 2023ElectroCardioGuard: Preventing Patient Misidentification in Electrocardiogram Databases through Neural Networkscited by 0
  20. 2025Contrastive Cross-Modal Learning for Infusing Chest X-ray Knowledge into ECGscited by 0
  21. 2026Beat-ssl: Capturing Local ECG Morphology through Heartbeat-level Contrastive Learning with Soft Targetscited by 0
  22. 2026Position: Evaluation of ECG Representations Must Be Fixedcited by 0
  23. 2026Information-theoretic Multimodal Representation Learning for Electrocardiogram Signalscited by 0
  24. 2026ReMAP: Self-supervised learning to unveil brain representations and vulnerabilitycited by 0