Pith. sign in

REVIEW 4 major objections 4 minor 69 references

Synthetic ECG Generation for Data Augmentation and Transfer Learning in Arrhythmia Classification

T0 review · 4 major / 4 minor · reviewed 2026-08-12 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read This paper tests whether synthetic ECG data generated by three deep generative models can augment real arrhythmia datasets and support transfer learning, and finds that gains appear only when the real datasets are merged.

desk verdict A useful but under-specified empirical comparison of three ECG generative models; the augmentation gains are marginal and the transfer-learning negative rests on a frozen-feature protocol. read the letter →

arxiv 2411.18456 v1 pith:ROIQRVS4 submitted 2024-11-27 cs.LG cs.AI

José Fernando Núñez , Jamie Arjona , Javier Béjar This is my paper · ORCID
classification cs.LGcs.AI
keywords syntheticECGdataaugmentationtransferlearningarrhythmiaclassificationdiffusionmodelsTime-VQVAEPTB-XLChapmandataset
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper asks whether synthetic electrocardiogram (ECG) traces generated by deep learning models can stand in for scarce, privacy-protected real recordings. The authors train three generative models—Diffweave, Time-Diffusion, and Time-VQVAE—on the PTB-XL and Chapman datasets, then use the synthetic samples to augment arrhythmia classification and to pretrain a classifier that is fine-tuned with increasing amounts of real data. The central finding is conditional: augmenting either dataset alone with synthetic samples gives barely noticeable improvements, but when both datasets are merged, every evaluated metric improves across classifiers. For transfer learning, synthetic pretraining never matches a classifier trained only on real data, with Time-VQVAE the least bad option. The paper also shows that common distribution-closeness measures such as MMD and dimensionality-reduction visualizations cannot separate synthetic from real ECGs, while classifier-based discrimination can.

What carries the argument

The machinery is a three-way comparison of generative architectures plus a split-train/evaluate protocol. Diffweave is a non-autoregressive diffusion model originally built for audio; Time-Diffusion is a U-Net-shaped diffusion model with a label-conditioning scheme; Time-VQVAE is a vector-quantized variational autoencoder that learns discrete low- and high-frequency latent codes in the STFT domain and samples them with a bidirectional transformer prior. Each model is trained per dataset and on the merged dataset. The classification evaluation uses a 1D residual CNN with hyperparameters tuned per dataset, and applies five train/test splits—real-only, synthetic-only, train-synthetic/evaluate-real, train-real/evaluate-synthetic, and real-plus-synthetic train/evaluate-real—plus a fine-tuning protocol that freezes feature layers and retrains only the dense classifier on real data in 20% increments. The splits are what make the boundary visible: synthetic-only classifiers look good on synthetic tests but collapse on real tests, while augmentation gains appear only in the merged-data real-plus-synthetic setting.

What would settle it

Unfreeze all layers during fine-tuning on real data and rerun the transfer experiments; if a synthetically pretrained model then reaches or exceeds the real-only classifier once it sees enough real data, the paper's negative transfer-learning conclusion is overturned.

Watch

Extended reading notes

Core claim

On the paper's own terms, the discovery is that current state-of-the-art generative models produce ECG samples that visually and statistically resemble real ones, yet the samples are only useful for data augmentation when the real training pool already spans diverse sources. In the train-on-real-plus-synthetic/evaluate-on-real setting, merging PTB-XL and Chapman before augmentation raises accuracy, precision, recall, f1, and ROC AUC across all three generators compared with real-only training; the same augmentation on either dataset individually leaves scores essentially unchanged. In transfer learning, a model pretrained on synthetic data and then fine-tuned on increasing fractions of real data—with all but the classifier layers frozen—consistently underperforms a model trained on real data alone, and Time-VQVAE's samples give the best fine-tuned scores of the three. The paper thus establishes a precise boundary: synthetic ECG data helps modestly as an additive to a merged real corpus, but cannot substitute for real data or bootstrap a classifier to real-data-level performance under this protocol.

Load-bearing premise

The claim that synthetic pretraining cannot match real-data training depends on the fine-tuning protocol that freezes all feature-extraction layers and retrains only the dense classifier; if those layers were allowed to adapt to real data, the gap might narrow or disappear.

Editorial extensions

If this is right

  • When both PTB-XL and Chapman are merged, adding synthetic samples from any of the three generators improves every reported classification metric under the real-plus-synthetic training setting.
  • Time-VQVAE is the generator of choice for transfer: it yields the highest fine-tuned scores across datasets and consistently reduces false negatives when synthetic-trained models are evaluated on real data.
  • None of the tested generators produces data that can replace real recordings: train-on-synthetic/evaluate-on-real scores remain far below real-only baselines.
  • Classifier-based discrimination is a stricter and more informative quality check than MMD or dimensionality-reduction visualization, which fail to separate synthetic from real ECGs.
  • Fine-tuning a synthetically pretrained model converges in about half the time of training from scratch on real data, so the practical payoff of synthetic pretraining is computational speed, not accuracy.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • The paper's frozen-feature fine-tuning protocol may understate transferability; a natural extension is to allow the convolutional feature extractor to adapt, which would test whether the negative result is a property of the data or of the protocol.
  • The fact that augmentation helps only on the merged dataset suggests sample diversity across recording sources matters more than raw sample count; this could be tested by holding the number of real training samples fixed while varying the number of datasets.
  • The classifier-based real-vs-synthetic discrimination that works in the quality evaluation could double as an authorship detector for generative physiological signals, a direction the paper mentions only as a closing remark.
  • The same comparison could be run on EEG or other periodic physiological signals to determine whether the gap between synthetic and real data is specific to ECG or general to time-series generation.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 4 minor

Summary. The paper studies whether synthetic ECG signals generated by three deep generative models (Diffweave, Time-Diffusion, and Time-VQVAE) can be used for data augmentation and transfer learning in arrhythmia classification. The authors train the generative models on PTB-XL, CHAPMAN, and a merged dataset, and evaluate the synthetic data by training a 1D CNN classifier under different train/test combinations (real, synthetic, and hybrid settings), as well as by fine-tuning a synthetically pretrained classifier with increasing proportions of real data. They report that augmentation with synthetic data gives barely noticeable improvements on individual datasets, but that on the merged dataset using synthetic samples as augmented data increases all metrics; they also report that synthetic pretraining followed by head-only fine-tuning does not reach the performance of real-data-only training, with Time-VQVAE performing best among the generative models.

Significance. If the results were fully supported, the paper would provide a useful empirical benchmark on the value of state-of-the-art generative models for ECG augmentation and a cautionary negative result on synthetic pretraining. The study is valuable in scope: it uses two widely used public datasets, compares three open-source generative models, and runs the classification experiments 25 times with multiple metrics. However, the absence of variance reporting, the unspecified synthetic-to-real mixing ratio, and the linear-probe-only transfer protocol mean that the two headline claims are not yet quantitatively established. The work could become a solid reference for the community after these issues are addressed, but in its current form the evidence is suggestive rather than conclusive.

major comments (4)
  1. [Section 9.3, Table 9] The central positive claim that merging PTB-XL and CHAPMAN with synthetic data increases all metrics under the TrRSTeR setting is not quantitatively supported: the manuscript never reports how many synthetic samples were added relative to the real training set, so the improvement from 0.8574 to 0.8673 accuracy could be due to dataset size or class-balance changes rather than to the quality of the synthetic data. In addition, despite stating in Section 8 that all experiments were repeated n=25 times, Tables 5-10 report only point estimates with no standard deviations or confidence intervals, making it impossible to judge whether the observed differences are statistically significant. The authors should report the mixing ratio and the variance of the 25 repeats.
  2. [Section 7.1, item 2; Tables 6, 8, 10] The transfer-learning conclusion is based on a protocol that freezes all layers except the dense classifier head during fine-tuning. This is a linear-probe evaluation: it tests how linearly separable the frozen synthetic-pretrained features are, not whether the pretrained model can be adapted to real data. The paper does not run the standard full fine-tuning alternative in which the feature extractor is also updated, so the negative claim that synthetic pretraining is 'not powerful enough to achieve results close to a classifier trained with real data only' is underdetermined. The authors should either report full fine-tuning results or substantially weaken the conclusion in the abstract and Section 10.1.
  3. [Section 8] The experimental setup omits details necessary for reproducibility and for assessing potential leakage or bias: there is no description of how the 25 repeats were constructed, how the train/test splits were stratified (e.g., per patient), whether the same test set was reused across repeats and across generative models, or whether the 500 Hz data were resampled to 100 Hz and how the signals were preprocessed (filtering, normalization, segmentation). These details are essential because the comparison between generative models and between real and synthetic settings depends on identical evaluation conditions.
  4. [Section 7, Section 10.3] The paper introduces several quality metrics (2-sample test classification score, MMD, and dimensionality-reduction visualizations) but never reports their quantitative results; Section 10.3 then draws conclusions such as 'MMD and visualization techniques lack the power to differentiate synthetic data from real one' without showing the corresponding data. If these metrics are part of the evaluation, their results should be presented; otherwise the claims in Section 10.3 should be removed or explicitly marked as qualitative impressions.
minor comments (4)
  1. [Section 5.1] The sentence listing the diffusion models is a fragment: 'We opted for three different architectures, namely Diffwave, [54] Diffusion-TS [55] and the Unet1D conditional model from Huggingface.' Also, the name 'Diffweave' in the abstract and Table 5 is inconsistent with 'Diffwave' used in the text.
  2. [Section 9.1] There is a typo in 'Excution Time' (should be 'Execution Time') in Table 6 and in the surrounding discussion.
  3. [Section 9.3] The sentence 'because PTB-XL and CHAPMAN have different number of samples in each class (4 and it is expected to have an effect...' is garbled and should be rewritten.
  4. [Section 7] The description of the 2-sample test classification score is ambiguous: it says 'A classifier trained with real data is used in order to discriminate between real data and synthetic data,' but a classifier trained only on real data cannot be used to discriminate real from synthetic without also training on synthetic labels; please clarify the intended procedure.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the evaluation is empirical and the generative models are fitted in the standard way, with only a protocol limitation affecting the transfer-learning claim.

full rationale

The paper's central claims are empirical measurements rather than derivations that reduce to their own inputs. The generative models (Diffweave, Time-Diffusion, Time-VQVAE) are fitted on real ECG training data, and the subsequent classification evaluations are independent out-of-sample measurements of generalization, not predictions forced by construction. The data-augmentation results compare classifiers trained on real, synthetic, and mixed data; the reported improvements on the merged PTB-XL+CHAPMAN dataset are empirical findings, not identities or fitted parameters renamed as predictions. The transfer-learning conclusion is based on a fine-tuning protocol in Section 7.1 that freezes all layers except the classifier head, which may understate the transferability of synthetic pretraining, but this is a validity limitation rather than a circular reduction. The only self-citation is reference [19], which concerns parking-system sensor data and is unrelated to the ECG claims. No uniqueness theorem, ansatz, or known result is smuggled in via citation, and no derivation step is equivalent to its input by definition.

Assumptions & free parameters 4 free parameters · 4 assumptions · 0 invented entities

The central claims depend on many unreported experimental choices: merged-dataset hyperparameters, generative training configurations, the augmentation ratio, and data splits. The paper also assumes the two datasets' class labels align and that the frozen-feature fine-tuning protocol is a fair test of transfer learning.

free parameters (4)
  • PTB-XL classifier hyperparameters = 6 conv blocks, 32 kernels, filter 7x7, 256 neurons, 3 layers, lr 0.000354, dropout 0.4743
    Obtained by tree-structured parzen estimator on real PTB-XL data (Table 3); these values are used for all subsequent synthetic-data evaluations on PTB-XL.
  • CHAPMAN classifier hyperparameters = 5 conv blocks, 16 kernels, filter 7x7, 256 neurons, 3 layers, lr 0.000158, dropout 0.3776
    Obtained by Optuna on real CHAPMAN data (Table 3); used for all CHAPMAN synthetic evaluations.
  • Merged dataset classifier hyperparameters = not reported
    No hyperparameters are given for the PTB-XL plus CHAPMAN merged dataset, so it is unclear which configuration was used and whether it was retuned.
  • Generative model training configurations = not reported
    Training details for Diffweave/SSSD-ECG, Time-Diffusion, and Time-VQVAE, including diffusion steps, learning rates, batch sizes, and conditioning embeddings, are not provided; these choices directly affect synthetic sample quality.
assumptions (4)
  • domain assumption The seven selected classes are semantically aligned between PTB-XL and CHAPMAN annotation schemes.
    Section 8 and Table 4 merge samples from both datasets by class ID; if the labeling criteria differ, the merged dataset and its augmentation results are confounded.
  • domain assumption Freezing all layers except the classifier's dense layers is a valid transfer-learning evaluation protocol.
    Section 7.1 item 2; this protocol prevents feature adaptation to real data and may understate the value of synthetic pretraining.
  • domain assumption The generative models produce correctly class-conditioned synthetic samples.
    The augmentation and transfer evaluations assume generated samples carry the requested class labels; conditioning accuracy is not evaluated.
  • domain assumption The unspecified preprocessing, including resampling and signal length, preserves the diagnostic content of the ECG signals.
    PTB-XL and CHAPMAN have different native sampling rates (100/500 Hz and 500 Hz); without a stated common preprocessing, the merged-data results could be an artifact of the pipeline.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Synthetic ECG Generation for Data Augmentation and Transfer Learning in Arrhythmia Classification." pith.science (2026). https://pith.science/paper/ROIQRVS4

@misc{pith2026241118456,
  author       = {Pith},
  title        = {Pith review of: Synthetic ECG Generation for Data Augmentation and Transfer Learning in Arrhythmia Classification},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/ROIQRVS4}},
  note         = {Machine review of arXiv:2411.18456}
}
read the original abstract

Deep learning models need a sufficient amount of data in order to be able to find the hidden patterns in it. It is the purpose of generative modeling to learn the data distribution, thus allowing us to sample more data and augment the original dataset. In the context of physiological data, and more specifically electrocardiogram (ECG) data, given its sensitive nature and expensive data collection, we can exploit the benefits of generative models in order to enlarge existing datasets and improve downstream tasks, in our case, classification of heart rhythm. In this work, we explore the usefulness of synthetic data generated with different generative models from Deep Learning namely Diffweave, Time-Diffusion and Time-VQVAE in order to obtain better classification results for two open source multivariate ECG datasets. Moreover, we also investigate the effects of transfer learning, by fine-tuning a synthetically pre-trained model and then progressively adding increasing proportions of real data. We conclude that although the synthetic samples resemble the real ones, the classification improvement when simply augmenting the real dataset is barely noticeable on individual datasets, but when both datasets are merged the results show an increase across all metrics for the classifiers when using synthetic samples as augmented data. From the fine-tuning results the Time-VQVAE generative model has shown to be superior to the others but not powerful enough to achieve results close to a classifier trained with real data only. In addition, methods and metrics for measuring closeness between synthetic data and the real one have been explored as a side effect of the main research questions of this study.

Figures

Figures reproduced from arXiv: 2411.18456 by the authors.

Figure 1
Figure 1. 1 dimensional deep residual convolutional classifier architecture [PITH_FULL_IMAGE:figures/full_fig_p009_1.png] view at source ↗

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

69 extracted references · 57 canonical work pages

  1. [1]

    A Systematic Review of Time Series Classification Techniques Used in Biomedical Applications,

    W. K. Wang, I. Chen, L. Hershkovich, J. Yang, A. Shetty, G. Singh, Y . Jiang, A. Kotla, J. Z. Shang, R. Yerrabelli, A. R. Roghanizad, M. M. H. Shandhi, and J. Dunn, “A Systematic Review of Time Series Classification Techniques Used in Biomedical Applications,”Sensors, vol. 22, p. 8016, Jan. 2022

    work page 2022

  2. [2]

    Ethical Challenges Posed by Big Data,

    E. G. Howe III and F. Elenberg, “Ethical Challenges Posed by Big Data,” Innov Clin Neurosci, vol. 17, pp. 24–30, Oct. 2020

    work page 2020

  3. [3]

    Biomedical Data Sharing and Reuse: Attitudes and Practices of Clinical and Scientific Research Staff,

    L. M. Federer, Y .-L. Lu, D. J. Joubert, J. Welsh, and B. Brandys, “Biomedical Data Sharing and Reuse: Attitudes and Practices of Clinical and Scientific Research Staff,” PLOS ONE, vol. 10, p. e0129506, June 2015

    work page 2015

  4. [4]

    Handling limited datasets with neural networks in medical appli- cations: A small-data approach,

    T. Shaikhina and N. A. Khovanova, “Handling limited datasets with neural networks in medical appli- cations: A small-data approach,” Artificial Intelligence in Medicine, vol. 75, pp. 51–63, Jan. 2017

    work page 2017

  5. [5]

    Commentary: The Problem of Class Imbalance in Biomedical Data,

    H. Ishwaran and R. O’Brien, “Commentary: The Problem of Class Imbalance in Biomedical Data,” J Thorac Cardiovasc Surg, vol. 161, pp. 1940–1941, June 2021

    work page 1940

  6. [6]

    Data Augmentation techniques in time series domain: A survey and taxonomy

    E. Talavera, G. Iglesias, Á. González-Prieto, A. Mozo, and S. Gómez-Canaval, “Data augmentation techniques in time series domain: A survey and taxonomy,”arXiv preprint arXiv:2206.13508, 2022

    work page Pith review arXiv 2022

  7. [7]

    On the Challenges and Opportunities in Generative AI,

    L. Manduchi, K. Pandey, R. Bamler, R. Cotterell, S. Däubener, S. Fellenz, A. Fischer, T. Gärt- ner, M. Kirchler, M. Kloft, Y . Li, C. Lippert, G. de Melo, E. Nalisnick, B. Ommer, R. Ranganath, M. Rudolph, K. Ullrich, G. V . den Broeck, J. E. V ogt, Y . Wang, F. Wenzel, F. Wood, S. Mandt, and V . Fortuin, “On the Challenges and Opportunities in Generative ...

    work page 2024

  8. [8]

    Deep generative modelling: A comparative review of vaes, gans, normalizing flows, energy-based and autoregressive models,

    S. Bond-Taylor, A. Leach, Y . Long, and C. G. Willcocks, “Deep generative modelling: A comparative review of vaes, gans, normalizing flows, energy-based and autoregressive models,” IEEE transactions on pattern analysis and machine intelligence, 2021

    work page 2021

Show all 69 references
  1. [9]

    A review on generative adversarial networks: Algorithms, the- ory, and applications,

    J. Gui, Z. Sun, Y . Wen, D. Tao, and J. Ye, “A review on generative adversarial networks: Algorithms, the- ory, and applications,”IEEE Transactions on Knowledge and Data Engineering, vol. 35, no. 4, pp. 3313– 3332, 2023

    2023

  2. [10]

    Autoregressive models: What are they good for?,

    M. Dalal, A. C. Li, and R. Taori, “Autoregressive models: What are they good for?,” 2019

    2019

  3. [11]

    Normalizing flows for probabilistic modeling and inference,

    G. Papamakarios, E. Nalisnick, D. J. Rezende, S. Mohamed, and B. Lakshminarayanan, “Normalizing flows for probabilistic modeling and inference,” J. Mach. Learn. Res., vol. 22, jan 2021

    2021

  4. [12]

    Diffusion models: A comprehensive survey of methods and applica- tions,

    L. Yang, Z. Zhang, and S. Hong, “Diffusion models: A comprehensive survey of methods and applica- tions,” arXiv preprint arXiv:2209.00796, 2022

    2022

  5. [13]

    Auto-encoding variational bayes,

    D. P. Kingma and M. Welling, “Auto-encoding variational bayes,” 2022

    2022

  6. [14]

    Generative adversarial networks,

    I. J. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair, A. Courville, and Y . Bengio, “Generative adversarial networks,” 2014

    2014

  7. [15]

    Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: dependence on recording region and brain state,

    R. G. Andrzejak, K. Lehnertz, F. Mormann, C. Rieke, P. David, and C. E. Elger, “Indications of nonlinear deterministic and finite-dimensional structures in time series of brain electrical activity: dependence on recording region and brain state,” Physical Review. E, Statistica...

    2001

  8. [16]

    Stock Market Analysis Using Time Series Relational Models for Stock Price Prediction,

    C. Zhao, P. Hu, X. Liu, X. Lan, and H. Zhang, “Stock Market Analysis Using Time Series Relational Models for Stock Price Prediction,” Mathematics, vol. 11, p. 1130, Jan. 2023. Number: 5 Publisher: Multidisciplinary Digital Publishing Institute

    2023

  9. [17]

    Time Series Prediction in Industry 4.0: A Com- prehensive Review and Prospects for Future Advancements,

    N. Kashpruk, C. Piskor-Ignatowicz, and J. Baranowski, “Time Series Prediction in Industry 4.0: A Com- prehensive Review and Prospects for Future Advancements,” Applied Sciences, vol. 13, p. 12374, Jan

  10. [18]

    Trend analysis of climate time series: A review of methods,

    M. Mudelsee, “Trend analysis of climate time series: A review of methods,” Earth-Science Reviews, vol. 190, pp. 310–322, Mar. 2019

    2019

  11. [19]

    Characterizing parking systems from sensor data through a data-driven approach,

    J. Arjona Martinez, M. P. Linares, and J. Casanovas, “Characterizing parking systems from sensor data through a data-driven approach,” Transportation Letters, vol. 13, pp. 183–192, Mar. 2021. Publisher: Taylor & Francis _eprint: https://doi.org/10.1080/19427867.2020.1866331

    2021

  12. [20]

    Electrocardiogram generation with a bidirectional LSTM- CNN generative adversarial network,

    F. Zhu, F. Ye, Y . Fu, Q. Liu, and B. Shen, “Electrocardiogram generation with a bidirectional LSTM- CNN generative adversarial network,”Sci Rep, vol. 9, p. 6734, May 2019. Number: 1 Publisher: Nature Publishing Group

    2019

  13. [21]

    Quick and Easy Time Series Generation with Established Image- based GANs,

    E. Brophy, Z. Wang, and T. E. Ward, “Quick and Easy Time Series Generation with Established Image- based GANs,” Oct. 2019. arXiv:1902.05624 [cs, stat]

    2019 arXiv

  14. [22]

    Synthesis of Realistic ECG using Generative Adversarial Networks,

    A. M. Delaney, E. Brophy, and T. E. Ward, “Synthesis of Realistic ECG using Generative Adversarial Networks,” Sept. 2019. arXiv:1909.09150 [cs, eess, stat]

    2019 arXiv

  15. [23]

    Pgans: Personalized generative adversarial networks for ecg synthesis to improve patient-specific deep ecg classification,

    T. Golany and K. Radinsky, “Pgans: Personalized generative adversarial networks for ecg synthesis to improve patient-specific deep ecg classification,” Proceedings of the AAAI Conference on Artificial Intelligence, vol. 33, pp. 557–564, July 2019. Number: 01

    2019

  16. [24]

    ECG Arrhythmias Detection Using Auxiliary Classifier Gen- erative Adversarial Network and Residual Network,

    P. Wang, B. Hou, S. Shao, and R. Yan, “ECG Arrhythmias Detection Using Auxiliary Classifier Gen- erative Adversarial Network and Residual Network,” IEEE Access, vol. 7, pp. 100910–100922, 2019. Conference Name: IEEE Access

    2019

  17. [25]

    Investigating Deep Convolution Conditional GANs for Electrocardio- gram Generation,

    D. Nankani and R. D. Baruah, “Investigating Deep Convolution Conditional GANs for Electrocardio- gram Generation,” in 2020 International Joint Conference on Neural Networks (IJCNN) , pp. 1–8, July

    2020

  18. [26]

    ECG signal generation based on conditional generative models,

    Y . Xia, W. Wang, and K. Wang, “ECG signal generation based on conditional generative models,” Biomedical Signal Processing and Control, vol. 82, p. 104587, Apr. 2023

    2023

  19. [27]

    Synthetic ECG Signal Generation Using Proba- bilistic Diffusion Models,

    E. Adib, A. S. Fernandez, F. Afghah, and J. J. Prevost, “Synthetic ECG Signal Generation Using Proba- bilistic Diffusion Models,” IEEE Access, vol. 11, pp. 75818–75828, 2023

    2023

  20. [28]

    Synthesis of standard 12-lead electrocardiograms using two- dimensional generative adversarial networks,

    Y .-H. Zhang and S. Babaeizadeh, “Synthesis of standard 12-lead electrocardiograms using two- dimensional generative adversarial networks,”J Electrocardiol, vol. 69, pp. 6–14, 2021

    2021

  21. [29]

    PTB-XL, a large publicly available electrocardiography dataset,

    P. Wagner, N. Strodthoff, R.-D. Bousseljot, W. Samek, and T. Schaeffter, “PTB-XL, a large publicly available electrocardiography dataset,” Apr. 2020

    2020

  22. [30]

    Chinese Cardiovascular Disease Database (CCDD) and Its Management Tool,

    J.-w. Zhang, L.-p. Wang, X. Liu, H.-h. Zhu, and J. Dong, “Chinese Cardiovascular Disease Database (CCDD) and Its Management Tool,” in 2010 IEEE International Conference on BioInformatics and BioEngineering, (Philadelphia, PA, USA), pp. 66–72, IEEE, May 2010

    2010

  23. [31]

    A 12-lead electrocardiogram database for arrhythmia research covering more than 10,000 patients,

    G. Hangyuan, “A 12-lead electrocardiogram database for arrhythmia research covering more than 10,000 patients,” Nov. 2019. Publisher: figshare

    2019

  24. [32]

    Deepfake electrocardiograms using generative adver- sarial networks are the beginning of the end for privacy issues in medicine,

    V . Thambawita, J. L. Isaksen, S. A. Hicks, J. Ghouse, G. Ahlberg, A. Linneberg, N. Grarup, C. Ellervik, M. S. Olesen, T. Hansen, C. Graff, N.-H. Holstein-Rathlou, I. Strümke, H. L. Hammer, M. M. Maleckar, P. Halvorsen, M. A. Riegler, and J. K. Kanters, “Deepfake electrocardio...

    2021

  25. [33]

    Multivariate Generative Adversarial Networks and Their Loss Functions for Synthesis of Multichannel ECGs,

    E. Brophy, M. De V os, G. Boylan, and T. Ward, “Multivariate Generative Adversarial Networks and Their Loss Functions for Synthesis of Multichannel ECGs,” IEEE Access, vol. 9, pp. 158936–158945,

  26. [34]

    TTS-CGAN: A Transformer Time-Series Conditional GAN for Biosignal Data Augmentation,

    X. Li, A. H. H. Ngu, and V . Metsis, “TTS-CGAN: A Transformer Time-Series Conditional GAN for Biosignal Data Augmentation,” June 2022. arXiv:2206.13676 [cs]

    2022 arXiv

  27. [35]

    Diffusion-based conditional ECG generation with structured state space models,

    J. M. L. Alcaraz and N. Strodthoff, “Diffusion-based conditional ECG generation with structured state space models,” Computers in Biology and Medicine, vol. 163, p. 107115, Sept. 2023

    2023

  28. [36]

    ECG Synthesis via Diffusion-Based State Space Augmented Trans- former,

    M. H. Zama and F. Schwenker, “ECG Synthesis via Diffusion-Based State Space Augmented Trans- former,” Sensors, vol. 23, p. 8328, Jan. 2023. Number: 19 Publisher: Multidisciplinary Digital Publish- ing Institute

    2023

  29. [37]

    Automatic classification of heartbeats using ECG mor- phology and heartbeat interval features,

    P. de Chazal, M. O’Dwyer, and R. B. Reilly, “Automatic classification of heartbeats using ECG mor- phology and heartbeat interval features,” IEEE Trans Biomed Eng, vol. 51, pp. 1196–1206, July 2004

    2004

  30. [38]

    Support vector machine-based expert system for reliable heartbeat recognition,

    S. Osowski, L. T. Hoai, and T. Markiewicz, “Support vector machine-based expert system for reliable heartbeat recognition,” IEEE Trans Biomed Eng, vol. 51, pp. 582–589, Apr. 2004

    2004

  31. [39]

    ECG arrhythmia classification based on optimum-path forest,

    E. J. d. S. Luz, T. M. Nunes, V . H. C. de Albuquerque, J. P. Papa, and D. Menotti, “ECG arrhythmia classification based on optimum-path forest,”Expert Systems with Applications, vol. 40, pp. 3561–3573, July 2013

    2013

  32. [40]

    Electrocardiogram Classifica- tion Using Reservoir Computing With Logistic Regression,

    M. A. Escalona-Moran, M. C. Soriano, I. Fischer, and C. R. Mirasso, “Electrocardiogram Classifica- tion Using Reservoir Computing With Logistic Regression,” IEEE J. Biomed. Health Inform. , vol. 19, pp. 892–898, May 2015

    2015

  33. [41]

    Patient-Specific ECG Classification Based on Recurrent Neural Networks and Clustering Technique,

    C. Zhang, G. Wang, J. Zhao, P. Gao, J. Lin, and H. Yang, “Patient-Specific ECG Classification Based on Recurrent Neural Networks and Clustering Technique,” in Biomedical Engineering, ACTA Press, Mar. 2017

    2017

  34. [42]

    Automated detection of atrial fibrillation using long short-term memory network with RR interval signals,

    O. Faust, A. Shenfield, M. Kareem, T. R. San, H. Fujita, and U. R. Acharya, “Automated detection of atrial fibrillation using long short-term memory network with RR interval signals,” Comput Biol Med, vol. 102, pp. 327–335, Nov. 2018

    2018

  35. [43]

    Multiclass classification of myocardial infarction with convolutional and recurrent neural networks for portable ECG devices,

    H. W. Lui and K. L. Chow, “Multiclass classification of myocardial infarction with convolutional and recurrent neural networks for portable ECG devices,”Informatics in Medicine Unlocked, vol. 13, pp. 26– 33, Jan. 2018

    2018

  36. [44]

    A deep learning approach for ECG-based heartbeat classification for arrhythmia detection,

    G. Sannino and G. De Pietro, “A deep learning approach for ECG-based heartbeat classification for arrhythmia detection,” Future Generation Computer Systems, vol. 86, pp. 446–455, Sept. 2018

    2018

  37. [45]

    Automatic QRS complex detection using two-level convolutional neural network,

    Y . Xiang, Z. Lin, and J. Meng, “Automatic QRS complex detection using two-level convolutional neural network,” Biomed Eng Online, vol. 17, p. 13, Jan. 2018

    2018

  38. [46]

    Automated arrhythmia classification based on a com- bination network of CNN and LSTM,

    C. Chen, Z. Hua, R. Zhang, G. Liu, and W. Wen, “Automated arrhythmia classification based on a com- bination network of CNN and LSTM,” Biomedical Signal Processing and Control, vol. 57, p. 101819, Mar. 2020

    2020

  39. [47]

    DeepArrNet: An Efficient Deep CNN Architecture for Au- tomatic Arrhythmia Detection and Classification From Denoised ECG Beats,

    T. Mahmud, S. A. Fattah, and M. Saquib, “DeepArrNet: An Efficient Deep CNN Architecture for Au- tomatic Arrhythmia Detection and Classification From Denoised ECG Beats,” IEEE Access , vol. 8, pp. 104788–104800, 2020

    2020

  40. [48]

    Deep Learning for ECG Analysis: Benchmarks and Insights from PTB-XL,

    N. Strodthoff, P. Wagner, T. Schaeffter, and W. Samek, “Deep Learning for ECG Analysis: Benchmarks and Insights from PTB-XL,” Apr. 2020. arXiv:2004.13701 [cs, stat]

    2020 arXiv

  41. [49]

    The impact of the MIT-BIH Arrhythmia Database,

    G. Moody and R. Mark, “The impact of the MIT-BIH Arrhythmia Database,” IEEE Engineering in Medicine and Biology Magazine, vol. 20, pp. 45–50, May 2001. Conference Name: IEEE Engineering in Medicine and Biology Magazine

    2001

  42. [50]

    An Open Access Database for Evaluating the Algorithms of Electrocardiogram Rhythm and Morphology Abnormality Detection,

    F. Liu, C. Liu, L. Zhao, X. Zhang, X. Wu, X. Xu, Y . Liu, C. Ma, S. Wei, Z. He, J. Li, and E. N. Yin Kwee, “An Open Access Database for Evaluating the Algorithms of Electrocardiogram Rhythm and Morphology Abnormality Detection,” j med imaging hlth inform, vol. 8, pp. 1368–1373...

    2018

  43. [51]

    Deep Learning- Based ECG Arrhythmia Classification: A Systematic Review,

    Q. Xiao, K. Lee, S. A. Mokhtar, I. Ismail, A. L. b. M. Pauzi, Q. Zhang, and P. Y . Lim, “Deep Learning- Based ECG Arrhythmia Classification: A Systematic Review,” Applied Sciences, vol. 13, p. 4964, Jan

  44. [52]

    Deep learning for ECG Arrhythmia detection and classification: an overview of progress for period 2017–2023,

    Y . Ansari, O. Mourad, K. Qaraqe, and E. Serpedin, “Deep learning for ECG Arrhythmia detection and classification: an overview of progress for period 2017–2023,”Front Physiol, vol. 14, p. 1246746, Sept. 2023

    2017

  45. [53]

    A 12-lead electrocardiogram database for arrhythmia research covering more than 10,000 patients,

    J. Zheng, J. Zhang, S. Danioko, H. Yao, H. Guo, and C. Rakovski, “A 12-lead electrocardiogram database for arrhythmia research covering more than 10,000 patients,” Sci Data, vol. 7, p. 48, Feb. 2020. Pub- lisher: Nature Publishing Group

    2020

  46. [54]

    Number: 8 Publisher: Multidisciplinary Digital Publishing Institute

  47. [55]

    Diffusion-ts: Interpretable diffusion for general time series generation,

    X. Yuan and Y . Qiao, “Diffusion-ts: Interpretable diffusion for general time series generation,” arXiv preprint arXiv:2403.01742, 2024

    2024 arXiv

  48. [56]

    U-net: Convolutional networks for biomedical image segmen- tation,

    O. Ronneberger, P. Fischer, and T. Brox, “U-net: Convolutional networks for biomedical image segmen- tation,” 2015

    2015

  49. [57]

    DiffWave: A Versatile Diffusion Model for Audio Synthesis,

    Z. Kong, W. Ping, J. Huang, K. Zhao, and B. Catanzaro, “DiffWave: A Versatile Diffusion Model for Audio Synthesis,” Mar. 2021. arXiv:2009.09761 [cs, eess, stat]

    2021 arXiv

  50. [58]

    Neural discrete representation learning,

    A. van den Oord, O. Vinyals, and K. Kavukcuoglu, “Neural discrete representation learning,” in Pro- ceedings of the 31st International Conference on Neural Information Processing Systems , NIPS’17, (Red Hook, NY , USA), p. 6309–6318, Curran Associates Inc., 2017

    2017

  51. [59]

    Time Series Classification from Scratch with Deep Neural Networks: A Strong Baseline,

    Z. Wang, W. Yan, and T. Oates, “Time Series Classification from Scratch with Deep Neural Networks: A Strong Baseline,” Dec. 2016. arXiv:1611.06455 [cs, stat]

    2016 arXiv

  52. [60]

    Vector quantized time series generation with a bidirectional prior model,

    D. Lee, S. Malacarne, and E. Aune, “Vector quantized time series generation with a bidirectional prior model,” in International Conference on Artificial Intelligence and Statistics , pp. 7665–7693, PMLR, 2023

    2023

  53. [61]

    GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilib- rium,

    N. B. Bynagari, “GANs Trained by a Two Time-Scale Update Rule Converge to a Local Nash Equilib- rium,” Asian j. appl. sci. eng., vol. 8, pp. 25–34, Apr. 2019

    2019

  54. [62]

    MMD GAN: Towards Deeper Understanding of Moment Matching Network,

    C.-L. Li, W.-C. Chang, Y . Cheng, Y . Yang, and B. Poczos, “MMD GAN: Towards Deeper Understanding of Moment Matching Network,”

  55. [63]

    Optuna: A next-generation hyperparameter op- timization framework,

    T. Akiba, S. Sano, T. Yanase, T. Ohta, and M. Koyama, “Optuna: A next-generation hyperparameter op- timization framework,” in Proceedings of the 25th ACM SIGKDD International Conference on Knowl- edge Discovery and Data Mining, 2019

    2019

  56. [64]

    Umap: Uniform manifold approximation and pro- jection,

    L. McInnes, J. Healy, N. Saul, and L. Großberger, “Umap: Uniform manifold approximation and pro- jection,” Journal of Open Source Software, vol. 3, no. 29, p. 861, 2018

    2018

  57. [65]

    Understanding how dimension reduction tools work: An empirical approach to deciphering t-sne, umap, trimap, and pacmap for data visualization,

    Y . Wang, H. Huang, C. Rudin, and Y . Shaposhnik, “Understanding how dimension reduction tools work: An empirical approach to deciphering t-sne, umap, trimap, and pacmap for data visualization,” Journal of Machine Learning Research, vol. 22, no. 201, pp. 1–73, 2021

    2021

  58. [66]

    Visualizing data using t-sne,

    L. van der Maaten and G. Hinton, “Visualizing data using t-sne,” Journal of Machine Learning Research, vol. 9, no. 86, pp. 2579–2605, 2008

    2008

  59. [69]

    Monitoring ai-modified content at scale: A case study on the impact of chatgpt on ai conference peer reviews,

    W. Liang, Z. Izzo, Y . Zhang, H. Lepp, H. Cao, X. Zhao, L. Chen, H. Ye, S. Liu, Z. Huang, D. A. McFarland, and J. Y . Zou, “Monitoring ai-modified content at scale: A case study on the impact of chatgpt on ai conference peer reviews,” 2024

    2024

  60. [2021]

    Conference Name: IEEE Access

  61. [2023]

    Number: 22 Publisher: Multidisciplinary Digital Publishing Institute

Pith tools

Reviewed August 12, 2026 · model on record in the stance chip above.