Manifold-Aware Diffusion-Augmented Contrastive Learning for Noise-Robust Biosignal Representation
DOI:
https://doi.org/10.3991/ijoe.v22i04.59841Keywords:
Contrastive Learning, Latent Diffusion Models, Wavelet scattering, Scattering TransformerAbstract
Learning robust representations for physiological time-series signals continues to pose a substantial challenge in developing efficient few-shot learning applications. This is largely due to the complex pathological variations in bio signals. In this context, this paper introduces a manifold-aware Diffusion-Augmented Contrastive Learning (DACL) framework, which efficiently leverages the generative structure of latent diffusion models (LDMs) with the discriminative power of supervised contrastive learning. The proposed framework operates within a contextualized scattering latent space derived from Scattering Transformer (ST) features. Within a contrastive learning framework, we employ a forward diffusion process in the scattering latent space as a structured manifold-aware feature augmentation technique. We assessed the proposed framework using the PhysioNet 2017 Electrocardiogram (ECG) benchmark dataset. The proposed method achieved a competitive AUROC of 0.9741 in the task of detecting atrial fibrillation (AF) from a single-lead ECG signal. The proposed framework achieved performance on par with relevant state-of-the-art related works. In-depth evaluation findings suggest that early-stage diffusion serves as an ideal “local manifold explorer,” producing embeddings with greater precision than typical augmentation methods while preserving inference efficiency.
References
[1] A. Jaiswal, A. R. Babu, M. Z. Zadeh, D. Banerjee, and F. Makedon, “A survey on contrastive self-supervised learning,” Technologies, vol. 8, no. 2, p. 36, 2020. https://doi.org/10.3390/technologies9010002
[2] T. Chen, S. Kornblith, M. Norouzi, and G. Hinton, “A simple framework for contrastive learning of visual representations,” in Proceedings of the 37th International Conference on Machine Learning, vol. 119, Article 149, 2020, pp. 1597–1607.
[3] K. He, H. Fan, Y. Wu, S. Xie and R. Girshick, “Momentum Contrast for Unsupervised Visual Representation Learning,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Seattle, WA, USA, 2020, pp. 9726-9735. https://doi.org/10.1109/CVPR42600.2020.00975
[4] T. T. Um, F. M. J. Pfister, D. Pichler, S. Endo, M. Lang, S. Hirche, U. Fietzek, and D. Kulić, “Data augmentation of wearable sensor data for parkinson’s disease monitoring using convolutional neural networks,” in Proceedings of the ACM ICMI, Nov. 2017, pp. 216–220. https://doi.org/10.1145/3136755.3136817
[5] E. Eldele, M. Ragab, Z. Chen, M. Wu, C. K. Kwoh, X. Li, C. Guan, “Time-Series Representation Learning via Temporal and Contextual Contrasting,” in Proceedings of the IJCAI, Aug. 2021, pp. 2352–2359. https://doi.org/10.24963/ijcai.2021/324
[6] Z. Yue, Y. Wang, J. Duan, T. Yang, C. Huang, Y. Tong, and B. Xu, “TS2Vec: Towards Universal Representation of Time Series,” AAAI, vol. 36, no. 8, pp. 8980–8987, June 2022. https://doi.org/10.1609/aaai.v36i8.20881
[7] J. Ho, A. Jain, and P. Abbeel, “Denoising diffusion probabilistic models,” in Proceedings of the 34th International Conference on Neural Information Processing Systems, Article 574, 2020, pp. 6840–6851.
[8] J. Sohl-Dickstein, E. A. Weiss, N. Maheswaranathan, and S. Ganguli, “Deep unsupervised learning using nonequilibrium thermodynamics,” in Proceedings of the International Conference on Machine Learning, vol. 37, 2015, pp. 2256–2265.
[9] P. Vincent, H. Larochelle, P.-A. Manzagol, and Y. Bengio, “Extracting and composing robust features with denoising autoencoders,” in Proceedings of the ICML, Jan. 2008, pp. 1096–1103. https://doi.org/10.1145/1390156.1390294
[10] R. Zewail, “Scattering Transformer: A Training-Free Transformer Architecture for Heart Murmur Detection,” presented at MEDI 2025, to be published in Lecture Notes in Computer Science. [Online]. Available: arXiv:2509.18424.
[11] R. Rombach, D. Lorenz, A. Blattmann, P. Esser, and B. Ommer, “High Resolution Image Synthesis with Latent Diffusion Models,” in Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), June 2022, pp. 10674–10685. https://doi.org/10.1109/cvpr52688.2022.01042
[12] K. Yin, L. Yang, Y. Zhang, F. Ou, C. Chen, and B. Mu, “Diffusion augmented contrastive learning framework for quantitative diagnosis under limited data conditions,” Advanced Engineering Informatics, vol. 69, p. 103930, Jan. 2026, https://doi.org/10.1016/j.aei.2025.103930
[13] Y. Tan, H. Zhang, Y. Chang, Q. Wu, K. Qian, and B. Hu, “A latent diffusion model for heart sound synthesis,” Proceedings of the 4th International Symposium on Computer Applications and Information Technology (ISCAIT), Mar. 2025, pp. 1314–1319.
[14] F. Wu, G. Zhao, X. Qian, and L.-W. H. Lehman, “A diffusion model with contrastive learning for ICU false arrhythmia alarm reduction,” in Proceedings of the IJCAI, Aug. 2023, pp. 4912–4920. https://doi.org/10.24963/ijcai.2023/546
[15] F. Zhang, X. Xie, and K. Guo, “ASD-diffusion: Anomalous sound detection with diffusion models,” arXiv:2409.15957, Sept. 24, 2024, https://doi.org/10.48550/arxiv.2409.15957
[16] G. Luo, L. Dunlap, D. H. Park, A. Holynski, and T. Darrell, “Diffusion hyper features: searching through time and space for semantic correspondence,” in Proceedings of The 37th International Conference Neural Information Processing Systems, Article 2057, 2023, pp. 47500–47510.
[17] Q. Song, J. Hu, L. Xiao, B. Sun, X. Gao, and S. Li, “DiffCL: A diffusion-based contrastive learning framework with semantic alignment for multimodal recommendations,” arXiv:2501.01066, Jan. 02, 2025. https://doi.org/10.48550/arxiv.2501.01066
[18] J. Li, J. Wang, and Z. Cao, “Contrastdm: Combining contrastive learning and diffusion model for hyperspectral image classification,” in Proceedings of the International Geoscience and Remote Sensing Symposium (IGARSS), July 2024, pp. 9966–9970, https://doi.org/10.1109/igarss53475.2024.10641234
[19] D. Zeng, X. Hu, Y.-J. Chen, Y. Wu, X. Xu, and Y. Shi, “Contrastive learning with diffusion features for weakly supervised medical image segmentation,” arXiv:2506.23460, June 30, 2025. https://doi.org/10.48550/arxiv.2506.23460
[20] Z. Xing, W. A. P. Smith, and O. Mehmood, “Unsupervised anomaly detection with a temporal continuation, confidence-aware VAE-GAN,” Pattern Recognition, vol. 166, p. 111699, Oct. 2025. https://doi.org/10.1016/j.patcog.2025.111699
[21] G. D. Clifford, C. Liu, B. Moody, L.-w. H. Lehman, I. Silva, Q. Li, A. E. Johnson, and R. G. Mark, “AF classification from a short single lead ECG recording: The PhysioNet/computing in cardiology challenge 2017,” in Proceedings of the Computing in Cardiology (CinC), Rennes, France, 2017, pp. 1-4. https://doi.org/10.22489/CinC.2017.065-469
[22] S. Hong, H. Li, J. Sun, C. Xiao, and T. Ma, “MINA: Multilevel knowledge guided attention for modeling electrocardiography signals,” in Proceedings of the IJCAI, Aug. 2019, pp. 5888–5894. https://doi.org/10.24963/ijcai.2019/816
[23] J. Xie, B. Yao, and S. Stavrakis, “Automated identification of atrial fibrillation from single-lead ECGs using multi-branching ResNet.,” Frontiers in Physiology, vol. 15, Apr. 2024, doi: 10.3389/fphys.2024.1362185.
[24] C. B. Chandrakala, S. Pooja, C. Pujari, S. Katavarapu, S. Awatramani, and S. Gohil, “Health-lens: A health diagnosis companion,” International Journal of Interactive Mobile Technologies (iJIM), vol. 19, no. 12, pp.68–102, 2025.
[25] R. Jeya, H. A. Karim, and S. B. Mansor, “Artificial Intelligence and Mobile Apps Support Intelligent Healthcare Systems for Mental Health Services,” International Journal of Interactive Mobile Technologies (iJIM), vol. 18, no. 20, pp. 157–168, 2024.
Downloads
Published
How to Cite
Issue
Section
License
Copyright (c) 2026 Rami Zewail
This work is licensed under a Creative Commons Attribution 4.0 International License.
