SGERA: Stein-Guided ECG-Report Alignment for ECG Representation Learning

Jian Chen, Yipeng Du, Wenhao Yuan, Shuai Wang, Jinfeng Xu, Zewei Liu, Running Zhao, Edith Cheuk-Han Ngai
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:14099-14117, 2026.

Abstract

Electrocardiogram (ECG) representation learning via ECG-report alignment is often hindered by the inherent structural and statistical divergence between signals and natural language. Existing methods struggle to bridge this gap with simple contrastive objectives, but struggle with distribution dependencies between heterogeneous features. To address this, we propose SGERA (Stein-Guided ECG-Report Alignment), which leverages the unique properties of Stein kernels to provide a more rigorous geometric alignment in the latent space: instance-level alignment via a Stein-RBF kernel enforces pairwise consistency between ECG and report embeddings and distribution-level alignment via a Stein-Score kernel captures higher-order interactions for global alignment. Furthermore, we introduce an ECG-Report matching task with a Hard Sample Mining strategy to refine discriminative boundaries. Experiments across three public datasets demonstrate that SGERA significantly outperforms state-of-the-art SSL methods in zero-shot classification, linear probing, and transfer learning, proving the superiority of Stein-guided alignment in handling complex medical modalities. Code is available at supplementary materials.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-chen26y, title = {{SGERA}: Stein-Guided {ECG}-Report Alignment for {ECG} Representation Learning}, author = {Chen, Jian and Du, Yipeng and Yuan, Wenhao and Wang, Shuai and Xu, Jinfeng and Liu, Zewei and Zhao, Running and Ngai, Edith Cheuk-Han}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {14099--14117}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/chen26y/chen26y.pdf}, url = {https://proceedings.mlr.press/v306/chen26y.html}, abstract = {Electrocardiogram (ECG) representation learning via ECG-report alignment is often hindered by the inherent structural and statistical divergence between signals and natural language. Existing methods struggle to bridge this gap with simple contrastive objectives, but struggle with distribution dependencies between heterogeneous features. To address this, we propose SGERA (Stein-Guided ECG-Report Alignment), which leverages the unique properties of Stein kernels to provide a more rigorous geometric alignment in the latent space: instance-level alignment via a Stein-RBF kernel enforces pairwise consistency between ECG and report embeddings and distribution-level alignment via a Stein-Score kernel captures higher-order interactions for global alignment. Furthermore, we introduce an ECG-Report matching task with a Hard Sample Mining strategy to refine discriminative boundaries. Experiments across three public datasets demonstrate that SGERA significantly outperforms state-of-the-art SSL methods in zero-shot classification, linear probing, and transfer learning, proving the superiority of Stein-guided alignment in handling complex medical modalities. Code is available at supplementary materials.} }
Endnote
%0 Conference Paper %T SGERA: Stein-Guided ECG-Report Alignment for ECG Representation Learning %A Jian Chen %A Yipeng Du %A Wenhao Yuan %A Shuai Wang %A Jinfeng Xu %A Zewei Liu %A Running Zhao %A Edith Cheuk-Han Ngai %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-chen26y %I PMLR %P 14099--14117 %U https://proceedings.mlr.press/v306/chen26y.html %V 306 %X Electrocardiogram (ECG) representation learning via ECG-report alignment is often hindered by the inherent structural and statistical divergence between signals and natural language. Existing methods struggle to bridge this gap with simple contrastive objectives, but struggle with distribution dependencies between heterogeneous features. To address this, we propose SGERA (Stein-Guided ECG-Report Alignment), which leverages the unique properties of Stein kernels to provide a more rigorous geometric alignment in the latent space: instance-level alignment via a Stein-RBF kernel enforces pairwise consistency between ECG and report embeddings and distribution-level alignment via a Stein-Score kernel captures higher-order interactions for global alignment. Furthermore, we introduce an ECG-Report matching task with a Hard Sample Mining strategy to refine discriminative boundaries. Experiments across three public datasets demonstrate that SGERA significantly outperforms state-of-the-art SSL methods in zero-shot classification, linear probing, and transfer learning, proving the superiority of Stein-guided alignment in handling complex medical modalities. Code is available at supplementary materials.
APA
Chen, J., Du, Y., Yuan, W., Wang, S., Xu, J., Liu, Z., Zhao, R. & Ngai, E.C.. (2026). SGERA: Stein-Guided ECG-Report Alignment for ECG Representation Learning. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:14099-14117 Available from https://proceedings.mlr.press/v306/chen26y.html.

Related Material