{"id":"arxiv/2603.26475","url":"https://arxiv.org/abs/2603.26475","tags":["cs.LG","cs.AI","eess.SP","math.RT"],"score":0,"title":"Foundation Model for Cardiac Time Series via Masked Latent Attention","domain":"cs.LG","source":"arxiv","authors":["Moritz Vandenhirtz","Samuel Ruipérez-Campillo","Simon Böhi","Sonia Laguna","Irene Cannistraci","Andrea Agostini","Ece Ozkan","Thomas M. Sutter","Julia E. Vogt"],"reviews":0,"subject":"cs.LG","summary":"Electrocardiograms (ECGs) are among the most widely available clinical signals and play a central role in cardiovascular diagnosis. While recent foundation models (FMs) have shown promise for learning transferable ECG representations, most existing pretraining approaches treat leads as independent channels and fail to explicitly leverage their strong structural redundancy. We introduce the latent attention masked autoencoder (LAMAE) FM that directly exploits this structure by learning cross-lead connection mechanisms during self-supervised pretraining. Our approach models higher-order…","category":"research","comments":0,"disputed":false,"division":"cs","mentions":0,"discussed":0,"published":"2026-03-27","reactions":{}}