Dual Attention Heads for Personalized Federated Learning in ECG Classification

2026-07-07

AI Summary

This work addresses the performance degradation in electrocardiogram (ECG) classification within federated learning settings caused by data heterogeneity across medical institutions. To tackle this challenge, the authors propose FedDualAtt, a novel method that integrates a dual-attention-head mechanism into the Transformer architecture. Specifically, a global attention head leverages FedAvg aggregation to capture shared features across clients, while a local attention head preserves client-specific characteristics to accommodate individual institutional data distributions. A dynamic weighting strategy adaptively balances the contributions of these two heads, enabling architecture-level personalization in federated learning. Experimental results on the FedCVD benchmark demonstrate that FedDualAtt significantly outperforms existing federated and personalized approaches, confirming the effectiveness and superiority of architectural personalization for ECG classification tasks.

Abstract

Federated learning (FL) enables collaborative model training across institutions without sharing sensitive patient data. However, the inherent heterogeneity of electrocardiogram (ECG) data across healthcare providers presents significant technical challenges for robust classification. We propose FedDualAtt, a personalized federated learning approach that splits transformer attention heads into global and local branches. Global heads are aggregated via FedAvg to capture shared cross-site patterns, while local heads remain client-specific to adapt to institution-level recording characteristics. Experiments on FedCVD, an FL benchmark for cardiovascular disease detection, demonstrate that FedDualAtt outperforms existing FL and personalized FL methods in ECG classification tasks. Analysis of global-local head ratios reveals that different clients benefit from varying levels of architectural personalization.

Problem

• Federated Learning
• ECG Classification
• Data Heterogeneity
• Personalization
• Electrocardiogram

Innovation

• personalized federated learning
• dual attention heads
• ECG classification
• data heterogeneity
• transformer architecture