What question did this study set out to answer?

The aim is to improve the robustness and accuracy of federated learning models in healthcare while ensuring data privacy.

March 16, 2026ICT ExpressOpen Access

Adversarial-aware federated learning with dynamic client weighting for healthcare record security

View Full Paper

Ask AI

Bookmark

Authors

Federated Learning enhances prediction accuracy in healthcare, addressing data privacy and adversarial threats.

The aim is to improve the robustness and accuracy of federated learning models in healthcare while ensuring data privacy.
Implemented an attack-aware federated learning pipeline with preprocessing and feature selection.
Utilized Deep Particle Swarm Optimization for discriminative feature selection.
Trained a temporal Deep Convolutional Neural Network model with per-round logging of accuracy.
Applied dynamic aggregation based on accuracy and sample counts for client weighting.
Achieved 94.44% accuracy on the UCL Heart Disease dataset.
Attained 97.36% accuracy on the UCL Breast Cancer dataset.
Demonstrated improved robustness against adversarial attacks.

Naeem et al. (2026) studied this question.

View Full Paper

Ask AI

Bookmark