Asynchronous Federated Learning for Improved Cardiovascular Disease Prediction Using Artificial Intelligence.

Area of Science:

• Cardiology
• Artificial Intelligence
• Machine Learning

Background:

• Predicting heart disease is crucial for healthcare professionals.
• Deep learning (DL) shows promise for accurate cardiac prediction.
• Existing methods may face challenges in efficiency and accuracy.

Purpose of the Study:

• To introduce a novel asynchronous federated deep learning approach for cardiac prediction (AFLCP).
• To enhance the accuracy and convergence of deep neural network (DNN) models for heart disease prediction.
• To reduce communication costs in federated learning for cardiac applications.

Main Methods:

• Developed the asynchronous federated deep learning approach for cardiac prediction (AFLCP).
• Utilized a heart disease dataset and deep neural networks (DNNs).
• Implemented asynchronous parameter updates and temporally weighted aggregation for DNNs.

Main Results:

• The AFLCP method demonstrated superior performance over baseline methods.
• AFLCP achieved higher model accuracy in cardiac prediction.
• The proposed approach showed a reduction in communication costs.

Conclusions:

• The asynchronous federated deep learning approach for cardiac prediction (AFLCP) is effective.
• AFLCP offers improved accuracy and communication efficiency for heart disease prediction.
• This novel method advances the application of deep learning in cardiology.