FEI-Hi: Federated Edge Intelligence for Healthcare Informatics.

Area of Science:

• Smart Healthcare
• Artificial Intelligence (AI)
• Internet of Things (IoT)

Background:

• Rapid evolution of IoT and AI technologies.
• Need for enhanced healthcare quality and resource allocation.
• Emergence of smart healthcare as a transformative solution.

Purpose of the Study:

• Introduce FEI-Hi, a federated edge intelligence paradigm.
• Enable secure and efficient medical data processing.
• Integrate edge computing with federated learning (FL).

Main Methods:

• Utilize a three-layer structure: FL, aggregation, and edge layers.
• Employ distributed intelligent computation, model parameter compression, and node clustering.
• Apply Wasserstein distance for clustering and parameter selection.

Main Results:

• Achieve a 30% improvement in model training speed.
• Attain an F1-score exceeding 90%.
• Surpass state-of-the-art benchmarks in efficiency, speed, and accuracy.

Conclusions:

• FEI-Hi enhances accuracy and efficiency in medical data processing.
• The paradigm ensures model convergence and stability.
• Demonstrates significant advancements over existing benchmarks.