New pipeline enhances privacy and accuracy for clinical AI models

Researchers have developed a robust pipeline for differentially private federated learning on imbalanced clinical data, specifically for cardiovascular risk prediction. The pipeline integrates the SMOTETomek technique to address data imbalance and the FedProx algorithm to handle non-IID data, outperforming standard FedAvg. The study identifies an optimal region on the privacy-utility frontier, demonstrating that strong privacy guarantees (epsilon 9.0) can be maintained with high clinical utility (recall > 77%). This work provides a practical framework for secure and accurate diagnostic tools using heterogeneous healthcare data. AI

IMPACT Enhances the feasibility of secure and effective AI models for sensitive clinical applications.

RANK_REASON The cluster contains an academic paper detailing a new methodology for AI model development. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• differential privacy
• FedAvg
• federated learning
• FedProx
• Rodrigo Tertulino
• SMOTETomek

• paper
• safety
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →