MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202641083248 A) filed by Dr Gopirajan Pv; and Dr. G. Venkatesan on July 07, 2026, for Federated Learning Framework For Privacy-Preserving Healthcare Data Analytics.
Inventors include Dr Gopirajan Pv; and Dr. G. Venkatesan.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: The present invention discloses a federated learning framework for privacy-preserving healthcare data analytics, enabling geographically distributed healthcare institutions to collaboratively train high-performance Al diagnostic models without sharing sensitive patient data. The framework employs differential privacy noise injection, homomorphic encryption for secure parameter aggregation, and an adaptive Federated Averaging algorithm capable of handling non-IID data distributions across heterogeneous clinical environments. A personalization module fine-tunes the global model for each institution's specific patient population, while a governance layer generates compliance-ready audit trails. Explainability mechanisms including attention maps and Integrated Gradients support transparent clinical decision- making. This invention addresses a critical gap in responsible Al deployment for healthcare, enabling collaborative intelligence while upholding the highest standards of patient privacy and regulatory compliance.
Disclaimer: Curated by HT Syndication.
