MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202611101626 A) filed by Ankit Awasthy on August 22, 2026, for Federated Learning System For Privacy-Preserving Rare Disease Diagnostics.
Inventor includes Ankit Awasthy.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: The present invention relates to a federated learning system and method for privacy-preserving rare disease diagnostics. The invention enables multiple healthcare institutions to collaboratively train an artificial intelligence-based diagnostic model without sharing raw patient data. Each participating hospital performs local model training using internally stored clinical records and transmits only encrypted model parameters after adaptive differential privacy noise injection. A federated coordination server securely aggregates the received model updates using a secure aggregation algorithm to generate an optimized global diagnostic model while preventing reconstruction of individual institutional data. The adaptive differential privacy mechanism dynamically adjusts injected noise according to model sensitivity and convergence characteristics, thereby maintaining an optimal balance between privacy protection and diagnostic accuracy. The trained global model is periodically redistributed to participating institutions for subsequent learning iterations. The invention enhances diagnostic performance for rare diseases using geographically distributed clinical datasets while ensuring compliance with healthcare privacy regulations, improving cybersecurity, preserving institutional data ownership, and preventing unauthorized disclosure of sensitive patient information. Figure-1
Disclaimer: Curated by HT Syndication.
