MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202611060550 A) filed by Manipal University Jaipur on May 13, 2026, for A Privacy-Preserving Recommender Systems: A Cybersecurity Perspective.

Inventors include Shivendra Dubey; Shreja Mishra; Om Satyam Panda; and Sakshi Dubey.

The application for the patent was published on July 17, 2026, under issue no. 29/2026.

Abstract: The present invention relates to a hybrid privacy-preserving recommendation system and method for generating personalized recommendations while reducing exposure of sensitive user information and mitigating adversarial attacks. The system comprises the integration of Differential Privacy (DP), Federated Learning (FL), and Secure Multiparty Computation (SMC) within a unified decentralized recommendation framework. A plurality of client devices perform local recommendation model training upon locally stored interaction data, while an adaptive privacy controller dynamically allocates client-specific privacy budgets and noise parameters according to data sensitivity, utility contribution, residual model error, and user privacy preferences. A secure aggregation and robust update module performs encrypted aggregation of client updates using secure multiparty computation and applies robust aggregation procedures comprising Trimmed Mean, Median, and Krum aggregation for mitigation of malicious updates, poisoning attacks, and shilling attacks. A monitoring and anomaly detection unit identifies anomalous update patterns, inference attacks, and adversarial behaviour. Communication-efficient procedures comprising gradient sparsification, quantization, compression, and selective client participation improve scalability and deployment efficiency. The invention provides enhanced privacy preservation, recommendation robustness, attack resistance, and regulatory compliance for recommendation systems deployed in electronic commerce, healthcare, media, fintech, cybersecurity, and large-scale personalization environments.

Disclaimer: Curated by HT Syndication.