Intellectual Property India Publishes Patent Application for 'Privacy-Preserving Federated Learning System For Intrusion Detection In Internet Of Things' Filed by Lovely Professional University

MUMBAI, India, Jan. 8 -- Intellectual Property India has published a patent application (202511102909 A) filed by Lovely Professional University, Phagwara, Punjab, on Oct. 25, 2025, for 'privacy-preserving federated learning system for intrusion detection in internet of things.'

Inventor(s) include Manish Kumar.

The application for the patent was published on Dec. 12, under issue no. 50/2025.

According to the abstract released by the Intellectual Property India: "A privacy-preserving federated learning system for intrusion detection in Internet of Things (IoT) networks comprising a distributed client module 101 configured to train local intrusion detection models on heterogeneous IoT data within resource-constrained devices, a SHAP computation module 102 integrated in each client derives feature importance scores using Shapley values to capture local data characteristics, a feature aggregation module 103 at a central server securely integrates distilled feature knowledge from multiple clients without collecting raw data, a global model optimization module 104 iteratively generates a unified intrusion detection model via federated averaging, an interpretability module 105 visualizes global feature relevance, enabling explainable intrusion classification across non-IID client environments, including detection of specific botnet attack patterns, the framework achieves over 99% intrusion detection accuracy while preserving data privacy and adapting to diverse IoT ecosystems."

Disclaimer: Curated by HT Syndication.