MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202511113211 A) filed by Goutham Sunkara; Priyanshu Sharma; Priyanshi Bhatnagar; Konidena Sowjanya Bharathi; Dr. Sreejith R; Jyothsna Devi Dontha; Dr. S. Saravanakumar; Dr. Preethi Jeevan; D. Balakrishnan; and Dr. Anusha Ampavathi on November 18, 2025, for System And Method For Federated Machine Learning-Based Dynamic Cybersecurity Threat Prediction.
Inventors include Goutham Sunkara; Priyanshu Sharma; Priyanshi Bhatnagar; Konidena Sowjanya Bharathi; Dr. Sreejith R; Jyothsna Devi Dontha; Dr. S. Saravanakumar; Dr. Preethi Jeevan; D. Balakrishnan; and Dr. Anusha Ampavathi.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: The present disclosure describes a system and method for federated machine learning-based real-time and dynamic prediction of cybersecurity threats. The invention enables distributed computing nodes—such as enterprise servers, IoT gateways, and edge devices—to collaboratively train localized threat detection models on contextual cybersecurity telemetry, including network traffic logs, access control data, and user behavior traces, without sharing raw data. The locally trained models are securely aggregated using homomorphic encryption and differential privacy to generate a global threat prediction model that continuously evolves through drift detection and adaptive retraining. Based on classification, the system issues severity-specific alerts—ranging from reassurance logs for normal activity, advisory alerts for suspicious anomalies, to emergency escalation alerts for critical cyberattacks—and transmits them securely to administrators, automated response systems, or centralized coordination servers. The architecture further incorporates blockchain-anchored audit logging and trusted execution environments to ensure integrity, transparency, and compliance. Collectively, the invention establishes a privacy-preserving, intelligent, and adaptive cybersecurity ecosystem capable of early detection, accurate classification, and real-time mitigation of emerging and evolving cyber threats.
Disclaimer: Curated by HT Syndication.
