MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641112708 A) filed by Rajeev Gandhi Memorial College Of Engineering & Technology Autonomous on September 20, 2026, for Adaptive Machine Learning System For Automated Detection And Classification Of Cybersecurity Threats In Computer Networks.
Inventors include Dr. K. Narasimhulu; and Ms. Madisetty Venkata Abhigna.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: An adaptive machine learning system for automated detection and classification of cybersecurity threats in computer networks is disclosed. A monitoring and feature module monitors network traffic and events and processes them into features, such as flow volumes, rates, ports, protocols, and patterns. A detection and classification module, comprising a machine learning model trained on threats and traffic, detects a threat from the features and classifies it into a category, denial of service, malware, intrusion, reconnaissance, data exfiltration, or another, with a confidence, so the type is identified to inform the response. An adaptation module adapts the detection to the evolving threat landscape by retraining the model on newly observed threats and traffic, by detecting novel previously-unseen threats through an anomaly-detection model that flags deviations from the normal even where they match no known threat, complementing the classification, and by incorporating analyst feedback of confirmed detections and corrected misclassifications. A novel threat detected as an anomaly is, upon confirmation, incorporated into training to become a known classified category, so the system maintains detection against evolution and novelty where static and signature-based detection progressively fail.
Disclaimer: Curated by HT Syndication.
