MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202641089405 A) filed by Vellore Institute Of Technology on July 22, 2026, for System And Method For Configuration-Driven Hybrid Machine Learning-Based Zero-Day Threat And Anomaly Detection.
Inventors include Dr. Dega Nagaraju; Ananya Shah; and Saanya Varshney.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: ABSTRACT The present invention discloses a non-intrusive, configuration-driven hybrid machine learning framework for real-time detection of zero-day attacks, cyber threats, fraudulent transactions, and anomalous activities. The framework comprises a hybrid detection engine integrating supervised learning models for identifying known attack patterns and unsupervised anomaly detection models for recognizing previously unseen threats. A configuration-driven architecture enables dynamic selection, deployment, and management of detection models without modifying existing applications or infrastructure. The framework further includes a synthetic attack generation module for creating simulated zero-day scenarios, a risk fusion engine for generating composite risk scores from multiple detection outputs, and a threshold-based decision module configured to classify events into ALLOW, REVIEW, or BLOCK categories. Secure webhook-based communication enables seamless integration with existing systems. The invention provides a scalable, model-agnostic, and adaptable cybersecurity solution capable of enhancing threat detection accuracy while minimizing deployment complexity and operational disruption.
Disclaimer: Curated by HT Syndication.
