MUMBAI, India, March 13 -- Intellectual Property India has published a patent application (202631014086 A) filed by C. V. Raman Global University, Bhubaneswar, Orissa, on Feb. 9, for 'an artificial intelligence based system and method for adversary emulation and adaptive intrusion.'
Inventor(s) include Raj Vikram; Amit Kumar Satapathy; Subrat Kumar Jena; Aditya Dash; Mrinalini Patnaik; and Manoranjan Sahoo.
The application for the patent was published on March 13, under issue no. 11/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to an enterprise-grade machine learning-based intrusion detection and prevention system that integrates multiple artificial intelligence paradigms including supervised learning models (Random Forest, XGBoost, LightGBM), deep learning architectures (Convolutional Neural Networks, Long Short-Term Memory networks, Transformer models), unsupervised anomaly detection mechanisms (Variational Autoencoders, Isolation Forest), and generative adversarial networks for synthetic attack data generation. The system further incorporates reinforcement learning for adaptive threat prevention policies, federated learning with differential privacy for privacy-preserving distributed training across multiple organizational sensors, and explainable artificial intelligence tools (SHAP, LIME) for transparent decisionmaking. The integrated system achieves superior performance metrics with 95.34% overall detection accuracy, 94.21% precision, and exceptionally low false positive rates of 0.5%, while maintaining real-time inference latency below 10 milliseconds suitable for enterprise network traffic analysis at scale. Specialized detector models achieve accuracy rates exceeding 96-98% for specific attack categories including Distributed Denial of Service attacks at 98.23%, web exploitation attacks at 97.56%, and Advanced Persistent Threat indicators at 96.12%. The system provides complete production-grade deployment capabilities through containerized Docker environments, Kubernetes orchestration, cloud platform integration, and comprehensive monitoring via Prometheus, Grafana, and the ELK stack, making it suitable for organizations of all sizes requiring advanced, adaptive, and explainable threat detection and automated prevention capabilities."
Disclaimer: Curated by HT Syndication.
