Intellectual Property India Publishes Patent Application for Ai-Powered Multi-Layer Security Architecture For Advanced Persistent Threat Detection

MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641112725 A) filed by Dr. Saradha Swaminathan; Mr. Mohan Bangalore Somashekar; Dr. Kannadasan Suriyan; Dr. Nandkumar Wagh; M. Poornima Devi; Bommaraju Srinivasa Rao; Prasanna Potnuru Lakshmi; Ranga Rao Maturi; Dr. S. Gopikha; and Gayathri Devi Subramanian on September 20, 2026, for Ai-Powered Multi-Layer Security Architecture For Advanced Persistent Threat Detection.

Inventors include Dr. Saradha Swaminathan; Mr. Mohan Bangalore Somashekar; Dr. Kannadasan Suriyan; Dr. Nandkumar Wagh; M. Poornima Devi; Bommaraju Srinivasa Rao; Prasanna Potnuru Lakshmi; Ranga Rao Maturi; Dr. S. Gopikha; and Gayathri Devi Subramanian.

The application for the patent was published on September 25, 2026, under issue no. 39/2026.

Abstract: An artificial-intelligence-powered multi-layer security architecture for detection of advanced persistent threats is disclosed. A plurality of detection layers each monitors a respective layer of a computing environment, the network, endpoint or host, user-and-identity, and application-and-data layers, each applying a machine learning model that learns the normal behavior of its layer and detects anomalies and indicators of compromise within it, including the marginal anomalies produced by the stealth of the threat. A correlation and analysis layer receives these indicators and, by a machine learning model trained on patterns of advanced persistent threats, correlates them across the layers and over time to detect an advanced persistent threat manifested as a coordinated pattern, corresponding to a progression through intrusion, persistence, privilege escalation, lateral movement, collection, and exfiltration stages, of individually inconspicuous indicators distributed across the layers, each of which evades single-layer detection but which together reveal the intrusion. The layer distinguishes the coordinated pattern from coincidental legitimate anomalies, computes a confidence, correlates indicators dispersed over weeks or months, and assembles the evidence for an analyst, detecting the stealthy, distributed, multi-stage, long-duration intrusions that signature-based and single-layer means miss

Disclaimer: Curated by HT Syndication.