MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202641092418 A) filed by Sr University on July 30, 2026, for An Edge-Ai Based Behavioral Profiling System With Deep Learning And Encrypted Audit Logging For Detection Of Insider Threats And Synthetic User Activity In Privileged Access Environments.
Inventors include Deepthi Bolukonda; Dr. Rupesh Kumar Mishra; and Dr. Indrajeet Gupta.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: The present invention discloses an edge-AI based behavioral profiling system for detecting insider threats and synthetic user activity within privileged access environments. The system comprises a plurality of edge computing nodes deployed proximate to privileged access terminals, each configured to capture multi- dimensional behavioral telemetry including keystroke dynamics, mouse movement patterns, command sequences and privilege escalation events. A behavioral baseline module establishes an individualized behavioral profile for each authorized user, against which a deep learning inference engine, comprising recurrent neural network and autoencoder architectures, computes real-time anomaly and reconstruction-error scores. A dedicated synthetic activity classifier further distinguishes genuine human interaction from automated, scripted or bot-driven activity based on statistical micro-variability features. An encrypted, hash-chained audit logging module records every behavioral inference, anomaly score and access event within a tamper-evident ledger, ensuring forensic integrity even where the underlying host system is compromised. A centralized aggregation and alerting layer correlates anomalies across multiple sessions and generates prioritized alerts, supported by a feedback mechanism for continuous model refinement, thereby providing a low-latency, privacy-preserving and evidentiarily robust solution to insider threat and synthetic activity detection.
Disclaimer: Curated by HT Syndication.
