MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202641053263 A) filed by Sharon Sunil; Dr. M. Anand; Dr. G. Sujatha; and Dr. T. M. Sheeba on April 27, 2026, for Advanced Cyber Threat Detection Based On Event Profiles Using Machine Language Algorithms.
Inventors include Sharon Sunil; Dr. M. Anand; Dr. G. Sujatha; and Dr. T. M. Sheeba.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: The present invention discloses an artificial intelligence-based Security Information and Event Management (AT-SIEM) system for automated and advanced cyber threat detection. The system mitigates the high false alarm rates typical of traditional network intrusion detection systems by condensing massive volumes of raw security events into distinct event profiles. The framework comprises a data preprocessing stage utilizing a Term Frequency- Inverse Document Frequency (TF-IDF) mechanism to generate event vectors, which are subsequently transformed into dimensional event profiles. An Artificial Neural Network-based learning engine, integrating Fully Connected Neural Networks (FCNN), Convolutional Neural Networks (CNN), and Long Short-Term Memory (LSTM) algorithms, processes these profiles to learn normal and threat patterns. Finally, a real-time detection module categorizes incoming security events using the trained models to accurately distinguish legitimate cyber threats from erroneous alarms, displaying only verified true positive alerts to security analysts on a dashboard.
Disclaimer: Curated by HT Syndication.
