MUMBAI, India, Feb. 27 -- Intellectual Property India has published a patent application (202541129891 A) filed by P. T. Santhiya; Mr. T. Gowrishankar; G. Santhiya; S. Sitharth; S. Subash; N. Baranidharan; M. Bharathkumar; B. Deepak; P. Dhaarani; and A. Chithra Kaviya, Salem, Tamil Nadu, on Dec. 22, 2025, for 'real time zero day attack prediction engine using transformer graph neural network hybrid model.'
Inventor(s) include P. T. Santhiya; Mr. T. Gowrishankar; G. Santhiya; S. Sitharth; S. Subash; N. Baranidharan; B. Deepak; P. Dhaarani; A. Chithra Kaviya; and M. Bharathkumar.
The application for the patent was published on Feb. 27, under issue no. 09/2026.
According to the abstract released by the Intellectual Property India: "The increasing sophistication of cyber-attacks, particularly zero-day attacks, poses a serious threat to modern digital infrastructures such as cloud systems, enterprise networks, IoT platforms, and critical information systems. Traditional intrusion detection and prevention systems rely on known signatures or predefined rules, which are ineffective against previously unseen attack patterns. Even existing machine learning-based security solutions struggle io detect coordinated, multi-stage, and evolving zero-day threats in real time. The present invention proposes a Real-Time Zero-Day Attack Prediction Engine using a Transformer-Graph Neural Network (GNN) Hybrid Model. The system models network entities, traffic flows, system calls, and user behaviors as dynamic graphs, while transformers capture long-range temporal dependencies and contextual relationships. By combining graphbased structural learning with transformer-based sequence modeling, the invention predicts zero-day attacks proactively before full exploitation occurs. The proposed system enables real-time threat intelligence, early warning alerts, and adaptive cyber defense mechanisms."
Disclaimer: Curated by HT Syndication.
