MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641124626 A) filed by Mahendra M; Sukeethi S; Nayan N Rai; Ch. Veena; Mrs. T. Mallika Devi; Vijayalaxmi Bindla; Kommineni Madhavi; and Guna Santhoshi on October 17, 2026, for Privacy-Preserving Cross-Bank Threat Intelligence Using Federated Learning, Differential Privacy, And Permissioned Blockchain.
Inventors include Mahendra M; Sukeethi S; Nayan N Rai; Ch. Veena; Mrs. T. Mallika Devi; Vijayalaxmi Bindla; Kommineni Madhavi; and Guna Santhoshi.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: 9. Abstract The present invention relates to a privacy-preserving cross-bank threat intelligence system using federated learning, differential privacy, and permissioned blockchain technology. The invention enables multiple banking institutions to collaboratively detect and mitigate cybersecurity threats without directly exchanging raw customer data, financial transaction records, or confidential security information. Each participating bank trains a deep learning model locally and generates protected model updates using differential privacy and encryption mechanisms. A permissioned blockchain network maintains tamper-evident records of participant authentication, model-update submissions, verification outcomes, and model versions. A secure aggregation mechanism validates eligible updates and generates a global threat intelligence model while incorporating measures to reduce the influence of malicious contributions, including model poisoning and backdoor attacks. The updated model supports the identification of fraudulent transactions, phishing activities, account takeover attempts, malware-related threats, and coordinated cyberattacks. The invention aims to enhance collaborative cybersecurity, preserve banking data confidentiality, improve threat detection capabilities, and support accountable intelligence sharing among participating financial institutions.
Disclaimer: Curated by HT Syndication.
