MUMBAI, India, Feb. 6 -- Intellectual Property India has published a patent application (202541084322 A) filed by Presidency University, Bengaluru, Karnataka, on Sept. 4, 2025, for 'a method for detecting credit card fraud using hybrid deep learning models.'
Inventor(s) include Dr. M. Renuka Devi; S. Devi; Dr. Vinutha C B; and Dr. Pradeep Bhaskar.
The application for the patent was published on Feb. 6, under issue no. 06/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to a computer-implemented method for detecting fraudulent credit card activities using hybrid deep learning models. The method includes receiving multimodal data comprising transactional, behavioral, and external intelligence (1), preprocessing the data (2), and generating synthetic fraud samples by means of a conditional Generative Adversarial Network (GAN). Sequential transaction behavior is analyzed using a Transformer-based model, while relational links among users, devices, and merchants are modelled with a Graph Neural Network (3). The outputs are dynamically fused (4) through an adaptive ensemble to compute a fraud risk score (5), with thresholds adjusted based on contextual modifiers. The said method further employs federated learning with differential privacy to enable multi-institution collaboration without exposing raw data. An explainability module (6) generates feature-level attributions for flagged transactions. The invention provides scalable, accurate, real-time, and interpretable fraud detection with enhanced data security."
Disclaimer: Curated by HT Syndication.
