MUMBAI, India, July 7 -- Intellectual Property India has published a patent application (202641080541 A) filed by Narsimha Reddy Engineering College Nrcm on June 30, 2026, for Graph Neural Network-Based Financial Fraud Detection Framework For Real-Time Digital Payment Systems.
Inventors include Dr. Nettem Adithya Valli; Bantu Mahesh; M Suresh Babu; Konakala Srinivasa Rao; Kurnapally Anuradha; and Ponnaganti Bhagya Lakshmi.
The application for the patent was published on July 03, 2026, under issue no. 27/2026.
Abstract: A graph neural network-based fraud detection framework for real-time digital payment systems is disclosed. The system includes a transaction ingestion module (110), streaming graph construction engine (120), feature encoding module (130), graph neural network inference engine (140), risk scoring module (150), decision orchestration module (160), model update module (170), and storage subsystem (180). The invention dynamically constructs a heterogeneous transaction graph from streaming payment data and performs low-latency neighborhood aggregation to generate fraud probability scores within authorization time limits. Adaptive threshold recalibration enhances detection accuracy while reducing false positives. Incremental model updating ensures resilience against evolving fraud patterns. The framework enables relational fraud detection in high-throughput digital payment infrastructures while maintaining compliance and operational efficiency.
Disclaimer: Curated by HT Syndication.
