MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202611109876 A) filed by Avinash Grandhi on September 13, 2026, for Cloud-Based System For Detecting Coordinated Financial Fraud Using Cross-Account Transaction Graphs And Real-Time Risk Propagation.
Inventor includes Avinash Grandhi.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: A cloud-based financial fraud detection system is disclosed for identifying coordinated fraudulent activity across multiple accounts and correlated entities. The system receives heterogeneous transaction events, normalizes the events with event-time information, resolves identifiers into graph-compatible tokens, and incrementally constructs a temporal cross-account transaction graph. The graph is distributed across cloud graph shards using affinity-aware partitioning and boundary summaries. A real-time risk propagation engine propagates seed risk through eligible graph edges according to edge type, temporal decay, hop attenuation, source confidence and path novelty while enforcing a bounded propagation envelope and suppressing cyclic re-use of evidence. An adaptive evidence and confidence engine combines local anomaly evidence, propagated evidence and structural graph features to identify coordinated subgraphs while reducing double counting of correlated evidence. A fraud decision and response orchestrator generates a transaction response, and a provenance ledger records graph epoch, contributing evidence, propagation factors and policy version for reproducibility. The architecture reduces distributed graph traversal and uncontrolled risk amplification while enabling low-latency relational fraud analysis. Accompanied Drawing [FIGS. 1-2].
Disclaimer: Curated by HT Syndication.
