MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202641068874 A) filed by Shiek Ruksana on June 01, 2026, for Lyapunov-Stable Bayesian Trust Propagation Graph Neural Network For Uncertainty-Aware Credibility Assessment In Dynamic Interaction Networks.
Inventors include Shiek Ruksana; and Pavan Kumar Karedla.
The application for the patent was published on June 12, 2026, under issue no. 24/2026.
Abstract: A computer-implemented system and method for trust, risk, and credibility assessment in dynamic interaction networks is disclosed. The invention constructs a graph representation comprising entities, interactions, transactions, content, resources, and associated relationships. A Bayesian Graph Neural Network (BGNN) performs uncertainty-aware message passing to estimate trustworthiness, credibility, risk, anomaly likelihood, and confidence measures for graph entities and interactions. A trust propagation mechanism updates node states using probabilistic belief updates derived from local and multi-hop neighborhood information. To ensure stable trust evolution and prevent manipulation-induced oscillations, a Lyapunov stability constraint is incorporated into the graph learning and propagation process. The system generates calibrated trust scores, risk scores, credibility indicators, anomaly probabilities, and confidence intervals for decision support. The framework is applicable to digital platforms including admissions systems, educational portals, e-commerce marketplaces, financial transaction systems, social networks, healthcare ecosystems, recruitment platforms, food delivery services, and enterprise applications. The invention provides improved robustness, interpretability, uncertainty quantification, and resistance to coordinated adversarial behavior compared to conventional reputation and fraud-detection systems.
Disclaimer: Curated by HT Syndication.
