MUMBAI, India, July 24 -- Intellectual Property India has published a patent application (202611061223 A) filed by Mr. Somasekhar Gubbala; and Mr. Ramachandra Reddy Vangala on May 14, 2026, for A System And A Method For Unified Spatio-Temporal Knowledge Validation And Forecasting Using Computer Vision, Knowledge Graphs, Graph Neural Networks, And Gpu-Accelerated Query Processing.
Inventors include Mr. Somasekhar Gubbala; and Mr. Ramachandra Reddy Vangala.
The application for the patent was published on July 17, 2026, under issue no. 29/2026.
Abstract: ABSTRACT A System and a Method for Unified Spatio-Temporal Knowledge Validation and Forecasting using Computer Vision, Knowledge Graphs, Graph Neural Networks, and GPU-Accelerated Query Processing The present disclosure relates to a system and method for executing unified spatio-temporal knowledge validation, causal discrepancy identification, and predictive forecasting across diverse geospatial data sources. The system gets map images and spatial metadata from different mapping services and uses computer vision to pull out text and symbol labels. The data that was taken out is organized and put into a heterogeneous knowledge graph that shows spatial, semantic, and temporal relationships. A neuro-symbolic reasoning module uses graph neural networks and rule-based constraints to find inconsistencies, which is called semantic validation. A digital twin simulation module shows how things are likely to be arranged in space, and a causal discrepancy attribution module finds the root causes of problems. A temporal forecasting module uses past patterns to guess what will happen in the future. A GPU-edge hybrid processing module speeds up query execution, and an uncertainty-aware correction module improves the results. The system can accurately, scale, and predict geospatial validation and anomaly detection.
Disclaimer: Curated by HT Syndication.
