MUMBAI, India, July 13 -- Intellectual Property India has published a patent application (202641082495 A) filed by Sr University on July 03, 2026, for An Explainable Vision Transformer System For Uncertainty-Aware Disaster Forecasting And Resilient Decision Support.
Inventors include Dannuri Mounika; Johnson Kolluri; and Dr. B. Sunil Srinivas.
The application for the patent was published on July 10, 2026, under issue no. 28/2026.
Abstract: AN EXPLAINABLE VISION TRANSFORMER SYSTEM FOR UNCERTAINTY-AWARE DISASTER FORECASTING AND RESILIENT DECISION SUPPORT An explainable vision transformer enterprise computing system is disclosed for uncertainty- aware disaster forecasting and automated resilient decision support. The system features a network telemetry ingestion interface that collects multimodal geospatial streams, which are sliced and embedded by a spatiotemporal patching engine into structured token vectors. An explainable vision transformer core processor deployed on dedicated hardware accelerators processes the tokens via self- attention registers to forecast hazard progression. An integrated attention attribution module maps internal attention registers to generate human-interpretable spatial relevance maps explaining system focus. Concurrently, a probabilistic uncertainty engine executes variational inference dropout loops during inference to write a distinct disaster prediction matrix and an explicit uncertainty variance matrix to a split memory cache. A downstream resilient decision support controller reads these matrices alongside logistical infrastructure graphs, running a risk-aware optimization solver to automatically output adaptive vehicle routing vectors and civil evacuation alerts that avoid high-hazard and high-uncertainty geographic zones.
Disclaimer: Curated by HT Syndication.
