MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641112638 A) filed by Aditya Institute Of Technology And Management; Meesala Sudhir Kumar; N. Asha Sri; T. Srikanya; K. Ananda Kumar; K. Sai Shri; Omkar Pattnaik; and Naresh Tangudu on September 19, 2026, for Machine-Learning-Based Dynamic Discovery And Validation Of Invariant Physical Representations.
Inventors include Meesala Sudhir Kumar; N. Asha Sri; T. Srikanya; K. Ananda Kumar; K. Sai Shri; Omkar Pattnaik; and Naresh Tangudu.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: An intelligent machine-learning-based invariant physical representation system (100) for adaptive monitoring of a physical system includes a multi-modal sensor network (102) configured to acquire time-series physical-state data under a plurality of operating conditions. An operating-regime identification engine (104) identifies operating regimes of the physical system. A candidate representation generation engine (106) generates candidate physical representations comprising relationships between physical variables. An invariant evaluation engine (108) determines invariance scores based on consistency of the candidate physical representations across the operating regimes. An invariant selection engine (110) selects validated invariant physical representations for machine-learning-based physical-state inference. An invariant monitoring engine (114) detects deviations from the validated invariant physical representations. A deviation classification engine (116) classifies the deviations as physical-system changes, sensor-related changes, operating-regime changes, or model-representation changes. An invariant adaptation controller (118) selectively retains, modifies, or replaces the invariant physical representations based on the classified deviations.
Disclaimer: Curated by HT Syndication.
