MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641049520 A) filed by J. B. Institute Of Engineering And Technology, Hyderabad, Telangana, on April 17, for 'an adaptive edge-aware artificial intelligence inference engine for resource-constrained computing devices.'
Inventor(s) include Dr. Gurrampally Kumar; Mr. S. Sathish Kumar; Mr. U Chandrashekhar Reddy; Mr. M. Mahesh Kumar; Mrs. Swapna Ganna; and Ms. Maria Divya Teja Pasala.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to an adaptive edge-aware artificial intelligence inference engine designed for efficient execution of machine learning models on resource-constrained devices. The system comprises a context monitoring module configured to continuously evaluate system parameters including processor utilization, memory availability, energy status, and thermal conditions. A dynamic model adaptation module modifies model execution by selectively activating layers, adjusting computational precision, and prioritizing critical features based on real-time conditions. A decision optimization unit determines optimal inference pathways by balancing latency, accuracy, and energy consumption. A resource allocation controller distributes computational tasks across available processing units, while a runtime scheduler manages execution priorities. The system further enables progressive inference execution with intermediate outputs and early termination based on confidence levels. The invention facilitates scalable, energy-efficient, and real-time deployment of artificial intelligence applications in resource-constrained edge environments."
Disclaimer: Curated by HT Syndication.
