MUMBAI, India, March 13 -- Intellectual Property India has published a patent application (202641025354 A) filed by Bala Bhoomeshwar, Ghatkesar, Telangana, on March 3, for 'adaptive resource aware multimodal generative ai system for secure on device deployment.'
Inventor(s) include Tasleema; Bala Bhoomeshwar; Muntha. Raju; Deyyapu Bhargav; Paripelly Pavan Kumar; and Chintalapalli Rishender.
The application for the patent was published on March 13, under issue no. 11/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to an adaptive resource-aware multimodal generative artificial intelligence system configured for secure on-device deployment. The system enables coordinated execution of multiple generative models including text generation, image synthesis, audio generation, and executable code generation on a local computing device without reliance on external cloud infrastructure. The invention comprises a resource monitoring module configured to continuously monitor hardware parameters including CPU utilization, GPU utilization, and memory availability; an adaptive scheduling engine configured to dynamically allocate computational workloads based on real-time hardware conditions; a multimodal orchestration module configured to analyze user input and determine required output modalities; and a secure execution environment configured to prevent external data transmission during inference. The adaptive scheduling engine selectively enables parallel or sequential execution of generative models and dynamically adjusts execution parameters including model selection, inference precision, and resolution based on available computational resources. All processing occurs locally, thereby enhancing data privacy and reducing latency. The invention provides a technically improved framework for stable and efficient multimodal generative AI execution in resource-constrained environments and is suitable for deployment in educational, research, industrial, healthcare, and secure enterprise applications."
Disclaimer: Curated by HT Syndication.
