MUMBAI, India, Jan. 9 -- Intellectual Property India has published a patent application (202521109633 A) filed by Mr. Shrikumar Pramod Ugale; Ms. Anjali Deepak Hazari; Mr. Ritesh Gangasingh Bais; and Ms. Sneha Anil Sahare, Pune, Maharashtra, on Nov. 11, 2025, for 'method and system for adaptive prompt engineering in generative ai models using feedback loop.'
Inventor(s) include Mr. Shrikumar Pramod Ugale; Ms. Anjali Deepak Hazari; Mr. Ritesh Gangasingh Bais; and Ms. Sneha Anil Sahare.
The application for the patent was published on Dec. 12, under issue no. 50/2025.
According to the abstract released by the Intellectual Property India: "The present invention relates to an adaptive prompt engineering system and method for generative artificial intelligence models using a feedback loop. The system is designed to autonomously optimize prompt representations based on continuous evaluation of generated outputs, thereby improving the contextual coherence, semantic precision, and factual accuracy of responses produced by large-scale generative models. The system comprises a prompt formulation unit configured to generate structured task-specific prompts, a generative processing unit for producing corresponding outputs, a feedback acquisition and evaluation unit for analyzing output performance using semantic, syntactic, and factual metrics, and a prompt adaptation control unit that modifies the prompt structure in response to computed deviation values derived from feedback results. A contextual memory and reinforcement storage unit maintains historical records of prior prompt configurations and associated performance scores, enabling cumulative learning across multiple iterations."
Disclaimer: Curated by HT Syndication.
