MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202541080258 A) filed by International Institute Of Information Technology on August 25, 2025, for System And Method For Optimizing An Attribute Value Dynamically In Finite-Horizon Environments Using Bayesian Inference.
Inventors include Dr. Tejas Bodas; Anush Anand; and Pranav Agrawal.
The application for the patent was published on June 19, 2026, under issue no. 25/2026.
Abstract: A system for optimizing an attribute value dynamically in finite-horizon environments using Bayesian inferenceis provided. The system(100) includesdata acquisition device (102) that collects historical data specific to attribute, and optimization server 104 that segmentsattributes into buckets; generates predictive model (110) using historical data; estimates cumulative objective value for each candidate from each bucket using predictive model; computesacquisition score for each value using acquisition function; determines new attribute by identifying candidate that maximizes acquisition score across buckets; transmits the new attribute value to external control device (108) through network (106) to apply the new attribute value in external control device to collect resulting observed performance metric; receives the observed performance metric associated with new attribute and generates new input-output data pair for retraining; updates the predictive model (110)with the new input-output; and iteratively performs all the steps until a termination condition is met. FIG. 1
Disclaimer: Curated by HT Syndication.
