MUMBAI, India, July 13 -- Intellectual Property India has published a patent application (202621059337 A) filed by Mritunjay Kr. Ranjan on May 10, 2026, for An Automata-Driven Self-Repairing Machine Learning Architecture For Autonomous Detection And Correction Of Model Degradation.
Inventors include Ankita Patil; Mritunjay Kumar Ranjan; Shreya Tiwari; Pranjal Sonje; Rajashri Rikame; and Disha Arsude.
The application for the patent was published on July 10, 2026, under issue no. 28/2026.
Abstract: The present invention relates to a computer-implemented system and method for autonomous detection and correction of machine learning model degradation in production environments. This system is designed to constantly observe performance of a deployed machine learning model by using a multiplicity of performance measures, identify degradation under a multi-metric threshold condition, and automatically perform model repair by retraining with one or more datasets. The invention also includes a validation mechanism that is capable to compare retrained models to predefined improvement criteria and a deployment module that is configured to deploy the retrained model conditionally when the validation criteria are met. In one implementation, the system uses a finite state machine to control deterministic transitions among the operational states such as training, deployment, monitoring, degradation detection, repair, validation and stable operation. The system is a closed-loop system, which allows adapting the system to varying data conditions within a continuous cycle and sustaining a model without human intervention.
Disclaimer: Curated by HT Syndication.
