MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641051707 A) filed by Sai Mounika Chintalapudi; Dr. J. Peremalatha; Mrs. S. Spandana; Mrs. A. Kanimozhi; Dr. M Santhosh Kumar; Mrs. Vunnam Asha Latha; and Dr. C. R. Rene Robin, Madison, U.S.A., on April 23, for 'automated machine learning model optimization system using meta-learning techniques.'
Inventor(s) include Sai Mounika Chintalapudi; Dr. J. Peremalatha; Mrs. S. Spandana; Mrs. A. Kanimozhi; Dr. M Santhosh Kumar; Mrs. Vunnam Asha Latha; and Dr. C. R. Rene Robin.
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 automated machine learning model optimization system using meta-learning techniques, wherein the system is configured to intelligently optimize machine learning models by leveraging historical learning experiences stored in a meta-knowledge repository. The system comprises a task characterization module for extracting dataset-specific meta-features, a meta-learning module for predicting optimal model architectures and hyperparameters based on prior tasks, and an optimization engine for refining predicted configurations through adaptive optimization strategies. The invention further includes a feedback mechanism that continuously updates the meta-knowledge repository, enabling lifelong learning and progressive improvement in optimization accuracy. By utilizing meta-learning, the system significantly reduces computational cost, minimizes manual intervention, accelerates convergence, and enhances model performance and generalization across diverse datasets and application domains, thereby providing an efficient, scalable, and intelligent framework for automated machine learning optimization."
Disclaimer: Curated by HT Syndication.
