MUMBAI, India, June 13 -- Intellectual Property India has published a patent application (202517044329 A) filed by Google Llc, Mountain View, U.S.A., on May 7, for 'modular training for flexible attention based end-to-end asr.'
Inventor(s) include Audhkhasi, Kartik; Ramabhadran, Bhuvana; and Farris, Brian.
The application for the patent was published on June 13, under issue no. 24/2025.
According to the abstract released by the Intellectual Property India: "A method (600) for training a modular neural network model (200) includes training only a backbone model (302) to provide a first model configuration (300) of the modular neural network model. The first model configuration includes only the trained backbone model. The method also includes adding an intrinsic sub-model (410) to the trained backbone model. During a fine-tuning training stage (502), the method includes freezing parameters of the trained backbone model and fine-tuning parameters of the intrinsic sub-model added to the trained backbone model while the parameters of the trained backbone model are frozen to provide a second model configuration (400) that includes the backbone model initially trained during the initial training stage and the intrinsic sub-model having the parameters fine-tuned during the fine-tuning stage."
The patent application was internationally filed on Dec. 01, 2023, under International application No.PCT/US2023/082036.
Disclaimer: Curated by HT Syndication.
