MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202611026115 A) filed by Indian Institute Of Technology, New Delhi, on March 5, for 'method and system for training a framework agnostic quantum neural network.'
Inventor(s) include Sivasubramani, Santhosh.
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: "Disclosed herein is a method (700) for training a framework agnostic quantum neural network. The method includes receiving (702) an input and generating (704) an intermediate representation by processing the received quantum circuit specification based on the target execution framework selection. Furthermore, the method includes translating (706) the intermediate representation into a framework-native quantum layer based on the target execution framework selection. Moreover, the method includes encoding (708) the classical input data into a quantum state based on an encoding strategy. In addition, the method includes determining (710) one or more parameter gradients based on executing the framework-native quantum layer. Further, the method includes updating (712) the one or more model parameters based on the one or more parameter gradients. Furthermore, the method includes training (714) the framework agnostic quantum neural network based on the updated one or more model parameters."
Disclaimer: Curated by HT Syndication.
