Batch Parameter Shift Training Reduces the Number of Cloud Jobs for Quantum Neural Networks

Written by

in

Researchers from LUMS and KAUST proposed a batch parameter shift training method that reframes hybrid quantum-classical training as an execution efficiency issue for quantum processors. Simulations show that, for models with 24 to 48 parameters, the method requires only one-fifth to one-half of the jobs needed by SPSA, while gradient error decreases as the number of shots per job increases.

Source:Quantum Zeitgeist · Read Original

Comments

Leave a Reply

Your email address will not be published. Required fields are marked *