Quantum Circuit Structure Learning
We propose an efficient method for simultaneously learning both the structure and parameter values of quantum circuits with only a small computational overhead. Shallow circuits trained using structure learning perform significantly better than circuits trained using parameter updates alone, making this method particularly suitable for use on noisy intermediate-scale quantum computers. We demonstrate the method for training a variational quantum eigensolver for finding the ground states of Lithium Hydride and the Heisenberg model.
Mateusz Ostaszewski, Edward Grant, Marcello Benedetti