Description
Classical simulation of quantum circuits is a computationally expensive task. However, physics inspired algorithms are often based in properties of the system, such as their symmetries. For the case of the Variational Quantum Eigensolvers, when the expressivity of the Ansatz is reduced to satisfy symmetry constrains, this reduces the space that the circuit can explore.
This constrains, in additional to backpropagation, can be used to reduce both, the number of evaluations of the circuit, as well as their dimensionality. With the motivation from classical machine learning, backpropagation can also be used to reduced the number of evaluation of the circuit.
In this flash talk, we will show how these two features can be combined to reduce the computational time of classical VQE simulations for the Nuclear Shell Model.