Speaker
Nikita Astrakhantsev
Description
Preparation of quantum states in form of a variational quantum circuit plays a crucial role in quantum computing. We show that, in addition to circuit architecture, the fidelity of the prepared state depends non-trivially on the number of circuit shots $N_s$ used in gradient descent. Namely, we observe that fidelity shows a critical behavior in $N_s$, giving rise to the notion of critical effective temperature, below which, the resource demand of optimization grows as $1 \sim \Delta^2$ with the system gap.
We analyze the effect of this dependence on the of large-scale simulations of frustrated magnets and provide a symmetry-enhanced simulation protocol reducing the computational costs in near-term quantum computers.
Authors
Prof.
Giuseppe Carleo
(EPFL)
Nikita Astrakhantsev