Quantum annealing of 3D spin glasses with neural network wavefunctions

D-Wave is at the forefront of building quantum annealers, which are quantum computing devices that can potentially solve an important class of problems more efficiently than classical computers. One important goal of D-Wave is to claim that they have attained quantum supremacy, which means they have built a machine that can do something that no classical computer can achieve in a reasonable amount of time (multiple weeks/months on a supercomputing cluster).
To claim that their machine does solve these problems, they have to compare their machine with numerical simulations of said machine. For this, they employ tensor network approaches or quantum Monte Carlo simulations. The latest D-Wave machine operates on thousands of qubits, for which accurate simulations with the aforementioned numerical techniques will take millions of years on a supercomputing cluster.
In this research project, we will develop a simulation of a quantum annealing protocol for a three-dimensional spin glass. We intend to use neural network wave functions, which are a class of variational ansatz that has been used in the past to solve ground state problems, and perform dynamics on a variety of different systems. The latter comes with additional difficulties and requires advanced numerical methods to provide a stable…

Faculty Supervisor:

Roger Melko

Student:

Partner:

D-Wave Systems Inc.

Discipline:

Physics

Sector:

Manufacturing; Professional, scientific and technical services

University:

University of Waterloo

Program: