Systems, methods, and apparatus for solving problems
View Patent ↗A method of improving the accuracy and computation time of automatic image recognition by the implementation of association graphs and a quantum processor. A method of solving problems using a quantum processor by casting a problem as a quadratic unconstrained binary optimization (“QUBO”) problem, mapping the QUBO problem to the quantum processor, and evolving the quantum processor to produce a solution to the QUBO problem.
1. A method of problem solving using a quantum processor, the method comprising:
casting a problem as a quadratic unconstrained binary optimization (“QUBO”) problem via a classical digital computer;
mapping the QUBO problem to the quantum processor via a programming system; and
evolving the quantum processor to produce a solution to the QUBO problem.
2. The method of claim 1 wherein casting a problem as a QUBO problem includes casting a problem that includes comparing two objects as the QUBO problem.
3. The method of claim 2 wherein casting a problem as a QUBO problem includes casting a determining a measure of similarity between two objects problem as the QUBO problem.
4. The method of claim 2 wherein casting a problem as a QUBO problem includes casting an image matching problem as a QUBO problem that includes at least two variables that determine how features in one image map to features in another image.
5. The method of claim 1 wherein casting a problem as a QUBO problem includes casting a finding a maximum independent set of an association graph problem as the QUBO problem.
6. The method of claim 1 wherein casting a problem as a QUBO problem includes casting a finding a maximum clique of an association graph problem as the QUBO problem.
7. The method of claim 1 wherein casting a problem as a QUBO problem includes casting the problem into an n-vertex QUBO problem, where n is limited by a number of qubits in the quantum processor.
8. The method of claim 7 , and further comprising:
identifying each of the n vertices with a respective qubit in the quantum processor via a programming system.
9. The method of claim 8 wherein identifying each of the n vertices with a respective qubit in the quantum processor via programming system includes identifying each of the n vertices with a respective qubit in the quantum processor via an initialization module.
10. The method of claim 7 , further comprising:
identifying each of the n vertices with a respective effective qubit in the quantum processor via a programming system.
11. The method of claim 10 wherein identifying each of the n vertices with a respective effective qubit in the quantum processor via a programming system includes identifying each of the n vertices with a respective effective qubit in the quantum processor via an initialization module.
12. The method of claim 1 wherein evolving the quantum processor to produce a solution to the QUBO problem includes performing an adiabatic quantum computation.
13. The method of claim 1 wherein evolving the quantum processor to produce a solution to the QUBO problem includes performing quantum annealing.
14. The method of claim 1 wherein mapping the QUBO to a quantum processor includes mapping the QUBO to a superconducting quantum processor that comprises a plurality of superconducting qubits, and wherein evolving the quantum processor to produce a solution to the QUBO problem includes evolving the superconducting quantum processor to produce a solution to the QUBO problem.
15. The method of claim 1 wherein casting a problem as a QUBO problem includes casting a pattern matching problem as the QUBO problem.
16. The method of claim 1 wherein casting a problem as a QUBO problem includes casting a finding a match to a query within a database problem as the QUBO problem.
17. The method of claim 1 wherein casting the problem as a QUBO problem includes mapping the problem to the form:
x
→
opt
=
arg
min
{
∑
i
≤
j
=
1
N
Q
ij
x
i
x
j
}
,
x
i
ɛ
{
0
,
1
}
wherein x i and x j are binary optimization variables and Q is a matrix.
18. The method of claim 1 wherein casting the problem as a QUBO problem further comprises:
decomposing the problem into a set of smaller problems; and
casting at least one of the smaller problems as a QUBO problem.
19. The method of claim 1 wherein casting the problem as a QUBO problem via a classical digital computer includes casting the problem as a QUBO problem via a user interface module.
20. The method of claim 1 wherein mapping the QUBO problem to the quantum processor via a programming system includes mapping the QUBO problem to the quantum processor via a mapper module.
21. The method of claim 1 wherein evolving the quantum processor to produce a solution to the QUBO problem includes evolving the quantum processor via an evolution module.
22. The method of claim 1 wherein evolving the quantum processor to produce a solution to the QUBO problem includes evolving the quantum processor to produce an exact solution to the QUBO problem.
23. The method of claim 1 wherein evolving the quantum processor to produce a solution to the QUBO problem includes evolving the quantum processor to produce an approximate solution to the QUBO problem.