Systems and methods for improving performance of an analog processor
In a hybrid computing system including at least one analog processor and at least one digital processor an embedded problem is repeatedly run or executed on the analog processor(s) to generate a first plurality of candidate solutions to the computational problem, the candidate solutions are returned to the digital processor(s) which determine a value for at least one statistical feature of the candidate solutions, at least one programmable parameter of the plurality of analog devices in the analog processor(s) is adjusted to at least partially compensate for deviations from an expected value of the at least one statistical feature, the expected value of the at least one statistical feature inferred from the structure of the embedded problem, the embedded problem is again repeatedly run or executed on the analog processor(s) to generate a second plurality of candidate solutions to the computational problem.
1. A method of operation of a hybrid computing system that comprises an analog processor and at least one digital processor, the analog processor and the at least one digital processor communicatively coupled to one another, the analog processor comprising a plurality of analog devices, the plurality of analog devices characterized by values of at least one programmable parameter, the at least one programmable parameter programmable by the at least one digital processor, the method comprising:
embedding, by the at least one digital processor, a computational problem on the analog processor to generate an embedded problem having one or more symmetries;
causing, by the at least one digital processor, a first repeated running of the embedded problem on the analog processor to generate a first plurality of candidate solutions to the computational problem;
receiving, by the at least one digital processor, the first plurality of candidate solutions to the computational problem;
determining, by the at least one digital processor, a value for at least one statistical feature of the first plurality of candidate solutions to the computational problem;
adjusting, by the at least one digital processor, the values of the at least one programmable parameter of the plurality of analog devices in the analog processor to at least partially compensate for deviations from an expected value of the at least one statistical feature, the expected value of the at least one statistical feature inferred from the one or more symmetries of the embedded problem; and
causing, by the at least one digital processor, a second repeated running of the embedded problem on the analog processor to generate a second plurality of candidate solutions to the computational problem.
2. The method of claim 1 , wherein embedding, by the at least one digital processor, a computational problem on the analog processor to generate an embedded problem includes embedding, by the at least one digital processor, a computational problem on a quantum processor.
3. The method of claim 2 , wherein embedding, by the at least one digital processor, a computational problem on a quantum processor includes embedding, by the at least one digital processor, a computational problem on a superconducting quantum processor.
4. The method of claim 3 , wherein adjusting the values of the at least one programmable parameter of the plurality of analog devices in the analog processor to at least partially compensate for deviations from an expected value of the at least one statistical feature includes adjusting the values of the at least one programmable parameter of a plurality of superconducting flux qubits and superconducting coupling devices in the superconducting quantum processor.
5. The method of claim 4 , wherein adjusting the values of the at least one programmable parameter of a plurality of superconducting flux qubits and superconducting coupling devices in the superconducting quantum processor includes adjusting values of at least one of a flux, a flux bias offset, a coupling strength, and an anneal offset.
6. The method of claim 1 , wherein embedding, by the at least one digital processor, a computational problem on the analog processor to generate an embedded problem includes embedding, by the at least one digital processor, an optimization problem on the analog processor.
7. The method of claim 1 , wherein adjusting values of the at least one programmable parameter of the plurality of analog devices in the analog processor to at least partially compensate for deviations from an expected value of the at least one statistical feature, the expected value of the at least one statistical feature inferred from the structure of the embedded problem includes adjusting the at least one programmable parameter of the plurality of analog devices in the analog processor to at least partially compensate for deviations from an expected value of the at least one statistical feature, the expected value of the at least one statistical feature inferred from one or more graph automorphisms of the embedded problem.
8. The method of claim 1 , wherein determining, by the at least one digital processor, a value for at least one statistical feature of the first plurality of candidate solutions to the computational problem includes determining, by the at least one digital processor, a value for at least one of a magnetization and a spin-spin correlation.
9. The method of claim 1 , wherein embedding, by the at least one digital processor, a computational problem on the analog processor to generate an embedded problem includes embedding, by the at least one digital processor, a computational problem on a topology comprising a repeating lattice.
10. The method of claim 9 , wherein embedding, by the at least one digital processor, a computational problem on a topology comprising a repeating lattice includes embedding, by the at least one digital processor, a computational problem on a topology comprising at least one of a triangular lattice and a square lattice.
11. The method of claim 1 , further comprising:
receiving, by the at least one digital processor, the second plurality of candidate solutions to the computational problem;
determining, by the at least one digital processor, a value for at least one statistical feature of the second plurality of candidate solutions to the computational problem;
adjusting, by the at least one digital processor, values of the at least one programmable parameter of the plurality of analog devices in the analog processor to at least partially compensate for deviations from the expected value of the at least one statistical feature; and
causing, by the at least one digital processor, a third repeated running of the embedded problem on the analog processor to generate a third plurality of candidate solutions to the computational problem.
12. A hybrid computing system, comprising:
an analog processor comprising a plurality of analog devices;
at least one digital processor communicatively coupled to the at least one analog processor; and
at least one non-transitory computer-readable storage medium that stores processor-executable instructions, which when executed causes the at least one digital processor to:
embed a computational problem on the analog processor to generate an embedded problem having one or more symmetries;
cause a first repeated running of the embedded problem on the analog processor to generate a first plurality of candidate solutions to the computational problem;
receive the first plurality of candidate solutions to the computational problem;
determine a value for at least one statistical feature of the first plurality of candidate solutions to the computational problem;
adjust values of at least one programmable parameter of the plurality of analog devices in the analog processor to at least partially compensate for deviations from an expected value of the at least one statistical feature, the expected value of the at least one statistical feature inferred from the one or more symmetries of the embedded problem; and
cause a second repeated running of the embedded problem on the analog processor to generate a second plurality of candidate solutions to the computational problem.
13. The hybrid computing system of claim 12 , wherein the analog processor includes a quantum processor.
14. The hybrid computing system of claim 13 , wherein the quantum processor includes a superconducting quantum processor.
15. The hybrid computing system of claim 14 , wherein the at least one programmable parameter of the plurality of analog devices includes at least one programmable parameter of a plurality of superconducting flux qubits and superconducting coupling devices in the superconducting quantum processor.
16. The hybrid computing system of claim 15 , wherein the processor-executable instructions, which when executed causes at least one processor-based device to adjust the values of the at least one programmable parameter of a plurality of superconducting flux qubits and superconducting coupling devices in the superconducting quantum processor include instructions, which when executed causes at least one processor-based device to adjust at least one of a flux, a flux bias offset, a coupling strength and an anneal offset.
17. The hybrid computing system of claim 12 , wherein the computational problem includes an optimization problem.
18. The hybrid computing system of claim 12 , wherein the structure of the embedded problem includes one or more graph automorphisms of the embedded problem.
19. The hybrid computing system of claim 12 , wherein the at least one statistical feature of the first plurality of candidate solutions includes at least one of a magnetization and a spin-spin correlation.
20. The hybrid computing system of claim 12 , wherein the analog processor includes a topology comprising a repeating lattice.
21. The hybrid computing system of claim 20 , wherein the repeating lattice is at least one of a triangular lattice and a square lattice.
22. The hybrid computing system of claim 12 , the at least one non-transitory computer-readable storage medium that stores processor-executable instructions, which when executed further causes the at least one digital processor to:
receive the second plurality of candidate solutions to the computational problem;
determine a value for at least one statistical feature of the second plurality of candidate solutions to the computational problem;
adjust values of the at least one programmable parameter of the plurality of analog devices in the analog processor to at least partially compensate for deviations from the expected value of the at least one statistical feature; and
cause a third repeated running of the embedded problem on the analog processor to generate a third plurality of candidate solutions to the computational problem.