Granted Patent B2

US 12,265,770 · App. 17/524,194 · Granted Apr 1, 2025

Inventors: Shixin Zhang (Shenzhen, CN); Changyu Hsieh (Shenzhen, CN); Shengyu Zhang (Shenzhen, CN)

Assignee: TENCENT TECHNOLOGY (SHENZHEN) COMPANY LIMITED

G06F30/337G06N10/00

Abstract

Claims (71)

1. A method for determining a quantum circuit, comprising:

sampling an initial circuit unit pool according to an initial sampling manner to obtain initial K groups of circuit units;

constructing and generating initial K candidate quantum circuits based on the initial K groups of circuit units, each group of the circuit units comprising at least one circuit unit for constructing and generating a candidate quantum circuit, and K being a positive integer;

determining a performance evaluation index corresponding to the initial K candidate quantum circuits;

updating the initial sampling manner and a circuit unit in the initial circuit unit pool based on the performance evaluation index, to obtain an updated sampling manner and an updated circuit unit pool;

sampling the updated circuit unit pool according to the updated sampling manner to obtain updated K groups of circuit units;

constructing and generating updated K candidate quantum circuits based on the updated K groups of circuit units; and

in response to a cycle termination condition being satisfied, determining a target quantum circuit from the updated K candidate quantum circuits.

2. The method according to claim 1 , wherein the sampling the initial circuit unit pool comprises:

constructing an initial probability model, the probability model comprising p×c parameters, p representing a maximum number of circuit units, and c representing a total number of circuit units in the initial circuit unit pool; and

sampling the initial circuit unit pool K times based on the initial probability model, to obtain the initial K groups of circuit units, each time of sampling being performed to obtain one group of circuit units.

3. The method according to claim 2 , wherein the updating the sampling manner and the circuit unit in the initial circuit unit pool comprises:

computing first gradient information and second gradient information based on the performance evaluation index, the first gradient information being gradient information of a model parameter of the initial probability model, and the second gradient information being gradient information of a circuit parameter of the circuit unit in the circuit unit pool;

updating the model parameter of the initial probability model based on the first gradient information, to obtain the updated probability model; and

updating the circuit parameter of the circuit unit in the circuit unit pool based on the second gradient information, to obtain the updated circuit unit pool.

4. The method according to claim 3 , wherein the performance evaluation index is an operation result of an objective function, and the computing the first gradient information and the second gradient information based on the performance evaluation index comprises:

computing a derivative of the objective function relative to the model parameter of the initial probability model, to obtain the first gradient information; and

computing a derivative of the objective function relative to the circuit parameter of the circuit unit in the circuit unit pool, to obtain the second gradient information.

5. The method according to claim 1 , wherein the performance evaluation index is an operation result of an objective function, and the determining the performance evaluation index corresponding to the initial K candidate quantum circuits comprises:

computing K operation results of the objective function respectively corresponding to the K candidate quantum circuits; and

obtaining the performance evaluation index based on the K operation results.

6. The method according to claim 1 , wherein the determining the target quantum circuit from the updated K candidate quantum circuits comprises:

determining a candidate quantum circuit with a largest generation probability from the updated K candidate quantum circuits; and

determining the candidate quantum circuit as the target quantum circuit.

7. The method according to claim 1 , wherein the method further comprises:

obtaining, for a j th circuit unit in an i th candidate quantum circuit, a circuit parameter of the j th circuit unit from a circuit parameter pool according to a position of the j th circuit unit in the circuit unit pool and a position of the j th circuit unit in the i th candidate quantum circuit,

the circuit parameter pool comprising a circuit parameter of each circuit unit in the circuit unit pool at each filling position, i being a positive integer less than or equal to K, and j being a positive integer.

8. The method according to claim 1 , wherein the method further comprises:

fixing a structure of the target quantum circuit, and adjusting a circuit parameter of each circuit unit in the target quantum circuit; and

in response to an adjustment termination condition being satisfied, obtaining the target quantum circuit post parameter tuning.

9. The method according to claim 1 , wherein a circuit unit is repeatable in each group of circuit units.

10. A method for determining quantum circuit, comprising:

selecting N circuit units from a circuit unit pool, the circuit unit pool comprising a plurality of circuit units, each circuit unit being an equivalent quantum circuit corresponding to a unitary matrix, and N being an integer greater than 1;

determining circuit parameters respectively corresponding to the N circuit units, the circuit parameter being for defining an operation performed by the circuit unit, and the circuit parameter being updatable; and

constructing and generating a quantum circuit based on the N circuit units and the circuit parameters respectively corresponding to the N circuit units.

11. The method according to claim 10 , wherein the determining the circuit parameters respectively corresponding to the N circuit units comprises:

obtaining, for an i th circuit unit in the N circuit units, a circuit parameter of the i th circuit unit from a circuit parameter pool according to a filling position of the i th circuit unit,

the circuit parameter pool comprising a circuit parameter of each circuit unit in the circuit unit pool at each filling position, i being a positive integer less than or equal to N.

12. The method according to claim 10 , wherein the method further comprises:

determining a performance evaluation index corresponding to the quantum circuit;

computing target gradient information based on the performance evaluation index, the target gradient information being gradient information of a circuit parameter of the quantum circuit; and

updating the circuit parameter of the quantum circuit based on the target gradient information, to obtain an updated quantum circuit.

13. The method according to claim 12 , wherein the performance evaluation index is an operation result of an objective function, and the computing the target gradient information based on the performance evaluation index comprises:

computing a derivative of the objective function relative to the circuit parameter of the quantum circuit, to obtain the target gradient information.

14. The method according to claim 10 , wherein a circuit unit is repeatable in the N circuit units.

15. An apparatus for determining a quantum circuit, comprising:

a memory operable to store computer-readable instructions; and

a processor circuitry operable to read the computer-readable instructions, the processor circuitry when executing the computer-readable instructions is configured to:

sample an initial circuit unit pool according to an initial sampling manner to obtain initial K groups of circuit units;

construct and generate initial K candidate quantum circuits based on the initial K groups of circuit units, each group of the circuit units comprising at least one circuit unit for constructing and generating a candidate quantum circuit, and K being a positive integer;

determine a performance evaluation index corresponding to the initial K candidate quantum circuits;

update the initial sampling manner and a circuit unit in the initial circuit unit pool based on the performance evaluation index, to obtain an updated sampling manner and an updated circuit unit pool;

sample the updated circuit unit pool according to the updated sampling manner to obtain updated K groups of circuit units;

construct and generate updated K candidate quantum circuits based on the updated K groups of circuit units; and

in response to a cycle termination condition being satisfied, determine a target quantum circuit from the updated K candidate quantum circuits.

16. The apparatus according to claim 15 , wherein the processor circuitry is configured to:

construct an initial probability model, the probability model comprising p×c parameters, p representing a maximum number of circuit units, and c representing a total number of circuit units in the initial circuit unit pool; and

sample the initial circuit unit pool K times based on the initial probability model, to obtain the initial K groups of circuit units, each time of sampling being performed to obtain one group of circuit units.

17. The apparatus according to claim 16 , wherein the processor circuitry is configured to:

compute first gradient information and second gradient information based on the performance evaluation index, the first gradient information being gradient information of a model parameter of the initial probability model, and the second gradient information being gradient information of a circuit parameter of the circuit unit in the circuit unit pool;

update the model parameter of the initial probability model based on the first gradient information, to obtain the updated probability model; and

update the circuit parameter of the circuit unit in the circuit unit pool based on the second gradient information, to obtain the updated circuit unit pool.

18. The apparatus according to claim 17 , wherein the performance evaluation index is an operation result of an objective function, and the processor circuitry is configured to:

compute a derivative of the objective function relative to the model parameter of the initial probability model, to obtain the first gradient information; and

compute a derivative of the objective function relative to the circuit parameter of the circuit unit in the circuit unit pool, to obtain the second gradient information.

19. The apparatus according to claim 15 , wherein the performance evaluation index is an operation result of an objective function, and the processor circuitry is configured to:

compute K operation results of the objective function respectively corresponding to the K candidate quantum circuits; and

obtain the performance evaluation index based on the K operation results.

20. The apparatus according to claim 15 , wherein the processor circuitry is configured to:

determine a candidate quantum circuit with a largest generation probability from the updated K candidate quantum circuits; and

determine the candidate quantum circuit as the target quantum circuit.

Assignments (1)

Continuity (2)

References Cited (31)