Constructing quantum processes for quantum processors
In a general aspect, a quantum process for execution by a quantum processor is generated. In some instances, test data representing a test output of a quantum process are obtained. The test data are obtained based on a value assigned to a variable parameter of the quantum process. An objective function is evaluated based on the test data, and an updated value is assigned to the variable parameter based on the evaluation of the objective function. The quantum process is provided for execution by a quantum processor, and the quantum process provided for execution has the updated value assigned to the variable parameter.
1. A method comprising:
identifying a variable parameter of a quantum process;
obtaining test data representing a test output of the quantum process, wherein the test data are generated by a quantum simulator or quantum computer executing the quantum process with a first value assigned to the variable parameter;
by operation of a data processing apparatus, evaluating an objective function based on the test output of the quantum process, the evaluating being executed by an optimization engine or machine learning engine;
modifying the quantum process by assigning a second, different value to the variable parameter based on the evaluation of the objective function; and
providing, for execution by a quantum processor, the quantum process having the second value assigned to the variable parameter, wherein modifying the quantum process improves the execution of the quantum process by the quantum processor.
2. The method of claim 1 , wherein the objective function is a cost function.
3. The method of claim 1 , wherein the optimization engine or machine learning engine comprises a neural network.
4. The method of claim 3 , wherein the neural network is a convolution neural network.
5. The method of claim 1 , wherein the modifying the quantum process comprises minimizing the objective function.
6. The method of claim 1 , wherein the quantum processor comprises one of a superconducting circuit system, an ion trap system, a spin system, or a photonic system.
7. The method of claim 1 , wherein assigning the second, different value to the variable parameter is performed using an optimization algorithm.
8. The method of claim 7 , wherein the optimization algorithm is a gradient descent algorithm.
9. The method of claim 1 , wherein the data processing apparatus comprises multiple programmable processors.
10. The method of claim 1 , wherein the objective function is evaluated based on the test data and training data for the quantum process, the training data comprising ground states for molecular structures.
11. The method of claim 1 , wherein the objective function is evaluated based on the test data and training data for the quantum process, the training data comprising ground state energies for molecular structures.
12. The method of claim 1 , wherein the objective function is evaluated based on the test data and training data for the quantum process, the training data comprising a quantum simulation benchmark.
13. The method of claim 1 , wherein the quantum process comprises a quantum logic gate set, and assigning the second, different value to the variable parameter is performed by shaping microwave pulses used to implement gates of the quantum logic gate set.
14. The method of claim 1 , wherein the quantum process comprises a quantum logic gate set, and assigning the second, different value to the variable parameter is performed by appending, inserting or removing gates from the quantum logic gate set.
15. A computer system comprising:
a quantum processor unit; and
means for constructing quantum logic circuits for execution by the quantum processor unit, wherein the means for constructing quantum logic circuits comprises an optimization engine or machine learning engine.
16. The computer system of claim 15 , wherein the means for constructing quantum logic circuits further comprises a classical processor unit programmed to perform a construction process.
17. The computer system of claim 16 , further comprising a memory shared by the quantum processor unit and the classical processor unit.
18. The computer system of claim 16 , wherein the classical processor unit comprises multiple programmable processors.
19. The computer system of claim 15 , wherein the optimization engine or machine learning engine comprises a neural network.
20. The computer system of claim 19 , wherein the neural network is a convolution neural network.
21. The computer system of claim 15 , wherein the quantum processor unit comprises one of a superconducting circuit system, an ion trap system, a spin system, or a photonic system.
22. The computer system of claim 15 , wherein the optimization engine or machine learning engine is configured to execute an optimization algorithm.
23. The computer system of claim 22 , wherein the optimization algorithm is a gradient descent algorithm.
24. The computer system of claim 22 , wherein the optimization algorithm is configured to evaluate an objective function based on training data comprising ground states for molecular structures.
25. The computer system of claim 22 , wherein the optimization algorithm is configured to evaluate an objective function based on training data comprising ground state energies for molecular structures.
26. The computer system of claim 22 , wherein the optimization algorithm is configured to evaluate an objective function based on training data comprising a quantum simulation benchmark.