Performing quantum error mitigation at runtime using trained machine learning model
A method, system, and computer program product for runtime quantum error mitigation. Training data, which includes noisy expectation values and target expectation values (noiseless expectation values), is generated. A machine learning model is then trained using the training data to perform quantum error mitigation based on learning the relationships between target and noisy expectation values. That is, such a machine learning model is trained to generate target expectation values based on inputted noisy expectation values. Upon executing a quantum circuit on a quantum computer creating quantum results, quantum error mitigation is performed on the quantum results at runtime using the trained machine learning model. In this manner, there are significant savings in quantum execution time while improving the accuracy of the results in performing quantum error mitigation on quantum results at runtime without additional mitigation circuits.
1 . A method for runtime quantum error mitigation, the method comprising:
generating training data;
training a machine learning model using said generated training data to perform quantum error mitigation;
performing an optimization of said machine learning model using a loss function;
executing a quantum circuit on a quantum computer creating quantum results; and
performing quantum error mitigation on said quantum results at runtime using said trained quantum machine learning model.
2 . The method as recited in claim 1 further comprising:
generating noisy expectation values for said training data by sampling parameters of parameterized circuits with a fixed structure.
3 . The method as recited in claim 2 , wherein said parameterized circuits are ansatz circuits.
4 . The method as recited in claim 2 , wherein said noisy expectation values of said training data are generated by sampling parameters of parameterized circuits using parameterized Hamiltonian time evolution.
5 . The method as recited in claim 2 further comprising:
implementing dynamic decoupling and/or twirling in a quantum circuit that generates said noisy expectation values.
6 . The method as recited in claim 5 further comprising:
encoding features of said quantum circuit on a noisy quantum device.
7 . The method as recited in claim 6 further comprising:
generating said target expectation values using said noisy expectation values and said encoded features, wherein said training data comprises said target expectation values generated using a noiseless simulator or an error mitigated quantum processing unit.
8 . The method as recited in claim 1 , wherein said machine learning model is selected from the group consisting of a graph neural network model, a multilayer perceptron model, a random forest model, and an ordinary least squares model.
9 . The method as recited in claim 1 further comprising:
training said machine learning model using said generated training data to perform quantum error mitigation based on learning relationships between target and noisy expectation values.
10 . A computer program product for runtime quantum error mitigation, the computer program product comprising one or more computer readable storage mediums having program code embodied therewith, the program code comprising programming instructions for:
generating training data;
training a machine learning model using said generated training data to perform quantum error mitigation;
performing an optimization of said machine learning model using a loss function;
executing a quantum circuit on a quantum computer creating quantum results; and
performing quantum error mitigation on said quantum results at runtime using said trained quantum machine learning model.
11 . The computer program product as recited in claim 10 , wherein the program code further comprises the programming instructions for:
generating noisy expectation values for said training data by sampling parameters of parameterized circuits with a fixed structure.
12 . The computer program product as recited in claim 11 , wherein said parameterized circuits are ansatz circuits.
13 . The computer program product as recited in claim 11 , wherein said noisy expectation values of said training data are generated by sampling parameters of parameterized circuits using parameterized Hamiltonian time evolution.
14 . The computer program product as recited in claim 11 , wherein the program code further comprises the programming instructions for:
implementing dynamic decoupling and/or twirling in a quantum circuit that generates said noisy expectation values.
15 . The computer program product as recited in claim 14 , wherein the program code further comprises the programming instructions for:
encoding features of said quantum circuit on a noisy quantum device.
16 . The computer program product as recited in claim 15 , wherein the program code further comprises the programming instructions for:
generating target expectation values using said noisy expectation values and said encoded features, wherein said training data comprises said target expectation values generated using a noiseless simulator or an error mitigated quantum processing unit.
17 . The computer program product as recited in claim 10 , wherein said machine learning model is selected from the group consisting of a graph neural network model, a multilayer perceptron model, a random forest model, and an ordinary least squares model.
18 . The computer program product as recited in claim 10 , wherein the program code further comprises the programming instructions for:
training said machine learning model using said generated training data to perform quantum error mitigation based on learning relationships between target and noisy expectation values.
19 . A system, comprising:
a memory for storing a computer program for runtime quantum error mitigation; and
a processor connected to said memory, wherein said processor is configured to execute program instructions of the computer program comprising:
generating training data;
training a machine learning model using said generated training data to perform quantum error mitigation;
performing an optimization of said machine learning model using a loss function;
executing a quantum circuit on a quantum computer creating quantum results; and
performing quantum error mitigation on said quantum results at runtime using said trained quantum machine learning model.
20 . The system as recited in claim 19 , wherein the program instructions of the computer program further comprise:
generating noisy expectation values for said training data by sampling parameters of parameterized circuits with a fixed structure.
21 . The system as recited in claim 20 , wherein said parameterized circuits are ansatz circuits.
22 . The system as recited in claim 20 , wherein said noisy expectation values of said training data are generated by sampling parameters of parameterized circuits using parameterized Hamiltonian time evolution.
23 . The system as recited in claim 20 , wherein the program instructions of the computer program further comprise:
implementing dynamic decoupling and/or twirling in a quantum circuit that generates said noisy expectation values.
24 . The system as recited in claim 23 , wherein the program instructions of the computer program further comprise:
encoding features of said quantum circuit on a noisy quantum device.
25 . The system as recited in claim 24 , wherein the program instructions of the computer program further comprise:
generating target expectation values using said noisy expectation values and said encoded features, wherein said training data comprises said target expectation values generated using a noiseless simulator or an error mitigated quantum processing unit.