Denoising qubit calibration data with deep learning

1 . A method for calibrating a qubit in operating a quantum computing system, the method comprising:

obtaining, by one or more computing devices, data associated with a set of one or more qubit parameters for a qubit included in quantum computing system, wherein obtaining the data associated with a set of one or more qubit parameters comprises calibrating the qubit using a quantum circuit;

obtaining, by the one or more computing devices, calibration data associated with at least one qubit parameter in the set of one or more qubit parameters, wherein obtaining the calibration data includes at least measuring a value associated with the qubit;

performing, by the one or more computing devices, a first calibration test on the calibration data;

determining, by the one or more computing devices, the calibration data fails the first calibration test;

in response to determining, by the one or more computing devices, the calibration data fails the first calibration test, processing, by the one or more computing devices, the calibration data using the de-corrupting autoencoder to generate reconstructed calibration data;

performing, by the one or more computing devices, a second calibration test on the reconstructed calibration data;

determining, by the one or more computing devices, the value for the at least one qubit parameter based at least in part on the reconstructed calibration data when the reconstructed calibration data passes the second calibration test, and

operating the quantum computing system based on the reconstructed calibration data, wherein operating the quantum computing system includes tuning the qubit in accordance with the determined value for the at least one qubit parameter.

2 . The method of claim 1 , wherein the de-corrupting autoencoder is operable to perform operations, the operations comprising:

encoding, by the one or more computing devices, the calibration data using an encoder network to generate a latent representation of the calibration data; and

decoding, by the one or more computing devices, the latent representation of the calibration data using a decoder network to generate reconstructed calibration data.

3 . The method of claim 2 , wherein the encoder network and the decoder network each comprise a machine learned model trained at least in part using a corrupted training dataset.

4 . The method of claim 3 , wherein the corrupted training dataset comprises images having random speckle noise or masking noise.

5 . The method of any of claim 2 , wherein determining, by the one or more computing devices, the value for the at least one qubit parameter comprises determining, by the one or more computing devices, the value for the at least one qubit parameter based at least in part on the reconstructed calibration data.

6 . The method of claim 1 , wherein prior to processing, by the one or more computing devices, the calibration data using the de-corrupting autoencoder to generate reconstructed calibration data, the method comprises determining, by the one or more computing devices, that the calibration data has not previously been reconstructed using the de-corrupting autoencoder.

7 . The method of claim 1 , wherein determining, by the one or more computing devices, the value for the at least one qubit parameter comprises:

determining, by the one or more computing devices, the calibration data is corrupted;

in response to determining, by the one or more computing devices, the calibration data is corrupted, processing, by the one or more computing devices, the calibration data using the de-corrupting autoencoder to generate reconstructed calibration data; and

performing, by the one or more computing devices, a calibration test on the reconstructed calibration data.

8 . The method of claim 7 , wherein the calibration data is determined to be corrupted using a signal-to-noise based classifier or a machine learned classifier model.

9 . The method of claim 1 , wherein the calibration data is represented in multiple dimensions.

10 . The method of claim 9 , wherein the calibration data is represented as a two-dimensional image.

11 . The method of claim 1 , wherein the data associated with a set of one or more qubit parameters for a qubit in a quantum computing system is represented by a directed graph, the directed graph comprising a node for each qubit parameter in the set of one or more qubit parameters and a directed edge for each dependency between qubit parameters in the set of one or more qubit parameters;

wherein determining, by the one or more computing devices, the value for the at least one qubit parameter is implemented according to a node ancestry ordering associated with the directed graph.

12 . A quantum computing system, comprising:

a quantum system comprising a qubit;

one or more processors;

one or more memory devices, the one or more memory devices storing computer-readable instructions that when executed by the one or more processors cause the one or more processors to perform operations for calibrating the qubit, the operations comprising:

obtaining data associated with a set of one or more qubit parameters for the qubit, wherein obtaining the data associated with a set of one or more qubit parameters comprises calibrating the qubit using a quantum circuit;

obtaining calibration data associated with at least one qubit parameter in the set of one or more qubit parameters, wherein obtaining the calibration data includes at least measuring a value associated with the qubit;

performing a first calibration test on the calibration data;

determining the calibration data fails the first calibration test;

in response to determining the calibration data fails the first calibration test, processing the calibration data using a classical de-corrupting autoencoder to generate reconstructed calibration data;

performing a second calibration test on the reconstructed calibration data;

determining a value for the at least one qubit parameter when the reconstructed calibration data passes the second calibration test; and

operating the quantum computing system based on the reconstructed calibration data, wherein operating the quantum computing system includes tuning the qubit in accordance with the determined value for the at least one qubit parameter.

13 . The quantum computing system of claim 12 , wherein the de-corrupting autoencoder is operable to perform operations, the operations comprising:

encoding the calibration data using an encoder network to generate a latent representation of the calibration data; and

decoding the latent representation of the calibration data using a decoder network to generate reconstructed calibration data.

14 . The quantum computing system of claim 13 , wherein the encoder network and the decoder network each comprise a machine learned model trained at least in part using a corrupted training dataset.

15 . The quantum computing system of any of claim 12 , wherein prior to processing the calibration data using the de-corrupting autoencoder to generate reconstructed calibration data, the operations comprise determining that the calibration data has not previously been reconstructed using the de-corrupting autoencoder.

16 . One or more tangible, non-transitory computer-readable media storing computer-readable instructions for execution by one or more processors to cause the one or more processors to perform operations for calibrating a qubit in a quantum computing system, the operations comprising:

obtaining data associated with a set of one or more qubit parameters for a qubit included in a quantum computing system, wherein obtaining the data associated with a set of one or more qubit parameters comprises calibrating the qubit using a quantum circuit;

obtaining calibration data associated with at least one qubit parameter in the set of one or more qubit parameters, wherein obtaining the calibration data includes at least measuring a value associated with the qubit;

performing a first calibration test on the calibration data;

determining the calibration data fails the first calibration test;

in response to determining the calibration data fails the first calibration test, processing, the calibration data using the de-corrupting autoencoder to generate reconstructed calibration data;

performing a second calibration test on the reconstructed calibration data;

determining the value for the at least one qubit parameter based at least in part on the reconstructed calibration data when the reconstructed calibration data passes the second calibration test; and

operating the quantum computing system based on the reconstructed calibration data, wherein operating the quantum computing system includes tuning the qubit in accordance with the determined value for the at least one qubit parameter.

17 . The non-transitory computer-readable media of claim 16 , wherein the de-corrupting autoencoder is operable to perform operations, the operations comprising:

encoding the calibration data using an encoder network to generate a latent representation of the calibration data; and

decoding the latent representation of the calibration data using a decoder network to generate reconstructed calibration data.

18 . The non-transitory computer-readable media of claim 17 , wherein the encoder network and the decoder network each comprise a machine learned model trained at least in part using a corrupted training dataset.

19 . A method for operating a quantum computing system, comprising:

calibrating one or more qubit parameters of the quantum computing system according to the method of claim 1 to generate calibrated qubit parameters; and

operating the quantum computing system using the calibrated qubit parameters to implement one or more quantum computing circuits.

20 . The one or more tangible non-transitory computer-readable media storing computer-readable instructions of claim 16 , wherein the calibration data is represented in multiple dimensions.