Short depth circuits as quantum classifiers

Patent metadata

Field	Value
Publication number	US-10977546-B2
Application number	US-201715826327-A
Country	US
Kind code	B2
Filing date	Nov 29, 2017
Priority date	Nov 29, 2017
Publication date	Apr 13, 2021
Grant date	Apr 13, 2021

Abstract

Techniques using short depth circuits as quantum classifiers are described. In one embodiment, a system is provided that comprises: quantum hardware, a memory that stores computer-executable components and a processor that executes computer-executable components stored in the memory. In one implementation, the computer-executable components comprise a calibration component that calibrates quantum hardware to generate a short depth quantum circuit. The computer-executable components further comprise a cost function component that determines a cost function for the short depth quantum circuit based on an initial value for a parameter of a machine-learning classifier. The computer-executable components further comprise a training component that modifies the initial value for the parameter during training to a second value for the parameter based on the cost function for the short depth quantum circuit.

First claim

What is claimed is:

1. A system, comprising:
quantum hardware;
a memory that stores computer executable components; and
a processor that executes the computer executable components stored in the memory, wherein the computer executable components comprise:
a calibration component that calibrates the quantum hardware to generate a short depth quantum circuit;
a cost function component that determines a cost function for the short depth quantum circuit based on an initial value for a parameter of a machine-learning classifier; and a training component that modifies the initial value for the parameter during training to a second value for the parameter based on the cost function for the short depth quantum circuit.

2. The system of claim 1 , wherein the computer executable components further comprise:
a machine learning component that determines an output of an activation function of the machine-learning classifier based on the second value of the parameter.

3. The system of claim 1 , wherein the training component encodes labeled training information into at least one qubit via a quantum feature map.

4. The system of claim 1 , wherein the short depth quantum circuit utilizes at least one gate that is naturally accessible on the quantum hardware.

5. The system of claim 1 , wherein the training component utilizes a fixed-frequency superconducting qubit architecture to generate an architecture of the machine-learning classifier.

6. The system of claim 1 , wherein the training component implements a measurement scheme for binary label classification of training information.

7. The system of claim 1 , wherein the training component implements a commuting measurement scheme facilitating multi-label classification of training information.

8. A computer-implemented method, comprising:
calibrating, by a computing system, quantum hardware to generate a short depth quantum circuit;
determining, by the system, a first cost function for the short depth quantum circuit based on an initial value for a parameter of a machine-learning classifier; and
modifying, by the system, the initial value for the parameter during training to a second value for the parameter based on the first cost function for the short depth quantum circuit.

9. The computer-implemented method of claim 8 , further comprising:
determining, by the system, a second cost function that indicates a measurement for a quantum binary state discrimination.

10. The computer-implemented method of claim 8 , further comprising:
determining, by the system, a second cost function that identifies a high-probability multi-label classification scheme for k-ary quantum state discrimination.

11. The computer-implemented method of claim 8 , further comprising:
determining, by the system, a second cost function that identifies a Hilbert space and feature space partition for binary data classification.

12. The computer-implemented method of claim 8 , further comprising:
determining, by the system, a second cost function that identifies a Hilbert space and feature space partition for k-ary data classification.

13. The computer-implemented method of claim 8 , further comprising:
determining, by the system, a feature map that prepares a simple input state for the short depth quantum circuit based on training information.

14. The computer-implemented method of claim 8 , further comprising:
selecting, by the system, the short depth quantum circuit from a hardware-efficient circuit family.

15. A computer program product, the computer program product comprising a computer readable storage medium having program instructions embodied therewith, the program instructions executable by a computing system to cause the computing system to at least:
calibrate quantum hardware to generate a short depth quantum circuit;
determine a cost function for the short depth quantum circuit based on an initial value for a parameter of a machine-learning classifier; and
modify the initial value for the parameter during training to a second value for the parameter based on the cost function for the short depth quantum circuit.

16. The computer program product of claim 15 , wherein the program instructions are further executable by the computing system to cause the computing system to at least:
measure an output statistic of the short depth quantum circuit;
determine a second cost function of a plurality of cost functions based on the output statistic; and
assign the output statistic to the second cost function.

17. The computer program product of claim 16 , wherein the program instructions are further executable by the computing system to cause the computing system to at least:
determine a total cost function based on the second cost function.

18. The computer program product of claim 15 , wherein the program instructions are further executable by the computing system to cause the computing system to at least:
prepare a sample to be classified as an input quantum state with a feature map.

19. The computer program product of claim 18 , wherein the program instructions are further executable by the computing system to cause the computing system to at least:
operate the short depth quantum circuit with the second value of the parameter.

20. The computer program product of claim 19 , wherein the program instructions are further executable by the computing system to cause the computing system to at least:
measure an output from operating the short depth quantum circuit; and
assigning a classification label for the sample to be classified based on the output from operating the short depth quantum circuit.

Assignees

• IBM

Inventors

• Gambetta Jay M
• Havlicek Vojtech
• Temme Paul Kristan

Classifications

• G06N3/048

  Activation functions · CPC title

• G06N3/047

  Probabilistic or stochastic networks · CPC title

• G06N10/60 Primary

  Quantum algorithms, e.g. based on quantum optimisation, quantum Fourier or Hadamard transforms · CPC title

• G06N3/09

  Supervised learning · CPC title

• G06N3/08 Primary

  Learning methods · CPC title