Information processing program, information processing method, and information processing device

Abstract

To reduce the processing time required when performing VQE calculations. [Solution] When an iteration has been performed a predetermined number of times, the information processing device 100 temporarily suspends the VQE calculation 150 and determines the number of level values ​​that can be set as the value of each first parameter 121 based on the first contribution of each first parameter 121. Based on the determined number, the information processing device 100 calculates the second contribution of each first parameter 121 based on two or more patterns 130 that represent the combinations of level values ​​to be set as the value of each first parameter 121. The information processing device 100 sets each of the two second parameters 141 that are judged to have a relatively high second contribution to the target parameter 151.

Description

[Technical Field]

[0001] The present invention relates to an information processing program, an information processing method, and an information processing apparatus. [Background technology]

[0002] Traditionally, in fields such as materials development or drug discovery research, there is a method called VQE (Variational Quantum Eigensolver) for performing quantum chemical calculations to investigate the properties of a target molecule or atom. VQE involves performing a series of operations called iterations repeatedly until the convergence conditions are met. In the following explanation, this calculation may be referred to as "VQE calculation." An iteration is, for example, the execution of a variational quantum circuit, the calculation of the Hamiltonian expectation value based on the quantum state obtained by executing the variational quantum circuit, and the updating of the variational quantum circuit parameters to minimize the Hamiltonian expectation value.

[0003] Prior art includes, for example, the recursive removal of gates from quantum circuits. Another technique involves determining the coefficient values ​​used in updating the parameters of variational quantum circuits to values ​​that periodically change between higher and lower than a predetermined reference value as the number of updates increases. Furthermore, there is a technique for optimizing the parameters of variational quantum circuits within a cluster while fixing other parameters of variational quantum circuits outside the cluster. Finally, there is a technique for updating parameter importance based on information regarding changes in the position of material data points. [Prior art documents] [Patent Documents]

[0004] [Patent Document 1] U.S. Patent Application Publication No. 2021 / 0133617 [Patent Document 2] International Publication No. 2023 / 243011 [Patent Document 3] U.S. Patent No. 011645442 [Patent Document 4] Japanese Patent Publication No. 2020-128962 [Overview of the project] [Problems that the invention aims to solve]

[0005] However, conventional techniques have a problem in that the processing time required to perform VQE calculations increases. For example, the more parameters that define a variational quantum circuit there are, the more the processing time required for each iteration increases, or the more iterations are required, which increases the processing time required to perform VQE calculations.

[0006] In one aspect, the present invention aims to reduce the processing time required when performing VQE calculations. [Means for solving the problem]

[0007] According to one embodiment, when iterations are repeatedly performed to update a set target parameter among a plurality of parameters defining a variational quantum circuit according to a variational quantum eigenvalue solver method, when the iterations are performed a predetermined number of times, the number of level values ​​that can be set as the value of each first parameter is determined based on the first contribution of each of the one or more first parameters to the amount of energy change corresponding to the quantum state represented by the variational quantum circuit, and one of the water values ​​of the number of level values ​​determined for each first parameter An information processing program, information processing method, and information processing device are proposed that, when setting standard values, calculate a second contribution of each of the first parameters to the variational quantum eigenvalue solver method based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of two or more patterns representing the combination of level values ​​to be set as the values ​​of each of the first parameters, and set each of the second parameters of the one or more first parameters that are judged to have a relatively high calculated second contribution based on the calculated second contribution to the target parameter. [Effects of the Invention]

[0008] According to one embodiment, it becomes possible to reduce the processing time required when performing VQE calculations. [Brief explanation of the drawing]

[0009] [Figure 1] Figure 1 is an explanatory diagram showing one embodiment of the information processing method according to the embodiment. [Figure 2] Figure 2 is an explanatory diagram showing an example of the information processing system 200. [Figure 3] Figure 3 is a block diagram showing an example of the hardware configuration of the information processing device 100. [Figure 4] Figure 4 is a block diagram showing an example of the hardware configuration of the quantum computing device 201. [Figure 5] Figure 5 is a block diagram showing an example of the functional configuration of the information processing device 100. [Figure 6] Figure 6 is an explanatory diagram (part 1) showing an example of the operation of the information processing device 100. [Figure 7] Figure 7 is an explanatory diagram (part 2) showing an example of the operation of the information processing device 100. [Figure 8] Figure 8 is an explanatory diagram (part 3) showing an example of the operation of the information processing device 100. [Figure 9] Figure 9 is an explanatory diagram (part 4) showing an example of the operation of the information processing device 100. [Figure 10] Figure 10 is an explanatory diagram (part 5) showing an example of the operation of the information processing device 100. [Figure 11] Figure 11 is an explanatory diagram (part 6) showing an example of the operation of the information processing device 100. [Figure 12] Figure 12 is an explanatory diagram (part 7) showing an example of the operation of the information processing device 100. [Figure 13] Figure 13 is an explanatory diagram (part 1) showing a specific example of the operation of the information processing device 100. [Figure 14] Figure 14 is an explanatory diagram (part 2) showing a specific example of the operation of the information processing device 100. [Figure 15] Figure 15 is an explanatory diagram (part 3) showing a specific example of the operation of the information processing device 100. [Figure 16] Figure 16 is an explanatory diagram (part 4) showing a specific example of the operation of the information processing device 100. [Figure 17] Figure 17 is an explanatory diagram (part 5) showing a specific example of the operation of the information processing device 100. [Figure 18] Figure 18 is an explanatory diagram (part 6) showing a specific example of the operation of the information processing device 100. [Figure 19] Figure 19 is an explanatory diagram (part 7) showing a specific example of the operation of the information processing device 100. [Figure 20] Figure 20 is a flowchart (part 1) showing an example of the overall processing procedure. [Figure 21]Figure 21 is a flowchart (part 2) showing an example of the overall processing procedure. [Figure 22] Figure 22 is a flowchart showing an example of the first review process. [Figure 23] Figure 23 is a flowchart showing an example of the second review process. [Modes for carrying out the invention]

[0010] Embodiments of the information processing program, information processing method, and information processing apparatus according to the present invention will be described in detail below with reference to the drawings.

[0011] (An embodiment of the information processing method according to the embodiment) Figure 1 is an explanatory diagram showing one embodiment of the information processing method according to the embodiment. The information processing device 100 is a computer for performing VQE calculations. The information processing device 100 is, for example, a server or a PC (Personal Computer).

[0012] VQE calculation involves repeatedly performing a series of operations called iterations until the convergence conditions are met. An iteration is a series of operations such as executing a variational quantum circuit, calculating the expectation value of the Hamiltonian based on the quantum state obtained from executing the variational quantum circuit, and updating the parameters of the variational quantum circuit to minimize the expectation value of the Hamiltonian. The Hamiltonian corresponds to energy. The convergence condition is, for example, that the expectation value of the Hamiltonian falls below a threshold.

[0013] One possible scenario for performing VQE calculations is a collaborative effort between a quantum computer and a classical computer. For example, the quantum computer calculates the expectation value of the Hamiltonian based on the quantum states obtained by executing a variational quantum circuit. Meanwhile, the classical computer updates the parameters of the variational quantum circuit. In this case, the quantum computer and the classical computer communicate with each iteration.

[0014] Herein lies a problem: conventional methods for performing VQE calculations have resulted in increased processing time. For example, as the size of the system performing the VQE calculation increases, the number of parameters defining the variational quantum circuit tends to increase. The more parameters defining the variational quantum circuit there are, the more processing time is required for each iteration, and the more iterations are needed to satisfy the convergence condition, thus increasing the processing time for VQE calculations. Furthermore, the more parameters defining the variational quantum circuit there are, the more communication occurs between the quantum computer and the classical computer, further increasing the processing time for VQE calculations.

[0015] Therefore, it is desirable to reduce the processing time required when performing VQE calculations. However, if we attempt to reduce the processing time required when performing VQE calculations by reducing the number of parameters that define the variational quantum circuit, there is a problem in that the accuracy of the VQE calculation decreases.

[0016] In contrast, a first method can be considered in which, for example, based on the results of the first iteration, the parameters to be set as the target parameters for updating among the multiple parameters that define the variational quantum circuit are limited, and then subsequent iterations are performed. For this first method, see, for example, reference 1 below.

[0017] Reference 1: International Publication No. 2023 / 144884

[0018] In this first method, the parameters to be set as target parameters are limited based solely on the results of the first iteration. This means that useful parameters in the VQE calculation may be excluded by not being set as target parameters. Consequently, this first method may converge to a local minimum, which can reduce the accuracy of the VQE calculation.

[0019] Furthermore, a second method can be considered in which, for example, after performing an iteration a predetermined number of times, the parameter set as the target parameter whose value is to be updated among the multiple parameters that define the variational quantum circuit is reviewed, and the iteration continues. Specifically, in this second method, multiple level values ​​common to multiple parameters are prepared. Specifically, in this second method, two or more patterns are prepared, each representing a combination of level values ​​to be set for each parameter when one of the level values ​​is set for each parameter according to an orthogonal array. Specifically, in this second method, the parameter set as the target parameter whose value is to be updated among the multiple parameters is reviewed based on the calculation result of a predetermined cost function for each of the two or more prepared patterns.

[0020] This second method may reduce the accuracy of the VQE calculation. For example, this second method uses multiple level values ​​common to multiple parameters, thus failing to consider the contribution of each parameter to the cost function, which can reduce the accuracy of the VQE calculation. Specifically, this second method cannot properly evaluate parameters for which it is preferable to consider a relatively large number of level values. Specifically, this second method increases the processing time required to perform the VQE calculation if the number of level values ​​is increased.

[0021] Therefore, this embodiment describes an information processing method that can reduce the processing time required when performing VQE calculations. Specifically, the information processing method can reduce the processing time required when performing VQE calculations while maintaining the accuracy of the VQE calculations.

[0022] In Figure 1, the information processing device 100 stores a variational quantum circuit 110 related to the VQE calculation 150. The information processing device 100 stores, for example, a plurality of parameters 111 that define the variational quantum circuit 110. The parameters 111 relate to the quantum gates that form the variational quantum circuit 110. The parameters 111 are, for example, the rotation angles of the quantum gates.

[0023] The information processing device 100 can control the arithmetic unit 101. The arithmetic unit 101 can execute the variational quantum circuit 110. The arithmetic unit 101 is, for example, a real quantum computer. The arithmetic unit 101 may also be a quantum simulator that simulates a quantum computer. The quantum simulator may exist outside the information processing device 100, for example. The quantum simulator may be included in the information processing device 100, for example. The information processing device 100 uses the arithmetic unit 101 to start the VQE calculation 150. Specifically, the information processing device 100 uses the arithmetic unit 101 to start repeatedly performing the iteration 152.

[0024] The VQE calculation 150 involves repeatedly performing iteration 152 until the convergence condition is met. Iteration 152 is a series of processes that update the set target parameters. For example, iteration 152 is a series of processes that involves executing a variational quantum circuit, calculating the expectation value of the Hamiltonian based on the quantum state obtained by executing the variational quantum circuit, and updating the parameters of the variational quantum circuit to minimize the expectation value of the Hamiltonian. The Hamiltonian corresponds to energy. The convergence condition is, for example, that the expectation value of the Hamiltonian becomes less than or equal to a threshold. The target parameters are, for example, each of the multiple parameters 111 set.

[0025] When iteration 152 has been performed a predetermined number of times, the information processing device 100 temporarily suspends the VQE calculation 150, and after resetting the target parameters by performing the following processes (1-1) and (1-2), resumes the VQE calculation 150. The predetermined number of iterations is, for example, set in advance by the user. Multiple predetermined number of iterations may be set. The predetermined number of iterations is, for example, ax + b. a is a coefficient. x is an integer. b is a constant. a is, for example, 5. b is, for example, 1.

[0026] (1-1) The information processing device 100 obtains a first contribution of one or more first parameters 121 from among the multiple parameters 111 to the change in energy corresponding to the quantum state represented by the variational quantum circuit 110. The first contribution corresponds, for example, to the ratio of the change in energy corresponding to the quantum state represented by the variational quantum circuit to the change in the first parameter. The change in the first parameter is, for example, the difference between the value of the first parameter in the previous iteration 152 and the value of the first parameter in the current iteration 152. The change in energy is, for example, the difference between the energy value in the previous iteration 152 and the energy value in the current iteration 152. The information processing device 100 obtains the first contribution by, for example, calculating the first contribution.

[0027] The information processing device 100 determines, based on the acquired first contribution, the number of level values ​​122 that can be set as values ​​for each first parameter 121. The level values ​​122 are, for example, discrete values ​​that can be set as values ​​for the first parameter 121. The information processing device 100 determines, for example, the number of level values ​​122 that can be set as values ​​for each first parameter 121, such that the larger the first contribution, the greater the number of level values ​​122 that can be set as values ​​for the first parameter 121. As a result, the information processing device 100 can prepare a relatively large number of level values ​​122 for first parameters that have a large first contribution and are judged to be of high importance in the VQE calculation 150, and can examine in detail whether or not to set them as target parameters.

[0028] (1-2) The information processing device 100 prepares a determined number of level values ​​122 for each first parameter 121. The information processing device 100 identifies two or more patterns 130 that represent the combinations of level values ​​122 to be set as the value of each first parameter 121 when setting one of the prepared level values ​​122 for each first parameter 121. From the viewpoint of suppressing an increase in processing load, it is preferable for the information processing device 100 to identify two or more of the patterns 130 out of all patterns 130 that represent the combinations of level values ​​122 to be set as the value of each first parameter 121. The information processing device 100 may, for example, identify all patterns 130. Specifically, the information processing device 100 identifies two or more patterns 130 by referring to an orthogonal array that limits the combinations of level values ​​122 to be set as the value of each first parameter 121 in accordance with experimental design. This allows the information processing device 100 to identify which patterns 130 should be tested to evaluate the contribution of each first parameter 121 to the VQE calculation 150.

[0029] The information processing device 100 obtains the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit 110 in each of the two or more identified patterns 130. The energy calculation results correspond to the calculation results of the cost function. The cost function is a function that returns the expectation value of the Hamiltonian. The information processing device 100 obtains the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit 110 in each of the patterns 130, for example, by using the arithmetic unit 101.

[0030] The information processing device 100 calculates the second contribution to the VQE calculation 150 based on the acquired calculation results, for each first parameter 121. The information processing device 100 evaluates the second contribution by, for example, calculating the correlation coefficient between the dependent variable corresponding to energy and the independent variables corresponding to each first parameter 121, based on the acquired calculation results. The second contribution is the correlation coefficient. For example, a larger value of the correlation coefficient is considered to indicate a higher second contribution to the VQE calculation 150.

[0031] Based on the calculated second contribution, the information processing device 100 selects one or more second parameters 141 from among one or more first parameters 121 whose calculated second contribution is judged to be relatively high. For example, the information processing device 100 selects one or more second parameters 141 from among one or more first parameters 121 whose calculated second contribution is above a threshold. The information processing device 100 sets each of the selected one or more second parameters 141 as a target parameter 151. This allows the information processing device 100 to identify useful parameters 111 in the VQE calculation 150 and to accurately set the target parameter 151 whose value is updated in the VQE calculation 150.

[0032] In this way, the information processing device 100 can narrow down the useful parameters 111 in the VQE calculation 150 and set them as target parameters 151. Furthermore, the information processing device 100 can deal with changes in the useful parameters 111 in the VQE calculation 150 while the VQE calculation 150 is being performed. In addition, when narrowing down the useful parameters 111, the information processing device 100 can adjust the number of level values ​​122 that can be set as the values ​​of each parameter, and adjust the degree of detail in examining the usefulness of each parameter on a parameter-by-parameter basis. As a result, the information processing device 100 can reduce the processing time required to perform the VQE calculation 150 while maintaining the accuracy of the VQE calculation 150. Therefore, the information processing device 100 can achieve both accuracy and efficiency in the VQE calculation 150.

[0033] Here, we have described the case where the functions of the information processing device 100 are realized by a single computer, but this is not the only case. For example, the functions of the information processing device 100 may be realized through the collaboration of multiple computers. For example, the functions of the information processing device 100 may be realized on the cloud.

[0034] (An example of information processing system 200) Next, using Figure 2, we will describe an example of an information processing system 200 to which the information processing device 100 shown in Figure 1 is applied.

[0035] Figure 2 is an explanatory diagram showing an example of an information processing system 200. In Figure 2, the information processing system 200 includes an information processing device 100, a quantum computing device 201, and a client device 202.

[0036] In the information processing system 200, the information processing device 100 and the quantum computing device 201 are connected via a wired or wireless network 210. The network 210 is, for example, a LAN (Local Area Network), a WAN (Wide Area Network), or the Internet. Also in the information processing system 200, the information processing device 100 and the client device 202 are connected via a wired or wireless network 210.

[0037] The information processing device 100 is a computer that controls VQE calculations. The information processing device 100 receives a processing request that requests the solution of a target problem. The processing request includes, for example, information defining the problem. The processing request includes, for example, information that makes it possible to identify a predetermined variational quantum circuit to be used in the VQE calculation. Specifically, the processing request includes information that makes it possible to identify a plurality of parameters that define the predetermined variational quantum circuit. The information processing device 100 receives the processing request, for example, by receiving it from the client device 202. The information processing device 100 may also receive the processing request, for example, based on user input.

[0038] The information processing device 100, in response to a processing request, starts a VQE calculation in cooperation with the quantum computing device 201. The information processing device 100, for example, resets the target parameter whose value is to be updated when the iteration has been performed a predetermined number of times, as in Figure 1. The predetermined number of iterations is, for example, set in advance by the user. The information processing device 100, for example, when the iteration has been performed a predetermined number of times, selects from among several parameters which parameters to set as the target parameter and which parameters not to set as the target parameter, and resets the target parameter. The information processing device 100 outputs the result of the VQE calculation. The information processing device 100, for example, sends the result of the VQE calculation to the client device 202. The information processing device 100 may, for example, output the result of the VQE calculation so that the user can refer to it. The information processing device 100 is, for example, a server or a PC.

[0039] The quantum computing device 201 is a computer that performs the requested computational processing. The quantum computing device 201 is capable of performing quantum computations. The quantum computing device 201 may also be capable of performing classical computations. The quantum computing device 201 performs quantum computations according to the control of the information processing device 100. The quantum computing device 201 returns the results of the quantum computations to the information processing device 100. The quantum computing device 201 is, for example, a physical quantum computer. The quantum computing device 201 may also be, for example, a classical computer that starts a quantum simulator. A classical computer is, for example, a server or a PC.

[0040] The client device 202 is a computer used by a user who wishes to perform a VQE calculation. Based on the user's input, the client device 202 generates a processing request that requests the solution of the target problem and sends it to the information processing device 100. The client device 202 receives the results of the VQE calculation from the information processing device 100. The client device 202 outputs the results of the VQE calculation so that the user can refer to them. The client device 202 may be, for example, a PC, a tablet terminal, or a smartphone.

[0041] This explanation describes a case where the information processing device 100 and the quantum computing device 201 are different devices, but the explanation is not limited to this case. For example, the information processing device 100 may have the functionality of a quantum computing device 201 and may operate as a quantum computing device 201. In this case, the information processing system 200 does not need to include a quantum computing device 201.

[0042] Furthermore, although we have described the case where the information processing device 100 and the client device 202 are different devices, this is not the only case. For example, the information processing device 100 may have the functionality of a client device 202 and may operate as a client device 202. In this case, the information processing system 200 does not need to include a client device 202.

[0043] (Examples of applications of Information Processing System 200) The information processing system 200 can be applied to fields such as materials development and pharmaceutical development. Specifically, the information processing system 200 can be applied to perform VQE calculations to solve problems related to molecules.

[0044] (Example of hardware configuration of information processing device 100) Next, an example of the hardware configuration of the information processing device 100 will be described using Figure 3.

[0045] Figure 3 is a block diagram showing an example of the hardware configuration of the information processing device 100. In Figure 3, the information processing device 100 includes a CPU (Central Processing Unit) 301, memory 302, network interface 303, recording medium interface 304, and recording medium 305. Each component is connected by a bus 300.

[0046] Here, the CPU 301 is responsible for the overall control of the information processing device 100. The memory 302 includes, for example, ROM (Read Only Memory), RAM (Random Access Memory), and flash ROM. Specifically, for example, flash ROM and ROM store various programs, and RAM is used as the work area for the CPU 301. Programs stored in memory 302 are loaded into the CPU 301, causing the CPU 301 to execute the coded processes.

[0047] The network interface 303 is connected to network 210 via a communication line, and then connects to other computers via network 210. The network interface 303 manages the internal interface with network 210 and controls the input and output of data from other computers. The network interface 303 is, for example, a modem or a LAN adapter.

[0048] The recording medium interface (I / F) 304 controls the reading and writing of data to the recording medium 305 according to the control of the CPU 301. The recording medium interface (I / F) 304 is, for example, a disk drive, an SSD (Solid State Drive), or a USB (Universal Serial Bus) port. The recording medium 305 is a non-volatile memory that stores the data written under the control of the recording medium interface (I / F) 304. The recording medium 305 is, for example, a disk, semiconductor memory, or USB memory. The recording medium 305 may be detachable from the information processing device 100.

[0049] In addition to the components described above, the information processing device 100 may also have, for example, a keyboard, mouse, display, printer, scanner, microphone, speaker, etc. Furthermore, the information processing device 100 may have multiple recording medium interfaces 304 and recording mediums 305. Alternatively, the information processing device 100 may not have recording medium interfaces 304 and recording mediums 305.

[0050] (Example hardware configuration of quantum computing device 201) In the case where the quantum computing device 201 is a classical computer that starts a quantum simulator, the hardware configuration example of the quantum computing device 201 is specifically the same as the hardware configuration example of the information processing device 100 shown in Figure 3, so the explanation is omitted.

[0051] On the other hand, it is possible that the quantum computing device 201 is a physical quantum computer. Here, using Figure 4, we will explain an example of the hardware configuration of the quantum computing device 201 when it is a physical quantum computer.

[0052] Figure 4 is a block diagram showing an example of the hardware configuration of the quantum computing device 201. In Figure 4, the quantum computing device 201 includes a CPU 401, a memory 402, a network interface 403, a recording medium interface 404, and a recording medium 405. The quantum computing device 201 further includes a computing chassis interface 406 and a quantum computing chassis 407. Each component is connected by a bus 400.

[0053] Here, the CPU 401 is responsible for the overall control of the quantum computing device 201. Memory 402 includes, for example, ROM, RAM, and flash ROM. Specifically, for example, flash ROM and ROM store various programs, and RAM is used as the work area for CPU 401. Programs stored in memory 402 are loaded into CPU 401, causing CPU 401 to execute the coded processes.

[0054] The network interface 403 is connected to network 210 via a communication line, and then connects to other computers via network 210. The network interface 403 manages the internal interface with network 210 and controls the input and output of data from other computers. The network interface 403 is, for example, a modem or a LAN adapter.

[0055] The recording medium interface (I / F) 404 controls the reading and writing of data to the recording medium (SSD) 405 according to the control of the CPU 401. The recording medium interface (I / F) 404 is, for example, a disk drive, SSD, or USB port. The recording medium (SSD) 405 is a non-volatile memory that stores the data written under the control of the recording medium interface (I / F) 404. The recording medium (SSD) 405 is, for example, a disk, semiconductor memory, or USB memory. The recording medium (SSD) 405 may be detachable from the quantum computing device (SSD) 201.

[0056] The computing chassis interface 406 controls access to the quantum computing chassis 407 according to the control of the CPU 401. The computing chassis interface 406 uses a microwave pulse generator to convert the output signal from the CPU 401 into an input signal for the quantum computing chassis 407 and transmits it to the quantum computing chassis 407. The computing chassis interface 406 uses a microwave pulse demodulator to convert the output signal from the quantum computing chassis 407 into an input signal for the CPU 401 and transmits it to the CPU 401. The quantum computing chassis 407 is a computing device equipped with one or more qubit chips cooled to an extremely low temperature of 10 mK. A qubit chip represents, for example, a logical qubit. The quantum computing chassis 407 uses one or more qubit chips to perform a predetermined operation in response to an input signal and outputs an output signal corresponding to the result of the predetermined operation.

[0057] In addition to the components described above, the quantum computing device 201 may also have, for example, a keyboard, mouse, display, printer, scanner, microphone, speaker, etc. Furthermore, the quantum computing device 201 may have multiple recording medium interfaces 404 and 405. Alternatively, the quantum computing device 201 may not have recording medium interfaces 404 and 405. Also, the qubit chip in the quantum computing chassis 407 may be controlled by methods other than microwaves. For example, the qubit chip in the quantum computing chassis 407 may implement optical qubits.

[0058] (Example hardware configuration for client device 202) The hardware configuration example for client device 202 is specifically the same as the hardware configuration example for information processing device 100 shown in Figure 3, so a detailed explanation is omitted.

[0059] (Example of the functional configuration of the information processing device 100) Next, an example of the functional configuration of the information processing device 100 will be described using Figure 5.

[0060] Figure 5 is a block diagram showing an example of the functional configuration of the information processing device 100. The information processing device 100 includes a storage unit 500, an acquisition unit 501, a determination unit 502, a setting unit 503, an implementation unit 504, and an output unit 505.

[0061] The information processing device 100 can utilize an arithmetic unit 510. The arithmetic unit 510 may, for example, be located outside the information processing device 100. The arithmetic unit 510 may, for example, be located inside the information processing device 100. The arithmetic unit 510 may, for example, be a quantum computing device 201.

[0062] The storage unit 500 is implemented by a storage area such as the memory 302 or recording medium 305 shown in Figure 3. The following description will focus on the case where the storage unit 500 is included in the information processing device 100, but is not limited to this case. For example, the storage unit 500 may be included in a device different from the information processing device 100, and the contents of the storage unit 500 may be accessible from the information processing device 100.

[0063] The acquisition unit 501 to the output unit 505 function as an example of a control unit. Specifically, the acquisition unit 501 to the output unit 505 realize their functions, for example, by having the CPU 301 execute a program stored in a storage area such as the memory 302 or recording medium 305 shown in Figure 3, or by using the network I / F 303. The processing results of each functional unit are stored in a storage area such as the memory 302 or recording medium 305 shown in Figure 3.

[0064] The memory unit 500 stores various information that is referenced or updated during the processing of each functional unit. The memory unit 500 stores, for example, circuit information that defines a variational quantum circuit used for VQE calculation. The circuit information includes, for example, information that identifies a plurality of parameters that define the variational quantum circuit. The parameters relate to the quantum gates that form the variational quantum circuit. The parameters are, for example, the rotation angles of the quantum gates. The parameters may take any of a plurality of discrete values. The circuit information is acquired, for example, by the acquisition unit 501. The circuit information may be pre-set by, for example, the user.

[0065] The memory unit 500 stores, for example, the number of level values ​​that can be set as the value of each of the first parameters, one or more of the multiple parameters. The level values ​​are, for example, discrete values ​​that can be set as the value of the first parameter. The number is determined, for example, by the determination unit 502.

[0066] The memory unit 500 stores, for example, level values ​​that can be set as the values ​​of one or more first parameters among a plurality of parameters. The level values ​​are set, for example, by the determination unit 502. Specifically, the level value is ((maximum value - minimum value) / (m-1)) × i + minimum value. The maximum value is the maximum value that can be set as the value of the first parameter. The minimum value is the minimum value that can be set as the value of the first parameter. m is the number of level values. i is the index of the level value. i is 1, 2, ..., m.

[0067] The memory unit 500 stores, for example, an orthogonal array that limits the combinations of level values ​​to be set as the values ​​of one or more first parameters among multiple parameters. The orthogonal array is, for example, information defined according to experimental design. The orthogonal array is generated, for example, by the setting unit 503.

[0068] The acquisition unit 501 acquires various types of information used in the processing of each functional unit. The acquisition unit 501 stores the acquired information in the storage unit 500 or outputs it to each functional unit. The acquisition unit 501 may also output the information stored in the storage unit 500 to each functional unit. The acquisition unit 501 acquires various types of information, for example, based on user input. The acquisition unit 501 may also receive various types of information from a device other than the information processing device 100, for example.

[0069] The acquisition unit 501 acquires, for example, a processing request that requests to perform a VQE calculation to solve the target problem. The processing request may include, for example, circuit information that defines the variational quantum circuit to be used in the VQE calculation. Specifically, the acquisition unit 501 acquires a processing request by receiving the input of a processing request. Specifically, the acquisition unit 501 may acquire a processing request by receiving a processing request from another computer. The other computer is, for example, a client device 202.

[0070] The acquisition unit 501 acquires, for example, circuit information that defines a variational quantum circuit used in VQE calculations. Specifically, the acquisition unit 501 acquires circuit information by extracting it from a processing request. Specifically, the acquisition unit 501 may acquire circuit information by accepting circuit information as input. Specifically, the acquisition unit 501 may acquire circuit information by receiving circuit information from another computer. The other computer is, for example, a client device 202.

[0071] The acquisition unit 501 may receive a start trigger to initiate processing in any of the functional units. A start trigger may be, for example, a predetermined operation input by a user. A start trigger may also be, for example, the receipt of predetermined information from another computer. A start trigger may also be, for example, the output of predetermined information by any of the functional units. The acquisition unit 501 accepts, for example, the acquisition of a processing request as a start trigger to initiate processing in the decision unit 502, the setting unit 503, and the implementation unit 504.

[0072] The implementation unit 504 repeatedly performs iterations to update a set target parameter among several parameters defining the variational quantum circuit, in accordance with the VQE, until the termination condition is met. The iterations include, for example, executing the variational quantum circuit, calculating the expectation value of the Hamiltonian based on the quantum state obtained by executing the variational quantum circuit, and updating the parameters of the variational quantum circuit to minimize the expectation value of the Hamiltonian. The update is, for example, by changing the value of the parameter.

[0073] When the iteration is performed for the first time, the target parameter is, for example, each of the parameters of a group of parameters. When the iteration is performed for the first time, the target parameter may also be, for example, one or more parameters randomly selected from the group of parameters. The implementation unit 504 repeatedly performs the iteration by, for example, having the calculation unit execute a variational quantum circuit. The calculation unit is capable of executing variational quantum circuits. As a result, the implementation unit 504 can perform VQE calculations.

[0074] The implementation unit 504 may temporarily suspend the VQE calculation when iterating a predetermined number of times. The predetermined number of times may be set to multiple values, for example. The predetermined number of times may include, for example, a first number and a second number. The second number is greater than the first number. The second number may be set to multiple values. When the VQE calculation is temporarily suspended, the information processing device 100 resets the target parameters using the determination unit 502 and the setting unit 503, as described later. The implementation unit 504 resumes the VQE calculation in accordance with the reset of the target parameters. This allows the implementation unit 504 to reduce the processing time required to perform the VQE calculation while maintaining the accuracy of the VQE calculation.

[0075] The decision unit 502 selects one or more first parameters from among multiple parameters when the iteration has been performed a predetermined number of times. For example, the decision unit 502 may select each of the multiple parameters as the first parameter. For example, the decision unit 502 may select a parameter other than the target parameter as the first parameter.

[0076] Specifically, the decision unit 502 may select a parameter other than the target parameter from among a plurality of parameters as the first parameter each time the iteration is performed a predetermined number of times. Specifically, when the iteration is performed for the first time, the decision unit 502 may select each of the plurality of parameters as the first parameter. Specifically, when the iteration is performed for the second time, the decision unit 502 may select each of the plurality of parameters as the first parameter. Specifically, when the iteration is performed for the second time, the decision unit 502 may select a parameter other than the target parameter from among a plurality of parameters as the first parameter.

[0077] The determination unit 502 obtains a first contribution for each of the one or more selected first parameters to the change in energy corresponding to the quantum state represented by the variational quantum circuit. The first contribution corresponds to the ratio of the change in energy corresponding to the quantum state represented by the variational quantum circuit to the change in the first parameter. The determination unit 502 obtains the first contribution, for example, by calculating the first contribution.

[0078] The determination unit 502 determines the number of level values ​​that can be set as the value of each first parameter for each first parameter, based on the acquired first contribution. For example, based on the first contribution of each first parameter, the determination unit 502 determines the number of level values ​​that can be set as the value of each first parameter such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter. Specifically, each time the iteration is performed a predetermined number of times, the determination unit 502 determines the number of level values ​​that can be set as the value of each first parameter such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter.

[0079] Specifically, when the iteration is performed for the first time, the decision unit 502 may determine the number of level values ​​that can be set as the value of each first parameter such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of each first parameter. Specifically, when the iteration is performed for the second time, the decision unit 502 determines the number of level values ​​that can be set as the value of each first parameter such that the number of level values ​​for the target parameter (first parameter) is smaller than that for the other parameters. Specifically, the decision unit 502 may further determine the number of level values ​​that can be set as the value of each first parameter based on the first contribution of each first parameter.

[0080] The determination unit 502 associates and sets the determined number of level values ​​for each first parameter. For example, the determination unit 502 associates the determined number of level values ​​((maximum value - minimum value) / (m-1))×i + minimum value for each first parameter. This allows the determination unit 502 to prepare a relatively large number of level values ​​for first parameters that have a large first contribution and are judged to be of high importance in the VQE calculation, and to consider in detail whether or not to set them for the target parameter.

[0081] The setting unit 503 identifies two or more patterns representing each combination of level values ​​to be set as the value of each first parameter. The level value to be set as the value of the first parameter is one of the determined number of level values. The setting unit 503 obtains or generates an orthogonal array that limits the combinations of level values ​​to be set as the value of each first parameter, for example, according to experimental design. The setting unit 503 identifies two or more patterns from among all patterns representing the combinations of level values ​​to be set as the value of each first parameter, for example, by referring to the orthogonal array. This allows the setting unit 503 to identify which patterns should be tested in order to evaluate the contribution of each first parameter to the VQE calculation. Therefore, the setting unit 503 can appropriately select the patterns to be tested and avoid comprehensively testing all patterns representing the combinations of level values ​​to be set as the value of each first parameter. The setting unit 503 can efficiently evaluate the contribution of each first parameter to the VQE calculation and reduce the amount of processing required.

[0082] The setting unit 503 obtains a second contribution to the variational quantum eigenvalue solver method based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more identified patterns, using each first parameter. The setting unit 503 obtains the second contribution, for example, by calculating the second contribution.

[0083] Specifically, the setting unit 503 obtains the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more patterns. Specifically, based on the obtained calculation results, the setting unit 503 calculates the second contribution with respect to the first parameter from the correlation coefficient between the objective variable corresponding to the energy and the explanatory variable corresponding to each first parameter. The second contribution is, for example, the correlation coefficient itself.

[0084] Specifically, the setting unit 503 obtains calculation results for the energy corresponding to the quantum state represented by the variational quantum circuit for each of two or more patterns. Specifically, based on the obtained calculation results, the setting unit 503 identifies a regression model that includes a target variable corresponding to the energy, explanatory variables corresponding to each first parameter, and coefficients relating to the explanatory variables. The setting unit 503 calculates a second contribution for each first parameter from the coefficients relating to the explanatory variables corresponding to each first parameter in the identified regression model. The second contribution is, for example, the coefficient itself.

[0085] Based on the acquired second contribution, the setting unit 503 sets the second parameter of one or more second parameters that are judged to have a relatively high calculated second contribution from among the one or more first parameters as the target parameter. This allows the setting unit 503 to identify parameters that are useful in the VQE calculation and to appropriately set the target parameters whose values ​​will be updated in the VQE calculation.

[0086] Based on the acquired second contribution, the setting unit 503 may remove from the target parameter one or more third parameters that are currently set as the target parameter and whose acquired second contribution is judged to be relatively low. This allows the setting unit 503 to identify useful parameters in the VQE calculation and to appropriately set the target parameter whose value will be updated in the VQE calculation.

[0087] The output unit 505 outputs the processing result of at least one of the functional units. The output format can be, for example, display on a screen, print to a printer, transmit to an external device via the network interface 303, or store in a storage area such as the memory 302 or recording medium 305. This allows the output unit 505 to notify the user of the processing result of at least one of the functional units, thereby improving the usability of the information processing device 100.

[0088] The output unit 505 outputs, for example, the results of the VQE calculation performed by the implementation unit 504. The output unit 505 transmits, for example, the results of the VQE calculation to another computer. The other computer is, for example, the client device 202. The output unit 505 may also output, for example, the results of the VQE calculation so that a user can refer to them.

[0089] (An example of the operation of the information processing device 100) Next, an example of the operation of the information processing device 100 will be explained using Figures 6 to 12.

[0090] Figures 6 to 12 are explanatory diagrams showing an example of the operation of the information processing device 100. In Figures 6 to 12, the information processing device 100 selects useful parameters in the VQE calculation and limits the parameters whose values ​​are updated in order to reduce the processing time required to perform the VQE calculation while maintaining the accuracy of the VQE calculation.

[0091] Here, when the information processing device 100 selects useful parameters in the VQE calculation, it evaluates the first contribution of each of the one or more parameters to the change in energy. The first contribution is, for example, the ratio of the change in energy to the change in the parameter. The first contribution corresponds to the energy contribution rate. Based on the evaluated first contribution, the information processing device 100 accurately evaluates the differences in importance between the parameters and adjusts the number of level values ​​that can be set as the values ​​of each parameter.

[0092] Furthermore, when selecting useful parameters in the VQE calculation, the information processing device 100 evaluates the second contribution of one or more parameters to the VQE calculation based on an orthogonal array, in accordance with experimental design. The orthogonal array, for example, limits the combinations of the adjusted number of level values ​​that can be set as the values ​​of each parameter.

[0093] The information processing device 100 identifies two or more patterns representing combinations of level values ​​to be set as the values ​​of one or more parameters, based on an orthogonal array, in accordance with experimental design. Here, the information processing device 100 can accurately determine which patterns representing combinations of level values ​​to be set as the values ​​of each parameter are preferable to test. For example, the information processing device 100 tests each of the two or more identified patterns to calculate the cost function and efficiently evaluates the second contribution of the parameters to the VQE calculation.

[0094] The information processing device 100 can appropriately select useful parameters in the VQE calculation based on the evaluated second contribution. Below, an example of the operation of the information processing device 100 will be explained in detail using Figures 6 to 12. First, let's move on to the explanation of Figure 6.

[0095] In Figure 6, the information processing device 100 identifies x parameters that define the variational quantum circuit. Here, x = 12. The 12 parameters are, for example, θ i i = 1, 2, ..., 12. The information processing device 100 controls each parameter θ. i The initial values ​​are set for each parameter θ. The information processing device 100 sets the initial values ​​for each parameter θ. i This is set as the parameter whose value will be updated. The information processing device 100 starts the VQE calculation using the quantum computing device 201.

[0096] The information processing device 100, for example, in the VQE calculation, each time it performs a review iteration, selects the parameter θ useful in the VQE calculation i and makes a selection, and re-sets the target parameter. The number of review iterations is, for example, (nk + 1) times. n is an integer greater than or equal to 0. k is a coefficient. k means the "number of review iterations" corresponding to the number of times the target parameter has been re-set. k is, for example, 5.

[0097] When the information processing device 100 performs the iteration (5×0 + 1) = 1 time, it interrupts the VQE calculation, selects the parameter useful in the VQE calculation, and re-sets the target parameter. First, the information processing device 100, for example, sets each parameter θ i to a candidate parameter to be set as the target parameter.

[0098] The information processing device 100, for example, evaluates the influence degree on the cost function representing energy by the first contribution degree representing the ratio of the change amount of energy to the change amount of each parameter θ i set to the candidate parameter. The energy corresponds to the quantum state represented by the variational quantum circuit. Specifically, the information processing device 100 calculates the first contribution degree representing the ratio of the change amount of energy to the change amount of each parameter θ i set to the candidate parameter. The first contribution degree is ΔE / Δθ i and so on.

[0099] ΔE is, for example, the difference between the energy in the current iteration and the energy in the previous iteration. ΔE is, for example, if the current iteration is the first time, it may be the difference between the energy in the current iteration and the energy corresponding to the initial value of each parameter θ i as well.

[0100] Also, Δθ i is, for example, the value of the parameter θ in the current iteration i and the parameter θ in the previous iteration iThis is the difference from the value of Δθ. i For example, if this is the first iteration, then the parameter θ in this iteration... i The value of the parameter θ i The difference from the initial value may also be used.

[0101] The information processing device 100, for example, calculates the first contribution and sets each of the candidate parameters θ i The number of level values ​​that can be set as the value of is adjusted. Here, the larger the first contribution, the parameter θ i It is considered preferable to increase the number of level values ​​that can be set as the value of . For example, graphs 600 and 610 show the parameter θ having a range of values ​​in which the contribution rate to energy is relatively large. i The value and its contribution rate to energy are shown.

[0102] Here, the parameter θ i Assume that the number of possible level values ​​for the parameter θ is 3. The dotted lines on Graph 600 show each level value when the number of level values ​​is 3. As shown in Graph 600, when the number of level values ​​is 3, there are no level values ​​that fall within a range of values ​​that are relatively important and have a relatively large contribution to energy. Therefore, the parameter θ i Regardless of which level value is set for this parameter, there is a problem in that it is not possible to consider the range of values ​​that are relatively important and have a relatively large contribution to energy.

[0103] On the other hand, the parameter θ i Let's assume that the number of possible level values ​​for the parameter θ is 5. The dotted line on Graph 610 shows each level value when the number of level values ​​is 5. As shown in Graph 610, when the number of level values ​​is 5, there are level values ​​that are included in a range of values ​​that are relatively important and have a relatively large contribution to energy. Therefore, the parameter θ i By setting level values ​​for each parameter, we can consider value ranges that are relatively important and have a relatively large contribution to energy.

[0104] Furthermore, if the first contribution is relatively small, the parameter θ i It is considered preferable to reduce the processing load by decreasing the number of level values ​​that can be set as the value of . Now, let's move on to the explanation of Figure 7. In Figure 7, for example, graph 700 shows the parameter θ that does not have a range of values ​​in which the contribution rate to energy is relatively large. i The value and its contribution rate to energy are shown.

[0105] Here, the parameter θ i Let's assume that there are 3 possible level values ​​to set for the value of . The dotted lines on Graph 700 show each level value when there are 3 level values. As shown in Graph 700, since we do not need to consider value ranges that are relatively important and have a relatively large contribution rate to energy, it is thought that even if the number of level values ​​is relatively small, the adverse effect on the VQE calculation is relatively small.

[0106] Therefore, the information processing device 100 adjusts each parameter θ such that the number of level values ​​increases as the first contribution increases. i The number of level values ​​that can be set as the value of is determined. The information processing device 100 refers to a table that associates, for example, the rank of the first contribution with the number of level values ​​that increase as the rank increases, and determines each parameter θ i Determine the number of level values ​​that can be set as the value of . Now, we move on to the explanation of Figure 8.

[0107] In Figure 8, the information processing device 100, for example, sets each of the candidate parameters θ according to the experimental design method. i ( i An orthogonal array is generated that limits the combinations of level values ​​to be set as values ​​(1~12). An example of an orthogonal array is Table 800 shown in Figure 8. Specifically, Table 800 is for the parameter θ i ( i This is an example of an orthogonal array that limits the combinations of level values ​​to be set as values ​​(1~9). The leftmost column of Table 800 shows the index of the combinations.

[0108] In Table 800, th(x) represents the x-th level value. In Table 800, the number of level values ​​that can be set for parameters θ1 and θ2 is 6. In Table 800, the number of level values ​​that can be set for parameters θ3, θ4, θ5, θ6, and θ7 is 3. In Table 800, the number of level values ​​that can be set for parameters θ8 and θ9 is 2. Next, we will move on to the explanation of Figure 9.

[0109] In Figure 9, the information processing device 100 controls each parameter θ i When setting one of the determined level values ​​for a given number of level values, the respective parameters θ i The information processing device 100 identifies two or more patterns that represent combinations of level values ​​to be set as the value of θ. For example, according to the generated orthogonal array, each parameter θ set as a candidate parameter is identified. i Identify two or more patterns from among all patterns representing combinations of level values ​​to be set as the value.

[0110] Specifically, the information processing device 100 uses an orthogonal array to determine the parameters θ1, θ2, θ3, θ4, θ5, θ6, θ7, θ8, θ9, and θ 10 and θ 11 and θ 12 Fifty patterns representing combinations of level values ​​to be set as the values ​​are identified. This allows the information processing device 100 to identify which patterns it is preferable to test by calculating the energy using a predetermined cost function. If the information processing device 100 calculates the energy using the cost function for all patterns, it may lead to an increase in processing load. Therefore, it is considered preferable for the information processing device 100 to reduce processing load by calculating the energy using the cost function for two or more of the patterns according to an orthogonal array.

[0111] The information processing device 100 calculates the energy for each of the two or more identified patterns using a cost function. In the example in Figure 9, for example, as shown in Table 900, the information processing device 100 uses the quantum computing device 201 to perform quantum calculations for each of the 50 identified patterns using a predetermined cost function, and calculates the energy E j We calculate E. j This shows the energy value calculated for the j-th pattern. θ j-k This is the kth parameter θ in the jth pattern. k This indicates the level value set as the value of [the variable].

[0112] The information processing device 100 is shown in Table 900 E j and θ j-k Based on this, parameters that have a relatively large impact on energy and contribute relatively large to the VQE calculation are searched for and selected as useful parameters in the VQE calculation. Next, we will move on to the explanation of Figure 100 and describe an example of how the information processing device 100 searches for useful parameters in the VQE calculation.

[0113] In Figure 10, the information processing device 100 is as shown in Table 1000, E j and θ j-k Based on this, the correlation coefficient between the dependent variable corresponding to energy and the explanatory variables corresponding to the parameters is calculated. The information processing device 100 selects from the candidate parameters those whose correlation coefficient is 0.7 or higher, which are parameters that have a relatively large influence on energy and are useful in VQE calculations. In the example in Figure 10, θ 11 The correlation coefficient for this is 0.7 or higher, which is the first threshold. This allows the information processing device 100 to narrow down the parameters that are useful in VQE calculations and which are preferable to set as the target parameters for updating their values.

[0114] Furthermore, the information processing device 100 is E j and θ j-kBased on this, useful parameters in the VQE calculation may be selected using a regression model that includes an objective variable corresponding to energy, explanatory variables corresponding to the parameters, and coefficients applied to the explanatory variables. The regression model may be, for example, E = Σ k=1 n a k θ k E is the dependent variable. θ k a is an explanatory variable. k is, θ k This is a coefficient related to the parameter. In this case, the information processing device 100 selects a from the candidate parameters. k Parameters whose absolute value is greater than or equal to the second threshold are selected as parameters that have a relatively large effect on energy and are useful in VQE calculations. The second threshold is, for example, a k Among the absolute values ​​of a, which of the largest is the most significant? k It could also mean the absolute value of [the given value].

[0115] Furthermore, the information processing device 100 may use both the correlation coefficient and the regression model to select parameters useful in the VQE calculation. In this case, for example, the information processing device 100 selects from among the candidate parameters the parameter whose correlation coefficient is 0.7 or higher as a first threshold parameter useful in the VQE calculation. Furthermore, the information processing device 100 selects from among the candidate parameters a k Parameters whose absolute value is greater than or equal to the second threshold are selected as useful parameters in the VQE calculation.

[0116] Furthermore, the information processing device 100, for example, among the candidate parameters, has a correlation coefficient that is equal to or greater than the first threshold and a k Parameters whose absolute value is greater than or equal to the second threshold may be selected as useful parameters in the VQE calculation. The information processing device 100 sets the selected set of useful parameters as the parameter group θ selection. The information processing device 100 selects parameters other than useful parameters and sets the set of parameters other than useful parameters as the parameter group θ exclusion.

[0117] The information processing device 100 selects each parameter θ of the selected parameter group θ. i The parameters to be kept as target parameters, and each of the selected parameter group θ to be excluded θ i The target parameters are reset to exclude the specified parameter. Once the target parameters are reset, the information processing device 100 resumes the VQE calculation.

[0118] The information processing device 100 interrupts the VQE calculation each time iterates (5 × k + 1) (k ≥ 1), selects useful parameters in the VQE calculation, and resets the target parameters. First, the information processing device 100 sets, for example, x parameters {θ i Each parameter θ of} i This is set as a candidate parameter to be set as the target parameter. Here, the information processing device 100 sets, for example, each parameter θ of the parameter group θ to be excluded. i You may also set this as a candidate parameter to be set as the target parameter.

[0119] The information processing device 100, for example, processes each parameter θ set as a candidate parameter. i The influence on the cost function representing energy is evaluated by the first contribution, which represents the ratio of the change in energy to the change in θ. Specifically, the information processing device 100 evaluates each parameter θ set in the candidate parameters. i The first contribution is calculated, which represents the ratio of the change in energy to the change in .

[0120] The information processing device 100, for example, calculates the first contribution and sets each of the candidate parameters θ i Adjust the number of level values ​​that can be set as the value of θ. Here, each parameter θ from the previous parameter group θ selection is adjusted. i Since it has been found to be of relatively high importance, the parameter θ i It is considered preferable to reduce the processing load by decreasing the number of level values ​​that can be set as the value of θ.i Regarding this, even if the number of level values ​​is relatively small, it is thought that the negative impact on VQE calculations will be relatively small.

[0121] Therefore, each parameter θ in the previous parameter group θ selection i Regarding this, each parameter θ of the previous parameter group θ removal i It is considered that the number of level values ​​can be smaller than the previous parameter group θ selection. For this reason, the information processing device 100 considers each parameter θ of the previous parameter group θ selection. i For this, we will determine a relatively small number of level values.

[0122] In this case, the information processing device 100 selects the parameter group θ from the previous selection so that the number of level values ​​increases as the first contribution is larger for each parameter θ. i The number of level values ​​that can be set as the value of may be determined. In addition, the information processing device 100 may determine the number of level values ​​for each parameter θ such that the larger the first contribution of the previous parameter group θ removal, the greater the number of level values. i You may also decide on the number of level values ​​that can be set as the value of .

[0123] The information processing device 100, in accordance with the experimental design method, sets each parameter θ as a candidate parameter. i The system generates an orthogonal array that limits the combinations of level values ​​to be set as the value of θ. The information processing device 100 then processes each parameter θ set as a candidate parameter according to the generated orthogonal array. i The information processing device 100 identifies two or more patterns from among all patterns representing combinations of level values ​​to be set as the value. For each of the two or more identified patterns, the information processing device 100 uses a cost function to calculate the energy E j We calculate E. j This shows the energy value calculated for the j-th pattern.

[0124] Information processing device 100 is E j and θ j-kBased on this, parameters that have a relatively large impact on energy and a relatively large contribution rate to the VQE calculation are explored and selected as useful parameters in the VQE calculation. Here, θ j-k represents the level value set as the value of the k-th parameter θ k in the j-th pattern. The information processing device 100, for example, E j and θ j-k Based on this, the information processing device 100 calculates the correlation coefficient between the objective variable corresponding to energy and the explanatory variable corresponding to the parameter. Among the candidate parameters, the information processing device 100 selects the parameters whose correlation coefficient is 0.7 or more, the first threshold value, as parameters that have a relatively large impact on energy and are useful in the VQE calculation.

[0125] In addition, the information processing device 100 may also select useful parameters in the VQE calculation by using a regression model that includes the objective variable corresponding to energy, the explanatory variable corresponding to the parameter, and the coefficient related to the explanatory variable, based on E j and θ j-k For example, the regression model is E = Σ k=1 n a k θ k Here, E is the objective variable. θ k is the explanatory variable. a k is the coefficient related to θ k In this case, the information processing device 100 selects, among the candidate parameters, the parameters whose absolute value of a k is equal to or greater than the second threshold value, as parameters that have a relatively large impact on energy and are useful in the VQE calculation. The second threshold value may, for example, represent the absolute value of a k which is the absolute value of the a k at what position from the largest among the absolute values of a

[0126] Furthermore, the information processing device 100 may use both the correlation coefficient and the regression model to select parameters useful in the VQE calculation. In this case, for example, the information processing device 100 selects from among the candidate parameters the parameter whose correlation coefficient is 0.7 or higher as a first threshold parameter useful in the VQE calculation. Furthermore, the information processing device 100 selects from among the candidate parameters a k Parameters whose absolute value is greater than or equal to the second threshold are selected as useful parameters in the VQE calculation.

[0127] Furthermore, the information processing device 100, for example, among the candidate parameters, has a correlation coefficient that is equal to or greater than the first threshold and a k Parameters whose absolute value is greater than or equal to the second threshold may be selected as useful parameters in the VQE calculation. The information processing device 100 sets the selected set of useful parameters as the parameter group θ selection. The information processing device 100 selects parameters other than useful parameters and sets the set of parameters other than useful parameters as the parameter group θ exclusion.

[0128] The information processing device 100 selects each parameter θ of the parameter group θ selected this time. i The parameters to be kept as target parameters, and the parameters θ excluded from the parameter group θ selected this time. i The target parameters are reset to exclude the target parameters. Once the target parameters are reset, the information processing device 100 resumes the VQE calculation. When the convergence conditions are met, the information processing device 100 completes the VQE calculation. Next, we will move on to the explanation of Figures 11 and 12.

[0129] Graph 1100 in Figure 11 shows the relationship between the number of iterations j and the energy E. As shown in Graph 1100, the information processing device 100 resets the target parameters each time iteration is performed (5 × k + 1). For example, when iteration is performed once, the information processing device 100 resets the parameters θ1, θ2, θ3, θ5, θ6, θ9, and θ 10 and θ 12The parameters are reset. For example, when the iteration is performed 6 times, the information processing device 100 sets the parameters θ1, θ2, θ3, θ5, θ7, θ9, and θ 10 and θ 11 The parameters are reset to their respective target parameters. The information processing device 100 can then minimize the energy E.

[0130] Graph 1200 in Figure 12 shows the relationship between the number of iterations j and the energy contribution rate. As shown in Graph 1200, for example, each time the information processing device 100 performs an iteration (5 × k + 1), the parameter θ whose energy contribution rate has become relatively higher in accordance with the change in the energy contribution rate. i This allows the target parameters to be reset. As a result, the information processing device 100 can narrow down the parameters useful in the VQE calculation and set them as target parameters, thereby reducing the processing time required when performing the VQE calculation while maintaining the accuracy of the VQE calculation.

[0131] Furthermore, the information processing device 100 can, for example, temporarily suspend the VQE calculation, update the parameter group θ selection, and update the target parameters each time the number of iterations reaches the number of revisions. Therefore, the information processing device 100 can deal with changes in useful parameters in the VQE calculation while it is being performed, and can reduce the processing time required to perform the VQE calculation while maintaining the accuracy of the VQE calculation.

[0132] Furthermore, the information processing device 100 can appropriately adjust the number of level values ​​that can be set as the values ​​of each parameter, according to the energy contribution rate, when updating the target parameters, for example. Therefore, the information processing device 100 can identify an appropriate pattern for testing to calculate energy while controlling how detailed the desirability of each parameter is examined according to the importance of each parameter. Accordingly, the information processing device 100 can reduce the processing time required when performing VQE calculations while maintaining the accuracy of the VQE calculations.

[0133] (Specific example of the operation of the information processing device 100) Next, we will explain specific examples of the operation of the information processing device 100 using Figures 13 to 19.

[0134] Figures 13 to 19 are explanatory diagrams illustrating specific examples of the operation of the information processing device 100. In Figure 13, the information processing device 100 is assumed to perform a VQE calculation to solve a problem concerning an H2 molecule. In this case, the VQE calculation will utilize a variational quantum circuit 1300. The variational quantum circuit 1300 includes, for example, rotation gates 1301 to 1312. The variational quantum circuit 1300 also includes controlled NOT gates 1313 to 1315. The parameters of the variational quantum circuit 1300 are θ1 to θ2 with respect to each of the rotation gates 1301 to 1312. 12 Therefore, θ1~θ 12 This indicates the rotation angle. Therefore, the number of parameters is 12. At this point, the information processing device 100 starts the VQE calculation. Next, we will move on to the explanation of Figures 14 and 15.

[0135] In Figures 14 and 15, the information processing device 100 controls parameters θ1 to θ 12 The parameters are set as target parameters, and an optimization calculation is performed once using the quantum computing device 201. The information processing device 100, for example, sets the parameters θ1 to θ as target parameters. 12 An optimization calculation is performed to update the value of the energy contribution coefficient ΔE / Δθ. kThe information processing device 100 calculates the absolute value of the energy contribution rate |ΔE / Δθ. k The larger the |, the more levels there will be, according to the parameters θ1~θ 12 Determine the number of levels. The number of levels is determined by the parameter θ. k This is the number of level values ​​that can be set as the value of [the variable].

[0136] In the examples shown in Figures 14 and 15, the information processing device 100 calculates the absolute value of the energy contribution rate |ΔE / Δθ k The top 6 parameters in descending order of | are θ2, θ5, θ7, θ8, θ 11 ,θ 12 The number of levels is determined to be 5. The information processing device 100 determines the top 6 parameters θ2, θ5, θ7, θ8, θ as shown in the level table 1400. 11 ,θ 12 For each of these, five level values ​​are determined. The leftmost column of the level table 1400 shows the index of the level values. The information processing device 100, for example, as shown in the level table 1400, determines the parameter θ k A random value is determined from the range [0, 2π] for the corresponding level value.

[0137] In the examples shown in Figures 14 and 15, the information processing device 100 calculates the absolute value of the energy contribution rate |ΔE / Δθ k The six lower parameters in descending order of | are θ1, θ3, θ4, θ6, θ9, θ 10 The number of levels is determined to be 2. The information processing device 100 has six lower parameters θ1, θ3, θ4, θ6, θ9, θ as shown in the level table 1500. 10 For each of these, two level values ​​are determined. The leftmost column of the level table 1500 shows the index of the level values. The information processing device 100, for example, as shown in the level table 1500, determines the parameter θ k A random value is determined from the range [0, 2π] for the corresponding level value. Next, we will move on to the explanation of Figure 16.

[0138] In Figure 16, the information processing device 100 obtains an orthogonal array 1600 of 5 levels and 6 parameters and 2 levels and 6 parameters. The orthogonal array 1600 is for parameters θ1 to θ12 This shows patterns representing combinations of level values ​​to be set as the value. The leftmost column of the orthogonal array 1600 shows the index of the pattern. In the orthogonal array 1600, xk[i] is the k-th parameter θ k This indicates setting the i-th level value. In the orthogonal array 1600, k is 1, 2, ..., 12. In the orthogonal array 1600, the top 6 parameters are θ2, θ5, θ7, θ8, θ 11 ,θ 12 For i, the values ​​are 1, 2, ..., 5. In the orthogonal array 1600, the lower six parameters are θ1, θ3, θ4, θ6, θ9, θ 10 For this, i is 1 or 2.

[0139] The information processing device 100 calculates the parameters θ1 to θ based on the level tables 1400, 1500 and the orthogonal array 1600. 12 The system identifies 36 patterns representing combinations of values. Based on this, the information processing device 100 determines the parameters θ1~θ 12 It is possible to identify which patterns representing combinations of values ​​are preferable to test for calculating energy using a predetermined cost function f.

[0140] The information processing device 100 uses the quantum computing device 201 to perform quantum computation for each of the 36 identified patterns using a predetermined cost function f, and calculates the energy E. Based on the calculated energy, the information processing device 100 determines the target variable corresponding to the energy and the parameters θ1~θ 12 A regression model E = Σ that includes the normalized explanatory variables and the coefficients applied to the explanatory variables. k=1 12 a k θ k The information processing device 100 generates a k The top 6 parameters θ in descending order of absolute value k Select and set the parameter group θs0. Here, the information processing device 100 sets the parameters θ8, θ7, θ2, θ5, θ 11 ,θ 12 Let's assume you selected this option.

[0141] The information processing device 100 sets only the selected parameter group θs0 as the target parameter whose value will be updated. In this way, the information processing device 100 determines the number of levels, identifies two or more patterns, and sets the parameter group θs p By selecting this option, you can perform the first step in a series of steps to set the target parameters. In the following explanation, this series of steps may be referred to as the "parameter review process." p is the number of times the parameter review process is performed.

[0142] The information processing device 100 resets the target parameters and then uses the quantum computing device 201 to perform up to 15 optimization calculations until predetermined convergence conditions are met. If the predetermined convergence conditions are met, the information processing device 100 completes the VQE calculation. The predetermined convergence conditions are, for example, when the magnitude of the gradient vector of the cost function f exceeds a threshold of 1 × 10⁻¹⁰. -8 The following is what will happen:

[0143] The information processing device 100 temporarily suspends the VQE calculation if the number of times the optimization calculation is performed without satisfying the predetermined convergence conditions reaches the number of revisions. The number of revisions is, for example, 15x + 1, where x is a positive integer. Since the number of revisions has been reached, the information processing device 100 will reconsider the target parameters. Here, the information processing device 100 considers parameters θ1, θ, and θ, other than the parameter group θs0. 10 Set θ9, θ6, θ4, and θ3 to the parameter group θs0'. Next, we will move on to the explanation of Figures 17 and 18.

[0144] In Figures 17 and 18, the information processing device 100 controls parameters θ1 to θ 12 The parameters are set as target parameters, and an optimization calculation is performed once using the quantum computing device 201. The information processing device 100, for example, sets the parameters θ1 to θ as target parameters. 12 An optimization calculation is performed to update the value of the energy contribution coefficient ΔE / Δθ. k Calculate.

[0145] The information processing device 100 processes the parameters θ1,θ of the parameter group θs0'. 10 The number of levels of θ9, θ6, θ4, θ3 corresponds to the parameters θ8, θ7, θ2, θ5, θ of the parameter group θs0. 11 ,θ 12 The parameters θ1~θ are set so that the number of levels is greater than the number of levels. 12 The number of levels is determined. In this process, the information processing device 100 determines the absolute value of the energy contribution rate |ΔE / Δθ from the parameter group θs0. k The larger the |, the more levels there are, so set the parameters θ8, θ7, θ2, θ5, θ 11 ,θ 12 Determine the number of levels.

[0146] In the examples shown in Figures 17 and 18, the information processing device 100 uses the absolute value of the energy contribution rate |ΔE / Δθ from the parameter group θs0. k The information processing device 100 determines 2 levels for each of the three lowest parameters θ2, θ7, and θ8, in descending order of |. The information processing device 100 determines 2 level values ​​for each of the three lowest parameters θ2, θ7, and θ8, as shown in the level table 1700. The leftmost column of the level table 1700 shows the index of the level values. For example, the information processing device 100 determines the parameter θ as shown in the level table 1700. k A random value is determined from the range [0, 2π] for the corresponding level value.

[0147] Furthermore, the information processing device 100 uses the absolute value of the energy contribution rate |ΔE / Δθ from the parameter group θs0. k The top 3 parameters θ5, θ in descending order of | 11 ,θ 12 The number of levels is determined to be 3. The information processing device 100 uses the top 3 parameters θ5,θ as shown in the level table 1710. 11 ,θ 12 Three level values ​​are determined for each. The leftmost column of the level table 1710 shows the index of the level values. The information processing device 100, for example, as shown in the level table 1710, determines the parameter θ k A random value is determined from the range [0, 2π] for the corresponding level value.

[0148] In the examples in Figures 17 and 18, the information processing device 100 processes the parameters θ1,θ of the parameter group θs0'. 10 For θ9, θ6, θ4, θ3, the number of levels is set for the parameters θ8, θ7, θ2, θ5, θ 11 ,θ 12 The number of levels is set to 5, which is more than the number of levels. The information processing device 100 sets the parameters θ1,θ as shown in the level table 1800. 10 For θ9, θ6, θ4, and θ3, five level values ​​are determined for each. The leftmost column of the level table 1800 shows the index of the level values. The information processing device 100, for example, as shown in the level table 1800, determines the parameter θ k A random value is determined from the range [0, 2π] for the corresponding level value. Next, we will move on to the explanation of Figure 19.

[0149] In Figure 19, the information processing device 100 obtains orthogonal arrays 1900 of 5 levels and 6 parameters, 3 levels and 3 parameters, and 2 levels and 3 parameters. The orthogonal array 1900 consists of parameters θ1 to θ 12 This shows patterns representing combinations of level values. The leftmost column of the orthogonal array 1900 shows the index of the pattern. In the orthogonal array 1900, xk[i] is the k-th parameter θ. k This indicates setting the i-th level value. In the orthogonal array 1900, k is 1, 2, ..., 12. In the orthogonal array 1900, there are 6 parameters 1, θ with 5 levels. 10 For θ9, θ6, θ4, θ3, i is 1, 2, ..., 5. In the orthogonal array 1900, there are three parameters θ5, θ6, θ7, θ8, θ9, θ6, θ4, θ34, θ3, θ4, θ3, θ4, θ3, θ4 11 ,θ 12 For this, i is 1, 2, 3. In the orthogonal array 1900, for the three parameters θ2, θ7, θ8 with 2 levels, i is 1, 2.

[0150] The information processing device 100 calculates the parameters θ1 to θ based on the level tables 1700, 1710, 1800 and the orthogonal array 1900. 12 The system identifies 36 patterns representing combinations of values. Based on this, the information processing device 100 determines the parameters θ1~θ 12It is possible to identify which patterns representing combinations of values ​​are preferable to test for calculating energy using a predetermined cost function f.

[0151] The information processing device 100 uses the quantum computing device 201 to perform quantum computation for each of the 36 identified patterns using a predetermined cost function f, and calculates the energy E. Based on the calculated energy, the information processing device 100 determines the target variable corresponding to the energy and the parameters θ1~θ 12 A regression model E = Σ that includes the normalized explanatory variables and the coefficients applied to the explanatory variables. k=1 12 a k θ k The information processing device 100 generates a k The top 6 parameters θ in descending order of absolute value k Select and set to parameter group θs1. Here, the information processing device 100 sets parameters θ8, θ7, θ2, θ5, θ 11 Let's assume we select θ6.

[0152] The information processing device 100 sets only the selected parameter group θs1 as the target parameters for updating the values. In this way, the information processing device 100 can perform the parameter review process a second time. After resetting the target parameters, the information processing device 100 uses the quantum computing device 201 to perform up to 15 optimization calculations until predetermined convergence conditions are met. If the predetermined convergence conditions are met, the information processing device 100 completes the VQE calculation. The predetermined convergence conditions are, for example, when the magnitude of the gradient vector of the cost function f is a threshold of 1 × 10⁻¹⁰. -8 The following is what will happen:

[0153] If the information processing device 100 performs 15 optimization calculations without satisfying the predetermined convergence conditions, it will repeat the operation described in Figures 17 to 19, which involves "performing a parameter review process and performing up to 15 optimization calculations until the predetermined convergence conditions are met."

[0154] Here, the information processing device 100 completes the VQE calculation because it has satisfied the predetermined convergence conditions after performing a total of 21 optimization calculations. When the information processing device 100 completes the VQE calculation, it obtains the total energy value E = -1.13615 and the last updated parameters θ1~θ 12 The combination of values ​​is output as the result of the VQE calculation. This allows the information processing device 100 to make the solution to the target problem accessible externally.

[0155] Thus, the information processing device 100 can limit the parameters whose values ​​are updated in order to reduce the processing time required when performing VQE calculations. In this case, the information processing device 100 can individually adjust the number of levels for each parameter. In order to maintain the accuracy of the VQE calculation, the information processing device 100 can efficiently test combinations of values ​​for two or more parameters based on the adjusted number of levels, in accordance with experimental design. As a result, the information processing device 100 can accurately select useful parameters in the VQE calculation. This allows the information processing device 100 to reduce the processing time required when performing VQE calculations while maintaining the accuracy of the VQE calculations.

[0156] Here, the information processing device 100 uses different criteria for each parameter θ in the first parameter review process and in the second and subsequent parameter review processes. k We have explained the case in which the number of levels for is determined, but it is not limited to this. For example, the information processing device 100 may use the same criteria for each parameter θ in the first parameter review process and in the second and subsequent parameter review processes. k In some cases, the number of levels for this may be determined. Specifically, the information processing device 100 determines the absolute value of the energy contribution rate |ΔE / Δθ in each of the parameter review processes. k The larger the |, the more levels there will be, according to the parameters θ1~θ 12 It may also be necessary to determine the number of levels.

[0157] Here, the information processing device 100 determines the number of levels for each parameter review process, using the same parameters θ1~θ. 12 The above describes the case where the parameters are set, but it is not limited to this. For example, the information processing device 100 may set the parameter to be determined for determining the number of levels in each iteration of the parameter review process to a different parameter. Specifically, the information processing device 100 may set a parameter other than the previous target parameter as the target parameter for the current iteration of the parameter review process, and determine the number of levels for the current target parameter. In this case, the information processing device 100 may specifically add useful parameters identified based on the determined number of levels from the current target parameter to the previous target parameter, and then set the target parameter for the next iteration.

[0158] (Overall processing procedure) Next, an example of the overall processing procedure executed by the information processing device 100 will be described using Figures 20 and 21. The overall processing is realized, for example, by the CPU 301 shown in Figure 3, storage areas such as memory 302 and recording medium 305, and network I / F 303.

[0159] Figures 20 and 21 are flowcharts illustrating an example of the overall processing procedure. In Figure 20, the information processing device 100 generates a variational quantum circuit (step S2101). Then, the information processing device 100 sets multiple parameters of the variational quantum circuit to the optimization parameter group θn (step S2002).

[0160] Next, the information processing device 100 creates an orthogonal array such that the number of level values ​​set as the values ​​of each parameter included in the optimization parameter group θn is different (step S2003). Then, the information processing device 100 uses the quantum computing device 201 to perform quantum computation based on the variational quantum circuit, referring to the created orthogonal array (step S2004).

[0161] Next, the information processing device 100 selects parameters to remain in the optimization parameter group θn and parameters to be removed from the optimization parameter group θn based on the results of the quantum computation, and updates the optimization parameter group θn (step S2005). Here, for example, the parameter review process shown in steps S2003 to S2005 is realized by the first review process shown in Figure 22, or the second review process shown in Figure 23. Specifically, the first parameter review process is realized by the first review process shown in Figure 22. Specifically, the second and subsequent parameter review processes are realized by the second review process shown in Figure 23. Then, the information processing device 100 proceeds to the process of step S2101 in Figure 21.

[0162] In Figure 21, the information processing device 100 uses the quantum computing device 201 to perform the j-th optimization calculation (step S2101). Then, based on the results of the optimization calculation, the information processing device 100 updates the values ​​of each parameter included in the optimization parameter group θn (step S2102).

[0163] Next, the information processing device 100 performs a quantum calculation using the quantum computing device 201 (step S2103). Then, the information processing device 100 calculates the energy E based on the results of the quantum calculation (step S2104).

[0164] Next, the information processing device 100 determines whether or not the convergence condition is met (step S2105). If the convergence condition is met (step S2105: Yes), the information processing device 100 terminates the entire process. On the other hand, if the convergence condition is not met (step S2105: No), the information processing device 100 proceeds to the process in step S2106.

[0165] In step S2106, the information processing device 100 sets j = j + 1 (step S2106). Next, the information processing device 100 determines whether j is the number of revisions (step S2107). If j is the number of revisions (step S2107: Yes), the information processing device 100 proceeds to the process in step S2108. On the other hand, if j is not the number of revisions (step S2107: No), the information processing device 100 returns to the process in step S2101.

[0166] In step S2108, the information processing device 100 updates the optimization parameter group θn so that the parameters that were removed from the optimization parameter group θn are temporarily returned to the optimization parameter group θn (step S2108). Then, the information processing device 100 returns to the process of step S2003 in Figure 20. As a result, the information processing device 100 can solve the optimization problem.

[0167] (First review procedure) Next, an example of the first review process procedure executed by the information processing device 100 will be described using Figure 22. The first review process corresponds to the first parameter review step. The first review process is implemented, for example, by the CPU 301 shown in Figure 3, storage areas such as memory 302 and recording medium 305, and network I / F 303.

[0168] Figure 22 is a flowchart of an example of the first review process procedure. In Figure 22, the information processing device 100 performs quantum computation using the quantum computing device 201 (step S2201). Then, the information processing device 100 calculates ΔE / Δθ for each parameter included in the optimization parameter group θn (step S2202).

[0169] Next, the information processing device 100 calculates the importance of each parameter in the optimization parameter group θn relative to E based on the calculated ΔE / Δθ (step S2203). Then, based on the calculated importance, the information processing device 100 determines the number of level values ​​for each parameter in the optimization parameter group θn (step S2204).

[0170] Next, the information processing device 100 creates an orthogonal array based on the determined number (step S2205). Then, the information processing device 100 uses the quantum computing device 201 to perform quantum computation by referring to the orthogonal array (step S2206).

[0171] Next, the information processing device 100 calculates the contribution of each parameter included in the optimization parameter group θn to the VQE calculation based on the results of the quantum computation (step S2207).

[0172] Then, based on the calculated contribution, the information processing device 100 selects which parameters to keep in the optimization parameter group θn and which parameters to remove from the optimization parameter group θn, and updates the optimization parameter group θn (step S2208). After that, the information processing device 100 terminates the first review process.

[0173] (Second Review Procedure) Next, an example of the second review process procedure executed by the information processing device 100 will be described using Figure 23. The second review process corresponds to the second and subsequent parameter review processes. The second review process is implemented, for example, by the CPU 301 shown in Figure 3, storage areas such as memory 302 and recording medium 305, and network I / F 303.

[0174] Figure 23 is a flowchart showing an example of the second review process. In Figure 23, the information processing device 100 performs quantum computation using the quantum computing device 201 (step S2301). Then, the information processing device 100 calculates ΔE / Δθ for each parameter included in the optimization parameter group θn (step S2302).

[0175] Next, the information processing device 100 calculates the importance of each parameter in the optimization parameter group θn with respect to E based on the calculated ΔE / Δθ (step S2303). Then, based on the calculated importance, the information processing device 100 distinguishes between the parameters that were retained last time and the parameters that were deleted last time from the optimization parameter group θn, and determines the number of level values ​​for each parameter (step S2304).

[0176] Next, the information processing device 100 creates an orthogonal array based on the determined number (step S2305). Then, the information processing device 100 uses the quantum computing device 201 to perform quantum computation by referring to the orthogonal array (step S2306).

[0177] Next, the information processing device 100 calculates the contribution of each parameter included in the optimization parameter group θn to the VQE calculation based on the results of the quantum computation (step S2307).

[0178] Then, based on the calculated contribution, the information processing device 100 selects which parameters to keep in the optimization parameter group θn and which parameters to remove from the optimization parameter group θn, and updates the optimization parameter group θn (step S2308). After that, the information processing device 100 terminates the second review process.

[0179] As explained above, the information processing device 100 can perform iteration a predetermined number of times according to VQE. During this process, the information processing device 100 can obtain the first contribution of one or more first parameters from among a plurality of parameters to the amount of energy change corresponding to the quantum state represented by the variational quantum circuit. Based on the obtained first contribution, the information processing device 100 can determine the number of level values ​​that can be set as the value of each first parameter. The information processing device 100 can identify two or more patterns that represent the combinations of level values ​​to be set as the value of each first parameter. Based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more identified patterns, the information processing device 100 can calculate the second contribution of each first parameter to VQE. Based on the calculated second contribution, the information processing device 100 can set each of the two or more second parameters from among the one or more first parameters that are judged to have a relatively high calculated second contribution as the target parameter. This allows the information processing device 100 to appropriately adjust the target parameters, thereby reducing the processing time required to perform VQE calculations while maintaining the accuracy of the VQE calculations.

[0180] According to the information processing device 100, when the iteration is performed for the first time, the target parameter can be set to each of the multiple parameters. This allows the information processing device 100 to perform the first iteration.

[0181] According to the information processing device 100, iteration can be performed a first time according to VQE. In this process, according to the information processing device 100, each of the multiple parameters can be set as the first parameter. According to the information processing device 100, the first contribution of each first parameter can be obtained. According to the information processing device 100, the number of level values ​​that can be set as the value of each first parameter can be determined such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter. In this way, the information processing device 100 can appropriately adjust the number of level values.

[0182] According to the information processing device 100, the iteration can be performed a second time according to VQE. In this case, according to the information processing device 100, each of the multiple parameters can be set as the first parameter. According to the information processing device 100, the first contribution of each first parameter can be obtained. According to the information processing device 100, the number of level values ​​that can be set as the value of each first parameter can be determined such that the number of level values ​​that can be set as the value of the target parameter (first parameter) is less than the number of level values ​​that can be set as the value of the first parameter that is not the target parameter. In this way, the information processing device 100 can appropriately adjust the number of level values.

[0183] According to the information processing device 100, the iteration can be performed a second time. According to the information processing device 100, among multiple parameters, parameters other than the target parameter can be designated as the first parameter. According to the information processing device 100, the first contribution of each first parameter can be obtained. According to the information processing device 100, the number of level values ​​that can be set as the value of each first parameter can be determined such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter. In this way, the information processing device 100 can appropriately adjust the number of level values.

[0184] According to the information processing device 100, among multiple parameters, parameters other than the target parameter can be designated as the first parameter. According to the information processing device 100, the first contribution of each first parameter can be obtained. According to the information processing device 100, the number of level values ​​that can be set as the value of each first parameter can be determined such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter. In this way, the information processing device 100 can appropriately adjust the number of level values.

[0185] According to the information processing device 100, among the one or more first parameters, the third parameters of one or more third parameters that are currently set as the target parameter and whose calculated second contribution is judged to be relatively low can be removed from the target parameter. As a result, the information processing device 100 can appropriately adjust the target parameter, thereby reducing the processing time required when performing the VQE calculation while maintaining the accuracy of the VQE calculation.

[0186] According to the information processing device 100, iteration can be repeatedly performed using an arithmetic unit capable of executing variational quantum circuits according to VQE. This allows the information processing device 100 to complete VQE calculations and solve optimization problems.

[0187] According to the information processing device 100, multiple predetermined number of times can be set. This allows the information processing device 100 to review the target parameters multiple times, thereby reducing the processing time required to perform the VQE calculation while maintaining the accuracy of the VQE calculation.

[0188] According to the information processing device 100, a first contribution can be obtained that corresponds to the ratio of the change in energy corresponding to the quantum state represented by the variational quantum circuit to the change in the first parameter. As a result, the information processing device 100 can utilize the first contribution, which accurately evaluates the contribution to the change in energy.

[0189] According to the information processing device 100, it is possible to obtain the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of two or more patterns. According to the information processing device 100, based on the calculation results, the second contribution with respect to the first parameter can be calculated from the correlation coefficient between the objective variable corresponding to the energy and the explanatory variable corresponding to each first parameter. In this way, the information processing device 100 can calculate the second contribution with high accuracy.

[0190] According to the information processing device 100, it is possible to obtain calculation results for the energy corresponding to the quantum state represented by the variational quantum circuit for each of two or more patterns. According to the information processing device 100, based on the calculation results, it is possible to identify a regression model that includes a target variable corresponding to the energy, explanatory variables corresponding to each first parameter, and coefficients applied to the explanatory variables. According to the information processing device 100, it is possible to calculate the second contribution with respect to the first parameter from the coefficients in the regression model. As a result, the information processing device 100 can calculate the second contribution with high accuracy.

[0191] The information processing method described in this embodiment can be implemented by executing a pre-prepared program on a computer such as a PC or workstation. The information processing program described in this embodiment is recorded on a computer-readable recording medium and executed by being read from the recording medium by the computer. The recording medium can be a hard disk, flexible disk, CD (Compact Disc)-ROM, MO (Magneto Optical Disc), DVD (Digital Versatile Disc), etc. Furthermore, the information processing program described in this embodiment may be distributed via a network such as the Internet.

[0192] With regard to the embodiments described above, the following additional information is disclosed.

[0193] (Note 1) In a case where an iteration is repeatedly performed to update a set target parameter among multiple parameters defining a variational quantum circuit according to the variational quantum eigenvalue solver method, when the iteration has been performed a predetermined number of times, Based on the first contribution of each of the one or more first parameters among the aforementioned plurality of parameters to the amount of energy change corresponding to the quantum state represented by the variational quantum circuit, the number of level values ​​that can be set as the value of each first parameter is determined. When setting one of the aforementioned number of level values ​​determined for each of the aforementioned first parameters, based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more patterns representing the combination of level values ​​set as the values ​​of each of the aforementioned first parameters, the second contribution of each of the aforementioned first parameters to the variational quantum eigenvalue solver method is calculated. Based on the calculated second contribution, the second parameter of each of the one or more second parameters among the one or more first parameters whose calculated second contribution is judged to be relatively high is set as the target parameter. An information processing program characterized by having a computer perform the processing.

[0194] (Note 2) The information processing program according to Note 1, characterized in that when the iteration is performed for the first time, the target parameter is each of the multiple parameters.

[0195] (Note 3) When the above iteration is performed for the first time, The process for making the aforementioned decision is: The information processing program according to Appendix 2, characterized in that each of the plurality of parameters is set as the first parameter, and the number of level values ​​that can be set as the value of each first parameter is determined based on the first contribution of each first parameter, such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter.

[0196] (Note 4) When the iteration is performed a second time, which is more than the first time, The process for making the aforementioned decision is: The information processing program according to Appendix 3, characterized in that each of the multiple parameters is set as the first parameter, and the number of level values ​​that can be set as the value of each of the one or more first parameters that are the target parameter is determined based on the first contribution of each of the first parameters, such that the number of level values ​​that can be set as the value of each of the one or more first parameters that are the target parameter is less than the number of level values ​​that can be set as the value of the first parameter that is not the target parameter.

[0197] (Note 5) When the iteration is performed a second time, which is more than the first time, The process for making the aforementioned decision is: The information processing program according to Appendix 3, characterized in that, among the plurality of parameters, all parameters other than the target parameter are set as the first parameter, and the number of level values ​​that can be set as the value of each first parameter is determined based on the first contribution of each first parameter, such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter.

[0198] (Note 6) The process to be determined above is: The information processing program according to Appendix 2, characterized in that, among the plurality of parameters, all parameters other than the target parameter are set as the first parameter, and the number of level values ​​that can be set as the value of each first parameter is determined based on the first contribution of each first parameter, such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter.

[0199] (Note 7) The process to be set above is: An information processing program according to any one of the appendices 1 to 6, characterized in that, based on the calculated second contribution, it removes from the target parameter one or more third parameters from the target parameter that are set in the target parameter and whose calculated second contribution is judged to be relatively low.

[0200] (Note 8) The iteration is repeatedly performed using an operation unit capable of executing the variational quantum circuit according to the variational quantum eigenvalue solver method. An information processing program according to any one of the appendices 1 to 7, characterized in that it causes the computer to perform the processing.

[0201] (Note 9) The information processing program according to any one of Notes 1 to 8, characterized in that the predetermined number of times is set to multiple values.

[0202] (Note 10) The information processing program according to any one of Notes 1 to 9, characterized in that the first contribution corresponds to the ratio of the change in energy corresponding to the quantum state represented by the variational quantum circuit to the change in the first parameter.

[0203] (Note 11) The calculation process described above is: Based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more patterns, calculate the second contribution degree regarding the first parameter from the correlation coefficient between the objective variable corresponding to the energy and the explanatory variable corresponding to each of the first parameters. The information processing program according to any one of Appendices 1 to 10, characterized in that.

[0204] (Appendix 12) The calculating process is Based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit for each of the two or more patterns, calculate the second contribution degree regarding the first parameter from the coefficient in the regression model including the objective variable corresponding to the energy, the explanatory variable corresponding to each of the first parameters, and the coefficient related to the explanatory variable. The information processing program according to any one of Appendices 1 to 10, characterized in that.

[0205] (Appendix 13) When iterations for updating a set target parameter among a plurality of parameters defining a variational quantum circuit are repeatedly performed according to the variational quantum eigenvalue solver method, when the iterations are performed a predetermined number of times, Based on the first contribution degree of each of one or more first parameters among the plurality of parameters to the change amount of the energy corresponding to the quantum state represented by the variational quantum circuit, determine the number of level values that can be set as the value of each of the first parameters for each of the first parameters, Based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more patterns representing each combination of the level values to be set as the values of the respective first parameters when setting any one of the determined number of level values for each of the first parameters, calculate the second contribution degree of each of the first parameters to the variational quantum eigenvalue solver method, Based on the calculated second contribution, the second parameter of each of the one or more second parameters among the one or more first parameters whose calculated second contribution is judged to be relatively high is set as the target parameter. An information processing method characterized in that the processing is performed by a computer.

[0206] (Note 14) In a case where an iteration is repeatedly performed to update a set target parameter among multiple parameters defining a variational quantum circuit according to the variational quantum eigenvalue solver method, when the iteration has been performed a predetermined number of times, Based on the first contribution of each of the one or more first parameters among the aforementioned plurality of parameters to the amount of energy change corresponding to the quantum state represented by the variational quantum circuit, the number of level values ​​that can be set as the value of each first parameter is determined. When setting one of the aforementioned number of level values ​​determined for each of the aforementioned first parameters, based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more patterns representing the combination of level values ​​set as the values ​​of each of the aforementioned first parameters, the second contribution of each of the aforementioned first parameters to the variational quantum eigenvalue solver method is calculated. Based on the calculated second contribution, the second parameter of each of the one or more second parameters among the one or more first parameters whose calculated second contribution is judged to be relatively high is set as the target parameter. An information processing device characterized by having a control unit. [Explanation of Symbols]

[0207] 100 Information Processing Devices 101 Arithmetic section 110,1300 Variational Quantum Circuits 111 parameters 121 First parameter 122 Level Values 130 patterns 141 Second parameter 150 VQE calculation 151 Target Parameters 152 Iterations 200 Information Processing Systems 201 Quantum computing device 202 Client Devices 210 Network 300,400 buses 301,401 CPU 302,402 memory 303,403 Network I / F 304,404 Recording medium I / F 305,405 recording media 406 Computational Enclosure I / F 407 Quantum Computing Chassis 500 storage section 501 Acquisition Department 502 Decision Section 503 Settings Section 504 Implementation Department 505 Output section 600, 610, 700, 1100, 1200 graph 800,900,1000 tables 1301-1312 Rotary gate 1313~1315 Controlled NOT gates 1400,1500,1700,1710,1800 level table 1600,1900 orthogonal array

Claims

1. In a variational quantum eigenvalue solver method, when iterations are repeatedly performed to update a set target parameter among multiple parameters defining a variational quantum circuit, when the iterations have been performed a predetermined number of times, Based on the first contribution of each of the one or more first parameters among the aforementioned plurality of parameters to the amount of energy change corresponding to the quantum state represented by the variational quantum circuit, the number of level values ​​that can be set as the value of each of the aforementioned first parameters is determined. When setting one of the aforementioned number of level values ​​determined for each of the aforementioned first parameters, the second contribution of each of the aforementioned first parameters to the variational quantum eigenvalue solver method is calculated based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more patterns representing the combination of level values ​​set as the values ​​of each of the aforementioned first parameters. Based on the calculated second contribution, the second parameter of each of the one or more first parameters whose calculated second contribution is judged to be relatively high is set as the target parameter. An information processing program characterized by having a computer perform the processing.

2. The information processing program according to claim 1, characterized in that when the iteration is performed for the first time, the target parameter is each of the multiple parameters.

3. When the aforementioned iteration is performed for the first time, The process for making the aforementioned decision is: The information processing program according to claim 2, characterized in that each of the plurality of parameters is set as the first parameter, and the number of level values ​​that can be set as the value of each first parameter is determined based on the first contribution of each first parameter, such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter.

4. When the aforementioned iteration is performed a second time, which is greater than the first time, The process for making the aforementioned decision is: The information processing program according to claim 3, characterized in that each of the plurality of parameters is set as the first parameter, and the number of level values ​​that can be set as the value of each of the one or more first parameters that are the target parameter is determined based on the first contribution of each of the first parameters, such that the number of level values ​​that can be set as the value of each of the one or more first parameters that are the target parameter is less than the number of level values ​​that can be set as the value of the first parameter that is not the target parameter.

5. When the aforementioned iteration is performed a second time, which is greater than the first time, The process for making the aforementioned decision is: The information processing program according to claim 3, characterized in that, among the plurality of parameters, the parameters other than the target parameter are set as the first parameter, and the number of level values ​​that can be set as the value of each first parameter is determined based on the first contribution of each first parameter, such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter.

6. The process for making the aforementioned decision is: The information processing program according to claim 2, characterized in that, among the plurality of parameters, all parameters other than the target parameter are set as the first parameter, and the number of level values ​​that can be set as the value of each first parameter is determined based on the first contribution of each first parameter, such that the larger the first contribution, the greater the number of level values ​​that can be set as the value of the first parameter.

7. The process to be set is, An information processing program according to any one of claims 1 to 6, characterized in that, based on the calculated second contribution, the third parameter of each of the one or more first parameters that is set as the target parameter and whose calculated second contribution is judged to be relatively low is removed from the target parameter.

8. In a variational quantum eigenvalue solver method, when iterations are repeatedly performed to update a set target parameter among multiple parameters defining a variational quantum circuit, when the iterations have been performed a predetermined number of times, Based on the first contribution of each of the one or more first parameters among the aforementioned plurality of parameters to the amount of energy change corresponding to the quantum state represented by the variational quantum circuit, the number of level values ​​that can be set as the value of each of the aforementioned first parameters is determined. When setting one of the aforementioned number of level values ​​determined for each of the aforementioned first parameters, the second contribution of each of the aforementioned first parameters to the variational quantum eigenvalue solver method is calculated based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more patterns representing the combination of level values ​​set as the values ​​of each of the aforementioned first parameters. Based on the calculated second contribution, the second parameter of each of the one or more first parameters whose calculated second contribution is judged to be relatively high is set as the target parameter. An information processing method characterized in that the processing is performed by a computer.

9. In a variational quantum eigenvalue solver method, when iterations are repeatedly performed to update a set target parameter among multiple parameters defining a variational quantum circuit, when the iterations have been performed a predetermined number of times, Based on the first contribution of each of the one or more first parameters among the aforementioned plurality of parameters to the amount of energy change corresponding to the quantum state represented by the variational quantum circuit, the number of level values ​​that can be set as the value of each of the aforementioned first parameters is determined. When setting one of the aforementioned number of level values ​​determined for each of the aforementioned first parameters, the second contribution of each of the aforementioned first parameters to the variational quantum eigenvalue solver method is calculated based on the calculation results of the energy corresponding to the quantum state represented by the variational quantum circuit in each of the two or more patterns representing the combination of level values ​​set as the values ​​of each of the aforementioned first parameters. Based on the calculated second contribution, the second parameter of each of the one or more first parameters whose calculated second contribution is judged to be relatively high is set as the target parameter. An information processing device characterized by having a control unit.

Images