Prediction device, prediction method, and program

Abstract

The present invention predicts a prescribed physical property in the reaction path of a multimolecular reaction. This prediction device comprises: an acquisition unit that acquires a trained model trained using a data set including, in a multimolecular reaction in which a formation system including two or more types of formation molecules is formed from a reaction system including two or more types of reaction molecules, the reaction system, the formation system, intermediate structures of the reaction molecules generated in the process of forming the formation molecules from the reaction molecules by the multimolecular reaction, and a prescribed physical property of at least one of the reaction system and the formation system; and a prediction unit that predicts the physical property using the trained model.

Description

Prediction device, prediction method, and program

[0001] The present disclosure relates to a prediction device, a prediction method, and a program.

[0002] Conventionally, reaction path searches have been performed in molecular reactions in which a product system containing one or two types of product molecules is produced from a reaction system containing only one type of reactant molecule, or in which a product system containing only one type of product molecule is produced from a reaction system containing two types of reactant molecules (see, for example, Patent Document 1).

[0003] International Publication No. 2022 / 260179

[0004] Here, although reaction path searches for unimolecular reactions have been conducted in the past, a method for predicting predetermined physical properties in the reaction path of a multimolecular reaction in which a product system containing two or more product molecules is produced from a reaction system containing two or more reactant molecules has not been established.

[0005] In view of the above-described technical problems, one aspect of the present disclosure aims to predict a predetermined physical property in a reaction pathway of a multimolecular reaction.

[0006] The present disclosure has the following configuration.

[0007] <1> A prediction device for a multimolecular reaction in which a product system including two or more types of product molecules is produced from a reaction system including two or more types of reactant molecules, the prediction device comprising: an acquisition unit that acquires a trained model trained using a dataset including the reaction system, the product system, intermediate structures of the reactant molecules generated in the process of producing the product molecules from the reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system; and a prediction unit that predicts the physical properties using the trained model.

[0008] <2> The prediction device according to <1> above, wherein the intermediate structure is obtained by optimizing a transition state structure of the reactant molecules in the multimolecular reaction and searching for a reaction path from the transition state structure to the reaction system or the product system.

[0009] <3> The prediction device according to <2>, wherein the transition state is optimized using a trained model.

[0010] <4> The prediction device according to <1> above, wherein the intermediate structure is obtained by determining the reaction system and the product system, describing a correspondence between each atom contained in the reactant molecule of the reaction system and each atom contained in the product molecule of the product system, and searching for a reaction path from the reaction system to the product system based on the correspondence.

[0011] <5> The prediction device according to <4>, wherein the reaction pathway is searched for using a trained model.

[0012] <6> The prediction device according to <1>, wherein the intermediate structure is obtained by determining the reaction system and searching for a reaction path from the reaction system to the product system using a trained model.

[0013] <7> The prediction device according to any one of <1> to <6>, wherein the physical property includes energy in the multimolecular reaction.

[0014] <8> A prediction method in which a computer executes the following steps: a step of acquiring a trained model trained using a dataset including a reaction system, in which a product system containing two or more product molecules is generated from a reaction system containing two or more reactant molecules, the product system, intermediate structures of the reactant molecules generated in the process of generating the product molecules from the reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system; and a step of predicting the physical properties using the trained model.

[0015] <9> A program for causing a computer to execute the following steps: a step of acquiring a trained model trained using a dataset including a reaction system, the product system, an intermediate structure of the reactant molecules generated in the process of generating the product molecules from the reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system, in a multimolecular reaction in which a product system containing two or more product molecules is generated from the reaction system containing two or more reactant molecules; and a step of predicting the physical properties using the trained model.

[0016] According to one aspect of the present disclosure, it is possible to predict a predetermined physical property in a reaction pathway of a multimolecular reaction.

[0017] 1 is a block diagram showing an example of the overall configuration of a prediction system; FIG. 2 is a block diagram showing an example of the hardware configuration of a computer; FIG. 3 is a block diagram showing an example of the functional configuration of a prediction system; FIG. 4 is a flowchart showing an example of a prediction method; FIG. 5 is a flowchart showing an example of a method for generating teacher data; FIG. 6 is a schematic diagram showing an example of arranging different molecules in a reaction system and a production system in the same space; FIG. 7 is a schematic diagram showing an example describing the correspondence between each atom in a reaction system and a production system; and FIG. 8 is a flowchart showing an example of a method for generating teacher data.

[0018] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. In this specification and drawings, components having substantially the same functional configurations are designated by the same reference numerals, and redundant description will be omitted.

[0019] [Embodiments] A prediction device according to an embodiment of the present disclosure will be described. The prediction device according to an embodiment of the present disclosure is a device that predicts a reaction path in a multimolecular reaction in which a product system containing two or more product molecules is produced from a reaction system containing two or more reactant molecules. Before describing the prediction device according to an embodiment of the present disclosure, a prediction system used in the prediction device according to an embodiment of the present disclosure will be described.

[0020] <Prediction System> Fig. 1 is a schematic diagram showing an example of a prediction system used in a prediction device according to an embodiment of the present disclosure. As shown in Fig. 1, the prediction system 1000 includes a prediction device 10 and a terminal device 20. The prediction device 10 and the terminal device 20 are connected to each other so as to be able to communicate data with each other via a communication network N1 such as a local area network (LAN) or the Internet.

[0021] The prediction device 10 is an information processing device such as a personal computer, workstation, or server that predicts a reaction path in a multimolecular reaction in which a product system containing two or more product molecules is produced from a reaction system containing two or more reactant molecules. The prediction device 10 receives data from a terminal device 20. The data is electronic data that represents either the reaction system, the product system, or an intermediate structure of the reactant molecules in the multimolecular reaction. The prediction device 10 predicts a reaction path in a multimolecular reaction in which a product system containing two or more product molecules is produced from a reaction system containing two or more reactant molecules, and transmits the prediction result to the terminal device 20.

[0022] The terminal device 20 is an information processing terminal such as a personal computer, smartphone, or tablet terminal operated by a user of the prediction system 1000. The terminal device 20 receives electronic data indicating any of the reaction system, product system, and intermediate structures of reactant molecules in a multimolecular reaction, and transmits the electronic data to the prediction device 10. The terminal device 20 displays the prediction results received from the prediction device 10 to the user.

[0023] The overall configuration of the prediction system 1000 shown in Fig. 1 is an example, and various system configuration examples are possible depending on the application and purpose. For example, the prediction system 1000 may include multiple prediction devices 10 and one or more terminal devices 20. For example, the prediction device 10 may be realized by multiple computers, or may be realized as a cloud computing service. The classification of devices such as the prediction device 10 and the terminal device 20 shown in Fig. 1 is an example.

[0024] <Hardware Configuration> The hardware configuration of the prediction system 1000 in this embodiment will be described with reference to FIG.

[0025] <Computer> The prediction device 10 and the terminal device 20 in this embodiment are realized by, for example, a computer. Fig. 2 is a block diagram showing an example of the hardware configuration of a computer 500 in this embodiment.

[0026] 2, the computer 500 includes a CPU (Central Processing Unit) 501, a ROM (Read Only Memory) 502, a RAM (Random Access Memory) 503, a HDD (Hard Disk Drive) 504, an input device 505, a display device 506, a communication I / F (Interface) 507, and an external I / F 508. The CPU 501, the ROM 502, and the RAM 503 form a so-called computer. The hardware components of the computer 500 are connected to each other via a bus line 509. The input device 505 and the display device 506 may be connected to the external I / F 508 for use.

[0027] The CPU 501 is a computing device that reads programs and data from a storage device such as the ROM 502 or the HDD 504 onto the RAM 503 and executes processing to realize overall control and functions of the computer 500 .

[0028] The ROM 502 is an example of a non-volatile semiconductor memory (storage device) that can retain programs and data even when the power is turned off. The ROM 502 functions as a main storage device that stores various programs, data, etc. required for the CPU 501 to execute various programs installed in the HDD 504. Specifically, the ROM 502 stores boot programs such as a Basic Input / Output System (BIOS) and an Extensible Firmware Interface (EFI) that are executed when the computer 500 starts up, as well as data such as OS (Operating System) settings and network settings.

[0029] The RAM 503 is an example of a volatile semiconductor memory (storage device) in which programs and data are erased when the power is turned off. The RAM 503 is, for example, a dynamic random access memory (DRAM) or a static random access memory (SRAM). The RAM 503 provides a working area in which various programs installed in the HDD 504 are expanded when executed by the CPU 501.

[0030] The HDD 504 is an example of a non-volatile storage device that stores programs and data. The programs and data stored in the HDD 504 include an OS, which is basic software that controls the entire computer 500, and applications that provide various functions on the OS. Note that the computer 500 may use a storage device that uses flash memory as a storage medium (e.g., an SSD (Solid State Drive)) instead of the HDD 504.

[0031] The input device 505 includes a touch panel, operation keys and buttons, a keyboard and mouse, a microphone for inputting sound data such as voice, and the like, which are used by the user to input various signals.

[0032] The display device 506 is composed of a display such as a liquid crystal display or organic electroluminescence (EL) display for displaying a screen, a speaker for outputting sound data such as voice, and the like.

[0033] The communication I / F 507 is an interface that connects to a communication network and enables the computer 500 to perform data communication.

[0034] The external I / F 508 is an interface with external devices, such as a drive device 510.

[0035] The drive device 510 is a device for loading a recording medium 511. The recording medium 511 here includes media that record information optically, electrically, or magnetically, such as CD-ROMs, flexible disks, and magneto-optical disks. The recording medium 511 may also include semiconductor memories that record information electrically, such as ROMs and flash memories. This allows the computer 500 to read from and / or write to the recording medium 511 via the external I / F 508.

[0036] The various programs to be installed in the HDD 504 are installed, for example, by setting the distributed recording medium 511 in a drive device 510 connected to the external I / F 508 and reading the various programs recorded on the recording medium 511 by the drive device 510. Alternatively, the various programs to be installed in the HDD 504 may be installed by being downloaded via the communication I / F 507 from the communication network or a network different from the communication network.

[0037] <Functional Configuration> The functional configuration of the prediction system in this embodiment will be described with reference to Fig. 3. Fig. 3 is a block diagram showing an example of the functional configuration of the prediction system.

[0038] <Prediction Device> As shown in FIG. 3 , the prediction device 10 includes a teacher data storage unit 101, a trained model 102, an acquisition unit 103, a prediction unit 104, and an output unit 105.

[0039] The acquisition unit 103, the prediction unit 104, and the output unit 105 are realized by the processing that the CPU 501 executes by a program loaded from the HDD 504 onto the RAM 503 shown in FIG.

[0040] The teacher data storage unit 101 is realized by the HDD 504 shown in FIG.

[0041] Teacher data is pre-stored in the teacher data storage unit 101. The teacher data includes a reaction system, a product system, intermediate structures of reactant molecules generated in the process of generating product molecules from reactant molecules through a multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system in a multimolecular reaction.

[0042] The trained model 102 is a trained model trained using a dataset as training data recorded in the training data storage unit 101.

[0043] The acquisition unit 103 acquires a trained model 102 trained using a dataset.

[0044] The acquiring unit 103 acquires electronic data by receiving the electronic data from the terminal device 20. The acquiring unit 103 may acquire electronic data by accepting electronic data input to the input device 505. The electronic data is electronic data that indicates any of a reaction system, a product system, and an intermediate structure of reactant molecules in a multimolecular reaction.

[0045] The prediction unit 104 predicts a reaction path in the received electronic data of the multi-molecular reaction based on the acquired trained model 102.

[0046] The output unit 105 transmits the prediction result to the terminal device 20. The prediction result includes the reaction path in the multimolecular reaction predicted by the prediction unit 104. The output unit 105 may display the prediction result on the display device 506.

[0047] <Terminal Device> As shown in FIG. 3 , the terminal device 20 includes an input unit 201 and a display unit 202 .

[0048] The input unit 201 and the display unit 202 are realized by the processing that the CPU 501 executes by a program loaded onto the RAM 503 from the HDD 504 shown in FIG.

[0049] The input unit 201 receives input of electronic data in response to a user's operation on an input screen, and transmits the received electronic data to the prediction device 10.

[0050] The display unit 202 outputs a screen that serves as a user interface for the prediction device 10 to the display device 506. The screens output by the display unit 202 include an input screen and a result screen. The input screen is a screen for inputting electronic data. The result screen is a screen for displaying the prediction result received from the prediction device 10.

[0051] <Processing Procedure> A prediction method executed by the prediction system 1000 according to the first embodiment will be described with reference to Fig. 4. Fig. 4 is a flowchart showing an example of the prediction method.

[0052] In step S101, the acquisition unit 103 of the prediction device 10 acquires the trained model 102 trained using a dataset.

[0053] In step S102, in response to a user operation, the display unit 202 of the terminal device 20 outputs an input screen to the display device 506. The operation for displaying the input screen is, for example, an operation for starting a prediction program pre-installed in the terminal device 20.

[0054] In step S103, the acquisition unit 103 of the prediction device 10 receives electronic data from the terminal device 20. The electronic data is electronic data that indicates a reaction system, a product system, an intermediate structure of reactant molecules generated in the process of generating product molecules from reactant molecules through a multimolecular reaction, and any of predetermined physical properties of at least one of the reaction system and the product system.

[0055] In step S104, the prediction unit 104 of the prediction device 10 predicts predetermined physical properties of the received electronic data of the multi-molecular reaction using the acquired trained model 102.

[0056] In step S105, the output unit 105 of the prediction device 10 receives the predicted predetermined physical properties from the prediction unit 104. Next, the output unit 105 transmits the prediction result to the terminal device 20.

[0057] <Method for Generating Training Data> A method for generating training data using the prediction system 1000 in this embodiment will be described.

[0058] First Embodiment FIG. 5 is a flowchart showing an example of a method for generating teacher data according to a first embodiment.

[0059] In step S201, an initial structure of a transition state in a reaction system containing two or more types of reactant molecules is determined. It is preferable to confirm that the initial structure of the transition state has an appropriate imaginary vibration.

[0060] In step S202, the initial structure created in step S201 is optimized toward a transition state. The transition state of the reactant molecule can be optimized using a trained model. The trained model can be a pre-constructed neural network potential or the like.

[0061] In step S203, an IRC (Intrinsic Reaction Coordinate) calculation is performed to obtain the intermediate structure of the reacting molecule.

[0062] In step S204, first-principles calculations are performed to obtain energy as a predetermined physical property of the obtained intermediate structure. This obtains a data set including a reaction system, a product system, an intermediate structure of the reactant molecules generated in the process of generating the product molecules from the reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system in a multimolecular reaction in which a product system containing two or more product molecules is generated from a reaction system containing two or more reactant molecules. When first-principles calculations are used for the calculations in steps S202 and S203, the intermediate structure and predetermined physical properties obtained in each step can be used, and the first-principles calculation in step S204 can be omitted.

[0063] Second Embodiment FIG. 6 is a flowchart showing an example of a method for generating teacher data according to a second embodiment.

[0064] In step S301, a reaction system and a product system are determined for a reaction system containing two or more types of reactant molecules. Next, different molecules in the reaction system and the product system are arranged in the same space. In this case, it is preferable that the reaction sites of each atom face each other and that the molecular arrangement is a locally stable structure. Figure 7 is a schematic diagram showing an example of arranging different molecules in a reaction system 1001 and a product system 1002 in the same space.

[0065] In step S302, the correspondence between each atom contained in the reactant molecule in the reaction system and each atom contained in the product molecule in the product system is described. Figure 8 is a schematic diagram showing an example of describing the correspondence between each atom in reaction system 1001 and product system 1002. By assigning an ID to each atom contained in reaction system 1001 and product system 1002 and associating them, an intermediate structure in which the relationship between each atom in the reaction mechanism is clear is obtained.

[0066] In step S303, based on the correspondence between each atom, a reaction path from the reaction system to the product system is searched for using the Nudged Elastic Band (NEB) method, the double-ended growing string method (GSM) method, the Artificial Force Induced Reaction (AFIR) method, the Synchronous Transit-Guided Quasi-Newton (STQN) method, or the like. This allows for the acquisition of intermediate structures in multimolecular reactions. Furthermore, a pre-constructed neural network potential, etc., can be used to search for the reaction path.

[0067] In step S304, first-principles calculations are performed to obtain energy as a predetermined physical property in the obtained reaction path. This obtains a data set including a reaction system, a product system, an intermediate structure of the reactant molecules generated in the process of generating the product molecules from the reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system in a multimolecular reaction in which a product system containing two or more product molecules is generated from a reaction system containing two or more reactant molecules. When first-principles calculations are used for the calculation in step S303, the intermediate structure and predetermined physical properties obtained in step S303 can be used, and the first-principles calculation in step S304 can be omitted.

[0068] Third Embodiment FIG. 9 is a flowchart showing an example of a method for generating teacher data according to a third embodiment.

[0069] In step S401, a reaction system containing two or more types of reactant molecules is determined. Next, different molecules in the reaction system are arranged in the same space. At this time, it is preferable that the reactive sites of each atom face each other and form a locally stable structure.

[0070] In step S402, the trained model is used to search for a reaction path from the reaction system to the product system. This allows for the acquisition of intermediate structures in multi-molecular reactions. For example, the single-ended GSM method, the AFIR method, the ADDF (anharmonic downward distortion following) method, and the meta-dynamics method can be used alone or in combination to search for a reaction path.

[0071] In step S403, first-principles calculations are performed to obtain energy as a predetermined physical property in the obtained reaction path, thereby obtaining a data set including a reaction system, a product system, intermediate structures of reactant molecules generated in the process of generating product molecules from reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system in a multimolecular reaction in which a product system containing two or more product molecules is generated from a reaction system containing two or more reactant molecules.

[0072] Effect of the Embodiment The prediction device 10 in this embodiment predicts physical properties of a multimolecular reaction in which a product system containing two or more product molecules is produced from a reaction system containing two or more reactant molecules, using a trained model trained using a dataset including the reaction system, the product system, intermediate structures of the reactant molecules produced in the process of producing the product molecules from the reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system. According to this embodiment, it is possible to predict predetermined physical properties in the reaction path of a multimolecular reaction. In this embodiment, by constructing a trained model trained using a dataset of a multimolecular reaction in which a product system containing two or more product molecules is produced from a reaction system containing two or more reactant molecules, it is possible to predict predetermined physical properties in the reaction path of a multimolecular reaction that were not possible with conventional methods.

[0073] The intermediate structure in this embodiment may be obtained by optimizing the transition state structure of the reactant molecules in a multimolecular reaction and searching for a reaction path from the transition state structure to the reaction system or the product system. According to this embodiment, it is possible to search for a reaction path from the initial structure of the transition state to the reaction system or the product system.

[0074] The transition state in this embodiment may be optimized using a trained model.

[0075] The intermediate structure in this embodiment may be obtained by determining a reaction system and a product system, describing the correspondence between each atom contained in the reactant molecule of the reaction system and each atom contained in the product molecule of the product system, and searching for a reaction path from the reaction system to the product system based on the correspondence. According to this embodiment, it is possible to search for a reaction path from a reaction system and a product system to the reaction system or the product system in a multimolecular reaction.

[0076] The reaction pathway in this embodiment may be searched using a trained model.

[0077] The intermediate structure in this embodiment may be obtained by determining a reaction system and searching for a reaction path from the reaction system to a product system using a trained model. According to this embodiment, it is possible to search for a reaction system or a reaction path to the product system from only the reaction system in a multimolecular reaction.

[0078] The physical property in this embodiment may include the energy in a multi-molecular reaction.

[0079] Although the embodiments of the present disclosure have been described in detail above, the embodiments disclosed herein are illustrative in all respects and are not limiting. The embodiments can be modified and improved in various ways without departing from the scope and spirit of the appended claims. The matters described in the above embodiments can be configured in other ways as long as they are not inconsistent, and can be combined as long as they are not inconsistent.

[0080] This application claims priority based on Japanese Patent Application No. 2024-124450, filed on July 31, 2024, the entire contents of which are incorporated herein by reference.

[0081] REFERENCE SIGNS LIST 10 Prediction device 20 Terminal device 101 Teacher data storage unit 102 Trained model 103 Acquisition unit 104 Prediction unit 105 Output unit 201 Input unit 202 Display unit 1000 Prediction system

Claims

1. A prediction device for a multimolecular reaction in which a product system containing two or more types of product molecules is produced from a reaction system containing two or more types of reactant molecules, the prediction device comprising: an acquisition unit that acquires a trained model trained using a dataset including the reaction system, the product system, intermediate structures of the reactant molecules produced in the process of producing the product molecules from the reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system; and a prediction unit that predicts the physical properties using the trained model.

2. The prediction device according to claim 1, wherein the intermediate structure is obtained by optimizing the structure of a transition state of the reactant molecules in the multimolecular reaction and searching for a reaction path from the structure of the transition state to the reaction system or the product system.

3. The prediction device according to claim 2, wherein the transition state is optimized using a trained model.

4. The prediction device described in claim 1, wherein the intermediate structure is obtained by determining the reaction system and the product system, describing the correspondence between each atom contained in the reactant molecule of the reaction system and each atom contained in the product molecule of the product system, and searching for a reaction path from the reaction system to the product system based on the correspondence.

5. The prediction device according to claim 4, wherein the reaction pathway is searched for using a trained model.

6. The prediction device according to claim 1, wherein the intermediate structure is obtained by determining the reaction system and searching for a reaction path from the reaction system to the product system using a trained model.

7. The prediction device according to any one of claims 1 to 6, wherein the physical properties include energy in the multimolecular reaction.

8. A prediction method in which a computer executes the following steps: obtaining a trained model trained using a dataset including a reaction system in which a product system containing two or more product molecules is produced from a reaction system containing two or more reactant molecules, the product system, intermediate structures of the reactant molecules produced in the process of producing the product molecules from the reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system; and predicting the physical properties using the trained model.

9. A program for causing a computer to execute the following steps: obtaining a trained model trained using a dataset including a reaction system, the product system, an intermediate structure of the reactant molecules produced in the process of generating the product molecules from the reactant molecules by the multimolecular reaction, and predetermined physical properties of at least one of the reaction system and the product system in a multimolecular reaction in which a product system containing two or more product molecules is generated from the reaction system containing two or more reactant molecules; and predicting the physical properties using the trained model.

Images