Information processing apparatus, information processing method, program, and prediction apparatus

Abstract

To appropriately execute yield prediction.SOLUTION: The information processing apparatus includes a memory and a processor. The processor trains a model that uses at least one of an electronic descriptor and a steric descriptor of a substance involved in a chemical reaction as an explanatory descriptor to calculate information relating to a yield of a product of the chemical reaction, the information being associated with the explanatory descriptor.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The present disclosure relates to an information processing device, an information processing method, a program, and a prediction device. [Background technology]

[0002] To efficiently search for new materials, it is desirable to accurately predict yields (e.g., yields in coupling reactions), especially when using rare or expensive materials, and to perform calculations for a vast variety of catalysts.

[0003] However, existing techniques do not offer significant improvements in yield prediction accuracy. Furthermore, calculations for a vast number of catalysts require considerable time, and when conformational considerations are taken into account, it is difficult to perform substantial calculations within a reasonable time frame. [Prior art documents] [Patent documents]

[0004] [Patent Document 1] Japanese Patent Publication No. 2022-102726 Summary of the Invention [Problem to be solved by the invention]

[0005] One non-limiting problem that the embodiments of the present disclosure aim to solve is to generate a model that improves the accuracy of yield prediction. The problem that the embodiments of the present disclosure aim to solve is not limited to the above-described problem, and as a further example of some limited problems, it can also be a problem that corresponds to the effects described in the embodiments. In other words, a problem that corresponds to at least one of the effects described in the description of the embodiments of the present disclosure can be the problem that the present disclosure aims to solve. [Means for solving the problem]

[0006] According to one embodiment, an information processing device includes a memory and a processor, wherein the processor uses at least one of electronic descriptors and steric descriptors of substances involved in a chemical reaction as explanatory descriptors to train a model that calculates information related to a yield of a product of the chemical reaction, the information being associated with the explanatory descriptors. [Brief explanation of the drawings]

[0007] [Figure 1] 10 is a flowchart showing processing of an information processing apparatus according to an embodiment. [Figure 2] FIG. 2 is a diagram illustrating an example of a material conformation according to an embodiment. [Figure 3] FIG. 10 is a diagram for explaining an example of a stereoscopic descriptor according to an embodiment. [Figure 4] FIG. 10 is a diagram for explaining an example of a stereoscopic descriptor according to an embodiment. [Figure 5] FIG. 10 is a diagram for explaining an example of a stereoscopic descriptor according to an embodiment. [Figure 6] 10 is a flowchart showing processing of an information processing apparatus according to an embodiment. [Figure 7] FIG. 1 is a diagram showing a hardware implementation example of an information processing device according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0008] DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS The following describes embodiments of the present invention with reference to the drawings. The drawings and the description of the embodiments are given by way of example only and are not intended to limit the present invention.

[0009] The processes in the present disclosure are executed, for example, using an information processing device. The information processing device may be a part of an information processing system. When executed by multiple information processing devices, at least one of the processes according to the embodiments of the present disclosure may be executed by one information processing device.

[0010] 1 is a flowchart showing an example of processing by an information processing device according to an embodiment. The information processing device trains a model for predicting yields through processing based on this flowchart.

[0011] The information processing device first acquires training data (S100). The training data may be, for example, experimental data, literature data, or calculation data. The training data is input data and output data for a model to be trained.

[0012] The input data may include data on substances involved in a chemical reaction. For example, the input data may include at least one of data on reactants of the chemical reaction and data on products produced by the chemical reaction. Each of the data may be, for example, an arbitrary chemical formula.

[0013] The output data may include information related to the yield of a product of a chemical reaction. The output data may be, for example, numerical data indicating the yield. This numerical data indicating the yield is the yield when a product is produced from the reactants obtained as input data.

[0014] For example, if substance A and substance B are reactants and substance C is produced as a product, data relating to the reactants A and B and the product C are input data, and data indicating the yield when substance C is produced is output data.

[0015] The input data may include data relating to substances used in analyzing chemical reactions, for example, data relating to substances for evaluating properties of substances involved in chemical reactions.

[0016] As another example, the input data may include not only data on substances involved in a chemical reaction but also data on catalysts, etc. In this case, the output data may be a numerical value indicating the yield when a product is produced using the catalyst. The catalyst may be, for example, a heterogeneous catalyst or a homogeneous catalyst, but is not limited to these.

[0017] After acquiring the data, the information processing device generates various three-dimensional structures for each substance (S102). The various three-dimensional structures are represented by the same structural formula but are different from one another. The three-dimensional structures are, for example, multiple conformations of a certain material.

[0018] 2 is a diagram showing an example of the conformation of a substance. After acquiring data on the substance in S100, the information processing device generates data on various three-dimensional structures of the material. For example, conformation data corresponding to each of three-dimensional structures that are represented by the same structural formula but are different from each other, such as conformation A, conformation B, and conformation C in FIG. 2, is generated.

[0019] The information processing device can generate conformation data from material information using a common method. For example, the information processing device can acquire a SMILES representation of the material and generate various conformation data based on this SMILES representation. The method for generating conformation data is not particularly limited, and any known method for generating conformation data can be used. For example, the information processing device can generate conformation data using an RDKit.

[0020] The information processing device may generate conformation data in an energetically stable state as well as conformation data in an energetically metastable state or conformation data in an energetically unstable state. Conformation data in an energetically unstable state is, for example, conformation data in a transition state of a substance in a chemical reaction.

[0021] The information processing device may generate conformational data corresponding to a predetermined number of three-dimensional structures for each of the target materials. For example, the information processing device may generate conformational data corresponding to a plurality of three-dimensional structures that are all or at least a portion of each of the target materials, or may generate a predetermined number of conformational data selected from the three-dimensional structures of each of the target materials. The predetermined number of conformational data to be generated is, for example, 50, but is not limited to this and can be defined within a range that allows appropriate inference.

[0022] In this way, the information processing device calculates a plurality of three-dimensional structures of a substance involved in a chemical reaction. These three-dimensional structures may be, for example, represented by the same structural formula but different from each other.

[0023] Next, the information processing device performs structural optimization of the reactants and products based on the conformational data of the reactants generated in S102 and the product data acquired in S100 (S104). The information processing device performs structural optimization using, for example, a machine learning potential. The information processing device can perform structural optimization using Neural Network Potential (NNP) as the machine learning potential. By using NNP, it is possible to perform high-speed and highly accurate structural optimization for various combinations of reactant conformations and products.

[0024] The model used in NNP is a model that uses quantum chemical calculations as training data to learn the relationship between atomic coordinates and physical properties such as energy. By using NNP, it is possible to predict physical properties such as energy without solving strict equations as in quantum chemical calculations, thereby achieving high-speed calculations.

[0025] In this way, the information processing device performs structural optimization based on each of the multiple different three-dimensional structures, calculates multiple optimized structures of the substance, and then, in subsequent processing, performs calculations based on various descriptors related to these multiple optimized structures.

[0026] Next, the information processing device calculates an electronic descriptor for the structure acquired in S104 (S106). The information processing device may calculate multiple electronic descriptors.

[0027] The electronic descriptor is calculated based on the electronic state of a substance involved in a chemical reaction. Examples of the electronic descriptor include, but are not limited to, physical properties that can be calculated by quantum chemical calculations, values ​​that are converted into specific evaluation indices based on values ​​that can be calculated by quantum chemical calculations, etc.

[0028] Examples of electronic descriptors include, but are not limited to, the Tolman χ value (CA Tolman, Am. Chem. Rev., 77 313, 1977), metal-ligand bond distance, metal-ligand Bader charge, activation energy, reaction energy, and descriptors for bicoordination of a monodentate ligand.

[0029] The information processing device calculates electronic descriptors for the structurally optimized reactants and products. In addition, the information processing device may calculate electronic descriptors for the catalysts and the like.

[0030] For example, the information processing device may calculate an electronic descriptor based on a partial structure of a substance involved in a chemical reaction, rather than the substance itself. In one non-limiting example, if the partial structure of the substance involved in the chemical reaction includes a structure having symmetric stretching vibration, such as a carbonyl group, the electronic descriptor can be calculated based on the Tolman χ value for the symmetric stretching vibration. In another non-limiting example, the electronic descriptor may include the distance between atoms included in the partial structure, or the Bader charge of the atom.

[0031] The information processing device calculates a stereoscopic descriptor in addition to the electronic descriptor (S108). The information processing device may calculate a plurality of stereoscopic descriptors.

[0032] A steric descriptor describes the three-dimensional structure of a molecule. For example, a steric descriptor is an index that indicates the bulkiness of the structure of a substance involved in a chemical reaction. For example, a steric descriptor is %V, which indicates the proportion of the volume that a ligand occupies within a sphere within a certain range from the metal. bur (Buried Volume), Sterimol parameters indicating the width (B1, B5) and depth (L) when the substituent is observed from the bond direction or cross section, etc.

[0033] 3 to 5 are diagrams showing examples of stereoscopic descriptors. As shown in FIG. bur is defined as the ratio of the volume occupied by the ligand in a sphere centered on a certain atom. In the figure, the atom shown with diagonal lines overlaps with the sphere of radius r centered on the metal atom, and the %V bur The central atom can be a metal atom.

[0034] The Sterimol parameter is defined by the depth direction from a certain atom, or by the shortest and longest widths when observed from the direction of this atom. In this case, the central atom can also be a metal atom.

[0035] For example, the Sterimol L parameter is defined as the depth of the ligand along the depth direction of the axis of the Metal-Ligand bond (ML bond) for the atom shown with diagonal lines in FIG.

[0036] For example, the Sterimol B1 and B5 parameters are defined by the shortest width (B1) and longest width (B5) of the ligand for the atom of interest, measured along the direction perpendicular to the axis of the M-L bond of the atom indicated by the diagonal lines, as shown in Figure 5. For example, these parameters can be defined as the minimum and maximum lengths within (atom-specific parameter + half the van der Waals radius of each element) Å from the center of the atom, measured along the direction perpendicular to the axis of the M-L bond.

[0037] The information processing device calculates steric descriptors for the structurally optimized reactants and products. In addition, the information processing device may calculate steric descriptors for the catalysts and the like.

[0038] After calculating the electronic and steric descriptors, the information processing device calculates a structure group descriptor for each structure group including the steric structure (S110). The structure group descriptor may be, for example, a maximum value, a minimum value, or a Boltzmann weighted average.

[0039] The information processing device uses one or more electronic descriptors and / or one or more steric descriptors and structure group descriptors calculated for the generated three-dimensional structure (including conformation) as explanatory descriptors for the model to train a model that estimates the corresponding yield (S112). That is, the information processing device uses at least one of the electronic descriptors and the steric descriptors of the substances involved in the chemical reaction as explanatory descriptors to train a model that calculates information related to the yield of the product of the chemical reaction associated with the explanatory descriptor.

[0040] 6 is a flowchart showing an example of a model training process according to an embodiment. Prior to optimizing the model, the information processing device selects explanatory descriptors (features) to be used in the model (S200). The information processing device selects the features to be used in the model based on, for example, a commonly known feature selection algorithm.

[0041] In one example, when a random forest is used as a model, the information processing device can use feature selection based on the feature importance of the random forest, Boruta, which is a feature selection method based on the feature importance, etc. However, the information processing device does not exclude other feature selection methods.

[0042] The information processing device removes the exceptional value data after the process of S200 or in parallel with the process of S200 (S202). The information processing device can select a descriptor to be used as an explanatory descriptor from the structure group descriptors based on the selection criteria for the structure group descriptors calculated in S110, for example.

[0043] The information processing device optimizes the model based on the feature quantities extracted by the processes of S200 and S202 (S204). The model may be generated as, for example, a regression model or another model (e.g., a classification model). The information processing device trains the relationship between the yield data acquired in S100 and the feature quantities acquired in S200, with the objective variable y being the yield.

number

[0044] The information processing device sets the objective variable and explanatory variables as shown in Equation (1), and uses various machine learning techniques to train a model that estimates the yield from the feature values ​​expressed by the descriptors, where y is the objective variable, the yield, and x is the selected feature value expressed by the descriptors, which are explanatory variables.

[0045] The information processing device compares the yield obtained in S100 with the predicted value y of the yield obtained by equation (1) to optimize the function f(·), i.e., the model. The information processing device can generate the model using techniques such as support vector machines and random forests. However, these are given as non-limiting examples, and the information processing device can optimize the model using various appropriate machine learning techniques.

[0046] The information processing device determines whether or not the generation of the final model has been completed for the model optimized using the selected feature (S206). The information processing device compares the yield data acquired in S100 with the yield estimated from the model to evaluate the model optimized in S204.

[0047] The information processing device is, for example, R 2 Evaluation can be performed using indicators such as Average Absolute Error (MAE), Root Mean Square Error (RMSE), etc. However, other evaluation indicators are not excluded, and any indicator that can evaluate an appropriate model will suffice.

[0048] The information processing device can complete the optimization (S206: YES) depending on whether the evaluation value based on the above index exceeds a predetermined evaluation value, for example. If it is determined not to complete the optimization (S206: NO), the information processing device repeats the process from S200, which selects features. Note that when repeating, the process of removing outlier data in S202 can be omitted if the data used in the previous iterative calculation can be used.

[0049] The determination of the completion of optimization is not limited to the above conditions. For example, the completion condition can be determined by generating a predetermined number of models, generating models for a predetermined time, etc. If multiple models are generated, the model with the highest evaluation value using a predetermined index can be determined as the optimized model.

[0050] When the information processing device determines that the optimization is complete, it outputs parameters related to the model and then completes the processing. In this way, the information processing device can generate a model that predicts the yield.

[0051] According to the process of the present disclosure, by using descriptors, particularly electronic and steric descriptors, in combination, it is possible to generate a model with improved yield prediction accuracy. Furthermore, by using NNP in structural optimization, it is possible to achieve optimization of a large number of conformations in a short time, and by using information on these large number of conformations, it is possible to generate a yield inference model with higher prediction accuracy.

[0052] An information processing device identical to the above-described information processing device or a different information processing device can use the yield inference model generated by the above-described method as, for example, a prediction device. That is, by using the generated yield inference model, the prediction device can predict information related to the yield of a product of a chemical reaction to be predicted by inputting the explanatory descriptor of a substance involved in the chemical reaction to be predicted into a learning model that uses at least one of an electronic descriptor or a steric descriptor of a substance involved in a chemical reaction as an explanatory descriptor and calculates information related to the yield of a product of the chemical reaction associated with the explanatory descriptor.

[0053] The trained model in the above-described embodiment may be a concept that includes, for example, a model that has been trained as described and then further distilled using a general method.

[0054] Some or all of the devices (information processing devices) in the above-described embodiments may be configured as hardware, or may be configured as software (programs) executing information processing by a CPU (Central Processing Unit) or a GPU (Graphics Processing Unit), etc. When software information processing is configured, software that realizes at least some of the functions of each device in the above-described embodiments may be stored on a non-transitory storage medium (non-transitory computer-readable medium) such as a CD-ROM (Compact Disc-Read Only Memory) or a USB (Universal Serial Bus) memory, and the software information processing may be executed by loading the software into a computer. The software may also be downloaded via a communications network. Furthermore, all or part of the software processing may be implemented in a circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array), thereby allowing the software information processing to be executed by hardware.

[0055] The storage medium that stores the software may be a removable medium such as an optical disk, or a fixed medium such as a hard disk or memory. The storage medium may be located inside the computer (such as a main memory or auxiliary memory), or may be located outside the computer.

[0056] 7 is a block diagram showing an example of the hardware configuration of each device (information processing device) in the above-described embodiment. Each device may be realized as a computer 7 including, for example, a processor 71, a main storage device 72 (memory), an auxiliary storage device 73 (memory), a network interface 74, and a device interface 75, all of which are connected via a bus 76.

[0057] Although the computer 7 in FIG. 7 includes one of each component, it may include multiple of the same component. Also, while FIG. 7 shows one computer 7, the software may be installed on multiple computers, and each of the multiple computers may execute the same or different parts of the software. In this case, a distributed computing configuration may be used in which each computer communicates with the other computers via a network interface 74 or the like to execute the processing. In other words, each device (information processing device) in the above-described embodiment may be configured as a system in which one or more computers execute instructions stored in one or more storage devices to achieve its function. Furthermore, the system may be configured such that information sent from a terminal is processed by one or more computers located on a cloud, and the processing results are sent to the terminal.

[0058] The various calculations of each device (information processing device) in the above-described embodiments may be executed in parallel using one or more processors, or using multiple computers connected via a network. Furthermore, the various calculations may be distributed to multiple processor cores within a processor and executed in parallel. Furthermore, some or all of the processes, means, etc. disclosed herein may be implemented by at least one processor and storage device provided on a cloud that can communicate with computer 7 via a network. Thus, each device in the above-described embodiments may be implemented in the form of parallel computing using one or more computers.

[0059] The processor 71 may be an electronic circuit (CPU, GPU, FPGA, ASIC, etc.) that performs at least one of computer control and calculation. The processor 71 may also be a general-purpose processor, a dedicated processing circuit designed to perform a specific calculation, or a semiconductor device that includes both a general-purpose processor and a dedicated processing circuit. The processor 71 may also include an optical circuit or a calculation function based on quantum computing.

[0060] The processor 71 may perform arithmetic processing based on data or software input from each device, etc., configured inside the computer 7, and may output the calculation results or control signals to each device, etc. The processor 71 may control each component constituting the computer 7 by executing the OS (Operating System) of the computer 7, applications, etc.

[0061] Each device (information processing device) in the above-described embodiments may be realized by one or more processors 71. Here, the processor 71 may refer to one or more electronic circuits arranged on one chip, or may refer to one or more electronic circuits arranged on two or more chips or two or more devices. When multiple electronic circuits are used, the electronic circuits may communicate with each other via wire or wirelessly.

[0062] The main memory device 72 may store instructions executed by the processor 71 and various data, etc., and information stored in the main memory device 72 may be read by the processor 71. The auxiliary memory device 73 is a memory device other than the main memory device 72. Note that these memory devices refer to any electronic component capable of storing electronic information and may be semiconductor memory. The semiconductor memory may be either volatile or nonvolatile memory. The memory device for saving various data, etc. in each device (information processing device) in the above-described embodiments may be realized by the main memory device 72 or the auxiliary memory device 73, or may be realized by an internal memory built into the processor 71. For example, the memory unit in the above-described embodiment may be realized by the main memory device 72 or the auxiliary memory device 73. For example, at least some of the operations in the present disclosure may be implemented by the processor constructing a trained model by referring to data related to the trained model stored in a memory circuit. The memory device stores, for example, data related to a trained model that outputs physical property values ​​when molecular information is input. For example, the processor uses the trained model to perform a simulation in which multiple molecular models are adsorbed onto multiple adsorption sites. The trained model is, for example, a model used in NNP (Neural Network Potential). For example, the physical property values ​​include at least the energy or force of the molecules.

[0063] When each device (information processing device) in the above-described embodiments is configured with at least one storage device (memory) and at least one processor connected (coupled) to this at least one storage device, at least one processor may be connected to one storage device. At least one storage device may be connected to one processor. A configuration in which at least one processor among multiple processors is connected to at least one storage device among multiple storage devices may also be included. This configuration may also be realized by storage devices and processors included in multiple computers. Furthermore, a configuration in which a storage device is integrated with a processor (for example, a cache memory including an L1 cache and an L2 cache) may also be included.

[0064] The network interface 74 is an interface for connecting to the communication network 8 wirelessly or via a wire. The network interface 74 may be an appropriate interface, such as one that conforms to an existing communication standard. The network interface 74 may exchange information with an external device 9A connected via the communication network 8. The communication network 8 may be any one of a WAN (Wide Area Network), a LAN (Local Area Network), a PAN (Personal Area Network), etc., or a combination thereof, as long as information is exchanged between the computer 7 and the external device 9A. An example of a WAN is the Internet, an example of a LAN is IEEE 802.11 or Ethernet (registered trademark), and an example of a PAN is Bluetooth (registered trademark) or NFC (Near Field Communication), etc.

[0065] The device interface 75 is an interface such as USB that directly connects to the external device 9B.

[0066] The external device 9A is a device connected to the computer 7 via a network. The external device 9B is a device directly connected to the computer 7.

[0067] For example, the external device 9A or the external device 9B may be an input device. The input device may be a device such as a camera, a microphone, a motion capture device, various sensors, a keyboard, a mouse, or a touch panel, and provides acquired information to the computer 7. Alternatively, the external device 9A or the external device 9B may be a device equipped with an input unit, a memory, and a processor, such as a personal computer, a tablet terminal, or a smartphone.

[0068] Furthermore, the external device 9A or the external device 9B may be, for example, an output device. The output device may be, for example, a display device such as an LCD (Liquid Crystal Display) or an organic EL (Electro Luminescence) panel, or a speaker that outputs sound or the like. Alternatively, the external device 9A or the external device 9B may be a device including an output unit, a memory, and a processor, such as a personal computer, a tablet terminal, or a smartphone.

[0069] Furthermore, the external device 9A or the external device 9B may be a storage device (memory). For example, the external device 9A may be a network storage or the like, and the external device 9B may be a storage such as an HDD.

[0070] Furthermore, the external device 9A or the external device 9B may be a device having some of the functions of the components of each device (information processing device) in the above-described embodiment. That is, the computer 7 may transmit some or all of the processing results to the external device 9A or the external device 9B, or may receive some or all of the processing results from the external device 9A or the external device 9B.

[0071] The embodiments of the present disclosure can be summarized as follows as an example.

[0072] (1) A memory and a processor, The processor: training a model that uses at least one of electronic descriptors and steric descriptors of substances involved in a chemical reaction as an explanatory descriptor, and calculates information related to the yield of a product of the chemical reaction, the information being associated with the explanatory descriptor; Information processing device.

[0073] (2) The processor: calculating a plurality of three-dimensional structures of substances involved in the chemical reaction, the three-dimensional structures being represented by the same structural formula but different from each other; calculating at least one of the electronic descriptor or the three-dimensional descriptor based on a plurality of the three-dimensional structures; An information processing device according to (1).

[0074] (3) The processor: calculating a plurality of optimized structures of the substance by performing structural optimization based on each of the plurality of three-dimensional structures; calculating at least one of the electronic descriptor or the steric descriptor based on a plurality of the optimized structures; (2) An information processing device according to the present invention.

[0075] (4) The processor: calculating a structure group descriptor for a structure group including a plurality of the three-dimensional structures; training the model using the structure group descriptors as the explanatory descriptors; An information processing device according to (2) or (3).

[0076] (5) The processor: training the model using descriptors selected from the structure group descriptors based on a selection criterion as the explanatory descriptors; (4) An information processing device according to the present invention.

[0077] (6) The electronic descriptor is a descriptor calculated based on the electronic state of a substance involved in the chemical reaction. An information processing device according to any one of (1) to (5).

[0078] (7) The steric descriptors include at least a descriptor representing an index of bulkiness of the structure of a substance involved in the chemical reaction. An information processing device according to any one of (1) to (6).

[0079] (8) the explanatory descriptors include at least a descriptor calculated based on a partial structure of a substance involved in the chemical reaction; An information processing device according to any one of (1) to (7).

[0080] (9) The processor: optimizing the structure of the plurality of three-dimensional structures using a machine learning potential; An information processing device according to any one of (3) or (4) to (8) that cites (3).

[0081] (10) The machine learning potential is a neural network potential. (9) An information processing device according to (9).

[0082] (11) The processor: training a model that uses at least one of electronic descriptors and steric descriptors of substances involved in a chemical reaction as an explanatory descriptor, and calculates information related to the yield of a product of the chemical reaction, the information being associated with the explanatory descriptor; Information processing methods.

[0083] (12) The processor training a model that uses at least one of electronic descriptors and steric descriptors of substances involved in a chemical reaction as an explanatory descriptor, and calculates information related to the yield of a product of the chemical reaction, the information being associated with the explanatory descriptor; A program that makes things happen.

[0084] (13) A memory and a processor, The processor: At least one of an electronic descriptor and a steric descriptor of a substance involved in a chemical reaction is used as an explanatory descriptor, and information related to the yield of a product of the chemical reaction associated with the explanatory descriptor is input to a trained model, thereby predicting information related to the yield of the product of the chemical reaction to be predicted. Prediction device.

[0085] Although the embodiments of the present disclosure have been described in detail above, the present disclosure is not limited to the individual embodiments described above. Various additions, modifications, substitutions, and partial deletions are possible within the scope of the conceptual idea and spirit of the present disclosure, which is derived from the content defined in the claims and their equivalents. For example, when numerical values ​​or formulas are used in the above-described embodiments, they are shown for illustrative purposes and do not limit the scope of the present disclosure. Furthermore, the order of each operation shown in the embodiments is also illustrative and does not limit the scope of the present disclosure. [Explanation of symbols]

[0086] 7: Computer, 71: Processor, 72: Main storage, 73: Auxiliary storage, 74: Network interface, 75: Device Interface, 76: Bus, 8: Communication networks, 9A, 9B: External device

Claims

1. A memory and a processor, The processor: training a model that uses at least one of electronic descriptors and steric descriptors of substances involved in a chemical reaction as an explanatory descriptor, and calculates information related to the yield of a product of the chemical reaction, the information being associated with the explanatory descriptor; Information processing device.

2. The processor: calculating a plurality of three-dimensional structures of substances involved in the chemical reaction, the three-dimensional structures being represented by the same structural formula but different from each other; calculating at least one of the electronic descriptor or the three-dimensional descriptor based on a plurality of the three-dimensional structures; The information processing device according to claim 1.

3. The processor: calculating a plurality of optimized structures of the substance by performing structural optimization based on each of the plurality of three-dimensional structures; calculating at least one of the electronic descriptor or the steric descriptor based on a plurality of the optimized structures; The information processing device according to claim 2.

4. The processor: calculating a structure group descriptor for a structure group including a plurality of the three-dimensional structures; training the model using the structure group descriptors as the explanatory descriptors; The information processing device according to claim 2.

5. The processor: training the model using descriptors selected from the structure group descriptors based on a selection criterion as the explanatory descriptors; The information processing device according to claim 4.

6. The electronic descriptor is a descriptor calculated based on the electronic state of a substance involved in the chemical reaction. The information processing device according to claim 1.

7. The steric descriptors include at least a descriptor representing an index of bulkiness of the structure of a substance involved in the chemical reaction. The information processing device according to claim 1.

8. the explanatory descriptors include at least a descriptor calculated based on a partial structure of a substance involved in the chemical reaction; The information processing device according to claim 1.

9. The processor: optimizing the structure of the plurality of three-dimensional structures using a machine learning potential; The information processing device according to claim 3.

10. The machine learning potential is a neural network potential. The information processing device according to claim 9.

11. The processor: training a model that uses at least one of electronic descriptors and steric descriptors of substances involved in a chemical reaction as an explanatory descriptor, and calculates information related to the yield of a product of the chemical reaction, the information being associated with the explanatory descriptor; Information processing methods.

12. The processor training a model that uses at least one of electronic descriptors and steric descriptors of substances involved in a chemical reaction as an explanatory descriptor, and calculates information related to the yield of a product of the chemical reaction, the information being associated with the explanatory descriptor; A program that makes things happen.

13. A memory and a processor, The processor: At least one of an electronic descriptor and a steric descriptor of a substance involved in a chemical reaction is used as an explanatory descriptor, and information related to the yield of a product of the chemical reaction associated with the explanatory descriptor is input to a trained model, thereby predicting information related to the yield of the product of the chemical reaction to be predicted. Prediction device.

Images