Information processing device, information processing method and program

Abstract

To acquire data of a highly accurate product based on a synthesis condition.SOLUTION: An information processing device includes a storage part and a processing circuit. The processing circuit acquires a synthesis condition and time series data of an experiment value about one or more indexes corresponding to the synthesis condition about one or more samples being the samples with the synthesis condition set, inputs the acquired synthesis condition to a first model for acquiring arithmetic data of the time series data about the one or more indexes if the synthesis condition is inputted to acquire time series data of an arithmetic value, evaluates the time series data of the experiment value and the time series data of the arithmetic value by an evaluation function, performs fitting of a parameter about easiness-to-causes of a reaction by using the evaluation function, and updates the evaluation function on the basis of a fitting result. The evaluation function includes at least an error based on distance and other errors.SELECTED DRAWING: Figure 3

Description

[Technical Field]

[0001] The embodiments disclosed in the present specification and drawings relate to an information processing device, an information processing method, and a program. [Background technology]

[0002] In an effort to protect the environment, the machinery industry as a whole is working to reduce CO2 emissions. One effective solution is to reduce friction loss that occurs between machine parts. To reduce friction, chemical products such as lubricants that control friction are used. The additives that make up lubricants are subject to various constraints in balancing production costs and quality, making it necessary to optimize synthesis conditions. These constraints include not infringing on manufacturing methods patented by third parties, having appropriate friction performance, and reducing toxicity. This poses the challenge of needing to find appropriate synthesis conditions.

[0003] When searching for these appropriate synthesis conditions as an optimization problem, it is difficult to explore all possible combinations of synthesis conditions due to cost considerations, and it is necessary to narrow the search range for the conditions to be optimized. This results in local optimization, and there is a high probability that better synthesis conditions will be overlooked. One method for efficiently searching for good conditions is to build a simulator that predicts the component ratios at each time under certain synthesis conditions, and to search for the optimal synthesis conditions in a virtual space. Simulation is effective because it allows for the determination of the merits and demerits of many synthesis conditions at low cost.

[0004] For example, a simulation can be performed by simultaneously solving the reaction rate equations using each reaction rate constant (or activation energy) for the reaction being considered. One method for determining these reaction rate constants, etc., is to use electronic structure calculations to determine them theoretically alone. This method often has difficulty quantitatively predicting the component ratios of complex real-world reactions. As a result, an approach is taken in which each reaction rate constant, etc., is optimized to reproduce experimental component ratios. However, such methods are likely to be insufficiently accurate in predicting the time-series changes of minor components. [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Patent Publication No. 2021-163422 Summary of the Invention [Problem to be solved by the invention]

[0006] One non-limiting problem that the embodiments of the present disclosure aim to solve is to obtain highly accurate product data based on synthesis conditions. 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 corresponding 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 a problem that the present disclosure aims to solve. [Means for solving the problem]

[0007] According to one embodiment, an information processing device includes a storage unit and a processing circuit. The processing circuitry For one or more samples for which synthesis conditions have been set, time series data of experimental values ​​for the synthesis conditions and one or more indexes corresponding to the synthesis conditions is acquired; inputting the acquired synthesis conditions into a first model that acquires calculation data of time series data for the one or more indexes when the synthesis conditions are input, and acquiring time series data of the calculation values; evaluating the time series data of the experimental values ​​and the time series data of the calculated values ​​using an evaluation function; Using the evaluation function, fitting a parameter related to the likelihood of reaction occurring; updating the evaluation function based on the fitting result; The evaluation function includes at least a distance-based error and another error. [Brief explanation of the drawings]

[0008] [Figure 1] FIG. 1 is a diagram schematically illustrating an example of an information processing apparatus according to an embodiment. [Figure 2] FIG. 1 is a diagram schematically illustrating an example of an information processing system according to an embodiment. [Figure 3] 10 is a flowchart showing an example of processing of an information processing apparatus according to an embodiment. [Figure 4] 10 is a flowchart showing an example of processing of an information processing apparatus according to an embodiment. [Figure 5] FIG. 10 is a diagram schematically illustrating an example of interpolation of a response variable according to an embodiment. [Figure 6] FIG. 10 is a diagram schematically illustrating an example of interpolation of a response variable according to an embodiment. [Figure 7] FIG. 1 is a diagram showing an example of hardware implementation of an information processing device according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0009] Hereinafter, embodiments of the present invention will be described with reference to the drawings. The drawings and the description of the embodiments are provided as examples and are not intended to limit the present invention. In this disclosure, processes are described as being executed by an information processing device, more specifically, a processing circuit within the information processing device, but the present invention is not limited to this. For example, one or more processes may be implemented by multiple information processing devices, or one or more processes may be implemented by multiple various circuits.

[0010] In the present disclosure, expressions such as "less than or equal to" and "greater than or equal to" are used, but these are not used as strict definitions and can be appropriately read as "less than," "greater than," etc. Furthermore, the larger the various evaluation values, the higher the evaluation, and the smaller the evaluation, the lower the evaluation, but this is not limited to this, and depending on the evaluation function, it is also possible to conversely say that the smaller the evaluation, the higher the evaluation, and the larger the evaluation, the lower the evaluation.

[0011] 1 is a block diagram schematically illustrating an example of an information processing device 10 according to an embodiment. The information processing device 10 includes an interface (I / F 100), a storage unit 102, and a processing unit 104. Although not shown, the information processing device 10 may also include a control unit (not shown) for controlling each component, a power supply unit (not shown) for supplying power to each component, and other components as needed. Furthermore, at least a portion of the circuits constituting the storage unit 102 and the processing unit 104 may be provided outside the information processing device 10 and implemented by transmitting and receiving data via the I / F 100.

[0012] The I / F 100 is an interface that connects the inside and outside of the information processing device 10. The I / F 100 may have an interface that realizes data transmission and reception, such as a network interface, etc. The I / F 100 may also have a user interface, such as a display that presents data to a user.

[0013] The memory unit 102 has a memory circuit that stores data necessary for the operation of the information processing device 10. The memory unit 102 can store, as necessary, data necessary for the operation of the processing unit 104, intermediate data used in calculations, or data obtained as a result.

[0014] The processing unit 104 has a processing circuit that executes the operations of the information processing device 10. The information processing in the present disclosure can be specifically implemented mainly using the processing unit 104 as a hardware resource.

[0015] In the present disclosure, the information processing device 10 can be broadly classified as a device that performs two operations. One information processing device 10 performs information processing to estimate the component ratio of each material in a compound from given conditions. The other information processing device 10 performs information processing to search for given data, for example, optimal synthesis conditions for producing a product, such as the synthesis process (timing of adding materials, temperature, pressure, time until completion of processing, etc.).

[0016] The above two information processes may be executed in separate information processing devices 10 or may be executed in the same information processing device 10 .

[0017] 2 is a block diagram schematically illustrating an example of an information processing system including an information processing device 10 according to an embodiment. The information processing system 1 includes one or more first devices 12 as an example of the information processing device 10, and one or more second devices 14 as an example of the information processing device 10. As described above, the information processing system 1 may be formed from a single information processing device 10, and in this case, the first device 12 and the second device 14 may be the same information processing device 10.

[0018] The first device 12 executes a simulation based on the synthesis conditions for the compound (e.g., the ratio of materials, the timing of input, the environment, etc.), and acquires data such as the compound synthesized as a synthesis result, or the amount of a compound during synthesis, or the ratio of multiple synthesized compounds (S1).

[0019] The second device 14 appropriately evaluates the product based on data such as the component ratio of the compounds in the product, and searches for appropriate synthesis conditions (S2).

[0020] The first model used in the simulation of the first device 12 is preferably properly trained.

[0021] On the other hand, if the accuracy of the first model and the second model that may be used in the second device 14 is higher than a predetermined value, it is also possible to update the model by mutually feeding back the outputs of the first device 12 and the second device 14.

[0022] For example, the information processing system 1 can improve the accuracy of the second model by inputting the synthesis conditions acquired by the second device 14 into the first device 12 to obtain the synthesis result, and feeding this synthesis result back to the second device 14.

[0023] Furthermore, after optimizing the first and second models, the first device 12 uses the first model and the second device 14 uses the second model, and processing is performed using the outputs of each other, thereby enabling the search for synthesis conditions that will produce a product with desired properties.

[0024] The optimization of the first and second models will be described in detail below.

[0025] <Optimization of the first model>

[0026] The first model is a model that can operate as a simulator that obtains a synthesis result from synthesis conditions in the first device 12. In the present disclosure, the following evaluation function can be used as a non-limiting example.

number

number

number

number

[0027] where α1, α2, and α3 are weighting coefficients, and C exp n,i (t) is the experimental value (teaching data) for synthesis condition n, component i, and time t, and C calc n,i (t) is the output value of the simulator under synthesis condition n, component i, and time t, and ε is the noise level, e.g., 10 -14 ~10 -8 where τ represents a predetermined time. The parameters and the evaluation function of the first model are updated so that the evaluation value shown in equation (1) becomes smaller. C n,i (t) is a value indicating the ratio of component i in the product under synthesis conditions n and time t, for example.

[0028] E1 in equation (2) is the sum of squares error. This error E1 is specialized for fitting large variables at each time. E1 is not limited to the error shown in equation (2), and can be other distances that represent the difference between the experimental value and the simulation value at each time, such as the mean square, the root mean square, the L1 norm, or the L2 norm, or a combination of these indices (including the sum of squares).

[0029] E2 in equation (3) is the sum of the squares of the logarithmic difference. This logarithmic error E2 is specialized for fitting small variables. The noise level, represented by ε, is set so that the logarithm does not become 0 and to suppress the influence of noise.

[0030] E3 shown in equation (4) is the sum of the differences in the rate of change of the components over time. E3, which is shown as the difference in rate from this predetermined time, is specialized for incorporating time changes. The predetermined time represented by τ is, for example, a reference time for the initial synthesis (e.g., 5 where unit time is 1), and the influence of time changes can be incorporated by the ratio with this time value.

[0031] In addition, in formula (1), all of the errors E1, E2, and E3 are used, but this is not limited to this. For example, one of α2 and α3 may be 0. If either one is 0, it is possible to omit calculation of the corresponding error in the following processing.

[0032] The evaluation function may also include errors other than those mentioned above. In this case, it is desirable to include errors that take into account small values ​​and changes over time.

[0033] To implement this technique, one or more information processing devices are used. The one or more information processing devices may collectively include at least one processing circuit and at least one memory circuit. Hereinafter, the processing circuit of the information processing device may perform the processing.

[0034] 3 is a flowchart showing an example of optimization processing of the first model according to an embodiment. This processing may be executed by the first device 12 or another information processing device 10. Hereinafter, the entity that executes the processing will be referred to as a processing circuit.

[0035] The processing circuit acquires synthesis conditions for synthesizing a compound and experimental values ​​relating to a product actually obtained under the synthesis conditions (S100). The processing circuit uses these data as training data to perform optimization of the first model. It is preferable that the amount of data acquired is sufficient to perform training by learning.

[0036] The data relating to the experimental values ​​is time-series data, for example, time-series data of the ratio of each component i.

[0037] The processing circuit inputs the acquired synthesis conditions into the first model and executes a simulation (S102). In the simulation process, the processing circuit acquires and stores the calculated values ​​of time-series data related to necessary data from the acquired data. The necessary data may be data representing desired features at the timing when the search is executed in the first device 12.

[0038] The processing circuit uses the evaluation function to perform fitting of a parameter related to the likelihood of the reaction (reaction rate constant or activation energy) (S104). When the reaction is expressed in multiple stages, the processing circuit can fit a parameter related to the likelihood of each reaction.

[0039] The processing circuit performs fitting using E1, E2, and E3 shown in formulas (2), (3), and (4), respectively, instead of E shown in formula (1). The reaction pathways to be fitted may be determined based on general knowledge, estimation from products, or various calculations.

[0040] That is, the processing circuit performs fitting using E1 as the evaluation function, fitting using E2 as the evaluation function, and fitting using E3 as the evaluation function. If one of E2 and E3 is not used as E, fitting for the unused evaluation function does not need to be performed.

[0041] The following explanation is based on equations (1) to (4), but the calculation of the error and weighting coefficients can also be performed appropriately in other cases. For example, when E = E1 + E2, the same process can be performed for α1 and α2.

[0042] The processing circuit calculates the minimum error (E1 min , E2 min , E3 min ) is obtained (S106).

[0043] The processing circuit calculates the minimum error (E1 min , E2 min , E3 min ) and the corresponding weighting coefficients (α1, α2, α3) (α1E1 min , α2E2 min , α3E3 min ), weighting coefficients (α1, α2, α3) are set so that the ratio between the minimum and maximum values ​​is equal to or less than a predetermined ratio, and an evaluation function E is defined (S108). Through this process, the evaluation function used in the search is defined.

[0044] The processing circuit can set, for example, the weighting coefficient α2 for E2 to be larger (about 10 to 50 times larger) than the weighting coefficients α1 and α3 for E1 and E3.

[0045] The processing circuit performs fitting of parameters related to the likelihood of reaction (S110). The processing circuit performs a search for a reaction path using Bayesian optimization, a genetic algorithm, a simulated annealing method, or a gradient method. The processing circuit can also perform a search for a reaction system by combining any of these methods. These methods are realized by common methods.

[0046] After the fitting is completed, the processing circuit determines whether the fitting result has sufficient prediction accuracy (S112). If it is determined that the fitting result has sufficient prediction accuracy, i.e., is represented by a sufficiently small evaluation function value (S112: YES), the processing circuit outputs the defined evaluation function (more specifically, weighting coefficients) (S114) and completes the processing. Outputting may include storing the result in the memory unit 102 or outputting the result to the outside via the I / F 100.

[0047] If the fitting result does not have sufficient prediction accuracy, i.e., if it is not determined that it is represented by a sufficiently small evaluation function value (S112: NO), the processing circuit can repeatedly execute the processes of S108 and S110, repeatedly update the weighting coefficients, and optimize the evaluation function E.

[0048] As described above, according to this embodiment, by using at least the error shown in formula (3), it is possible to construct a simulation that can predict variables with small values ​​of teacher data. Also, by using at least the error shown in formula (4), it is possible to construct a simulation that can predict time-series changes in teacher data.

[0049] By defining the equations (1) to (4), the first device 12 can optimize the evaluation function for constructing a simulation that can predict the time series changes of variables with small values ​​of the training data.

[0050] <Search using the second model>

[0051] The second device 14 uses the first model to search for x that will result in a desired value of y by using a function f as a response variable y and an explanatory variable x, as in general optimization. The response variable y is an evaluation value that indicates the synthesis result, and the explanatory variable x may be a synthesis condition.

number

[0052] The explanatory variable x contains data related to synthesis conditions in each component of the vector. For example, this data is the amount of material to be input at a certain time.

[0053] One example of such an optimization problem, but not limited to it, is Bayesian optimization, which can be found with a small number of searches. This Bayesian search is a method that predicts a function f, which is a target variable y, based on machine learning techniques, and prioritizes searching for explanatory variables x that are likely to cause the target variable y to exceed the champion data.

[0054] When the objective variable y changes smoothly (continuously) with respect to the explanatory variable x, it is possible to search for the explanatory variable x that becomes the desired objective variable y with a small number of trials. On the other hand, when processing is performed using a threshold or the like and there is a region where the objective variable y becomes 0 (for example, the lowest evaluation) discontinuously, it is necessary to increase the number of trials in order to search for a good explanatory variable x.

[0055] For example, in the case of chemical manufacturing, thresholds may be used to determine whether a product contains ingredients that are subject to patent protection from others, whether the product's composition ratio does not infringe on other people's patents, whether the content of substances is below legally specified levels, whether toxicity is below a certain level, etc. When processing using such thresholds occurs, discontinuous regions where the objective variable is 0 may occur in the space of explanatory variables.

[0056] In an optimization problem in chemical manufacturing, it is necessary to perform a search based on the objective variable y in the explanatory variable space where the explanatory variable value becomes 0 discontinuously as described above. One embodiment of the present disclosure describes a method for performing a search quickly, i.e., with a small number of trials, when the connections are not smooth.

[0057] In one embodiment, the smooth connection may be a differentiable connection. This connection may be realized in any suitable manner.

[0058] FIG. 4 is a flowchart showing an example of processing by the second device 14 according to an embodiment.

[0059] The processing circuit first interpolates the response variables (S200). The response variables may be thresholded depending on the conditions. If there are regions where thresholding has been performed, the accuracy may be reduced depending on the optimization method.

[0060] 5 is a diagram showing an example of explanatory variables and response variables. For the sake of explanation, the explanatory variables are assumed to be one-dimensional, but this is not limiting and the dimensions of the explanatory variables can be any dimension. For example, in the portions indicated by dotted lines in the graph (region X0 where the explanatory variables are from th0 to th1 and region X1 where the explanatory variables are from th2 to th3), the response variables are set to 0 by threshold processing.

[0061] For example, consider an optimization method that performs sampling of optimal solutions, such as Bayesian optimization. In a region X2 (where explanatory variables are between th0 and th3) ​​where the objective variable is low (e.g., 0) due to thresholding, sampling is rarely performed and the optimization accuracy is likely to be low even if the optimal solution is within the region X2.

[0062] A non-limiting example of a chemical compound manufacturing process is the addition of glycerin (GOL). In the process, the objective variable may be 0 as a result of threshold processing when GOL is added before a predetermined time and after the predetermined time. For example, a condition may be added to set the explanatory variable to 0 for GOL added as the objective variable other than the initial GOL added before the predetermined time and for GOL added after the predetermined time.

[0063] In such cases, if a sampling and optimization technique is used, for example, the objective variable resulting from sampling the initial GOL input before a predetermined time may be 0, and depending on the behavior of the objective variable around the threshold time, it may be difficult to find the best explanatory variable. Also, unless the number of samples is increased, in other words, the number of optimization trials is significantly increased, it may be difficult to properly estimate the relationship between the appropriate explanatory variables and the objective variable as a function f.

[0064] On the other hand, optimization using a function f without threshold processing may result in an unnecessary increase in the number of samples in areas that should be excluded by the threshold, or may result in optimization of a function f that indicates an inappropriate target variable in areas that should be excluded. For this reason, it is necessary to perform optimization that allows appropriate estimation after threshold processing.

[0065] The second device 14 according to one embodiment appropriately performs this threshold processing to reduce the cost of optimization and improve accuracy.

[0066] In S200, the processing circuitry interpolates the response variable truncated by the threshold. For example, if a threshold is set for the domain of the explanatory variables, the processing circuitry can set the response variable based on the distance from the threshold. Preferably, the processing circuitry may interpolate the truncated response variable using a function that defines a response variable that smoothly decreases with distance from the threshold.

[0067] As an example, the processing circuit may calculate the distance d between the threshold value and the explanatory variable for a region of the explanatory variable where the dependent variable becomes 0 due to the threshold value, and calculate the dependent variable based on the following formula:

number

[0068] Here, β is a coefficient for smoothly connecting the value of the objective variable to the explanatory variable at the threshold, and the vector x th is a vector indicating explanatory variables that are threshold values. β may be, for example, the value of the function f (= f(x)) before threshold processing. The embodiment is not limited to simple exponential functions, and other functions such as trigonometric functions, inverse trigonometric functions, hyperbolic functions, and sigmoid functions may be used.

[0069] Figure 6 shows an example of setting a threshold for the response variable within a certain domain of explanatory variables. In this case, if a threshold is set for the response variable within a certain domain of explanatory variables, the threshold y th For this, the following formula can be used:

number

[0070] y th is the threshold for the objective variable. Again, this is not limited to a simple exponential function.

[0071] The threshold process may be performed based on, for example, indicators such as not infringing on patents of others, toxicity being within a predetermined value, etc. Alternatively, the process may be performed based on indicators such as requiring less raw material, or favorable conditions continuing.

[0072] The processing circuit may be configured to lower the evaluation when at least one of the patent of another person and toxicity exceeds a threshold value in obtaining the objective variables.

[0073] For low friction coefficient components, the processing circuit can use the output of the function f as the score for the component as is.

[0074] The processing circuitry can use qualitative determinations of conditions such as continued good conditions.

[0075] The process of S200 extrapolates the response variable so that it becomes smooth. Alternatively, the processing circuit may perform interpolation using various methods so that the response variable becomes smoothly connected.

[0076] After the variables are interpolated as in the above example, the processing circuit initializes the synthesis conditions (S202). This initialization is similar to the initialization in general search. For example, the processing circuit can initialize the synthesis conditions by making an initial estimate of a product having desired characteristics based on values ​​such as energy estimated by Neural Network Potential (NNP).

[0077] The processing circuitry acquires information about the product (synthesis result) based on the synthesis conditions (S204), and calculates the objective variable (S206).

[0078] The processing circuit determines whether the value of the objective variable satisfies the condition, for example, whether the value is sufficiently large (S208). If a sufficient objective variable has been obtained (S208: YES), the processing circuit outputs the result (S212) and can complete the processing. The output result is, for example, an explanatory variable (synthesis condition) that has obtained a good objective variable. For example, if an evaluation value higher than a predetermined evaluation value is obtained, the processing circuit determines that a sufficient objective variable has been obtained, that is, can complete the search.

[0079] If the value of the objective variable does not satisfy the condition (S208: NO), the processing circuit updates the synthesis conditions (S210) and repeats the search process. The synthesis conditions are updated as appropriate using an optimization technique, and can be achieved by updating explanatory variables based on Bayesian optimization, for example.

[0080] As described above, according to this embodiment, it is possible to realize a more accurate search with fewer trials. The synthesis method obtained by this search can provide an optimal dropping method and an optimal end timing.

[0081] <Operation as an information processing system>

[0082] The information processing system 1 can perform processing by combining the first device 12 and the second device 14. For example, the information processing system 1 uses the second device 14 to perform a search for synthesis conditions for a target having desired characteristics in order to obtain synthesis conditions having desired characteristics.

[0083] The first device 12 can optimize the second model into a more accurate model by feeding back the characteristics of the product to the function f. This feedback can be realized by any method.

[0084] As described above, according to the information processing system 1 of one embodiment, it is possible to automatically realize a more accurate search with desired characteristics, and to propose a process that meets the required specifications.

[0085] 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 configured as software information processing, 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 communication 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.

[0086] 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 provided inside the computer (such as a main storage device or auxiliary storage device) or outside the computer.

[0087] 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.

[0088] 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 realize its functions. 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.

[0089] 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 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.

[0090] 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.

[0091] 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.

[0092] 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.

[0093] 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.

[0094] 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. Also, at least one storage device may be connected to one processor. Also, a configuration in which at least one processor among multiple processors is connected to at least one storage device among multiple storage devices may be included. Also, this configuration may 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 be included.

[0095] 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.

[0096] The device interface 75 is an interface such as USB that directly connects to the external device 9B.

[0097] 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.

[0098] 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.

[0099] 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.

[0100] 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 device such as an HDD.

[0101] 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.

[0102] 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, partial deletions, etc. 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.

[0103] (1) A storage unit and a processing circuit, The processing circuitry For one or more samples for which synthesis conditions have been set, time series data of experimental values ​​for the synthesis conditions and one or more indexes corresponding to the synthesis conditions is acquired; inputting the acquired synthesis conditions into a first model that acquires calculation data of time series data for the one or more indexes when the synthesis conditions are input, and acquiring time series data of the calculation values; evaluating the time series data of the experimental values ​​and the time series data of the calculated values ​​using an evaluation function; Using the evaluation function, fitting a parameter related to the likelihood of reaction occurring; updating the evaluation function based on the fitting result; The evaluation function includes at least a distance-based error and another error. Information processing device.

[0104] (2) The at least distance-based error may be: Time series data of the experimental values; time-series data of the calculated values; is the sum of squared errors of The information processing device described in (1).

[0105] (3) The other errors are: a logarithm including components of the time series data of the experimental values; a logarithm including components of the time series data of the calculated value; is the sum of squared errors of (2) An information processing device according to the present invention.

[0106] (4) The other errors are: a ratio of a time point at which a component of the time series data of the experimental value is present to a predetermined time point; a ratio of a time when a component of the time series data of the calculated value is present to a predetermined time; is the absolute value of the difference between An information processing device according to (2) or (3).

[0107] (5) The evaluation function is the sum of squared errors; each of said other errors; and is a function obtained by weighting and adding The processing circuitry updating a weighting coefficient for each error to optimize the evaluation function; An information processing device according to any one of (2) to (4).

[0108] (6) The processing circuitry searching for a parameter related to the likelihood of the reaction occurring under the synthesis conditions based on each error constituting the evaluation function; obtaining the minimum value of each of the errors; updating the weighting coefficients so that a ratio between a minimum value and a maximum value of a product of the minimum value of each error and the corresponding weighting coefficient is equal to or less than a predetermined value; (5) An information processing device according to the present invention.

[0109] (7) The processing circuitry Searching for parameters related to the likelihood of the reaction occurring by Bayesian optimization, genetic algorithm, simulated annealing, gradient approximation, or a combination thereof; An information processing device according to any one of (1) to (6).

[0110] (8) The parameter relating to the likelihood of the reaction includes at least one of a reaction rate constant or an activation energy. An information processing device according to any one of (1) to (7).

[0111] (9) A storage unit and a processing circuit, The processing circuitry Using the evaluation function optimized by any one of the information processing devices (1) to (8), a synthesis result based on the synthesis conditions is obtained. Information processing device.

[0112] (10) The processing circuit For one or more samples for which synthesis conditions have been set, time series data of experimental values ​​for the synthesis conditions and one or more indexes corresponding to the synthesis conditions is acquired; inputting the acquired synthesis conditions into a first model that acquires calculation data of time series data for the one or more indexes when the synthesis conditions are input, and acquiring time series data of the calculation values; evaluating the time series data of the experimental values ​​and the time series data of the calculated values ​​using an evaluation function; Using the evaluation function, fitting a parameter related to the likelihood of reaction occurring; updating the evaluation function based on the fitting result; An information processing method, comprising: The information processing method, wherein the evaluation function includes at least a distance-based error and another error.

[0113] (11) The processing circuit For one or more samples for which synthesis conditions have been set, time series data of experimental values ​​for the synthesis conditions and one or more indexes corresponding to the synthesis conditions is acquired; inputting the acquired synthesis conditions into a first model that acquires calculation data of time series data for the one or more indexes when the synthesis conditions are input, and acquiring time series data of the calculation values; evaluating the time series data of the experimental values ​​and the time series data of the calculated values ​​using an evaluation function; Using the evaluation function, fitting a parameter related to the likelihood of reaction occurring; updating the evaluation function based on the fitting result; A program for executing an information processing method, The evaluation function includes at least a distance-based error and another error. program. [Explanation of symbols]

[0114] 1: Information processing system, 10: Information processing device, 100: I / F, 102: Memory section, 104: Processing unit, 12: first device, 14: Second device, 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 storage unit and a processing circuit, The processing circuitry For one or more samples for which synthesis conditions have been set, time series data of experimental values ​​for the synthesis conditions and one or more indexes corresponding to the synthesis conditions is acquired; inputting the acquired synthesis conditions into a first model that acquires calculation data of time series data for the one or more indexes when the synthesis conditions are input, and acquiring time series data of the calculation values; evaluating the time series data of the experimental values ​​and the time series data of the calculated values ​​using an evaluation function; Using the evaluation function, fitting a parameter related to the likelihood of reaction occurring; updating the evaluation function based on the fitting result; The evaluation function includes at least a distance-based error and another error. Information processing device.

2. The at least distance-based error may be: Time series data of the experimental values; time-series data of the calculated values; is the sum of squared errors of The information processing device according to claim 1.

3. The other errors are: a logarithm including components of the time series data of the experimental values; a logarithm including components of the time series data of the calculated value; is the sum of squared errors of The information processing device according to claim 2.

4. The other errors are: a ratio of a time point at which a component of the time series data of the experimental value is present to a predetermined time point; a ratio of a time when a component of the time series data of the calculated value is present to a predetermined time; is the absolute value of the difference between The information processing device according to claim 2.

5. The evaluation function is the sum of squared errors; each of said other errors; and is a function obtained by weighting and adding The processing circuitry updating a weighting coefficient for each error to optimize the evaluation function; The information processing device according to claim 2.

6. The processing circuitry searching for a parameter related to the likelihood of the reaction occurring under the synthesis conditions based on each error constituting the evaluation function; obtaining the minimum value of each of the errors; updating the weighting coefficients so that a ratio between a minimum value and a maximum value of a product of the minimum value of each error and the corresponding weighting coefficient is equal to or less than a predetermined value; The information processing device according to claim 5.

7. The processing circuitry Searching for parameters related to the likelihood of the reaction occurring by Bayesian optimization, genetic algorithm, simulated annealing, gradient approximation, or a combination thereof; The information processing device according to claim 1.

8. The parameter relating to the likelihood of the reaction includes at least one of a reaction rate constant or an activation energy. The information processing device according to claim 1.

9. A storage unit and a processing circuit, The processing circuitry Using the evaluation function optimized by the information processing device of any one of claims 1 to 8, a synthesis result based on synthesis conditions is obtained. Information processing device.

10. The processing circuit For one or more samples for which synthesis conditions have been set, time series data of experimental values ​​for the synthesis conditions and one or more indexes corresponding to the synthesis conditions is acquired; inputting the acquired synthesis conditions into a first model that acquires calculation data of time series data for the one or more indexes when the synthesis conditions are input, and acquiring time series data of the calculation values; evaluating the time series data of the experimental values ​​and the time series data of the calculated values ​​using an evaluation function; Using the evaluation function, fitting a parameter related to the likelihood of reaction occurring; updating the evaluation function based on the fitting result; An information processing method, comprising: The evaluation function includes at least a distance-based error and another error. Information processing methods.

11. The processing circuit For one or more samples for which synthesis conditions have been set, time series data of experimental values ​​for the synthesis conditions and one or more indexes corresponding to the synthesis conditions is acquired; inputting the acquired synthesis conditions into a first model that acquires calculation data of time series data for the one or more indexes when the synthesis conditions are input, and acquiring time series data of the calculation values; evaluating the time series data of the experimental values ​​and the time series data of the calculated values ​​using an evaluation function; Using the evaluation function, fitting a parameter related to the likelihood of reaction occurring; updating the evaluation function based on the fitting result; A program for executing an information processing method, The evaluation function includes at least a distance-based error and another error. program.

Images