Information processing device, information processing method, and program

Abstract

Regarding the claims of the patent document, information on the numerical range representing the characteristics of the substance is obtained. [Solution] An information processing device comprising: a prompt input unit that inputs information including at least one claim of one or more patent documents as a prompt to a generating AI device; a feature data acquisition unit that acquires feature data including a numerical range and units of one or more features of a substance relating to the claim from the generating AI device in response to the prompt; and a storage processing unit that stores the feature data acquired by the feature data acquisition unit, or data based on the feature data, in a storage device as storage data.

Description

[[Technical Field]]

[0001] The present disclosure relates to an information processing apparatus, an information processing method, and a program. [[Background Art]]

[0002] The advent of Large Language Models (LLM) has brought about major changes in the utilization of technical bibliographic information. Various services utilizing large language models have also emerged in intellectual property and patent-related services.

[0003] In the information processing apparatus described in Patent Document 1, a prompt is input to a large language model to extract an extraction target from a plurality of file data, and unstructured data is structured using the prompt (see Patent Document 1). [[Prior Art Documents]] [[Patent Documents]]

[0004] [[Patent Document 1]] Japanese Patent No. 7468960 [[Summary of the Invention]] [[Problem to be Solved by the Invention]]

[0005] However, with conventional techniques, there have been cases where sufficient information on numerical ranges representing characteristics of substances cannot be acquired with respect to claims of patent documents. With conventional natural language processing techniques, it has sometimes been difficult to automate extraction of numerical ranges representing characteristics of substances in claims.

[0006] The present disclosure has been made in consideration of such circumstances, and an object thereof is to provide an information processing apparatus, an information processing method, and a program capable of acquiring information on numerical ranges representing characteristics of substances with respect to claims of patent documents. [[Means for Solving the Problem]]

[0007] One aspect of the present disclosure is an information processing device comprising: a prompt input unit that inputs information including at least one claim of one or more patent documents as a prompt to a generating AI device; a feature data acquisition unit that acquires feature data including a numerical range and units of one or more features of a substance relating to the claim from the generating AI device in response to the prompt; and a storage processing unit that stores the feature data acquired by the feature data acquisition unit, or data based on the feature data, as storage data in a storage device.

[0008] One aspect of the present disclosure is an information processing method comprising: a prompt input process in which an information processing device inputs information including at least one claim of one or more patent documents as a prompt to a generating AI device; a feature data acquisition process in which the generating AI device acquires feature data including a numerical range and units of one or more features of a substance relating to the claim in response to the prompt; and a storage process in which the feature data acquired by the feature data acquisition process, or data based on the feature data, is stored in a storage device as storage data.

[0009] One aspect of this disclosure is a program for a computer to implement a prompt input function that inputs information including at least one claim of one or more patent documents as a prompt to a generating AI device; a feature data acquisition function that acquires feature data including a numerical range and units of one or more features of a substance relating to the claim from the generating AI device in response to the prompt; and a storage function that stores the feature data acquired by the feature data acquisition function, or data based on the feature data, as storage data in a storage device. [Effects of the Invention]

[0010] According to the information processing device, information processing device, information processing method, and program relating to this disclosure, it is possible to obtain information representing a numerical range of characteristics of a substance with respect to the claims of the patent document. [Brief explanation of the drawing]

[0011] [Figure 1] FIG. 1 is a diagram showing a configuration example of an information processing system according to an embodiment. [Figure 2] FIG. 2 is a diagram showing a configuration example of functional blocks of an information processing apparatus according to an embodiment. [Figure 3] FIG. 3 is a diagram showing a hardware configuration example of an information processing unit according to an embodiment. [Figure 4] FIG. 4 is a diagram showing a configuration example of a structuring processing unit according to an embodiment. [Figure 5] FIG. 5 is a diagram showing a configuration example of an applicability determination processing unit according to an embodiment. [Figure 6] FIG. 6 is a diagram showing a configuration example of a new substance proposal processing unit according to an embodiment. [Figure 7] FIG. 7 is a diagram schematically showing an example of a series of processes according to an embodiment. [Figure 8A] FIG. 8 is a diagram showing an example of results of cluster analysis according to an embodiment. [Figure 8B] FIG. 9 is a diagram showing an example of results of applicability determination according to an embodiment. [Figure 9A] FIG. 10 is a diagram showing an example of results of density calculation according to an embodiment. [Figure 9B] FIG. 11 is a diagram showing an example of results of refractive index calculation according to an embodiment. [Figure 10A] FIG. 12 is a diagram showing an example of a distribution of density and refractive index according to an embodiment. [Figure 10B] FIG. 13 is a diagram showing an example of results of liquidus temperature calculation according to an embodiment. [Figure 11A] FIG. 14 is a diagram showing an example of input and output of a learning model during learning according to an embodiment. [Figure 11B] FIG. 15 is a diagram showing an example of input and output of a trained learning model according to an embodiment. [Figure 11C] FIG. 16 is a diagram showing an example of a learning device and a determination device according to an embodiment. DETAILED DESCRIPTION OF THE INVENTION

[0012] Hereinafter, embodiments of the present disclosure will be described with reference to the drawings.

[0013] [Information processing system] Figure 1 is a diagram showing a configuration example of the information processing system 1 according to the embodiment. The information processing system 1 includes an information processing apparatus 11, a database 12, a generative AI apparatus 21, a patent document storage apparatus 22, and a network 31.

[0014] The information processing apparatus 11, the database 12, the generative AI apparatus 21, and the patent document storage apparatus 22 are each connected to the network 31. The information processing apparatus 11 can communicate with each of the database 12, the generative AI apparatus 21, and the patent document storage apparatus 22 via the network 31.

[0015] Here, the network 31 may be any network, and may be, for example, the Internet. Further, the network 31 may include, for example, a plurality of types of networks. Further, the network 31 may include a wired network, or may include a wireless network, or may include both of these networks.

[0016] <Generative AI Apparatus> The generative AI apparatus 21 has a generative AI function. In the present embodiment, generative AI using a large language model is used as the generative AI. Here, as the generative AI, for example, generative AI using a natural language model is used, but any generative AI may be used as long as processing similar to that of the present embodiment can be realized, and if a model with a different name is developed in the future, that model may be used. As the generative AI apparatus 21 having a generative AI function, for example, a server (e.g., a general-purpose server) for a generative AI service provided by any entity (e.g., a business operator) may be used, or an apparatus specially prepared for implementing the technology according to the present embodiment may be used.

[0017] <Patent document storage device> The patent document storage device 22 stores information from patent documents. The patent document storage device 22 stores, for example, information on multiple patent documents that have been made public in the past. The patent document storage device 22 may, for example, have a function to provide a search service for the stored patent documents. As the patent document storage device 22, for example, a server for a patent document provision service provided by any party (for example, a business operator, etc.) (for example, a general-purpose server) may be used, or a device specifically prepared for realizing the technology according to this embodiment may be used.

[0018] <database> Database 12 stores information. In this embodiment, the database 12 stores information used by the information processing device 11. In this embodiment, the database 12 is, for example, a relational database, but other databases may be used.

[0019] In this embodiment, the AI ​​generation device 21, the patent document storage device 22, and the database 12 are provided separately from the information processing device 11. However, as another example, one or more of the functions of the AI ​​generation device 21, the patent document storage device 22, and the database 12 may be provided in the information processing device 11. In this case, devices corresponding to the functions provided in the information processing device 11 (one or more of the AI ​​generation device 21, the patent document storage device 22, and the database 12) do not need to be provided.

[0020] <Information Processing Device> Figure 2 shows an example of the configuration of a functional block in the information processing device 11 according to the embodiment. The information processing device 11 is configured, for example, using a computer. The information processing device 11 includes an input unit 131, an output unit 132, a communication unit 133, a storage unit 134, a control unit 135, and a processing unit 136. The input unit 131 includes an operation unit 151. The output unit 132 includes a display unit 152. The processing unit 136 comprises a structuring processing unit 171, a matching determination processing unit 172, and a new material proposal processing unit 173.

[0021] The input unit 131 receives information from an external source. The input unit 131 includes, for example, an operation unit 151 that receives operations performed by the user, and inputs information corresponding to the operations received by the operation unit 151. The operation unit 151 may have, for example, a function to accept operation from a touch panel, or it may have a function to accept operation from physical keys. The operation unit 151 may also have a function to accept operation by voice (for example, the user's voice). Furthermore, the input unit 131 may be connected to an external device and receive information output from that external device. The external device may be, for example, a portable recording medium. The recording medium may also be called, for example, a storage medium.

[0022] The output unit 132 outputs information. The output unit 132, for example, has a display unit 152, which displays (outputs) information on the screen of the display unit 152. This screen may have touch panel functionality. Furthermore, the output unit 132 may be connected to an external device and output information to that external device. The external device may be, for example, a portable recording medium. The output unit 132 may output information in a manner other than display, such as outputting audio.

[0023] In this embodiment, the input unit 131 and the output unit 132 are shown as separate functional units, but the input unit 131 and the output unit 132 may be configured as a common functional unit (input / output unit). This input / output unit may be composed of a touch panel.

[0024] The communication unit 133 has the function of communicating information with external devices. This communication may be wired or wireless. In this embodiment, the communication unit 133 is shown as a separate functional unit from the input unit 131 and the output unit 132. However, the receiving function of the communication unit 133 may be considered as being included in the functions of the input unit 131, and the transmitting function of the communication unit 133 may be considered as being included in the functions of the output unit 132.

[0025] The memory unit 134 stores information. The memory unit 134 may store any information. Information may also be called data.

[0026] The control unit 135 performs various types of control. In this embodiment, the control unit 135 is equipped with a processor such as a CPU (Central Processing Unit), and performs the control defined in a predetermined program (control program) by executing the program using the processor. The program may be stored in, for example, the storage unit 134.

[0027] The processing unit 136 performs various processes. In this embodiment, for the sake of explanation, the processing unit 136 is shown to include a structuring processing unit 171, a determination processing unit 172, and a new material proposal processing unit 173, but each of these processes does not necessarily have to be clearly distinguished. Furthermore, the processing unit 136 may perform other processes.

[0028] In this embodiment, for the sake of explanation, the control unit 135 and the processing unit 136 are shown separately. However, for example, it may be understood that various processes are realized by performing various controls by the control unit 135, and the control unit 135 and the processing unit 136 may be considered as a single unit.

[0029] <Example of hardware configuration for information processing equipment> Figure 3 shows an example of the hardware configuration of the information processing unit 1001 according to the embodiment. The configuration of the information processing unit 1001 shown in Figure 3 may also be applied to the information processing unit 11 shown in Figure 2. In this case, the functions of the information processing unit 1001 shown in Figure 3 will realize the functions of the information processing unit 11. For example, the configuration of the information processing unit 1001 or a part of the configuration of the information processing unit 1001 may be used in any part of the information processing unit 11.

[0030] In the example shown in Figure 3, the information processing unit 1001 includes a processor 1011, an operating device 1012, a display device 1013, a storage device 1014, a memory 1015, an input / output interface 1016, a network interface 1017, and a bus 1021 connecting these. Furthermore, when the information processing unit 1001 is applied to the information processing device 11, it is not necessarily required that all the processing units shown in Figure 3 be provided, and any additional processing units not shown in Figure 3 may be added.

[0031] The processor 1011 consists of a CPU and other components, and executes the control and processing defined in a program by running that program. The operating device 1012 is equipped with one or more input devices such as a keyboard and a mouse, and accepts operations performed by a person (user). The display device 1013 has a screen and outputs information to that screen.

[0032] The storage device 1014 is a non-volatile storage unit, such as a hard disk, and stores information. Memory 1015 is a volatile storage unit, consisting of RAM (Random Access Memory) or the like, and temporarily stores information. For example, DRAM (Dynamic Random Access Memory) may be used as the RAM. The storage device 1014 or memory 1015 may store, for example, information about a program executed by the processor 1011.

[0033] The input / output interface 1016 is an interface for connecting to external recording media and other devices. Network interface 1017 is an interface that connects to an external network.

[0034] Here, the information processing unit 1001 may have one processor as the processor 1011, or it may have two or more processors. For example, the information processing unit 1001 may have multiple CPUs, with each CPU performing its own control and processing, and these multiple CPUs cooperating to achieve overall control and processing.

[0035] <Structured Processing> Figure 4 shows an example of the configuration of the structured processing unit 171 according to the embodiment. The structured processing unit 171 comprises a patent group collection unit 211, a prompt input unit 212, a structured data acquisition unit 213, and a structured data storage processing unit 214.

[0036] In this embodiment, processing is performed in the following order: patent group collection unit 211, prompt input unit 212, structured data acquisition unit 213, and structured data storage processing unit 214. This processing is just one example and is not necessarily limited to this example; various other processing methods may be used. Furthermore, the configurations of the patent group collection unit 211, the prompt input unit 212, the structured data acquisition unit 213, and the structured data storage processing unit 214 are merely examples and are not necessarily limited to this example; various other configurations may be used.

[0037] Furthermore, some or all of these processes may be performed, for example, in response to user operations, or automatically by the information processing device 11. Furthermore, some or all of these processes may be performed using, for example, a machine learning model. For example, the prompt input unit 212 may use a predetermined machine learning model when generating prompts. Also, for example, the structured data storage processing unit 214 may use a predetermined machine learning model in the processing when storing structured data after processing it rather than storing it as is. However, it is not always necessary to use a configuration that utilizes a machine learning model in this way.

[0038] The patent group collection unit 211 collects groups of patent documents (also referred to as patent groups for convenience of explanation). Here, the patent group may include, for example, multiple patent documents, but in this embodiment, it may also include only one patent document.

[0039] In this embodiment, the patent collection unit 211 accesses the patent document storage device 22 via the communication unit 133, retrieves (for example, downloads) one or more patent documents stored in the patent document storage device 22, and stores the retrieved patent documents in the storage unit 134 so that they can be included in the patent group. In this case, for example, the user may operate the operation unit 151, or the information processing device 11 may automatically use the search function of the patent document storage device 22 to search for the desired patent document.

[0040] As another example, the group of patents may be stored in the storage unit 134 in advance, in which case, for example, the process of collecting patent documents by the patent group collection unit 211 may be omitted. Each patent document is structured, for example, as text data or image data.

[0041] Here, various documents may be used as patent documents. For example, a patent gazette containing claims that have already been granted patents may be used, or a published patent gazette containing claims whose patentability is yet to be determined may be used, or both may be used. As another example, patent application documents may be used as patent documents. Furthermore, the claims may be, for example, individual claims within the scope of the patent claims, or the group of all claims included in the scope of the patent claims (for convenience of explanation, also called the group of claims) may be used. In other words, the claims may be, for example, the entire scope of the patent claims (i.e., all claims) may be treated as a processing unit, or the individual claims included in the scope of the patent claims (i.e., each claim) may be treated as a processing unit. Here, the group of claims may include, for example, multiple claims, but in this embodiment, it may also include only one claim.

[0042] The prompt input unit 212 processes information including at least one claim from one or more patent documents as a prompt and inputs it to the generation AI device 21. In this embodiment, the prompt input unit 212 inputs the prompt to the generation AI device 21 by transmitting the prompt to the generation AI device 21 via the communication unit 133. In this embodiment, the prompt includes content that specifies what information to request from the generating AI device 21. In this embodiment, the prompt includes content that requests predetermined structured data relating to a claim.

[0043] The structured data acquisition unit 213 (an example of a feature data acquisition unit) acquires predetermined structured data (an example of feature data) from the generating AI device 21 in response to a prompt. In other words, in this embodiment, the generation AI device 21 generates the structured data in response to a prompt received from the information processing device 11 and provides it to the information processing device 11.

[0044] Here, structured data is data that includes the numerical range and units of one or more characteristics of the substance relating to the claim. In other words, structured data includes information that represents a numerical range with units for one or more characteristics of the substance. The feature in question may be any feature of the substance, for example, a feature relating to any physical property of the substance, or a feature relating to the composition of the substance.

[0045] Structured data may be generated, for example, corresponding to the content of one claim in one patent document, or it may be generated corresponding to the content of two or more claims in one patent document (for example, the entire scope of the claims). Alternatively, structured data may be generated, for example, in relation to a combination of one or more claims from multiple patent documents (or, for example, a combination of the entire contents of the claims from each of the multiple patent documents).

[0046] As a result, structured data represents the patent rights, taking into account all the content of one or more claims used in generating the structured data, as expressed in the numerical range and units of one or more features. Furthermore, the patent rights may include not only rights that have already been granted a patent, but also rights related to claims whose patentability is yet to be determined.

[0047] A substance may be composed of, for example, one component, or it may be composed of a combination of multiple components. In other words, the substance may be a mixture of multiple components. Examples of mixtures of multiple components include glass, ceramics, and blended polymers. Furthermore, the components may be specified, for example, in units of material, or in units of other materials.

[0048] As a specific example, characteristics related to the physical properties of a substance may include, for instance, thermal properties, electrical properties, magnetic properties, optical properties, or mechanical properties, as well as density, refractive index, liquidus temperature, etc. As a specific example, the characteristics related to the composition of a substance may be, for example, the characteristics of each component, or, for two or more components, the ratio of the components, or the sum of the components, etc. For example, the unit may be a percentage (%) or a gram (g). If the substance is glass, for example, one or more of the following may be used: milligrams (mg) or grams (g), mass percentage (mass%), mole percentage (mol%), or cation percentage (cation%). As a concrete example, structured data may be data presented as a table in a relational database, or it may be tabular file data, or even CSV file data. Numerical ranges may be expressed using, for example, an equals sign (=) or an inequality sign (<, >, ≦, ≧).

[0049] The structured data storage processing unit 214 stores the structured data acquired by the structured data acquisition unit 213 in the database 12 as storage data. The structured data storage processing unit 214 may, for example, store the structured data acquired by the structured data acquisition unit 213 directly in the storage unit 134, or it may store data based on the structured data in the storage unit 134. The data based on the structured data is, for example, data obtained by processing the structured data.

[0050] Here, a specific example of the processing performed by the structured processing unit 171 is shown. Note that this processing is just one example and is not necessarily limited to this processing.

[0051] The patent group collection unit 211 collects the patent group to be used in this case using a patent search system, which is an example of a patent document storage device 22.

[0052] The prompt input unit 212 inputs prompts to the generation AI device 21 for extracting desired structured data. In this example, the prompt includes the group of patents to be used. The prompt may include, for example, the entire group of patents (one or more patent documents), or it may include at least the content of the claims to be used from the group of patents (one or more patent documents).

[0053] The structured data acquisition unit 213 acquires the structured data generated by the generation AI device 21 in response to the prompt. Here, the generating AI device 21 reads one or more claims of the patent group included in the prompt, extracts the patent scope of the substance defined by the entirety of the claims, using the numerical range and units of one or more features of the substance, and generates structured data in which the extraction results are structured. In this embodiment, the content of the prompt is described so that such structured data is generated by the generating AI device 21. In this type of structured data, the feature name, numerical value, range of the numerical value, and unit of the numerical value are structured. These may be output together with the patent number and claim number.

[0054] The structured data storage processing unit 214 stores the acquired structured data by storing it in the database 12 as storage data.

[0055] <Applicability Determination Processing Unit> Figure 5 shows an example of the configuration of the non-determination processing unit 172 according to the embodiment. The determination processing unit 172 comprises a determination target information acquisition unit 231, a database reference unit 232, a determination unit 233, and a determination result output unit 234.

[0056] In this embodiment, processing is performed in the following order: the unit for acquiring information to be judged 231, the database reference unit 232, the judgment unit 233, and the judgment result output unit 234. This processing is just one example and is not necessarily limited to this example; various other processing methods may be used. Furthermore, the configuration of the judgment target information acquisition unit 231, database reference unit 232, judgment unit 233, and judgment result output unit 234 is just one example and is not necessarily limited to this example; various other configurations may be used.

[0057] Furthermore, some or all of these processes may be performed, for example, in response to user operations, or automatically by the information processing device 11. Furthermore, some or all of such processing may be performed using, for example, machine learning models.

[0058] The determination target information acquisition unit 231 acquires information about the substance to be determined (for convenience of explanation, also referred to as the determination target substance) (for convenience of explanation, also referred to as the determination target information). The information to be judged includes numerical values ​​for one or more characteristics of the substance to be judged. Here, the substance to be determined may be specified, for example, by the user operating the control unit 151, or it may be set in advance.

[0059] The database reference unit 232 references the storage data stored in the database 12. For example, if the data to be stored is already stored in the storage unit 134, the database reference unit 232 does not need to refer to the database 12.

[0060] The determination unit 233 determines whether the numerical value of a predetermined characteristic of the substance to be determined (the numerical value of one or more characteristics to be determined) falls within the range of determination values ​​defined by the stored data. In this embodiment, if the numerical value of a predetermined characteristic of the substance to be judged (the numerical value of one or more characteristics to be judged) falls within the numerical range defined by the storage data, the numerical range of the substance to be judged is determined to fall within the numerical range. On the other hand, in this embodiment, if the numerical value of a predetermined characteristic of the substance to be judged (the numerical value of one or more characteristics to be judged) does not fall within the numerical range defined by the storage data, the numerical range of the substance to be judged is determined not to fall within the numerical range (i.e., it is not applicable). Note that one or more numerical ranges are defined for each characteristic. Here, the numerical range for determination is defined by structured data equivalent to memory data, and is the numerical range of one or more characteristics of the substance related to said structured data, with units attached to the numerical values. However, in the case of dimensionalless characteristics, units may not be attached.

[0061] Furthermore, when numerical values ​​for multiple characteristics of the substance to be judged are used, for example, if the numerical values ​​for all of these characteristics fall within the numerical range for judgment, the numerical range of the substance to be judged will be determined to fall within the numerical range for judgment. In this case, if the numerical value for at least one of these characteristics does not fall within the numerical range for judgment, the numerical range of the substance to be judged will be determined not to fall within the numerical range for judgment.

[0062] Here, the numerical range used for judgment regarding multiple features can be considered as a combination of the numerical ranges for each of these features (the numerical ranges used for judgment for each feature). As a concrete example, the numerical range for determining features A and B, which are multiple features, can be understood as a combination of the numerical range for feature A (numerical range for determination) and the numerical range for feature B (numerical range for determination). In other words, being in the numerical range for determining features A and B, which are multiple features, is equivalent to being in the numerical range for feature A (numerical range for determination) and also in the numerical range for feature B (numerical range for determination). On the other hand, if a value does not fall under the numerical range for feature A (numerical range for determination), or does not fall under the numerical range for feature B (numerical range for determination), or does not fall under either of these, then it will not be in the numerical range for determining features A and B, which are multiple features.

[0063] The judgment result output unit 234 outputs the judgment result from the judgment unit 233. In this embodiment, the judgment result output unit 234 outputs the judgment result information to the screen via the display unit 152. The user can understand the judgment result by viewing the information displayed on the screen. The reason for the judgment may also be output by the judgment result output unit along with the judgment result. The method of outputting the judgment result may include, for example, outputting it by printing, or outputting it by sound (voice).

[0064] Here, when the determination unit 233 makes a determination of whether or not a material is classified based on the storage data stored in the database 12, it may also use, for example, materials informatics (MI) processing, which utilizes information science such as AI in materials development, or a predetermined simulation process. For example, when there is insufficient information on the features necessary for determining whether or not a material is classified, numerical values ​​predicted by MI or simulation may be used.

[0065] As a specific example, the determination unit 233 may determine that the substance to be determined corresponds to a set of features in structured data corresponding to the said storage data if the numerical values ​​of some of the features of the substance to be determined (the numerical values ​​of one or more features to be determined) are within the range of numerical values ​​for determination defined by the storage data, and the other features are in accordance with the MI processing results. Here, the MI processing may also involve predicting numerical values ​​of the characteristics necessary for determination based on the composition ratio and physical properties.

[0066] Similarly, the determination unit 233 may determine that the substance to be determined corresponds to a set of features in structured data corresponding to the said storage data if the numerical values ​​of some of the features of the substance to be determined (the numerical values ​​of one or more features to be determined) are within the range of numerical values ​​for determination defined by the storage data, and the other features also match the processing results of the simulation. Here, the prescribed simulation process involves outputting characteristic values ​​of the substance to be evaluated based on its structure and component ratios, using theoretically-based numerical analysis.

[0067] Furthermore, various types of simulations may be performed. For example, a simulation performing quantum computation may be performed, a molecular dynamics simulation based on Newtonian mechanics may be performed, or a simulation using a machine learning simulator that mimics quantum computation using machine learning may be performed. For example, quantum computation may be performed after a molecular dynamics simulation based on Newtonian mechanics, or a simulation using a machine learning simulator that mimics quantum computation using machine learning. In a machine learning simulator, for example, it is possible to create information about the structure of a substance (or the structure of a material if the substance is a material) through simulation.

[0068] Thus, when the determination unit 233 determines whether the substance to be determined meets the requirements for storage data (requirements for structured data equivalent to said storage data), it may use MI or a predetermined simulation to determine whether some or all of the requirements are also met, and obtain a comprehensive determination result based on these. Furthermore, MI and the specified simulation may be performed either individually or both simultaneously.

[0069] Here, a specific example of the processing performed by the matching determination processing unit 172 is shown. Note that this processing is just one example and is not necessarily limited to this processing.

[0070] The judgment target information acquisition unit 231 acquires numerical values ​​and other information necessary for the judgment. The numerical value may, for example, include predetermined physical property values, or it may include predetermined ratios or sums of components. The numerical value in question may be, for example, a single point value, or it may be a numerical value with a range (a numerical range). One or more numerical ranges may be defined. Furthermore, units are included with the numerical values ​​if they are defined.

[0071] The database reference unit 232 accesses the storage data (essentially structured data) stored in the database 12. Here, the database reference unit 232 refers to the storage data used as the criterion for determining whether or not a substance is applicable (that is, the storage data used to determine whether or not the substance to be judged is applicable).

[0072] The determination unit 233 compares the substance to be determined with the storage data (essentially structured data) in the database 12 to determine whether or not it falls within the scope of the structured data (the numerical range and units of one or more characteristics defined by the structured data). In this case, the determination unit 233 may, for example, further utilize MI processing or predetermined simulation processing.

[0073] The determination result output unit 234 outputs the result of the determination by the determination unit 233.

[0074] <New Material Proposal Processing Section> Figure 6 shows an example of the configuration of the new material proposal processing unit 173 according to the embodiment. The new substance proposal processing unit 173 comprises a candidate group generation unit 251, a feature value calculation unit 252, a non-applicable detection unit 253, a validity verification processing unit 254, and a new substance proposal unit 255.

[0075] In this embodiment, processing is performed in the following order: candidate group generation unit 251, feature value calculation unit 252, non-applicable detection unit 253, validity verification processing unit 254, and new substance proposal unit 255. This processing is just one example and is not necessarily limited to this example; various other processing methods may be used. Furthermore, the configuration of the candidate group generation unit 251, feature value calculation unit 252, non-applicable detection unit 253, validity verification processing unit 254, and new material proposal unit 255 is just one example and is not necessarily limited to this example; various other configurations may be used.

[0076] Furthermore, some or all of these processes may be performed, for example, in response to user operations, or automatically by the information processing device 11. Furthermore, some or all of such processing may be performed using, for example, machine learning models or generative AI.

[0077] The candidate group generation unit 251 generates a group of candidate substances (for convenience of explanation, also called a candidate group). Here, the candidate group includes one or more candidate substances.

[0078] The feature value calculation unit 252 calculates predetermined feature values ​​for each candidate substance. The specified feature value is a numerical value for a specified feature (a numerical value for one or more features to be judged), and the numerical value is accompanied by a unit. Here, the feature value calculation unit 252 may calculate predetermined feature values ​​for each candidate substance by, for example, performing MI processing or predetermined simulation processing.

[0079] The non-applicable detection unit 253 detects from the candidate group any substance that does not fall within the numerical range for determination defined by the storage data that serves as the criterion for determining applicability (i.e., the numerical range for determination defined by the structured data corresponding to the storage data) (i.e., is not applicable). Here, the non-applicable detection unit 253, for example, determines whether the numerical value of a predetermined characteristic of each substance included in the candidate group (the numerical value of one or more characteristics to be judged) falls within the numerical range for judgment, and detects candidate substances that do not fall within the numerical range for judgment. Furthermore, the process for determining whether or not a product is classified may be the same as the process performed by the classification determination processing unit 172 shown in Figure 5.

[0080] The memory data used as the basis for determining whether or not an item is classified may be specified by the user through operation of the control unit 151, or it may be pre-set. As another example, the storage data used as the criterion for determining whether or not a data is applicable may be, for example, the storage data used in the processing immediately preceding the processing performed in the information processing device 11 (for example, the processing shown in Figure 4 or the processing shown in Figure 5).

[0081] The validation processing unit 254 performs a validation process for each of the candidate substances detected by the non-applicable detection unit 253. In this embodiment, the validation processing unit 254 determines, for each candidate substance detected by the non-applicable detection unit 253, whether it conforms to information that could actually exist (such as a numerical range) with respect to characteristics not defined in the numerical range for determination (e.g., physical properties or composition). If it determines that it conforms to information that could actually exist (such as a numerical range), it determines that the candidate substance is valid. In this way, the validation processing unit 254 confirms the validity of each candidate substance detected by the non-applicable detection unit 253. In addition, in such validation processes, for example, the MI process or a predetermined simulation process described in the assertion process performed by the assertion determination processing unit 172 shown in Figure 5 may be used.

[0082] The new material proposal unit 255 proposes candidate materials as new materials that have been confirmed (determined) as valid by the validation processing unit 254. If the candidate material is a material, it can propose a new material. In this embodiment, the new material proposal unit 255 displays information proposing a new material on the screen via the display unit 152. The user can understand the proposed new material by viewing the information displayed on the screen. Furthermore, the methods used to output information proposing new materials may include, for example, outputting by printing, or outputting by sound (voice).

[0083] In this way, the new substance proposal processing unit 173 can propose a substance to be proposed (for convenience of explanation, also called a proposed substance) from among candidate substances (examples of substances to be determined) that have been determined not to fall within the numerical range for determination by the function of determining whether or not a substance is a substance (in this embodiment, the function of the determination processing unit 172). In this case, the validation processing unit 254 can verify the numerical values ​​of one or more characteristics of the proposed substance using MI or a simulator (a functional unit that performs predetermined simulation processing).

[0084] The new material proposed in this way is a material that falls outside the numerical range for determination defined by structured data equivalent to the memory data that serves as the criterion for determining whether or not it is classified as a non-classifiable substance. Furthermore, in this embodiment, it is a material that is realistically likely to have other characteristics (e.g., physical properties, composition, etc.). Furthermore, MI and the specified simulation may be performed either individually or both simultaneously.

[0085] Here, we show a specific example of the processing performed by the new material proposal processing unit 173. Note that this processing is just one example and is not necessarily limited to this processing.

[0086] The candidate group generation unit 251 generates a candidate group containing a large number of possible substances. Here, the large number of possible substances may include, for example, substances having component ratios or sums of components of various materials, or substances having various physical property values.

[0087] The feature value calculation unit 252 calculates (for example, estimates) feature values ​​related to the components or physical properties of each substance included in the candidate group. Here, for example, MI processing or a predetermined simulation process may be used for this calculation.

[0088] The non-applicable detection unit 253, based on the information of the candidate group and the characteristic values ​​described above, compares it with the storage data (essentially structured data) constructed in the database 12 to detect new substances (candidate substances) that do not apply to the structured data.

[0089] The validation processing unit 254 calculates feature values ​​for the characteristics of the detected substance (for example, some of the various characteristics) and determines whether or not it is valid based on the calculation results. Here, for example, MI processing or a predetermined simulation process may be used for this calculation.

[0090] The New Material Proposal Unit 255 proposes as a new material a substance (candidate substance) that is deemed not to be applicable to the structured data used as the criteria for determining applicability and is also deemed to be valid.

[0091] As described above, the information processing device 11 according to this embodiment can acquire information on a numerical range representing the characteristics of a substance (structured data in this embodiment) with respect to the claims of the patent document. Furthermore, the information processing device 11 according to this embodiment can determine whether or not the substance to be judged falls within the scope of a predetermined patent based on the structured data. Furthermore, the information processing device 11 according to this embodiment can propose new materials that do not fall within the scope of a predetermined patent based on the structured data.

[0092] In this embodiment, the structured processing unit 171 can generate structured data that identifies the range and units of numerical values ​​related to substances extracted from patent documents. For example, it can significantly reduce the workload of simultaneously acquiring supplementary information (range and units) of numerical values ​​from a large number of patent documents. Furthermore, in this embodiment, for example, a reusable relational database can be maintained as a database for storing structured data.

[0093] In this embodiment, the processing of the matching determination processing unit 172 allows for accurate matching based on precise numerical ranges and units by referring to structured data, enabling efficient determination of matching against a large number of conditions, for example. Furthermore, in this embodiment, traceability can be ensured, for example, by keeping a record of the determination of whether or not the item is applicable.

[0094] In this embodiment, the new substance proposal processing unit 173 can propose non-applicable substances based on precise numerical ranges and units by referring to structured data, for example, by finding promising substances from a large number of candidate substances without experimentation. Furthermore, in this embodiment, the validity of the new material can be confirmed by, for example, using MI processing or a predetermined simulation process. While MI generally has weaknesses in extrapolation, these weaknesses can be compensated for by performing specific simulation processing.

[0095] In this embodiment, by devising instructions (prompts) for the generating AI (large-scale language model), it is possible to automatically structure (table) the compositional components, component formulas, maximum and minimum values ​​of physical properties, and their units with high accuracy, and to construct a database 12 (patent DB). Furthermore, in this embodiment, the constructed database 12 can be used to determine whether or not a material with a desired component ratio falls under an existing patent (non-applicability determination). For example, an algorithm can be constructed to automatically perform such determinations, and a determination system that uses this algorithm can be constructed. Furthermore, in this embodiment, by performing such a determination, it is possible to propose compositions that avoid patents (for example, patents related to the claims described in the patent claims). In this case, for example, screening can be performed quickly by using a machine learning model instead of experiments to estimate the target physical properties. Furthermore, in this embodiment, some of the material properties are determined by molecular dynamics simulation or quantum computation.

[0096] As a concrete example, inorganic materials such as glass or ceramics are defined by a range of components. In addition, there are many patent documents that define such inorganic materials within a range of desired physical properties. Therefore, it is effective to construct a database 12 that defines substances within these ranges, and for example, it is effective for automating the determination of whether a substance is applicable or not.

[0097] In this embodiment, the information processing device 11 is shown to have a function for processing MI. However, as another example, instead of having a function for processing MI, the information processing device 11 may access an external device (for example, a predetermined server) that has a function for processing MI and obtain the MI processing results by receiving the MI processing service provided by the external device. Similarly, in this embodiment, the information processing device 11 is shown to have a function for performing a predetermined simulation. However, in other examples, instead of having a function for performing a predetermined simulation, the information processing device 11 may access an external device (for example, a predetermined server) that has a function for performing a predetermined simulation and obtain the processing results of a predetermined simulation by receiving the predetermined simulation processing service provided by the external device.

[0098] [Example of a specific image] Referring to Figures 7, 8A and 8B, 9A and 9B, 10A and 10B, a concrete example of the process according to this embodiment is shown. These examples are just a few examples and are not necessarily the only ones that apply. Furthermore, the graphs schematically shown in these examples are not necessarily rigorous, but are merely illustrative examples for illustrative purposes.

[0099] <Example of a series of processes> Figure 7 is a schematic diagram illustrating an example of a series of processes according to the embodiment. Figure 7 illustrates an example where a novel high-refractive index glass is being searched for as a new material. Figure 7 shows the patent document content G1, the structured data content G2, the patent map G3, the MI / simulation processing unit G4, and the proposed content G5.

[0100] The content of the patent document G1 represents the content of the claims of a given patent document. In patents for inorganic materials, the inorganic material is often identified using a numerical range for its components. The patent document content G1 is included in the prompt and sent to the generating AI (in this embodiment, the generating AI device 21).

[0101] Structured data content G2 represents the content of the structured data generated by the generation AI. In this example, the structured data content G2 stores multiple components, associating each component's name, a note (note information) that identifies whether it represents the maximum or minimum value of a numerical range, the numerical value, and the unit (in this example, [mol%]).

[0102] Patent map G3 represents a patent map generated based on structured data. In this example, "A," "B," "C," "D," and "E" shown in patent map G3 each represent different applicants, and the scope of patents held by each applicant is clearly indicated. The patent map G3 may, for example, be generated by a generating AI and provided to the information processing device 11, or the information processing device 11 may generate the patent map G3 based on structured data provided by the generating AI.

[0103] The MI / simulation processing unit G4 assists in determining the validity of proposed candidate materials by performing MI processing or predetermined simulation processing. As an example, a materials simulation based on Density Functional Theory (DFT) may be performed.

[0104] Proposal G5 schematically shows proposed substance J1 within the same patent map as patent map G3. A user (in this example, a developer) can consider such a proposed substance J1 as a candidate for development.

[0105] <Example of how to create candidate compositions> For example, consider a case where there were 56 types of oxide components described in the claims of 892 patent documents, and after removing duplicates, there were 41 sets. While comprehensively searching for these items is a laborious task, the workload can be reduced by executing either "Search Pattern A" or "Search Pattern B" below.

[0106] In processing search pattern A, the combination of compositions in the claims is used as a reference. In this example, there are 41 unique combinations, and the information processing device 11 creates 1000 candidates for each of these combinations. Here, to ensure that non-applicable compositions are always included, for example, compositions are created that fall outside the numerical range of one or more features defined by the structured data. Specifically, for example, compositions may be generated using a numerical range obtained by multiplying the maximum value of a certain numerical range by 1.1 (or any other multiple greater than 1), or compositions where a certain numerical range is set to 0 to 100%.

[0107] In the processing of search pattern B, numerical values ​​related to the target component are determined in advance. For example, target components such as "X[P2O5,Nb2O5,TiO2]-(100-X)[Li2O,Na2O,K2O,MgO,CaO,SrO,BaO,Y2O3,ZrO2,WO3,ZnO,Al2O3,Ga2O3,In2O3,B2O3,SiO2,SnO2], X={10,20,...,90} may be predetermined. In the example above, the comma (,) represents AND / OR (AND or OR). Furthermore, in the example above, the variable X can take any of the following values: 10, 20, ..., or 90.

[0108] <Example of the result of determining whether a match or not using search pattern A> Using the compositional combinations described in the claims as a reference, cluster analysis was performed to search for non-applicable high-refractive-index glasses.

[0109] Figure 8A shows an example of the results of cluster analysis according to this embodiment. In the graph shown in Figure 8A (Analysis Result G11), the horizontal axis represents the 1st component, and the vertical axis represents the 2nd component. In this example, the results of projecting 41,000 compositions into two dimensions using UMAP (Uniform Manifold Approximation and Projection) are shown. In this example, the data is divided into 20 clusters.

[0110] Figure 8B shows an example of the results of the determination of whether or not the product / service is applicable according to the embodiment. In the graph shown in Figure 8B (Applicability Determination Result G12), the horizontal axis represents the 1st component, and the vertical axis represents the 2nd component. In this example, the relevant parts and the non-relevant parts are clearly indicated for identification. In this example, the number of non-applicable compositions was 2144.

[0111] <Density and refractive index> Using a machine learning model (e.g., GlassNet), we calculated (e.g., estimated) the density and refractive index of 2144 compositions that were not applicable. It was found that density and refractive index are strongly correlated, and that low density and high refractive index are fundamentally inverse relationships. Generally speaking, the reproducibility of experimental results for machine learning models in question (e.g., GlassNet) is very high. Here, GlassNet is a multi-task deep neural network capable of predicting the temperature dependence of 85 different glass properties and viscosity.

[0112] Figure 9A shows an example of the density calculation results according to the embodiment. In the graph shown in Figure 9A (density calculation result G21), the horizontal axis represents the 1st component, and the vertical axis represents the 2nd component.

[0113] Figure 9B shows an example of the results of the refractive index calculation according to this embodiment. In the graph shown in Figure 9B (refractive index calculation result G22), the horizontal axis represents the 1st component and the vertical axis represents the 2nd component.

[0114] <Example of creating a compositional structure> We created a glass structure through simulation and then performed quantum calculations. For example, a glass structure was created using a machine learning simulator, and the refractive index was calculated using quantum computing. Even using non-empirical methods, the refractive index can be estimated with high accuracy. Furthermore, accurate extrapolation is also possible.

[0115] <Example of narrowing down the composition> From the compositions that did not meet the criteria, we narrowed down the possibilities by density and refractive index, and further narrowed down the possibilities by liquidus temperature.

[0116] Figure 10A shows an example of the density and refractive index distribution according to the embodiment. In the graph shown in Figure 10A (density refractive index distribution G31), the horizontal axis represents the 1st component and the vertical axis represents the 2nd component. In this example, we narrowed down the compositions to those predicted to have {density < 4.5 [g / cm3] and refractive index > 2.0}. In the example in Figure 10A, some of these compositions are shown as compositions H1 to H5. Note that in the example in Figure 10A, other compositions may also exist (i.e., compositions H1 to H5), but their symbols are omitted from the illustration.

[0117] Figure 10B shows an example of the calculation results for the liquidus temperature according to the embodiment. In the graph shown in Figure 10B (liquid phase temperature calculation result G32), the horizontal axis represents the 1st component and the vertical axis represents the 2nd component.

[0118] In this example, we narrowed down the possibilities to 44 compositions based on density and refractive index, and then further narrowed it down to 26 compositions by excluding As, Sb, V, and Pb. Furthermore, the condition of {liquid phase temperature < 1500 [K]} narrowed the possibilities down to 11 compositions. Furthermore, searching for compositions containing three or more components narrows the options down to five.

[0119] [Explanation of machine learning learning models] In this embodiment, we have shown a case where a generation AI device 21 provided outside the information processing device 11 is used. However, in other examples, the information processing device 11 may have a generation AI function, in which case it is not necessary to use the external generation AI device 21.

[0120] Generative AI can be built, for example, based on deep learning techniques, machine learning techniques, or both. Here, we will explain the machine learning model by referring to Figures 11A, 11B, and 11C.

[0121] <Overview of machine learning training> Figure 11A shows an example of the input and output of the learning model C1 during training according to the embodiment. Here, the learning model C1 represents the state in which the learning model is undergoing machine learning. In general terms, the control unit that controls the learning process updates the parameters of the learning model C1 based on the output data when predetermined input data is input to the learning model C1. In this process, the control unit may perform supervised learning using training data. As another example, unsupervised learning may be performed without using training data. Note that the schematic shown in Figure 11A is a general illustration and is not necessarily rigorous. A training model that is currently being trained may be called, for example, a training model.

[0122] <Overview of decision-making using machine learning results> Figure 11B shows an example of the input and output of the trained model C2 according to the embodiment. Here, the learning model C2 represents the state in which the learning model has been trained. In general terms, the control unit that controls the decision (inference) obtains a predetermined decision result based on the output data when predetermined input data is input to the learning model C2. The determination result may be, for example, the output data from the learning model C2 itself, or it may be other data obtained based on the output data from the learning model C2. Note that the schematic shown in Figure 11B is a general illustration for illustrative purposes only and is not necessarily rigorous. A trained model may be called, for example, a pre-trained model.

[0123] <Examples where learning and judgment functions are provided in separate devices> Here, we illustrate a case where the same device (for example, an information processing device 11) has both a learning function for performing machine learning training and a determination function for performing determination (inference) based on the results of machine learning. As another example, the learning function and the judgment function may be provided in separate devices.

[0124] <Learning device and judgment device> Figure 11C shows an example of a learning device 511 and a determination device 512 according to an embodiment. In the example shown in Figure 11C, the learning device 511 and the determination device 512 are configured as separate devices. The learning device 511 performs machine learning training and provides the training results (e.g., the trained model) to the judgment device 512. The transfer of learning results (e.g., a trained model) from the learning device 511 to the determination device 512 may be performed, for example, by communication between the learning device 511 and the determination device 512, or by using a portable recording medium.

[0125] The determination device 512 obtains a predetermined determination result based on the learning results (e.g., a trained model) provided by the learning device 511. For example, if the information processing device 11 is equipped with the function of the determination device 512, the function of the learning device 511 may be provided in the information processing device 11, or it may be provided in a device other than the information processing device 11.

[0126] (Regarding the above embodiments) Furthermore, a program to realize the function of any component in any of the devices described above may be recorded on a computer-readable recording medium, and that program may be loaded into a computer system and executed. Here, "computer system" includes hardware such as operating systems and peripheral devices. "Computer-readable recording medium" refers to portable media such as flexible disks, magneto-optical disks, ROMs, CD (Compact Disc)-ROMs (Read Only Memory), and storage devices such as hard disks built into computer systems. Moreover, "computer-readable recording medium" also includes volatile memory within a computer system that retains a program for a certain period of time, such as volatile memory within a computer system that acts as a server or client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. Such volatile memory may be, for example, RAM (Random Access Memory). The recording medium may be, for example, a non-temporary recording medium.

[0127] Furthermore, the above program may be transmitted from a computer system that stores the program in a memory device or the like to another computer system via a transmission medium or by transmission waves within the transmission medium. Here, the "transmission medium" used to transmit the program refers to a medium that has the function of transmitting information, such as a network like the Internet or a communication line like a telephone line. Furthermore, the above program may be intended to implement only a portion of the functions described above. Moreover, the above program may be a so-called differential file, capable of implementing the aforementioned functions in combination with programs already recorded in the computer system. A differential file may also be called a differential program.

[0128] Furthermore, the functions of any component in any device described above may be implemented by a processor. For example, each process in the embodiment may be implemented by a processor that operates based on information such as a program, and a computer-readable recording medium that stores information such as a program. Here, the processor may be implemented by implementing the functions of each part in separate hardware, or by implementing the functions of each part in integrated hardware. For example, the processor includes hardware, and the hardware may include at least one of a circuit that processes digital signals and a circuit that processes analog signals. For example, the processor may be configured using one or more circuit devices or one or both of one or more circuit elements mounted on a circuit board. An IC (Integrated Circuit) may be used as the circuit device, and a resistor or capacitor may be used as the circuit element.

[0129] Here, the processor may be, for example, a CPU. However, the processor is not limited to a CPU; various types of processors may be used, such as a GPU (Graphics Processing Unit) or a DSP (Digital Signal Processor). The processor may also be, for example, a hardware circuit using an ASIC (Application Specific Integrated Circuit). Furthermore, the processor may be composed of, for example, multiple CPUs, or multiple hardware circuits using ASICs. The processor may also be composed of, for example, a combination of multiple CPUs and multiple hardware circuits using ASICs. The processor may also include, for example, one or more amplifier circuits or filter circuits that process analog signals.

[0130] While embodiments of this disclosure have been described in detail above with reference to the drawings, the specific configuration is not limited to these embodiments and may include designs and other elements that do not depart from the gist of this disclosure.

[0131] [Note] Examples of configurations 1 through 7 are shown.

[0132] (Configuration Example 1) A prompt input unit that inputs information containing at least one claim from one or more patent documents as a prompt to a generating AI device, A feature data acquisition unit acquires feature data from the generating AI device in response to the prompt, including a numerical range and units of one or more features of the substance relating to the claim. A storage processing unit that stores the feature data acquired by the feature data acquisition unit, or data based on the feature data, as storage data in a storage device, An information processing device equipped with the following features.

[0133] In this embodiment, the generation AI device 21, database 12, and information processing device 11 shown in Figure 1 are, respectively, an example of the generation AI device, an example of the storage device, and an example of the information processing device. Furthermore, the functions of the generation AI device may be provided, for example, in an information processing device. Furthermore, the memory function may be provided, for example, in an information processing device.

[0134] Furthermore, in this embodiment, the prompt input unit 212, structured data acquisition unit 213, and structured data storage processing unit 214 shown in Figure 4 are examples of the prompt input unit, feature data acquisition unit, and storage processing unit, respectively. In this embodiment, structured data is an example of the feature data described above. In this embodiment, the data to be stored corresponds, for example, to the data when storing structured data, and may be, for example, the structured data as is, or it may be data based on the structured data (for example, data into which the structured data has been processed).

[0135] (Configuration example 2) The substance is a mixture of multiple components. The information processing device described in (Configuration Example 1).

[0136] In this embodiment, the substance may be a substance consisting of its individual components.

[0137] (Configuration Example 3) The system includes a determination processing unit that determines whether the numerical values ​​of one or more characteristics of the substance to be determined fall within a numerical range defined by the storage data. An information processing device as described in (Configuration Example 1) or (Configuration Example 2).

[0138] In this embodiment, the determination processing unit 172 shown in Figure 5 is an example of the determination processing unit described above.

[0139] The determination processing unit estimates some or all of the numerical values ​​of the characteristics of the substance to be determined using MI or a simulator. The information processing device described in (Configuration Example 3).

[0140] (Configuration example 5) The system includes a proposal processing unit that proposes a proposed substance from among the substances to be determined that are determined not to be included in the numerical range for determination by the determination processing unit. The information processing device described in (Configuration Example 3) or (Configuration Example 4).

[0141] In this embodiment, the new material proposal processing unit 173 shown in Figure 6 is an example of the proposal processing unit described above.

[0142] (Configuration example 6) The aforementioned proposal processing unit includes a verification processing unit that verifies the numerical values ​​of one or more characteristics of the proposed substance using MI or a simulator. The information processing device described in (Configuration Example 5).

[0143] In this embodiment, the validation processing unit 254 shown in Figure 6 is an example of the validation processing unit described above.

[0144] It is also possible to provide a method for processing data using an information processing device. (Configuration example 7) Information processing device, A prompt input process that inputs information containing at least one claim from one or more patent documents as a prompt to an AI device, A feature data acquisition process is performed to acquire feature data from the generating AI device in response to the prompt, including the numerical range and units of one or more features of the substance relating to the claim. A storage process that stores the feature data obtained by the feature data acquisition process, or data based on the feature data, in a storage device as storage data, To do Information processing methods.

[0145] It is also possible to provide programs that are executed by computers (computer programs). (Configuration example 8) On the computer, A prompt input function that inputs information containing at least one claim from one or more patent documents as a prompt to a generating AI device, A feature data acquisition function that, in response to the prompt, acquires feature data from the generating AI device, including the numerical range and units of one or more features of the substance relating to the claim, A storage function that stores the feature data acquired by the feature data acquisition function, or data based on the feature data, as storage data in a storage device, A program to achieve this. [Explanation of Symbols]

[0146] 1...Information processing system, 11...Information processing device, 12...Database, 21...Generating AI device, 22...Patent document storage device, 31...Network, 131...Input unit, 132...Output unit, 133...Communication unit, 134...Storage unit, 135...Control unit, 136...Processing unit, 151...Operation unit, 152...Display unit, 171...Structured processing unit, 172...Applicability determination processing unit, 173...New material proposal processing unit, 211...Patent group collection unit, 212...Prompt input unit, 213...Structured data acquisition unit, 214...Structured data storage processing unit, 231...Determination target information acquisition unit, 232...Database reference unit, 233...Determination unit, 234...Determination result output unit, 251...Candidate group generation unit, 252...Feature value calculation unit, 253...Non-applicable detection unit, 254 ...Validation processing unit, 255...New material proposal unit, 511...Learning device, 512...Determination device, 1001...Information processing unit, 1011...Processor, 1012...Operating device, 1013...Display device, 1014...Storage device, 1015...Memory, 1016...Input / output interface, 1017...Network interface, 1021...Bus, C1~C2...Learning model, G1...Patent document content, G2...Structured data content, G3...Patent map, G4...MI / simulation processing unit, G5...Proposal content, G11...Analysis results, G12...Applicability determination results, G21...Density calculation results, G22...Refractive index calculation results, G31...Density refractive index distribution, G32...Liquid phase temperature calculation results, H1~H5...Composition, J1...Proposed material

Claims

1. A prompt input unit that inputs information including at least one claim of one or more patent documents as a prompt to a generating AI device, A feature data acquisition unit that acquires feature data from the generating AI device in response to the prompt, including the numerical range and units of one or more features of the substance relating to the claim, A storage processing unit that stores the feature data acquired by the feature data acquisition unit, or data based on the feature data, as storage data in a storage device, An information processing device equipped with the following features.

2. The substance is a mixture of multiple components. The information processing apparatus according to claim 1.

3. The system includes a determination processing unit that determines whether the numerical values ​​of one or more characteristics of the substance to be determined fall within a numerical range defined by the storage data. The information processing apparatus according to claim 1 or claim 2.

4. The determination processing unit estimates some or all of the numerical values ​​of the characteristics of the substance to be determined using MI or a simulator. The information processing apparatus according to claim 3.

5. The system includes a proposal processing unit that proposes a proposed substance from among the substances to be determined that are determined not to be included in the numerical range for determination by the determination processing unit. The information processing apparatus according to claim 3.

6. The aforementioned proposal processing unit includes a verification processing unit that verifies the numerical values ​​of one or more characteristics of the proposed substance using MI or a simulator. The information processing apparatus according to claim 5.

7. Information processing device, A prompt input process that inputs information including at least one claim of one or more patent documents as a prompt to a generating AI device, A feature data acquisition process is performed to acquire feature data from the generating AI device in response to the prompt, including the numerical range and units of one or more features of the substance relating to the claim. A storage process that stores the feature data obtained by the feature data acquisition process, or data based on the feature data, in a storage device as storage data, To do Information processing methods.

8. On the computer, A prompt input function that inputs information including at least one claim of one or more patent documents as a prompt to a generating AI device, A feature data acquisition function that, in response to the prompt, acquires feature data from the generating AI device, including the numerical range and units of one or more features of the substance relating to the claim. A storage function that stores the feature data acquired by the feature data acquisition function, or data based on the feature data, as storage data in a storage device, A program to achieve this.

Images