Information processing system, information processing method, and program

Abstract

Easily perform conversions between multiple data sets. [Solution] The data conversion system comprises: an input information acquisition unit that acquires input information including a first data item and input data; a conversion processing unit that inputs source schema information obtained by inputting source information including a first data item and source data into a large-scale language model; a conversion definition that indicates the processing content for converting source information into a data structure including a second data item of the destination information, obtained by inputting destination schema information obtained by inputting destination information including a second data item and destination data into a large-scale language model; and an output information acquisition unit that acquires output information from a large-scale language model that associates a second data item, output data related to the input data included in the input information, which has been analyzed by the large-scale language model based on the input information and the conversion definition.

Description

Technical Field

[0001] The present invention relates to an information processing system, an information processing method, and a program.

Background Art

[0002] An information processing system that outputs response information to a user's question has been disclosed (Patent Document 1).

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] The system described in Patent Document 1 acquires body information indicating the state of the body and question information indicating a question. The system stores the body information in association with history information indicating a work history. The system acquires similar body information similar to the body information related to the question information from the history information. The system inputs the question information and the similar body information into a large language model, and acquires response information indicating an answer regarding a predetermined method for the question information from the large language model. Thereby, an answer to the user's question can be obtained.

[0005] Conventionally, converting data having at least one data item (column) of a predetermined data format to data having at least one data item of the same data format requires, for example, the user to confirm the relationships between data items and the construction of a dedicated system to realize the conversion process of data items between data, which results in the problem of high costs. The system described in Patent Document 1 can obtain answers to questions using a large-scale language model, but it cannot identify the relationships between items in the data, and therefore cannot appropriately convert data items between data, and thus cannot solve the above problem.

[0006] Therefore, the present invention aims to easily perform conversions between multiple data sets in order to solve the above problems. [Means for solving the problem]

[0007] An information processing system according to one aspect of the present invention includes an input information acquisition unit that acquires input information including at least one first data item in a predetermined data format and input data which is associated with the first data item, and a conversion processing unit that inputs source information including the first data item and source data which is associated with the first data item into a large-scale language model, thereby acquiring source schema information that shows the data structure of the source information and destination information including at least one second data item in a predetermined data format and destination data which is associated with the second data item into a large-scale language model. The system comprises: a conversion processing unit that inputs to a predetermined large-scale language model: a conversion schema information indicating the data structure of the destination information, which is obtained from a large-scale language model by inputting to a word model; a conversion definition indicating the processing content for converting source information into a data structure including a second data item of the destination information, which is obtained from a large-scale language model by inputting to the large-scale language model; input information; and output information acquisition unit that acquires from a predetermined large-scale language model output information which associates a second data item, which has been analyzed by the predetermined large-scale language model based on the input information and the conversion definition, with output data, which is data relating to the input data contained in the input information.

[0008] An information processing method according to one aspect of the present invention involves a computer acquiring input information including at least one first data item in a predetermined data format and input data which is data associated with the first data item, and inputting source information including the first data item and source data which is data associated with the first data item into a large-scale language model, thereby obtaining source schema information indicating the data structure of the source information from the large-scale language model, and destination information including at least one second data item in a predetermined data format and destination data which is data associated with the second data item, into a large-scale language model. The process involves inputting the following into a predetermined large-scale language model: inputting the target schema information, which indicates the data structure of the target information obtained from the large-scale language model by inputting it into the model; inputting the source information into the large-scale language model; inputting the conversion definition, which indicates the processing content for converting the source information into a data structure that includes a second data item of the target information obtained from the large-scale language model; inputting the source information into a predetermined large-scale language model; and obtaining output information from the predetermined large-scale language model that associates the second data item, which has been parsed by the predetermined large-scale language model based on the input information and the conversion definition, with the output data, which is data related to the input data contained in the input information.

[0009] A program according to one aspect of the present invention provides a computer with input information including at least one first data item in a predetermined data format and input data which is data associated with the first data item, and inputs source information including the first data item and source data which is data associated with the first data item into a large-scale language model, thereby obtaining source schema information indicating the data structure of the source information from the large-scale language model, and destination information including at least one second data item in a predetermined data format and destination data which is data associated with the second data item into the large-scale language model. The process involves inputting the following into a predetermined large-scale language model: inputting the destination schema information, which indicates the data structure of the destination information obtained from the large-scale language model; inputting the source information into the large-scale language model; inputting the source information into the large-scale language model; inputting the source information into the large-scale language model; inputting the source information into the large-scale language model; and obtaining output information from the predetermined large-scale language model that associates the second data item, which has been analyzed by the predetermined large-scale language model based on the input information and the conversion definition, with the output data, which is data related to the input data contained in the input information. [Effects of the Invention]

[0010] According to the present invention, conversions between multiple data sets can be easily performed. [Brief explanation of the drawing]

[0011] [Figure 1] This diagram shows an overview of the data conversion system according to the first embodiment. [Figure 2] This figure shows an example of a conversion definition prompt. [Figure 3] This figure shows an example of the conversion between source information and destination information. [Figure 4] This database shows an example of the conversion definition database D181. [Figure 5] This figure shows an example of the input screen T10. [Figure 6] It is a diagram showing an example of an input screen T20 according to the first modification. [Figure 7] It is a diagram showing an example of an input screen T30 according to the third modification. [Figure 8] It is a diagram showing an example of a generation screen T40 according to the fourth modification. [Figure 9] It is a flowchart showing the processing procedure of the data conversion system according to the first embodiment. [Figure 10] It is a diagram showing an overview of the data conversion system according to the second embodiment. [Figure 11] It is a diagram showing an example of the similarities and differences between the output information and the structure of the modified header information. [Figure 12] It is a diagram showing an example of a re - conversion definition prompt. [Figure 13] It is a diagram showing an example of a re - output confirmation screen T50. [Figure 14] It is a diagram showing an example of a re - definition confirmation screen T60. [Figure 15] It is a flowchart showing the processing procedure of the data conversion system according to the second embodiment. [Figure 16] It is a diagram showing an example of the hardware configuration of a computer.

Modes for Carrying Out the Invention

[0012] Hereinafter, the data conversion system 100 in an embodiment of the present invention will be described in detail with reference to the drawings. However, the embodiments described below are merely examples, and there is no intention to exclude various modifications and applications of technologies not explicitly stated below. That is, the present invention can be implemented with various modifications or by combining each embodiment without departing from its gist. Also, in the following description of the drawings, the same or similar parts are denoted by the same or similar reference numerals.

[0013] In addition, in this embodiment, the "part", "device", and "system" do not simply mean physical means, but also include cases where the functions of the "part", "device", and "system" are realized by software. Also, the functions of one "part", "device", or "system" may be realized by two or more physical means or devices, and the functions of two or more "parts", "devices", or "systems" may be realized by one physical means or device. Furthermore, each of the various functions shown below of the plurality of devices constituting the data conversion system 100 may be configured to be executed by other devices in the plurality of devices.

[0014] ===Data Conversion System 100 According to the First Embodiment=== <<Overview of the System>> Referring to FIG. 1, the overview of the data conversion system 100 according to the first embodiment will be described. FIG. 1 is a diagram showing the overview of the data conversion system 100 according to the first embodiment.

[0015] The data conversion system 100 is a system that enables appropriate data conversion between a plurality of data by appropriately analyzing the data structure of each of the plurality of data in a predetermined data format using a large language model (hereinafter referred to as "generative AI (Artificial Intelligence)") having a huge dataset based on a deep learning model.

[0016] The generative AI is, for example, a text generation system AI that is trained to extract appropriate information from predetermined data and generate an appropriate answer when a prompt including information indicating the content of a user's question is input, such as ChatGPT or Google Bard.

[0017] The specified data formats include, for example, CSV (Comma-Separated Values), JSON (JavaScript Object Notation), and XML (Extensible Markup Language). In the following, as an example, information in the specified data format will be described as a CSV file consisting of rows and columns.

[0018] Furthermore, in the following explanation, the "columns" of a CSV file will be described as data items. The "rows" of a CSV file will be described as records, and the data in each cell that makes up a record will be described as "data related to a data item." Note that when referring to the "name of a data item," it means the name of the column.

[0019] Traditionally, when converting a CSV file, which associates the names of multiple data items (e.g., column names such as name and age) with related data (e.g., data indicating name and age), to a CSV file composed of different data items (e.g., column names such as phonetic spelling), the user typically reviews the data related to the data items and considers the conversion method. In this case, implementing the considered conversion method requires building a dedicated system for each conversion method to perform the data conversion. Building dedicated systems requires significant costs and effort, resulting in a considerable waste of resources for businesses.

[0020] Furthermore, for example, when converting CSV files, data conversion using tools such as Microsoft Excel macros requires a large amount of manual work, such as adjusting the number of characters, resulting in enormous costs.

[0021] Therefore, the data conversion system 100 achieves simple, rapid, and inexpensive data conversion by appropriately using a large-scale language model in the process of converting each piece of data composed of different data items.

[0022] The data conversion system 100 may be, for example, a cloud computer, a server computer, a personal computer (e.g., a desktop, laptop, tablet, etc.), a media computer platform (e.g., a cable, satellite set-top box, digital video recorder), a handheld computer device (e.g., a PDA, email client, etc.), or other types of computers or communication platforms. At least a portion of the processing in the data conversion system 100 may be implemented by one or more computers (not limited to, but for example, a cloud computing system consisting of one or more computers).

[0023] <<Process Overview>> Next, with reference to Figure 1, an overview of the processing of the data conversion system 100 will be described.

[0024] In step S10, the data conversion system 100 obtains a CSV file (hereinafter referred to as "source information") from the terminal device 200 that includes at least one data item (hereinafter referred to as "first data item") and data associated with the first data item (hereinafter referred to as "source data"). Hereinafter, for convenience, the data set consisting of the first data item and the source data associated with the first data item will be referred to as the "source data structure".

[0025] In step S11, the data conversion system 100 obtains a CSV file (hereinafter referred to as "destination information") from the terminal device 200 that includes at least one data item (hereinafter referred to as the "second data item") having a data structure different from the source data structure, and data associated with the second data item (hereinafter referred to as the "destination data"). Hereinafter, for convenience, the data set consisting of the second data item and the destination data associated with the second data item will be referred to as the "destination data structure".

[0026] In step S12, the data conversion system 100 inputs a prompt (hereinafter referred to as the "structural analysis prompt") (not shown in Figure 1) obtained from the terminal device 200 and the source information into the generating AI. The data conversion system 100 also inputs the structural analysis prompt and the destination information into the generating AI.

[0027] A structural analysis prompt is, for example, a prompt that instructs the generating AI to analyze the data structure of a CSV file. For example, a structural analysis prompt indicates the process to instruct the generating AI to analyze the data items, which are columns, and the data content associated with each of those data items. Specifically, a structural analysis prompt might say, for example, "Return all the column information of the given CSV in JSON format in the format shown in the sample."

[0028] In step S13, the data transformation system 100 obtains information indicating the source data structure (hereinafter referred to as "source schema information") and information indicating the destination data structure (hereinafter referred to as "destination schema information") from the generating AI.

[0029] The source schema information is the result of the generation AI analyzing at least one of the following: the name of each data item, or the source data associated with each data item. Specifically, the source schema information includes, for example, information indicating that the name of the data item in column A1 is "Last Name," the name of the data item in column A2 is "First Name," the source data for the first row of the "Last Name" data item is "Yamada," and the source data for the first row of the "First Name" data item is "Taro." Furthermore, the source schema information also includes the results of analyzing the relationships between each data item and the content of the source data for each data item. For example, it includes information indicating that the first row of the full name is "Yamada Taro," based on the source data associated with the "Last Name" data item and the source data associated with the "First Name" data item.

[0030] The destination schema information is the result of the generation AI's analysis of, for example, the names of each data item, or at least the destination data associated with each data item. Specifically, the destination schema information may include, as an example, information indicating that the name of the data item in column B1 is "URL", the name of the data item in column B2 is "Phonetic Name", the destination data for the first row of the "URL" data item is "https / / ··", and the destination data for the first row of the "Phonetic Name" data item is "Kawaguchi Hanako". The source schema information also includes the results of the analysis of the relationships between each data item and the content of the destination data for each data item, such as information indicating that the phonetic spelling of the first and last name is entered in the first row of the "Phonetic Name" data item.

[0031] In step S14, the data conversion system 100 inputs a prompt (hereinafter referred to as the "conversion analysis prompt") (not shown in Figure 1) obtained from the terminal device 200, along with source schema information and destination schema information, into the generating AI.

[0032] A transformation analysis prompt is a prompt that instructs the generating AI to analyze the process for transforming the data structure of the source information into the data structure of the target information. Specifically, a transformation analysis prompt instructs the generating AI to identify the relationship between each data item in the source information and each data item in the target information, and to generate a transformation definition that allows the generating AI to appropriately transform the data structure of the source information into the data structure of the target information. Specifically, the transformation analysis prompt may contain instructions such as, "We want to transform data from one CSV to another CSV format. Given the format of the transformation definition, the source CSV, and the target CSV, infer and create a JSON transformation definition based on them."

[0033] In this case, the conversion analysis prompt may contain at least one of the following formats for the conversion definition: "The conversion definition is defined from the following three types for each column of the target CSV format: 1. Fixed value: Insert the fixed value defined by value for all rows of the definition column. 2. Column shift: Insert the value of the specified column of the source specified by source_field_id as is. 3. Generate: Insert the result of the process specified by instruction as the value, while referring to the value of the specified column(s) of the source specified by source_field_id (multiple columns can be specified)."

[0034] In step S15, the data transformation system 100 obtains a transformation definition from the generated AI based on the transformation analysis prompt, source schema information, and destination schema information.

[0035] A conversion definition is information that indicates the processing steps for converting source information into a data structure that includes a second data item of the target information. In other words, a conversion definition is a prompt for the generating AI to convert the data of the source information's data structure (for example, a CSV file that has the same data item arrangement and names as the source CSV file) into the data of the target information's data structure (for example, a CSV file that has the same data item arrangement and names as the target CSV file). For convenience, the prompt indicating the conversion definition will be referred to as the "conversion definition prompt" below. The conversion definition prompt will be discussed later.

[0036] In step S16, the data conversion system 100 displays the input screen T10, which will be described later, on the terminal device 200. Based on the user's operation input to the input screen T10 on the terminal device 200, the system obtains the information to be converted (hereinafter referred to as "input information") from the terminal device 200.

[0037] The input information includes the same first data item as the source information, and data associated with the first data item (hereinafter referred to as "input data").

[0038] In step S17, the data conversion system 100 inputs the conversion definition prompt and the input information to the generating AI.

[0039] In step S18, the data conversion system 100 obtains from the generating AI information in which the data content of the input information is associated with a second data item (hereinafter referred to as "output information").

[0040] The output information is information obtained by converting the input information to the data structure of the target information. Specifically, for example, if the input information (information with the same data structure as the source information) includes information indicating that the name of the data item in column A is "Last Name", the name of the data item in column B is "First Name", the data content of the first row of the "Last Name" data item is "Yamada", and the data content of the first row of the "First Name" data item is "Taro", then the output information is information in which "YAMADA TARO" (output data) is entered into the "Phonetic Name" data item (second data item) which has the same data structure as the target information.

[0041] For example, steps S10 to S15 may be steps that have been pre-executed by the data conversion system 100 based on the operator's input. In this case, the data conversion system 100 can provide the desired output information to the user simply by receiving input information from the user who desires data conversion to the input screen T10.

[0042] Thus, the data conversion system 100 enables the conversion of data structures to the desired state by the user simply by accepting user input information from the generating AI, using the results of analyzing the data structure before and after data conversion using the generating AI (conversion definition prompt). As a result, the data conversion system 100 can perform data structure conversion inexpensively, quickly, and easily.

[0043] <<Structure>> Referring to Figure 1, the configuration of the data conversion system 100 will be described. As shown in Figure 1, the data conversion system 100 includes a source information processing unit 110, a destination information processing unit 120, an information transmission unit 130, a schema information acquisition unit 140, an input information acquisition unit 150, a conversion processing unit 160, an output information acquisition unit 170, a storage unit 180, and a display processing unit 190.

[0044] The source information processing unit 110 acquires source information. The source information processing unit 110 may acquire source information from the terminal device 200 by, for example, receiving input of source information from the user on a predetermined screen of the terminal device 200.

[0045] The destination information processing unit 120 acquires destination information. The destination information processing unit 120 may acquire destination information from the terminal device 200 by, for example, receiving input of destination information from the user on a predetermined screen of the terminal device 200.

[0046] The information transmission unit 130 inputs various information to the generating AI. Specifically, the information transmission unit 130 sends, for example, source information, destination information, source schema information, destination schema information, structure analysis prompt, transformation analysis prompt, etc., to the generating AI.

[0047] The schema information acquisition unit 140 acquires source schema information and destination schema information from the generating AI, which has received the structure analysis prompt, source information, and destination information as input.

[0048] The input information acquisition unit 150 acquires input information. Specifically, the input information acquisition unit 150 may acquire input information from the terminal device 200 by, for example, receiving input information from the user on the input screen T10 displayed on the display unit of the terminal device 200.

[0049] The conversion processing unit 160 generates a conversion definition prompt (conversion definition) using the generation AI. Specifically, the conversion processing unit 160 obtains a conversion definition prompt from the generation AI by inputting a conversion analysis prompt, source schema information, and destination schema information into the generation AI. Then, the conversion processing unit 160 inputs the conversion definition prompt and the input information into the generation AI.

[0050] Here, we will explain the conversion definition prompt with reference to Figures 2 and 3. Figure 2 is a diagram showing an example of a conversion definition prompt. In Figure 2, the conversion definition prompt is shown in JSON format as an example. Note that the conversion definition prompt does not have to be in JSON format. In Figure 2, the conversion definition prompt is explained as conversion definition prompt P10. Figure 3 is a diagram showing an example of conversion between source information (information in column A) and destination information (information in column B).

[0051] As shown in Figure 2, the conversion definition prompt P10 includes, for example, the processing content for each data item in the output information. For example, the conversion definition prompt P10 shown in Figure 2 includes processing content P11, processing content P12, processing content P13, and processing content P14.

[0052] For example, page P11 of the processing instructions describes a process that inputs the fixed value "https: / / ..." into column "B1" of the output information.

[0053] The process for generating processing content P11 is outlined below. First, the data conversion system 100 inputs the structural analysis prompt and destination information (for example, the destination information in Figure 3(b)) into the generating AI and has it analyze the destination information to obtain destination schema information indicating that a fixed value is entered in column "B1" of the destination information. Then, the data conversion system 100 has the generating AI analyze the conversion definition prompt and destination schema information to obtain a conversion definition prompt that includes processing content P11 indicating that a fixed value is entered in column "B1".

[0054] For example, processing description P12 describes a process in which data content created by combining the data content from column A1 and column A2 of the input information, with the phonetic readings converted to full-width katakana and a half-width space inserted between the last name and first name, is entered into column B2 of the output information.

[0055] The following outlines the process by which processing content P12 is generated. First, the data conversion system 100 inputs the structural analysis prompt and source information (for example, the source information in Figure 3(a)) into the generating AI and has the generating AI analyze it to obtain source schema information indicating that the "A1" column of the source information contains the surname in kanji (for example, "Yamada" in Figure 3(a)) as source data, and the "A2" column contains the given name in kanji (for example, "Taro" in Figure 3(a)) as source data. Next, the data conversion system 100 inputs the structural analysis prompt and destination information (for example, the destination information in Figure 3(b)) into the generating AI and has the generating AI analyze it to obtain destination schema information indicating that the "B2" column of the destination information contains the phonetic spelling of the surname and given name (for example, "Yamada Taro" in Figure 3(b)) as destination data. The data conversion system 100 then has the generating AI analyze the conversion analysis prompt, source schema information, and destination schema information, thereby obtaining a conversion definition prompt in column "B1" of the output information that includes processing details P12 indicating that the phonetic readings of the surname entered in column "A1" and the phonetic readings of the given name entered in column "A2" of the input information will be joined together with a half-width space in between.

[0056] For example, processing description P13 describes a process in which the data content from column A4 of the input information is converted from English to Japanese and entered into column B3 of the output information.

[0057] The following outlines the process by which processing content P13 is generated. First, the data conversion system 100 inputs the structural analysis prompt and source information (for example, the source information in Figure 3(a)) into the generating AI and has the generating AI analyze it to obtain source schema information indicating that the source data in column "A4" of the source information is gender (for example, "male" in Figure 3(a)) entered in English. Next, the data conversion system 100 inputs the structural analysis prompt and destination information (for example, the destination information in Figure 3(b)) into the generating AI and has the generating AI analyze it to obtain destination schema information indicating that the destination data in column "B3" of the destination information is gender in Japanese (for example, "male" in Figure 3(b)). The data conversion system 100 then has the generating AI analyze the conversion definition prompt, source schema information, and destination schema information to obtain a conversion definition prompt in column "B3" of the output information, which includes processing content P13 indicating that the gender entered in column "A4" of the input information in English should be converted to Japanese and entered.

[0058] For example, processing instructions P14 describe a process that directly references the data content from column A3 of the input information into column B4 of the output information.

[0059] The following outlines the process by which processing content P14 is generated. First, the data transformation system 100 inputs the structural analysis prompt and source information (for example, the source information in Figure 3(a)) into the generating AI and has the generating AI analyze it to obtain source schema information indicating that the "A3" column of the source information contains age (for example, "27" in Figure 3(a)) as source data. Next, the data transformation system 100 inputs the structural analysis prompt and destination information (for example, the destination information in Figure 3(b)) into the generating AI and has the generating AI analyze it to obtain destination schema information indicating that the "B4" column of the destination information contains age (for example, "27" in Figure 3(b)) as destination data. Finally, the data transformation system 100 has the generating AI analyze the source schema information and destination schema information to obtain a transformation definition prompt containing processing content P14, which indicates that the "B4" column of the output information contains the age entered in the "A3" column of the input information.

[0060] Thus, the conversion definition prompt indicates the processing content for converting from the data structure of the source information to the data structure of the destination information on a column-by-column basis in a predetermined data format. In other words, the data conversion system 100 appropriately identifies the relationships between data items (columns) and enables data conversion appropriately according to the relationships between data items, such as combining data related to data items, inputting fixed values ​​into data items, and converting the language of data related to data items. As a result, the data conversion system 100 enables data conversion simply, quickly, and inexpensively.

[0061] The output information acquisition unit 170 acquires output information from the generated AI. For example, if the input information is the same as the source information shown in Figure 3(a), the output information acquisition unit 170 acquires output information with the content of the destination information shown in Figure 3(b).

[0062] The storage unit 180 stores various types of information. The storage unit 180 includes, for example, a conversion definition database D181.

[0063] Refer to Figure 4 to explain the conversion definition database D181. Figure 4 shows an example of the conversion definition database D181. The conversion definition database D181 is a database that stores information about conversion definitions.

[0064] The conversion database D107a includes items such as [Conversion Definition ID], [Conversion Definition Prompt], [Source Information], [Source Schema Information], [Destination Information], [Destination Schema Information], [Structural Analysis Prompt], and [Conversion Analysis Prompt]. [Conversion Definition ID] stores identification information that uniquely identifies the conversion definition. [Conversion Definition Prompt] stores the conversion definition prompt (conversion definition). [Source Information] stores the source information, such as a CSV file. [Source Schema Information] stores the source schema information. [Destination Information] stores the destination information, such as a CSV file. [Destination Schema Information] stores the destination schema information. [Structural Analysis Prompt] stores the structural analysis prompt. [Conversion Analysis Prompt] stores the conversion analysis prompt.

[0065] The display processing unit 190 generates information (hereinafter referred to as "screens") for displaying various screens on the display unit of the terminal device 200, or, if the data conversion system 100 is a terminal device, on the display unit. For example, the display processing unit 190 generates an input screen T10.

[0066] The input screen T10 will be described with reference to Figure 5. Figure 5 is a diagram showing an example of the input screen T10. As shown in Figure 5, the input screen T10 includes a conversion definition selection area T11, an input reception area T12, and an execution button T13.

[0067] The conversion definition selection area T11 is an area where the user can select one of at least one conversion definitions based on user input. The conversion definition selection area T11 includes a pull-down button T11a and a conversion definition display area T11b. The pull-down button T11a is a button that displays a list of selectable conversion definitions when user input is received. The conversion definition display area T11b is an area that displays the selected conversion definition.

[0068] The input reception area T12 is an area for receiving input information. The input reception area T12 is an area where input information can be received, for example, by dragging. This allows the data conversion system 100 to perform data conversion easily and quickly.

[0069] The execution button T13 is a button that converts the input information into the generating AI using the selected conversion definition. When the data conversion system 100 receives user input via the execution button T13, it inputs the input information and the conversion definition prompt into the generating AI. This allows the data conversion system 100 to obtain output information from the generating AI.

[0070] In this way, the data conversion system 100 provides the user with an object in which the user can select the desired conversion definition prompt from among multiple conversion definition prompts (conversion definitions). This allows the data conversion system 100 to enhance user convenience.

[0071] <<First variation>> The conversion processing unit 160 may have a function to recommend one of several conversion definitions (hereinafter referred to as the "recommended conversion definition") to the user based on the input information entered by the user. This allows the data conversion system 100 to present the user with an appropriate conversion definition corresponding to the input information when the user is unable to identify the desired conversion definition, thereby improving user convenience.

[0072] Specifically, as described above, the conversion processing unit 160 inputs source information (hereinafter referred to as "first source information"), which includes a first data item and data related to the first data item (hereinafter referred to as "first source data"), into the large-scale language model based on user input. Then, the conversion processing unit 160 obtains information indicating the data structure of the first source information (hereinafter referred to as "first source schema information") from the large-scale language model based on the first source information.

[0073] Then, as described above, the conversion processing unit 160 inputs conversion destination information (hereinafter referred to as "first conversion destination information"), which includes the second data item and data related to the second data item (hereinafter referred to as "first conversion destination data"), into the large-scale language model based on user input. Then, the conversion processing unit 160 obtains information indicating the data structure of the first conversion destination information (hereinafter referred to as "first conversion destination source schema information") from the large-scale language model based on the first conversion destination information.

[0074] Similarly, the conversion processing unit 160 obtains second source schema information from the large-scale language model based on second source information, which includes second source data consisting of a third data item different from the first data item. The conversion processing unit 160 also obtains second destination schema information from the large-scale language model based on second destination information, which includes second destination data consisting of a fourth data item different from the second data item.

[0075] The conversion processing unit 160 inputs the first source schema information and the first destination schema information to the generating AI, thereby obtaining a conversion definition (hereinafter referred to as the "first conversion definition") from the generating AI. The conversion processing unit 160 stores the first conversion definition in the conversion definition database D181. Similarly, the conversion processing unit 160 inputs the second source schema information and the second destination schema information to the generating AI, thereby obtaining a conversion definition (hereinafter referred to as the "second conversion definition") from the generating AI. The conversion processing unit 160 stores the second conversion definition in the conversion definition database D181.

[0076] When the conversion processing unit 160 obtains input information, it refers to the conversion definition database D181, for example, based on the data items included in the input information, to identify a recommended conversion definition. Specifically, the conversion processing unit 160 may identify a conversion definition that is set to enable data conversion for data items whose names (e.g., names such as "Last Name," "First Name," "Age," and "Gender") and their order in the input information are the same as those of data items whose names and order are the same. In this case, the conversion processing unit 160 may also identify the conversion definition by referring to the source schema information corresponding to the conversion definition.

[0077] Furthermore, the conversion processing unit 160 may specifically identify recommended conversion definitions that can be used for data conversion when the content of the input data in the input information (for example, input content such as "Yamada", "Taro", "27", "male") and the arrangement of the input data in the record are the same as the content of the input data and the arrangement of the input data in the record. In this case, the conversion processing unit 160 may identify the conversion definition by referring to the source schema information corresponding to the conversion definition.

[0078] The conversion processing unit 160 may also identify recommended conversion definitions that allow data conversion for data items and their contents in the input information, as well as for their respective arrangements.

[0079] Referring to Figure 6, the input screen T20, which includes the recommended conversion definition, will be described. Figure 6 is a diagram showing an example of the input screen T20 according to the first modified example. As shown in Figure 6, the input screen T20 includes a conversion definition selection area T21, an input reception area T22, an execution button T23, and a recommendation area T24. Note that the conversion definition selection area T21, input reception area T22, and execution button T23 in the input screen T20 are the same as the conversion definition selection area T11, input reception area T12, and execution button T13 in the input screen T10, so their explanation will be omitted.

[0080] The recommendation area T24 includes the recommended conversion definition display area T24a and the recommended conversion definition explanation area T24b.

[0081] The recommended conversion definition display area T24a is an area that displays information that identifies the recommended conversion definition (e.g., conversion definition ID, conversion definition name, etc.). When the data conversion system 100 receives user input regarding the information that identifies the recommended conversion definition displayed in the recommended conversion definition display area T24a, for example, the recommended conversion definition is set in the conversion definition selection area T21. As shown in Figure 6, the conversion processing unit 160 may also display multiple recommended conversion definitions in the recommended conversion definition display area T24a in order of recommendation. This makes it possible for the data conversion system 100 to prevent the user from making a mistake in selecting a conversion definition prompt.

[0082] The recommended conversion definition explanation area T24b is an area that displays an overview of the data conversion of the recommended conversion definition. The overview of the data conversion of the recommended conversion definition may, for example, show the relationship between the data items of the source information and the data items of the destination information, or it may display the source information and destination information in a table format and illustrate the relationship between each data item. This enables the data conversion system 100 to provide a display that allows the user to intuitively understand the conversion definition prompt, thereby preventing the user from making a mistake in selecting the conversion definition prompt.

[0083] <<Second variation>> The conversion processing unit 160 may verify the output information obtained from the generating AI based on the constraint information for each column set in at least one of the source information and the destination information. The constraint information may include, for example, the length of a character or the type of character.

[0084] Specifically, the conversion processing unit 160 adds constraint information (for example, information indicating a character length limit such as "100 characters or less") to at least one of the source information and destination information obtained from the terminal device 200, based on user input.

[0085] The conversion processing unit 160 determines whether the data content of each data item in the output information satisfies the restrictions indicated by the constraint information. If the conversion processing unit 160 determines that there are data items that do not satisfy the restrictions, it may send information to the terminal device 200 indicating that the restrictions are not met.

[0086] Furthermore, if the conversion processing unit 160 determines that there are data items that do not satisfy the restrictions, it may repeat the process of inputting the conversion definition prompt and input information to the generating AI (for example, the process in step S107 shown in Figure 9). This makes it possible for the data conversion system 100 to obtain output information that satisfies the constraints related to the conversion.

[0087] Furthermore, if the conversion processing unit 160 determines that there are data items that do not satisfy the restrictions, it may generate a prompt that adds restriction information to the conversion definition prompt, and then input this prompt and the input information to the generation AI. This makes it possible for the data conversion system 100 to more reliably obtain output information that satisfies the constraints related to the conversion.

[0088] <<Third variation>> The data transformation system 100 may have a function that allows the user to select a generating AI from among multiple generating AIs to generate a transformation definition prompt. Furthermore, the data transformation system 100 may have a function that allows the user to select the columns to be transformed by the selected generating AI.

[0089] Referring to Figure 7, we will now describe the input screen T30, which includes a selection area where the generated AI can be selected (hereinafter referred to as the "AI selection area T34") and a selection area where columns can be selected (hereinafter referred to as the "column selection area T35"). Figure 7 is a diagram showing an example of the input screen T30 according to the third modified example.

[0090] As shown in Figure 7, the input screen T30 includes a conversion definition selection area T31, an input reception area T32, an execution button T33, an AI selection area T34, and a column selection area T35. Note that the conversion definition selection area T31, input reception area T32, and execution button T33 in the input screen T30 are the same as the conversion definition selection area T11, input reception area T12, and execution button T13 in the input screen T10, so their explanation is omitted.

[0091] The display processing unit 190, for example, makes the check buttons in the AI ​​selection area T34 selectable when input information is entered into the input reception area T32. When any of the check buttons is selected by the user, the conversion processing unit 160 identifies the generation AI corresponding to the selected check button. Then, when the conversion processing unit 160 receives user input for the execution button T33, it inputs the conversion definition prompt and input information to the identified generation AI. As a result, the data conversion system 100 enables the user to select a generation AI that can convert the input information more appropriately, thereby achieving appropriate data conversion.

[0092] Furthermore, when input information is entered into the input reception area T32, the display processing unit 190 makes the check buttons in the column selection area T35 selectable. When the user selects one of the check buttons in the AI ​​selection area T34 and then selects a check button in the column selection area T35, the conversion processing unit 160 identifies the generation AI and column (data item) corresponding to the selected check button. Then, when the conversion processing unit 160 receives user input for the execution button T33, it inputs the conversion definition prompt and the data related to the identified column from the input information into the identified generation AI. Data related to columns that have not been identified from the input information may be input into the identified generation AI or into a predetermined generation AI. As a result, the data conversion system 100 can allow the user to select a generation AI that can convert the input information more appropriately, and furthermore, it can convert the data into a generation AI that matches the characteristics of the column, thereby achieving appropriate data conversion.

[0093] In the above description, it was explained that the system has a function to select the generation AI and columns that generate the conversion definition prompt, but it is not limited to this. For example, the data conversion system 100 may have a generation AI pre-configured for each column in the conversion definition prompt to input data related to that column. Specifically, in the conversion definition prompt, "GPT-4o" may be pre-configured for a predetermined column (for example, "Column A1" in Figure 7) (i.e., no user selection operation for the generation AI and column is required). The predetermined column may be one column or multiple columns. In this case, at least one column and the generation AI to be applied to that at least one column are associated and stored within the "Conversion Definition Prompt" item of the conversion definition database D181 (embedded in the conversion definition prompt). This enables the data conversion system 100 to achieve high-speed and appropriate data conversion. Furthermore, the generation AI associated with each column may be a generation AI that is fine-tuned to specialize in processing columns. This enables the data conversion system 100 to achieve high-speed and more appropriate data conversion.

[0094] <<Fourth variation>> The data conversion system 100 may have a function that allows the user to select any of the multiple source information and any of the multiple destination information. Furthermore, the data conversion system 100 may have a function that allows the user to set information regarding examples of conversions (hereinafter referred to as "example conversion information") to improve the accuracy of the conversion definition prompt.

[0095] Referring to Figure 8, the generation screen T40 for generating the conversion definition prompt will be described. Figure 8 shows an example of the generation screen T40 according to the fourth modified example.

[0096] As shown in Figure 8, the generation screen T40 includes a source information input area T41, a source information selection area T42, a destination information input area T43, a destination information selection area T44, a conversion example information input area T45, a generate button T46, a generation result display area T47, and a conversion definition save button T48.

[0097] The source information input area T41 is an area for receiving source information. The source information input area T41 is an area where source information can be received, for example, by dragging.

[0098] The source information selection area T42 is an area that displays the source information entered in the source information input area T41 in a selectable format. When source information is entered in the source information input area T41, the display processing unit 190 makes the check button in the source information selection area T42 selectable. Note that the source information selection area T42 is not limited to displaying selectable source information. For example, the source information selection area T42 may display selectable source schema information, which is the result of the source information received in the source information input area T41 being input into the generated AI by the conversion processing unit 160.

[0099] The destination information input area T43 is an area for receiving destination information. The destination information input area T43 is an area where destination information can be received, for example, by dragging.

[0100] The destination information selection area T44 is an area that displays the destination information entered in the destination information input area T43 in a selectable format. When destination information is entered in the destination information input area T43, the display processing unit 190 makes the check button in the destination information selection area T44 selectable. Note that the destination information selection area T44 is not limited to displaying selectable destination information. For example, the destination information selection area T44 may display selectable destination schema information, which is the result of the destination information received in the destination information input area T43 being input into the generated AI by the conversion processing unit 160.

[0101] The conversion example information input area T45 is an area for receiving conversion example information. The conversion example information input area T45 includes, for example, a source example input area T45a, a destination example input area T45b, and an add conversion example button T45c.

[0102] The source example input area T45a is an area into which source example information (e.g., CSV) containing data items and data associated with those data items is input. The destination example input area T45b is an area into which destination example information (e.g., CSV) containing data items and data associated with those data items is input. The source example information and destination example information are, for example, information generated by the user. That is, the source example input area T45a is an area into which an example of a conversion process by the user can be input, and a set of source example information and destination example information is input. The set of source example information and destination example information is information for so-called few-shot learning.

[0103] The generate button T46 is a button that executes the process of generating a conversion prompt to convert the source information selected in the source information selection area T42 (hereinafter referred to as "selected destination information") to the destination information selected in the destination information selection area T44 (hereinafter referred to as "selected destination information").

[0104] Specifically, when the conversion processing unit 160 receives user input for the generation button T46, it inputs the selected source information and the structural analysis prompt to the generation AI. The conversion processing unit 160 obtains the source schema information from the generation AI. Similarly, when the conversion processing unit 160 receives user input for the generation button T46, it inputs the selected destination information and the structural analysis prompt to the generation AI. The conversion processing unit 160 obtains the destination schema information from the generation AI. Then, the conversion processing unit 160 inputs the source schema information and destination schema information obtained from the generation AI, along with the conversion analysis prompt, to the generation AI.

[0105] Specifically, if conversion example information is entered in the conversion example information input area T45, the conversion processing unit 160, upon receiving user input for the generate button T46, inputs the source schema information (information obtained from the generating AI based on the source information), destination schema information (information obtained from the generating AI based on the source information), a conversion analysis prompt, and a set of source example information and destination example information to the generating AI. By inputting this set of source example information and destination example information to the generating AI, it becomes possible to obtain a conversion definition prompt with higher accuracy from the generating AI.

[0106] The generation result display area T47 is an area that displays information about the conversion definition prompt generated by the generation AI when user input is received for the generation button T46.

[0107] The conversion definition save button T48 is a button for storing the conversion definition prompt generated by the generation AI in the storage unit 180. At this time, the data conversion system 100 allows the user to set the name of the conversion definition prompt.

[0108] Thus, according to the fourth modified data conversion system 100, the user can arbitrarily generate various conversion definition prompts with simple operations, making it possible to perform data conversion more appropriately.

[0109] ===Processing Procedure of Data Conversion System 100 According to the First Embodiment=== The processing procedure of the data conversion system 100 will be described with reference to Figure 9. Figure 9 is a flowchart showing the processing procedure of the data conversion system 100 according to the first embodiment. Note that in Figure 9, the content that overlaps with steps S10 to S18, which are an overview of the process described with reference to Figure 1, will be explained briefly.

[0110] In step S100, the data conversion system 100 obtains source information (see, for example, Figure 3(a)) and destination information (see, for example, Figure 3(b)) in the form of CSV files based on user input. The data conversion system 100 stores the source information and destination information in the conversion definition database D181. It is assumed that the data conversion system 100 obtains a structural analysis prompt at any time.

[0111] In Figure 9, the data conversion system 100 is shown to acquire source and destination information through the terminal device 200. However, the data conversion system 100 may also directly accept user input to acquire source and destination information.

[0112] In step S101, the data conversion system 100 inputs the structural analysis prompt and source information to the generating AI. The data conversion system 100 also inputs the structural analysis prompt and destination information to the generating AI.

[0113] In step S102, the data transformation system 100 obtains source schema information indicating the source data structure and destination schema information indicating the destination data structure from the generating AI. The data transformation system 100 stores the source schema information and destination schema information in the transformation definition database D181.

[0114] In step S103, the data conversion system 100 inputs the conversion analysis prompt, source schema information, and destination schema information into the generating AI.

[0115] In step S104, the data conversion system 100 obtains a conversion definition prompt (see, for example, Figure 2) from the generating AI. The data conversion system 100 stores the conversion definition prompt in the conversion definition database D181. The conversion definition prompt specifies the conversion process for each data item (column) of the CSV.

[0116] Note that the above procedure may be a processing procedure in which step S101 is not performed. In this case, for example, in step S100, a prompt containing the contents of the structural analysis prompt and the transformation analysis prompt, along with source information and destination information, may be input to the generating AI to obtain a transformation definition prompt.

[0117] In step S105, the data conversion system 100 receives a request for the input screen T10 (screen request) from the terminal device 200 and displays the input screen T10 on the display unit of the terminal device 200.

[0118] In step S106, the data conversion system 100 acquires input information from the terminal device 200. The data conversion system 100 stores the input information in the conversion definition database D181.

[0119] In step S107, the data conversion system 100 inputs the conversion definition prompt and the input information to the generating AI.

[0120] In step S108, the data conversion system 100 obtains output information from the generating AI. The data conversion system 100 stores the output information in the conversion definition database D181. The data conversion system 100 transmits the output information to the terminal device 200 so that it can be viewed.

[0121] In this way, the data conversion system 100 automatically identifies the relationships between data in each column of the CSV file and enables automatic and appropriate data conversion.

[0122] ===Data conversion system 100 according to the second embodiment=== Referring to Figure 10, the data conversion system 100a according to the second embodiment will be described. Figure 10 is a diagram showing an overview of the data conversion system 100a according to the second embodiment.

[0123] In the following, only the differences from the data conversion system 100 according to the first embodiment will be described. Unless otherwise specified, it is assumed that the system has the same configuration as the data conversion system 100 and achieves the same effects.

[0124] Compared to the data conversion system 100 according to the first embodiment, the data conversion system 100a has a configuration that allows the process of obtaining a conversion definition prompt from the generating AI to be omitted by inputting the source schema information and destination schema information obtained from the generating AI, along with a conversion analysis prompt, into the generating AI.

[0125] In other words, the data conversion system 100a has a configuration that allows it to obtain a conversion definition prompt from the generating AI by inputting, for example, a prompt containing the contents of a structural analysis prompt and a conversion analysis prompt (hereinafter referred to as the "first analysis prompt"), source information, and destination information to the generating AI. As a result, the data conversion system 100a can omit the process of inputting the conversion analysis prompt, source schema information, and destination schema information to the generating AI, thereby improving processing efficiency.

[0126] Furthermore, compared to the data conversion system 100 according to the first embodiment, the data conversion system 100a acquires output information obtained by converting input information into the data structure of the target information based on a conversion definition prompt, and when the output information is sent to the processing system 300, if the processing system 300 is unable to perform processing on the output information because the template has been changed in the processing system 300, the data conversion system 100a further includes a function that enables it to automatically convert the data structure of the output information to the template without human conversion work.

[0127] As a result, when a template is changed in the processing system 300, instead of the operator manually entering data into the changed template to create a CSV file, the data conversion system 100a can automatically generate output information that conforms to the changed format, thereby improving user convenience.

[0128] <<Process Overview>> First, with reference to Figure 10, an overview of the processing of the data conversion system 100a will be described. Below, the explanation of the processing in steps 10 and S11 in Figure 1 will be omitted, and the processing after these steps have been executed will be described.

[0129] In step S20, the data conversion system 100a inputs the first analysis prompt and the source information to the generating AI. The data conversion system 100a also inputs the first analysis prompt and the destination information to the generating AI.

[0130] The first analysis prompt is, for example, a prompt that causes the generating AI to generate a conversion definition prompt for converting source information into information having the data structure of destination information. For example, the first analysis prompt causes the AI ​​to identify source schema information and destination schema information by analyzing the data items, which are columns, and the data content associated with each of the data items in both the source and destination information, and generates a conversion definition for the generating AI to appropriately convert the data structure of the source information into the data structure of the destination. In other words, the first analysis prompt includes, for example, the contents of the "structure analysis prompt" and the "conversion analysis prompt" in the data conversion system 100 described above.

[0131] Specifically, the first analysis prompt states, "Identify the structure of all column information in the given CSV in the format shown in sample," and further, the format of the conversion definition is described as, for example, "The conversion definition is defined from the following three types for each column of the target CSV format," "1. Fixed value: Insert the fixed value defined in value for all rows of the definition column," "2. Column movement: Insert the value of the specified column of the source specified in source_field_id as is," "3. Generation: Insert the result of the process specified in instruction as a value, while referring to the value of the specified column (multiple can be specified) of the source specified in source_field_id."

[0132] Based on the first analysis prompt, the generating AI generates a conversion definition prompt that identifies, for example, as shown in Figure 2, that column B1 of the destination information is a fixed value, column B2 of the destination information is the result of combining the data in columns A1 and A2 of the source information and converting them to kana, column B3 of the destination information is the Japanese translation of column A4 of the source information, and column B4 of the destination information will receive column A3 of the source information as input.

[0133] In step S21, the data conversion system 100a obtains a conversion definition prompt from the generated AI based on the first analysis prompt, source information, and destination information.

[0134] In step S22, the data conversion system 100a inputs the conversion definition prompt and the input information to the generating AI.

[0135] In step S23, the data conversion system 100a obtains output information from the generating AI in which the data content of the input information is associated with a second data item.

[0136] In step S24, the data conversion system 100a transmits the output information to the processing system 300. The processing system 300 is a system of a business operator that performs predetermined processing based on the CSV output information. Predetermined processing is, for example, the process of listing products on an e-commerce (Electronic Commerce) site.

[0137] In this case, the processing system 300 may change the format of the CSV file in order to perform a predetermined process. In this case, the CSV output information output from the data conversion system 100a will not conform to the changed format, and therefore the predetermined process will not be performed in the processing system 300.

[0138] Therefore, in step S25, if the processing system 300 is unable to perform a predetermined process based on the output information, the data conversion system 100a obtains from the processing system 300 information indicating that the processing system 300 was unable to perform a predetermined process (hereinafter referred to as "error information") and header information of the changed format (hereinafter referred to as "modified header information").

[0139] The modified header information is a CSV file containing at least one data item (hereinafter referred to as the "third data item") that has a different data structure from the target data structure. The modified header information is information that clarifies what each column means (for example, name, address, age, etc.). In other words, the modified header information can be said to be an indicator for correctly mapping items in formats such as CSV or JSON.

[0140] In step S26, the data conversion system 100a inputs the second analysis prompt, the conversion definition prompt, and the converted header information to the generating AI. Alternatively, the data conversion system 100a may input the source information to the generating AI instead of, or together with, the conversion definition prompt.

[0141] The second analysis prompt is, for example, a prompt that indicates modifying the conversion definition prompt based on the modified header information. For example, the second analysis prompt is a prompt that causes the generation AI to analyze the content of the conversion definition prompt which converts the source information to the data structure of the converted information and the data structure of the third data item in the modified header information, and to generate a prompt (hereinafter referred to as the "re-conversion definition prompt") which causes the generation AI to appropriately convert the source information to the data structure of the modified header information.

[0142] In step S27, the data conversion system 100a obtains a re-conversion definition prompt from the generating AI to convert the source information into the data structure of the modified header information, based on the second analysis prompt, the conversion definition prompt, and the converted header information.

[0143] In step S28, the data conversion system 100a inputs the reconversion definition prompt and the input information to the generating AI.

[0144] In step S29, the data conversion system 100a obtains output information (hereinafter referred to as "re-output information") from the generating AI in which the data content of the input information is associated with the third data item of the converted header information.

[0145] In step S30, the data conversion system 100a transmits the re-output information to the processing system 300.

[0146] This allows the processing system 300 to perform predetermined processing on the re-output information. In other words, even if the format for performing predetermined processing in the processing system 300 is changed, the data conversion system 100a can automatically convert the format of the input information to the changed format and automatically send the re-output information with the content appropriately entered in that format to the processing system 300, thereby improving user convenience.

[0147] <<Structure>> Referring to Figure 10, the configuration of the data conversion system 100a will be described. As shown in Figure 10, the data conversion system 100a includes, for example, a source information processing unit 110, a destination information processing unit 120, an information transmission unit 130a, an input information acquisition unit 150, a conversion processing unit 160a, an information transmission / reception unit 170a, a storage unit 180, and a display processing unit 190a.

[0148] Note that the source information processing unit 110, destination information processing unit 120, input information acquisition unit 150, and storage unit 180 are the same as those in the data conversion system 100 according to the first embodiment, so their descriptions will be omitted.

[0149] The information transmission unit 130a transmits, for example, source information, destination information, and a first analysis prompt to the generating AI.

[0150] The conversion processing unit 160a generates a reconversion definition prompt (reconversion definition) using the generation AI. Specifically, the conversion processing unit 160a obtains the reconversion definition prompt from the generation AI by inputting the second analysis prompt, the conversion definition prompt, and the modified header information to the generation AI, for example, through the information transmission / reception unit 170a. Then, the conversion processing unit 160a inputs the reconversion definition prompt and the input information to the generation AI.

[0151] Here, with reference to Figure 11, we will explain the structure of the output information and the modified header information after the format has been changed. Figure 11 is a diagram showing an example of the differences between the structure of the output information and the modified header information.

[0152] The modified header information shown in Figure 11(d) has an added column X1 (the English name) between columns B2 and B3 of the destination information, compared to the output information (information shown in column B) shown in Figure 11(c). For example, the English name is entered in column X1. That is, the English name is entered in column X1 by concatenating the information in column A1 (surname) and column A2 (given name) of the source information shown in Figure 3(a).

[0153] In this case, since the converted header information has had columns added compared to the output information, the processing system 300 that acquired the output information determines that it cannot obtain the necessary information and does not perform the predetermined processing.

[0154] Therefore, based on the second analysis prompt, the conversion processing unit 160a causes the generation AI to identify data items in the conversion definition prompt that do not conform to the modified header information, and causes the generation AI to generate a re-conversion definition prompt for generating re-output information that conforms to the data structure of the modified header information based on the converted header information.

[0155] The re-conversion definition prompt P20 will be explained with reference to Figure 12. Note that only the differences from the conversion definition prompt P10 will be explained below. Figure 12 shows an example of a re-conversion definition prompt.

[0156] As shown in Figure 12, the re-conversion definition prompt P20 includes, for example, processing contents P21 to P24, which correspond to processing contents P11 to P14 of the conversion definition prompt P10, as well as processing contents P25. For example, processing contents P25 describes the process of inputting data into the "X1" column of the re-output information, where the phonetic readings of the data contents in column "A1" and the data contents in column "A2" of the input information are converted to English letters (Roman letters), and a half-width space is inserted between the last name and first name.

[0157] The following outlines the process by which processing content P25 is generated. First, the data conversion system 100a inputs the second analysis prompt, the conversion definition prompt, and the modified header information to the generation AI, causing the generation AI to analyze any inconsistencies between the data items in the modified header information and the conversion definition prompt. During the generation process of the re-conversion definition prompt P20, a structure is identified that indicates that the "X1" column (third data item) of the modified header information will contain the English letters of the first and last name (for example, "YAMADA TARO") (output data).

[0158] The data conversion system 100a then has the generation AI analyze the identified structure to obtain a re-conversion definition prompt in the "X1" column of the re-output information, which includes processing content P25 indicating that the English letters of the surname entered in column "A1" and the English letters of the given name entered in column "A2" of the input information are joined together with a half-width space in between. Note that the generation AI analyzes that processing content P21 to P24 does not need to be changed from the processing content P11 to P14 of the conversion definition prompt, so the content of processing content P11 to P14 is reflected in the re-conversion definition prompt without modification.

[0159] Thus, the reconversion definition prompt indicates the processing steps for converting the data structure of the source information to the data structure of the modified header information, column by column in a given data format.

[0160] In other words, when the format has been changed in the processing system 300, the data conversion system 100a appropriately identifies the relationship between each data item (column) in the input information and the changed header information based on the changed header information of the changed format, and enables appropriate data conversion according to the relationship between each data item, such as combining data related to the data items, inputting fixed values ​​into the data items, and converting the language of the data related to the data items. As a result, even when the format is frequently changed by the processing system 300, the data conversion system 100 enables simple, rapid, and inexpensive data conversion without requiring any work from the user.

[0161] The information transmission / reception unit 170a transmits and receives various types of information between the processing system 300 and the generating AI. Specifically, the information transmission / reception unit 170a transmits output information and re-output information to the processing system 300. The information transmission / reception unit 170a also obtains re-output information from the generating AI that associates a third data item (for example, the "English letters of name" item in column X1 of Figure 11) with output data related to the input data included in the input information (for example, "YAMADA TARO").

[0162] The display processing unit 190a may further generate, for example, a re-output confirmation screen T50 and a redefinition confirmation screen T60.

[0163] The re-output confirmation screen T50 will be described with reference to Figure 13. Figure 13 is a diagram showing an example of the re-output confirmation screen T50. As shown in Figure 13, the re-output confirmation screen T50 includes an error information display area T51, a re-output information display area T52, and a re-send button T53.

[0164] The error information display area T51 is an area that displays information indicating, for example, that error information has been obtained from the processing system 300. The re-output information display area T52 includes a confirmation object T52a that allows the user to check the contents of the re-output information. When the display processing unit 190a receives user input for the confirmation object T52a, it transitions the screen to the redefinition confirmation screen T60 shown in Figure 14, which will be described later. The re-send button T53 is a button for resending the created re-output information to the processing system 300.

[0165] In this way, the data conversion system 100 can quickly notify the user if processing fails in the processing system 300, and can also quickly execute the process again using the re-output information converted to the format of the processing system 300. This enhances user convenience for the data conversion system 100.

[0166] In the above description, the re-output confirmation screen T50 was explained as a screen on which the data conversion system 100a presents error information to the terminal device 200, but it is not limited to this. For example, if the processing system 300 is unable to perform a predetermined process, the processing system 300 may present error information to the terminal device 200 through a management screen (not shown) that it displays instead of the re-output confirmation screen T50. In this case, the processing system 300 displays a management screen on the terminal device 200 that allows the user to confirm the error information (for example, a screen that includes the error information display area T51 of the re-output confirmation screen), and also displays a button (hereinafter referred to as the "upload button") to cause the data conversion system 100a to generate the re-output information. Based on the user's input to the upload button, the processing system 300 uploads the error information and the modified header information to the data conversion system 100a. In other words, the terminal device 200 allows the user to check error information on a management screen provided by the processing system 300. When the processing system 300 receives user input for the upload button on the management screen, it sends the error information and modified header information (or modified header information only) to the data conversion system 100a. The upload button is, for example, a button that displays "Modify output information." The data conversion system 100a generates re-output information based on the uploaded modified header information.

[0167] The redefinition confirmation screen T60 will be described with reference to Figure 14. Figure 14 is a diagram showing an example of the redefinition confirmation screen T60. As shown in Figure 14, the redefinition confirmation screen T60 includes an output information display area T61 and a re-output information display area T62.

[0168] The output information display area T61 is an area that displays the data structure and data content of the output information before conversion.

[0169] The re-output information display area T62 is an area that displays the data structure and data content of the re-output information generated with the modified format structure. It is desirable that the re-output information display area T62 displays columns in a way that allows for comparison with the output information display area T61. This makes it easy for the user to understand which columns have been added or deleted.

[0170] Furthermore, the re-output information display area T62 may include a change identification object T62a that can identify which columns have been added or deleted from the data structure of the output information. The change identification object T62a may be made thicker by thickening the border of the added data item, or by coloring the area enclosed by the border, or any object that can be identified from other data items. In Figure 14, as an example, the re-output information has a change identification object T62a set to thicken the border surrounding the "English name" data item because the "English name" data item has been added from the output information. For example, if a format change is made in which a predetermined data item is deleted from among multiple data items included in the output information, the change identification object may be set to identify that predetermined data item in the output information. This allows the data conversion system 100a to easily confirm the format-converted data items, thereby improving user convenience.

[0171] <<Fifth variation>> The functions of the conversion processing unit 160a, the information transmission / reception unit 170a, and the display processing unit 190a in the data conversion system 100a may be added to the data conversion system 100 according to the first embodiment. That is, the data conversion system 100 according to the first embodiment may have a function to resend re-output information to the processing system 300 using the functions of the conversion processing unit 160a, the information transmission / reception unit 170a, and the display processing unit 190a described above when it transmits output information to the processing system 300 and receives error information from the processing system 300.

[0172] In this case, if the conversion processing unit 160a of the data conversion system 100 relating to the first embodiment obtains error information from the processing system 300, it obtains modified header information from the processing system 300. The information transmission unit 130 of the data conversion system 100 inputs the structure analysis prompt, the source information, and the modified header information to the generation AI. The schema information acquisition unit 140 obtains information indicating the data structure of the modified header information (hereinafter referred to as "modified schema information").

[0173] The modified schema information is, for example, the result of the generation AI analyzing the names of each data item. Specifically, the modified schema information includes, as an example, information indicating that the name of the data item in column X1 is "English letters of the name". Furthermore, the modified schema information also includes the result of analyzing the names of the data items, for example, information indicating that the data item in column X1 is entered by combining data related to the "surname" data item and data related to the "given name" data item and converting the resulting English letters.

[0174] The conversion processing unit 160a then inputs the source schema information and the modified schema information to the generating AI. The conversion processing unit 160 obtains a re-conversion definition prompt from the generating AI. The conversion processing unit 160a inputs the re-conversion definition prompt and the input information to the generating AI. The output information acquisition unit 170 obtains the re-output information from the generating AI.

[0175] Thus, by adding the functions of a conversion processing unit 160a, an information transmission / reception unit 170a, and a display processing unit 190a to the data conversion system 100 relating to the first embodiment, user convenience can also be improved.

[0176] ===Processing procedure of the data conversion system 100a according to the second embodiment=== The processing procedure of the data conversion system 100a will be described with reference to Figure 15. Figure 15 is a flowchart showing the processing procedure of the data conversion system 100a according to the second embodiment. Note that in Figure 15, content that overlaps with the content explained with reference to Figure 9 will be omitted or explained concisely.

[0177] In step S100, the data conversion system 100a acquires source information and destination information.

[0178] In step S101, the data conversion system 100a inputs the first analysis prompt, source information, and destination information to the generating AI.

[0179] In step S102, the data conversion system 100a obtains a conversion definition prompt from the generating AI.

[0180] In step S103, the data conversion system 100a acquires input information from the terminal device 200.

[0181] In step S104, the data conversion system 100a inputs the conversion definition prompt and the input information to the generating AI.

[0182] In step S105, the data conversion system 100a obtains output information from the generating AI. The data conversion system 100a transmits the output information to the processing system 300.

[0183] The data conversion system 100a may also transmit output information to the processing system 300 via the terminal device 200. In this case, for example, the data conversion system 100a transmits output information to the terminal device 200, and when the terminal device 200 receives user input for uploading output information on the management screen provided by the processing system 300 to the processing system 300, the terminal device 200 transmits the output information to the processing system 300.

[0184] In step S106, if the data conversion system 100a determines that the processing system 300 cannot perform a predetermined process based on the output information, it obtains error information and modified header information from the processing system 300. The data conversion system 100a stores the modified header information in the storage unit 180.

[0185] In step S106, a management screen (not shown) displaying error information provided by the processing system 300 may be displayed on the display unit of the terminal device 200, and the processing system 300 may transmit the modified header information to the data conversion system 100a by receiving user input for the upload button displayed on the management screen.

[0186] In step S107, the data conversion system 100a inputs the second analysis prompt, the conversion definition prompt, and the modified header information to the generating AI.

[0187] In step S108, the data conversion system 100a obtains a reconversion definition prompt from the generating AI.

[0188] In step S109, the data conversion system 100a inputs the reconversion definition prompt and the input information to the generating AI.

[0189] In step S110, the data conversion system 100a obtains the re-output information from the generating AI.

[0190] In step S111, the data conversion system 100a displays a re-output confirmation screen T50 (see Figure 3) regarding re-output information on the display unit of the terminal device 200.

[0191] In step S112, the data conversion system 100a obtains operation information from the terminal device 200 that indicates a user's operation input (for example, information indicating an operation input to select the "Resend button T53").

[0192] In step S113, the data conversion system 100a transmits re-output information to the processing system 300 based on the operation information.

[0193] Alternatively, instead of the processing in steps S112 and S113, for example, the data conversion system 100a may send the re-output information to the terminal device 200, and if the terminal device 200 receives user input for a button on the management screen provided by the processing system 300 to upload the re-output information to the processing system 300, the terminal device 200 may send the re-output information to the processing system 300.

[0194] As a result, even if the format for executing predetermined processing in the processing system 300 is changed, the data conversion system 100a can automatically send re-output information with the content appropriately entered into the changed format to the processing system 300 without requiring the user to manually generate re-output information, thereby improving user convenience.

[0195] ===Hardware Configuration=== Referring to Figure 16, an example of a hardware configuration when the data conversion system 100 is implemented using a computer will be described. Figure 16 is a diagram showing an example of a computer hardware configuration.

[0196] As shown in Figure 16, the computer 1000 includes a processor 1001, a memory 1002, a storage device 1003, an input I / F unit 1004, a data I / F unit 1005, a communication I / F unit 1006, and a display unit 1007.

[0197] The processor 1001 is a control unit that controls various processes in the computer 1000 by executing programs stored in the memory 1002.

[0198] Memory 1002 is a storage medium such as RAM (Random Access Memory). Memory 1002 temporarily stores the program code of the program executed by the processor 1001, as well as data required during program execution.

[0199] The storage device 1003 is a non-volatile storage medium such as a hard disk drive (HDD) or flash memory. The storage device 1003 stores the operating system and various programs for realizing the above configurations.

[0200] The input interface unit 1004 is a device for receiving input from the user. Specific examples of the input interface unit 1004 include keyboards, mice, touch panels, various sensors, and wearable devices. The input interface unit 1004 may be connected to the computer 1000 via an interface such as USB (Universal Serial Bus).

[0201] The data I / F unit 1005 is a device for inputting data from outside the computer 1000. Specific examples of the data I / F unit 1005 include drive devices for reading data stored on various storage media. The data I / F unit 1005 may also be located outside the computer 1000. In that case, the data I / F unit 1005 would be connected to the computer 1000 via an interface such as USB.

[0202] The communication interface unit 1006 is a device for performing data communication via the Internet N with external devices of the computer 1000, either via wired or wireless connection. The communication interface unit 1006 may also be located outside the computer 1000. In that case, the communication interface unit 1006 is connected to the computer 1000 via an interface such as USB.

[0203] The display unit 1007 is a device for displaying various types of information. Specific examples of the display unit 1007 include liquid crystal displays, organic EL (Electro-Luminescence) displays, and displays for wearable devices. The display unit 1007 may be located outside the computer 1000. In that case, the display unit 1007 is connected to the computer 1000 via, for example, a display cable. Furthermore, if a touch panel is used as the input I / F unit 1004, the display unit 1007 can be integrated with the input I / F unit 1004.

[0204] ===Summary=== <1> The data conversion system 100 in this embodiment includes an input information acquisition unit 150 that acquires input information including at least one first data item (e.g., a column) in a predetermined data format and input data (e.g., data entered into a column cell) which is associated with the first data item, and a conversion processing unit 160 that inputs source information including the first data item and source data which is associated with the first data item to a generating AI (large-scale language model), thereby acquiring source schema information indicating the data structure of the source information from the generating AI (large-scale language model), and destination information including at least one second data item in a predetermined data format and destination data which is associated with the second data item, which is then converted by the generating AI (large-scale language model). The data conversion system 100 includes: a conversion schema information indicating the data structure of the destination information, which is obtained from a generating AI (large-scale language model) by inputting it into the model; a conversion definition prompt (conversion definition) indicating the processing content for converting the source information into a data structure that includes a second data item of the destination information, which is obtained from the generating AI (large-scale language model) by inputting it into the generating AI (large-scale language model); a conversion processing unit 160 that inputs the input information and the conversion schema information to the generating AI (large-scale language model); and an output information acquisition unit 170 that acquires output information from the generating AI (large-scale language model) that associates the second data item, which has been analyzed by the generating AI (large-scale language model) based on the input information and the conversion definition prompt (conversion definition), with output data, which is data related to the input data contained in the input information. This enables data conversion to be performed simply, quickly, and inexpensively.

[0205] <2> Furthermore, in the data conversion system 100 of this embodiment, the predetermined data format is a data format that can be generated by spreadsheet software (e.g., Microsoft Excel) consisting of rows and columns (e.g., the format of a CSV file), and the conversion definition prompt (conversion definition) indicates the processing content for converting from the data structure of the source information to the data structure of the destination information on a column-by-column basis in the predetermined data format. <1> The data conversion system 100 described above. This allows the data conversion system 100 to appropriately identify the relationships between data items (columns) and to perform data conversion appropriately according to the relationships between data items, such as combining data related to data items, inputting fixed values ​​into data items, and converting the language of data related to data items. As a result, the data conversion system 100 enables data conversion simply, quickly, and inexpensively.

[0206] <3> Furthermore, in the data conversion system 100 of this embodiment, the conversion processing unit 160 inputs a conversion definition prompt (conversion definition) that sets a processing content (e.g., processing content P12) indicating the combination of input data related to a plurality of first data items for a predetermined column which is a second data item in a data format that can be generated by spreadsheet software (e.g., CSV file format), along with input information, to the generating AI (large-scale language model). <2> The data conversion system 100 described above. This allows the data conversion system 100 to appropriately identify the relationships between data items (columns) and to perform data conversion appropriately according to the relationships between data items, such as combining data related to data items, inputting fixed values ​​into data items, and converting the language of data related to data items. As a result, the data conversion system 100 enables data conversion simply, quickly, and inexpensively.

[0207] <4> Furthermore, the data conversion system 100 in this embodiment further includes a display processing unit 190 that displays an input reception area, which is an area for inputting input information into a generation AI (large-scale language model), on the display unit. <3> The data conversion system 100 described above. This enables the data conversion system 100 to perform data conversion easily and quickly.

[0208] <5> Furthermore, in the data conversion system 100 of this embodiment, the display processing unit 190 inputs first source information, including a first data item and first source data which is data associated with the first data item, into the generating AI (large-scale language model), thereby obtaining first source schema information which shows the data structure of the first source information, and inputs first destination information, including at least one second data item of a predetermined data format and first destination data which is data associated with the second data item, into the generating AI (large-scale language model), thereby obtaining first destination schema information which shows the data structure of the first destination information, and inputs these into the generating AI (large-scale language model), thereby obtaining first conversion definition which shows the processing content for converting the first source information into a data structure including a second data item of the first destination information, The display unit displays a screen with a selectable conversion definition selection area T11 (selection area) that includes: second source schema information indicating the data structure of the second source information, obtained from the generating AI (large-scale language model) by inputting second source information including a third data item and second source data which is data associated with the third data item into the generating AI (large-scale language model); second destination schema information indicating the data structure of the second destination information, obtained from the generating AI (large-scale language model) by inputting second destination information including at least one fourth data item of a predetermined data format and second destination data which is data associated with the fourth data item into the generating AI (large-scale language model); and second conversion definition indicating the processing content for converting the second source information to a data structure including the fourth data item of the second destination information, obtained from the generating AI (large-scale language model) by inputting second source information including a third data item and second source data which is data associated with the third data item into the generating AI (large-scale language model). <1> from <4> The data conversion system 100 is described in any one of the above. In this way, the data conversion system 100 provides the user with an object in which the user can select a desired conversion definition prompt from among multiple conversion definition prompts (conversion definitions).This allows the data conversion system 100 to enhance user convenience.

[0209] <6> Furthermore, in the data conversion system 100 of this embodiment, when input information is acquired by the input information acquisition unit 150, the conversion processing unit 160 identifies a recommended conversion definition, which is one of the first conversion definitions and the second conversion definition, based on at least one of the first data item of the input information or the input data, and the display processing unit 190 displays the recommended conversion definition on the display unit. <5> The data conversion system 100 described above. This allows the data conversion system 100 to present the user with an appropriate conversion definition corresponding to the input information when the user is unable to identify the desired conversion definition, thereby improving user convenience.

[0210] <7> Furthermore, the data conversion system 100 in this embodiment further includes a display processing unit 190 that displays a source information input area T41, which is an area for inputting source information into a generating AI (large-scale language model), and a destination information input area T43, which is an area for inputting destination information into a generating AI (large-scale language model). The display processing unit 190 displays a source information selection area T42, which allows selection of source information when source information is input into the source information input area T41, and a destination information selection area T44, which allows selection of destination information when destination information is input into the destination information input area T43. The conversion processing unit 160 displays the source information selection area T By inputting the selected source information, which is the information selected by the user from the source information displayed in 42, into the generating AI (large-scale language model), the generating AI (large-scale language model) obtains selected source schema information, which shows the data structure of the selected source information. Similarly, by inputting the selected destination information, which is the information selected by the user from the destination information displayed in the destination information selection area T44, into the generating AI (large-scale language model), the generating AI (large-scale language model) obtains selected destination schema information, which shows the data structure of the selected destination information. By inputting these into the generating AI (large-scale language model), a conversion definition prompt (conversion definition) is generated. As a result, the data conversion system 100 enables the user to arbitrarily generate various conversion definition prompts with simple operations, thereby enabling more appropriate data conversion.

[0211] <8> Furthermore, the data conversion system 100 further includes an information transmission / reception unit 170a that transmits output information to the processing system 300, and the conversion processing unit 160a acquires error information from the processing system 300 that acquired the output information, indicating that a predetermined process cannot be performed based on the output information, and modified header information including at least one third data item, and inputs the modified header information along with at least one of the source schema information and the conversion definition to the generating AI (large-scale language model), thereby obtaining the modified header information which shows the data structure of the modified header information obtained from the generating AI (large-scale language model). The system inputs schema information and a re-conversion definition, which indicates the processing content for converting source information into a data structure including a third data item of modified header information, into a generation AI (large-scale language model). The information transmission / reception unit 170a inputs this information and a re-output information, which associates the third data item, analyzed by the generation AI (large-scale language model) based on the input information and the re-conversion definition, with output data related to the input data contained in the input information. As a result, even if the format for executing a predetermined process in the processing system 300 is changed, the data conversion system 100 can automatically convert the format of the input information to the changed format and automatically send the re-output information with the content appropriately entered in that format to the processing system 300, thereby improving user convenience.

[0212] <9> Furthermore, the data conversion system 100a includes: an input information acquisition unit that acquires input information including at least one first data item in a predetermined data format and input data which is data associated with the first data item; a conversion processing unit that inputs source information including the first data item and source data which is data associated with the first data item, and destination information including at least one second data item in a predetermined data format and destination data which is data associated with the second data item, into a large-scale language model, thereby acquiring a conversion definition that indicates the processing content for converting the source information into a data structure including the second data item of the destination information, which is acquired from the large-scale language model; and the input information. This enables the data conversion system 100a to perform data conversion simply, quickly, and inexpensively.

[0213] <10> Furthermore, the data conversion system 100a further includes an information transmission / reception unit that transmits the output information to a processing system, and the conversion processing unit, upon obtaining error information from the processing system that has acquired the output information, indicating that it cannot perform a predetermined process based on the output information, obtains modified header information including at least one third data item from the processing system, inputs the source information and the conversion definition, along with the modified header information, into a large-scale language model, thereby obtaining a re-conversion definition from the large-scale language model that indicates the processing content for converting the source information into a data structure including the third data item of the modified header information, and inputs the input information, along with the input information, and the information transmission / reception unit obtains re-output information from the predetermined large-scale language model that associates the third data item, which has been analyzed by the predetermined large-scale language model based on the input information and the re-conversion definition, with the output data relating to the input data included in the input information. <9> The data conversion system described above. As a result, even if the format for executing a predetermined process in the processing system 300 is changed, the data conversion system 100a can automatically convert the format of the input information to the changed format and automatically send the re-output information with the content appropriately entered in that format to the processing system 300, thereby improving user convenience.

[0214] <11> Furthermore, the data conversion system 100a further includes a display processing unit that displays the second data item of the output information and the third data item of the re-output information in a comparable manner on the display unit. <10> The data conversion system described above. This allows the data conversion system 100a to easily verify the formatted data items, thereby improving user convenience. [Explanation of Symbols]

[0215] 100, 100a...Data conversion system, 110...Source information processing unit, 120...Destination information processing unit, 130, 130a...Information transmission unit, 140...Schema information acquisition unit, 150...Input information acquisition unit, 160, 160a...Conversion processing unit, 170...Output information acquisition unit, 170a...Information transmission / reception unit, 180...Storage unit, 190, 190a...Display processing unit, 200...Terminal device.

Claims

1. An input information acquisition unit that acquires input information including at least one first data item in a predetermined data format and input data which is data associated with the first data item, A conversion processing unit, By inputting source information, which includes the first data item and source data associated with the first data item, into a large-scale language model, source schema information indicating the data structure of the source information is obtained from the large-scale language model. By inputting destination information, which includes at least one second data item of the predetermined data format and destination data associated with the second data item, into the large-scale language model, destination schema information indicating the data structure of the destination information is obtained from the large-scale language model. By inputting the above large-scale language model, a conversion definition is obtained from the above large-scale language model, which shows the processing content for converting the source information into a data structure that includes the second data item of the destination information, The aforementioned input information and, A conversion processing unit that inputs the following into a predetermined large-scale language model, An output information acquisition unit acquires output information from the predetermined large-scale language model, which associates the second data item, analyzed by the predetermined large-scale language model based on the input information and the conversion definition, with output data, which is data relating to the input data included in the input information. An information processing system equipped with the following features.

2. The aforementioned predetermined data format is a data format consisting of rows and columns that can be generated by spreadsheet software. The conversion definition indicates the processing content for converting the data structure of the source information to the data structure of the destination information on a column-by-column basis in the predetermined data format. The information processing system according to claim 1.

3. The conversion processing unit inputs the conversion definition, which sets processing content indicating the combination of the input data related to a plurality of first data items for a predetermined column that is a second data item in a data format that can be generated by the spreadsheet software, and the input information into the predetermined large-scale language model. The information processing system according to claim 2.

4. The system further includes a display processing unit that displays an input reception area, which is an area for inputting the aforementioned input information into the predetermined large-scale language model, on a display unit. The information processing system according to claim 3.

5. The display processing unit, By inputting first source information, which includes the first data item and first source data associated with the first data item, into the large-scale language model, first source schema information, which indicates the data structure of the first source information, is obtained from the large-scale language model. By inputting first destination information, which includes at least one second data item in the predetermined data format and first destination data associated with the second data item, into the large-scale language model, first destination schema information indicating the data structure of the first destination information is obtained from the large-scale language model. By inputting the above large-scale language model, a first conversion definition is obtained from the above large-scale language model, which indicates the processing content for converting the first source information into a data structure that includes a second data item of the first destination information, By inputting second source information, which includes at least one third data item in the predetermined data format and second source data associated with the third data item, into a large-scale language model, a second source schema information is obtained from the large-scale language model, which shows the data structure of the second source information. By inputting second destination information, which includes at least one fourth data item in the predetermined data format and second destination data associated with the fourth data item, into the large-scale language model, a second destination schema information indicating the data structure of the second destination information is obtained from the large-scale language model. By inputting the above-mentioned large-scale language model, a second conversion definition is provided which indicates the processing content for converting the second source information obtained from the large-scale language model into a data structure that includes a fourth data item of the second destination information, The selectable selection area is displayed on the display unit on the same screen as the input reception area. The information processing system according to claim 4.

6. When the input information is acquired by the input information acquisition unit, the conversion processing unit identifies a recommended conversion definition, which is one of the first conversion definitions and the second conversion definition, based on at least one of the first data item of the input information or the input data. The display processing unit causes the recommended conversion definition to be displayed on the display unit. The information processing system according to claim 5.

7. The system further includes a display processing unit that displays a source information input area, which is an area for inputting the source information into the large-scale language model, and a destination information input area, which is an area for inputting the destination information into the large-scale language model. The display processing unit, When the source information is entered into the source information input area, a source information selection area is provided to display the source information in a selectable format. When the destination information is entered into the destination information input area, a destination information selection area is provided to display the destination information in a selectable format. Display and The aforementioned conversion processing unit By inputting the selected source information, which is the information selected by the user from the source information displayed in the source information selection area, into the large-scale language model, the selected source schema information, which indicates the data structure of the selected source information, is obtained from the large-scale language model. By inputting the selected destination information, which is the information selected by the user from among the destination information displayed in the destination information selection area, into the large-scale language model, the selected destination schema information, which indicates the data structure of the selected destination information, is obtained from the large-scale language model. By inputting the above large-scale language model, the above conversion definition is generated. The information processing system according to claim 1.

8. The system further comprises an information transmission / reception unit that transmits the output information to a processing system, The aforementioned conversion processing unit The processing system that acquired the output information outputs error information indicating that a predetermined process cannot be performed based on the output information, and modified header information including at least one third data item, At least one of the source schema information and the conversion definition, By inputting the modified header information into the large-scale language model, modified schema information indicating the data structure of the modified header information is obtained from the large-scale language model, A reconversion definition that indicates the processing content for converting the source information into a data structure including the third data item of the modified header information, obtained from the large-scale language model by inputting the source information into the large-scale language model, The aforementioned input information and, This is input into a predetermined large-scale language model, The information transmission and reception unit obtains re-output information from the predetermined large-scale language model, which associates the third data item, analyzed by the predetermined large-scale language model based on the input information and the re-conversion definition, with the output data relating to the input data included in the input information. The information processing system according to claim 1.

9. Computers To obtain input information including at least one first data item in a predetermined data format and input data which is data associated with the first data item, By inputting source information, which includes the first data item and source data associated with the first data item, into a large-scale language model, source schema information indicating the data structure of the source information is obtained from the large-scale language model. By inputting destination information, which includes at least one second data item of the predetermined data format and destination data associated with the second data item, into the large-scale language model, destination schema information indicating the data structure of the destination information is obtained from the large-scale language model. By inputting the above large-scale language model, a conversion definition is obtained from the above large-scale language model, which shows the processing content for converting the source information into a data structure that includes the second data item of the destination information, The aforementioned input information and, This involves inputting the data into a predetermined large-scale language model, To obtain output information from the predetermined large-scale language model that associates the second data item, which has been analyzed by the predetermined large-scale language model based on the input information and the conversion definition, with the output data, which is data relating to the input data included in the input information; An information processing method that performs the following.

10. On the computer, To obtain input information including at least one first data item in a predetermined data format and input data which is data associated with the first data item, By inputting source information, which includes the first data item and source data associated with the first data item, into a large-scale language model, source schema information indicating the data structure of the source information is obtained from the large-scale language model. By inputting destination information, which includes at least one second data item of the predetermined data format and destination data associated with the second data item, into the large-scale language model, destination schema information indicating the data structure of the destination information is obtained from the large-scale language model. By inputting the above large-scale language model, a conversion definition is obtained from the above large-scale language model, which shows the processing content for converting the source information into a data structure that includes the second data item of the destination information, The aforementioned input information and, This involves inputting the data into a predetermined large-scale language model, To obtain output information from the predetermined large-scale language model that associates the second data item, which has been analyzed by the predetermined large-scale language model based on the input information and the conversion definition, with the output data, which is data relating to the input data included in the input information; A program that executes the command.

Images