Information learning device, information learning method, and information learning program

Abstract

To efficiently train artificial intelligence using specific electronic documents. [Solution] An information learning device 10 comprising: an acquisition unit 12 that acquires an electronic document having predetermined attribute information; an extraction unit 14 that extracts text data from the electronic document acquired by the acquisition unit 12; and an AI control unit 15 that transmits a learning instruction to an artificial intelligence unit 20 to perform a learning process on a trained model owned by the artificial intelligence unit 20 using the text data.

Description

Technical Field

[0001] The present invention relates to an information learning device, an information learning method, and an information learning program.

Background Art

[0002] In the business of a company, etc., there are cases where responses, instructions, etc. are given to internal and external stakeholders of the company by e-mail, chat apps, etc. At this time, it becomes cumbersome if the same response, etc. is repeatedly given. Therefore, there is a need for a technology to easily manage the information included in responses, etc. as needed. In particular, it is convenient if the artificial intelligence can efficiently learn specific information such as responses, so that the artificial intelligence can give the same response on its behalf.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In Patent Document 1, the email content, etc. of a specific person is collected and analyzed, knowledge, skills, decision-making tendencies, etc. related to the business of that person are extracted, and a generative AI generates a persona. However, the email content, etc. of a specific person necessarily includes things that are not necessarily required for the generation of a persona.

[0005] An object of the present invention is to provide an information learning device that can efficiently cause a specific electronic document to be learned by an artificial intelligence.

Means for Solving the Problems

[0006] To achieve the above objective, an information learning device according to one aspect of the present invention comprises: an acquisition unit that acquires an electronic document having predetermined attribute information; an extraction unit that extracts text data from the electronic document acquired by the acquisition unit; and an AI control unit that transmits a learning instruction to an artificial intelligence unit to perform a learning process on a trained model owned by the artificial intelligence unit using the text data.

[0007] The attribute information may include information about the recipient of the electronic document.

[0008] The electronic document may be an email, and the predetermined attribute information may be a predetermined email address included in at least one of the TO, CC, and BCC fields of the email.

[0009] The electronic document may be a message in a chat tool application, and the predetermined attribute information may be predetermined account information that can view messages sent in the chat tool application.

[0010] The electronic document may be in HTML format, and the extraction unit may extract the main text data by excluding signature, header, or style information based on the HTML tags attached to the electronic document.

[0011] The electronic document may be an HTML-formatted email, and the extraction unit may detect citation-indicating tags or citation symbols in the email, exclude the citation portion, and extract the main text data.

[0012] The electronic document may include a predetermined HTML tag or a user-specified specific symbol, and the extraction unit may extract the main text data from within the area specified by the predetermined HTML tag or the specific symbol.

[0013] To achieve the above objective, an information learning method according to another aspect of the present invention is performed by a computer, which includes: an acquisition process for acquiring an electronic document having predetermined attribute information; an extraction process for extracting text data from the electronic document acquired by the acquisition process; and an AI control process for transmitting a learning instruction to an artificial intelligence unit to perform a learning process on a trained model owned by the artificial intelligence unit using the text data.

[0014] To achieve the above objective, an information learning program according to yet another aspect of the present invention causes a computer to execute an acquisition command for acquiring an electronic document having predetermined attribute information, an extraction command for extracting text data from the electronic document acquired by the acquisition command, and an AI control command for transmitting a learning instruction to an artificial intelligence unit to perform a learning process on a trained model owned by the artificial intelligence unit using the text data. Computer programs can be provided by storing them on various data-readable storage media, or by making them available for download via networks such as the Internet. [Effects of the Invention]

[0015] According to the present invention, specific electronic documents can be used to efficiently train artificial intelligence. [Brief explanation of the drawing]

[0016] [Figure 1] This figure shows the overall configuration and functional configuration of an information learning device according to an embodiment of the present invention. [Figure 2] This is a sequence diagram showing the processing flow executed by the information learning device described above. [Modes for carrying out the invention]

[0017] Hereinafter, embodiments of the information learning device according to the present invention will be described with reference to the drawings.

[0018] ●Information Learning System● The information learning system 1 is a system that gives an instruction to let artificial intelligence learn the text data of a specific electronic document specified by a user, so that the content instructed by the user to other users or appropriate information recorded by the user can be learned by the artificial intelligence. That is, the information learning system 1 enables the artificial intelligence to learn the responses of the user who is the object of learning by the artificial intelligence, that is, the target person, so that it can give responses similar to those of the target person.

[0019] As shown in FIG. 1, the information learning system 1 is configured such that an information learning device 10 and a user terminal 30 used by a user can communicate with each other via a network NW. The information learning device 10 may be constituted by a hardware device, or some or all of its functions may be realized by a cloud computer. Also, each component of the information learning device 10 may be realized by an API (Application Programming Interface). In this embodiment, the mutual communication between the information learning device 10 and the user terminal 30 is wireless, but part or all of the connections may be wired. Also, the information learning device 10 may be constituted by a plurality of hardware components. In this case, the plurality of hardware components may be connected by wire or wirelessly, and information may be transmitted and received between them.

[0020] Furthermore, the information learning device 10 is communicably connected to various databases via an appropriate network NW. For example, it is connected to an attribute information database DB1 and an in-company information database DB2. The contents of each database will be described later.

[0021] Also, the information learning device 10 is configured to be able to communicate with an artificial intelligence unit 20 having an AI (Artificial Intelligence) function. The information learning device 10 transmits an appropriate instruction to the artificial intelligence unit 20 and acquires the information output from the artificial intelligence unit 20.

[0022] ●User Terminal 30 The user terminal 30 is a terminal operated by the user, such as a personal computer, a tablet terminal, or a smartphone. The user terminal 30 mainly constitutes a functional block including a display unit 31, an input unit 32, and a communication processing unit 33 by a CPU (Central Processing Unit), a computer program executed by the CPU, a RAM (Random Access Memory) and a ROM (Read Only Memory) that store the computer program and predetermined data.

[0023] The display unit 31 is realized by a display or the like for outputting data. The display unit 31 displays, for example, a creation screen of a transmission email that the user sends to other users and a display screen of a received email received from other users, which are displayed by email software for sending and receiving emails. The email may be in an appropriate format such as HTML (Hyper Text Markup Language) format in addition to the text format.

[0024] Further, the display unit 31 may display a chat screen displayed by a so-called chat tool application in which the user sends and receives messages to and from one or more other users. The chat screen is configured such that, for example, a message input by the user operating the user terminal 30 and a message input by other users participating in the chat from an appropriate terminal are displayed on the same screen as a single chat room. Account information of users who can view the messages posted in the chat room is linked to the chat room. Also, the chat tool application may be a so-called SNS (Social Networking Service), and in this case, the messages may be so-called posts or comments on the posts. Further, the display unit 31 may display an electronically recorded document file. Further, the display unit 31 may display various digital contents such as video data and audio data.

[0025] The input unit 32 is a functional unit for inputting data, and is, for example, a keyboard or touch panel display that accepts text input. The input unit 32 may also be a microphone for picking up sound or a camera for capturing images or videos. The user uses the input unit 32 to, for example, specify the recipient of an email, specify a chat room to post the answer, and input body data including the answer.

[0026] The communication processing unit 33 is a processing unit that enables the transmission and reception of data with the information learning device 10 in accordance with a predetermined protocol via a network NW such as the Internet, and is implemented by an application or a web browser.

[0027] ● Information learning device 10 As shown in Figure 1, the information learning device 10 consists mainly of a display control unit 11, an acquisition unit 12, an information storage unit 13, an extraction unit 14, an AI control unit 15, and a communication processing unit 16, comprising a CPU (Central Processing Unit), a computer program executed by the CPU, and RAM (Random Access Memory) and ROM (Read Only Memory) for storing the computer program and predetermined data.

[0028] ●Display control unit 11 The display control unit 11 executes processing to display an appropriate screen on the user terminal 30.

[0029] ●Acquisition part 12 The acquisition unit 12 is a functional unit that acquires electronic documents having predetermined attribute information from a user terminal 30 connected to the network NW via the communication processing unit 16. The electronic documents to be acquired are all electronic data generated, stored, or transmitted and received by the user, and include, for example, emails, messages in chat tool applications, or electronically recorded document files. The electronic documents may also include various digital content such as video data and audio data. In this case, the acquisition unit 12 may acquire text data obtained from the video data or audio data through appropriate transcription processing. Note that the electronic documents to be acquired are not limited to these examples and may be any data format that can be recorded electronically.

[0030] The acquisition unit 12 is associated with a predetermined email address. This email address is one prepared for inputting electronic documents into the information learning device 10, and the acquisition unit 12 accesses the receiving mail server and receives emails that specify this email address as the recipient. The acquisition unit 12 may also be associated with predetermined account information in a chat tool application. The acquisition unit 12 uses this account information to log in to the chat tool application and acquires messages displayed on the screen that the account can view. The acquisition unit 12 may also acquire messages that are sent by specifying the account information associated with the acquisition unit 12, i.e., messages sent via mention, from among the messages viewable in the chat tool application. With this configuration, users can easily have the information learning device 10 acquire information simply by sending emails or messages to the email address or account information associated with the information learning device 10 in applications that they normally use to interact with other users. The electronic documents acquired by the acquisition unit 12 may be stored in the information storage unit 13.

[0031] ●Information storage section 13 The information storage unit 13 is a functional unit that controls the appropriate storage device provided by the information learning device 10 and performs data writing and reading. The information storage unit 13 stores electronic documents acquired by the acquisition unit 12 and text data extracted from electronic documents by the extraction unit 14 described later. It may also store part or all of the attribute information database DB1 or the company information database DB2.

[0032] ●Extraction part 14 The extraction unit 14 is a functional unit that extracts body data from electronic documents acquired by the acquisition unit 12. The extraction unit 14 performs appropriate extraction processing according to the type and format of the electronic document. For example, if the acquired electronic document is an email, the extraction unit 14 may perform extraction processing according to the format of the email. If the electronic document is in HTML format, the extraction unit 14 may identify the body data based on HTML tags, or it may identify parts other than the body data, such as signatures, headers, or style information, and extract the body data by excluding these. Furthermore, if an HTML email contains quotation suggestion tags or quotation marks, the extraction unit 14 may detect the quotation suggestion tags or quotation marks, exclude quoted portions from other emails, and extract body data only from the most recent sender's text. In addition, if the electronic document has predetermined HTML tags or user-specified specific symbols attached, the extraction unit 14 may extract body data only from the area enclosed by the predetermined HTML tags or specific symbols. With these configurations, only the important parts of the body can be targeted for extraction.

[0033] The extraction unit 14 may extract text data from electronic documents that have predetermined attribute information, and may not extract text data from electronic documents that do not have such attribute information. Attribute information is identification information attached separately from the text data that contains the actual content of the electronic document, and may include, for example, the sender and recipient of the electronic document, the creator, the date and time of creation, the date of transmission, tags, labels, etc.

[0034] More specifically, for example, attribute information in email is the email address entered in the TO, CC, BCC, or FROM fields. The attribute information recorded in the attribute information database DB1 includes the recipient information and the sender information of the data. More specifically, the recipient information is the email address entered in the TO, CC, and BCC fields, and the sender information is the email address entered in the FROM field.

[0035] Furthermore, attribute information in a chat tool application may include, for example, account information that identifies the user who sent the message, or identification information for the chat room where the message was posted. Attribute information may also include account information that allows access to messages sent to that chat room.

[0036] Furthermore, attribute information may be part of information stored in a predetermined field, such as the subject line of an email, the content of a message, or the tag field in a document file such as a Word file. In this case, the attribute information may be a label or tag such as "[For Learning]". The acquisition unit 12 acquires emails sent from a predetermined user from the sending mail server, and then the extraction unit 14 extracts messages that have the predetermined attribute information. With this configuration, even if the email address and other information included in the attribute information cannot be referenced on the spot, the data can be easily made into data to be extracted simply by making appropriate inputs in the predetermined fields.

[0037] The extraction unit 14 may refer to the attribute information database DB1 and extract body data from electronic documents containing attribute information specified in the attribute information database DB1. The attribute information database DB1 is a database in which the types of attribute information are pre-stored, and the attribute information to be extracted is specified in advance by appropriate administrators, including users.

[0038] In a configuration where the artificial intelligence unit 20 is trained on electronic documents identified based on an email address specified by the user, training data can be sent to the artificial intelligence unit 20 via an email application or chat tool application. Therefore, the artificial intelligence unit 20 can be easily trained by adding predetermined attribute information to the necessary information without having to set up a separate system for training processing. Furthermore, in a configuration where the artificial intelligence unit 20 is trained on the body data of an email simply by specifying a predetermined email address when sending an email, it is convenient because training instructions to the artificial intelligence unit 20 can be given at the same time as the task of conveying instructions to others. In addition, in a configuration where training content can be sent to the artificial intelligence unit 20 via email or a chat tool application, instructions and insights from daily work can be immediately entered into electronic documents as soon as they come to mind, making it easy to use.

[0039] Furthermore, if the extraction unit 14 is configured to extract body data only from electronic documents that include a predetermined sender in their attribute information, then even if an electronic document is mistakenly sent by another user, it is possible to reliably extract only the body data from the specific sender. In addition, if the electronic document is an appropriate document file, the extraction unit 14 can also be configured to extract the contents of the document file if the tag information attached to the document file contains predetermined content.

[0040] The extraction unit 14 may refer to information in the internal information database DB2. The internal information database DB2 includes personnel information, company calendars, meeting minutes, manuals, various contracts, and project management tools, and may also include other internal information as needed.

[0041] The extraction unit 14 may refer to the internal information database DB2 and identify a person corresponding to the attribute information contained in the electronic document. The extraction unit 14 may also extract personal information such as name, department, and job title for the identified person. This configuration allows for accurate understanding of the instructions and related individuals contained in the electronic document, thereby improving the accuracy of learning. In addition, the internal information database DB2 may also extract information such as schedules, tasks, and meeting contents related to the person. This configuration, by integrating personal information with surrounding information, clarifies the circumstances under which the electronic document was created, contributing to improved learning accuracy.

[0042] Furthermore, the extraction unit 14 may comprehensively analyze and extract the word structure and reference information of the text, and extract information that should be referenced together with the text data of the acquired electronic document. The extraction unit 14 may also perform natural language processing analysis on the content of the electronic document. Natural language processing analysis is performed by appropriate processing such as morphological analysis, syntactic analysis, or semantic analysis. For example, if the text is "For this inquiry, please refer to Chapter 3 of the manual when responding. Chapter 3 contains content particularly suitable for customers using this service for the first time, so we recommend referring to it in cases like this," then natural language processing will be used to analyze words such as "this inquiry," "Chapter 3 of the manual," "this service," "first-time users," "cases like this," and "reference," and the text data will be extracted. Alternatively, the text data may be extracted by structuring the relationship between instructions, target audience, reference materials, and advice based on the analyzed words.

[0043] The extraction unit 14 may extract the contents of any attached documents if they exist in the electronic document. If another document is cited in the electronic document but that document is not attached, the extraction unit 14 searches the internal information database DB2 or an external database to obtain the relevant document. Furthermore, if a specific part of a document is cited in the electronic document, the extraction unit 14 may extract that part.

[0044] Alternatively, the extraction process of the extraction unit 14 may be carried out in a manner in which the AI ​​control unit 15 (described later) transmits an extraction instruction to the artificial intelligence unit 20 via the communication processing unit 16 to extract the main text data, and receives the main text data from the artificial intelligence unit 20.

[0045] ● AI control unit 15 The AI ​​control unit 15 is primarily a functional unit that controls input to the artificial intelligence unit 20. The various instructions generated by the AI ​​control unit 15 are, for example, prompts, but are not limited to any instructions that the artificial intelligence unit 20 can interpret. For example, the AI ​​control unit 15 sends a learning instruction to the artificial intelligence unit 20 via the communication processing unit 16, which executes a learning process on a trained model that the artificial intelligence unit 20 possesses using text data. The AI ​​control unit 15 also receives information output from the artificial intelligence unit 20. The received information is stored in the information storage unit 13.

[0046] Furthermore, the AI ​​control unit 15 may specify the content of the learning instructions in detail. For example, the AI ​​control unit 15 may send a learning instruction to the artificial intelligence unit 20 via the communication processing unit 16, instructing it to learn the content of answers and instructions contained in the text data. In addition, the AI ​​control unit 15 may input the source information attached to the electronic document and the text data in association with each other, and instruct the artificial intelligence unit 20 to learn the content for each source information. With such a configuration, the artificial intelligence unit 20 can learn the response tendencies of each target person who is the sender of the electronic document.

[0047] Furthermore, the AI ​​control unit 15 may also associate the recipient information attached to the electronic document with the text data and input it to the artificial intelligence unit 20, instructing it to learn the content for each recipient of the information. With this configuration, the artificial intelligence unit 20 can learn the response tendencies and other factors transmitted to each recipient of the electronic document in a subdivided manner.

[0048] Furthermore, the AI ​​control unit 15 may also send a learning instruction to the artificial intelligence unit 20 via the communication processing unit 16, in which the AI ​​control unit 15 learns the content of instruction for each target person by referring to the person information contained in the text data, the past context of the situation in which the target person gave instructions, and the relationship with the recipient of the instructions. Furthermore, the AI ​​control unit 15 may also send a learning instruction to the artificial intelligence unit 20 via the communication processing unit 16, in which the AI ​​control unit 15 abstracts and generalizes the correspondence between the content of instructions selected by the target person under specific circumstances contained in the text data and conditions such as the circumstances and the attributes of the recipient of the instructions, in which the AI ​​control unit 15 learns the characteristics of instruction tendencies and instruction content. With these configurations, fragmented information for each target person can be used as instruction data for that person.

[0049] ● Artificial Intelligence Department 20 The Artificial Intelligence Unit 20 is an artificial intelligence (AI) equipped with language models such as transformers including BART (Bidirectional and Auto-regressive Transformer), BERT (Bidirectional Encoder Representations from Transformers), or GPT (Generative Pretrained Transformer, including GPT-1, GPT-2, GPT-3, GPT-4, etc., all registered trademarks), and particularly learning models such as Large Language Models (LLM). A learning model (also called a machine learning model) refers to a learning model based on a machine learning algorithm. Specific machine learning algorithms include nearest neighbors, naive Bayes, decision trees, and support vector machines. Deep learning, which uses neural networks to generate features and joint weighting coefficients for learning, is another example. The Artificial Intelligence Unit 20 can appropriately apply the above algorithms.

[0050] The artificial intelligence unit 20 has a pre-trained model that has been appropriately machine-learned. The artificial intelligence unit 20 has a pre-trained model for answering questions that takes questions as input and outputs answers to those questions. The artificial intelligence unit 20 may also have a pre-trained model for extracting data that takes electronic documents as input and outputs the body data of those electronic documents. The training data may be provided by an administrator or the like, or it may include information collected from the internet or elsewhere. The pre-trained model has obtained appropriate training data in advance and can also undergo additional training as needed.

[0051] The artificial intelligence unit 20 executes a learning process to train the pre-trained model for answering text data in response to learning instructions from the AI ​​control unit 15.

[0052] Furthermore, the artificial intelligence unit 20 may, in response to an extraction instruction from the AI ​​control unit 15, use a pre-trained model for extraction as input to the data acquired by the acquisition unit 12 and output the main text data.

[0053] Information output from the artificial intelligence unit 20 is transmitted to the AI ​​control unit 15 via the communication processing unit 16 and stored in the information storage unit 13.

[0054] The communication processing unit 16 is a processing unit that enables the transmission and reception of data with the user terminal 30 via a network NW such as the Internet, in accordance with a predetermined protocol. For example, the communication processing unit 16 receives data input from the user terminal 30.

[0055] According to the configuration described above, it is possible to efficiently train artificial intelligence on specific electronic documents.

[0056] Furthermore, the AI ​​control unit 15 may use the trained model for answering questions, as described above, to send an answer instruction to the artificial intelligence unit 20. In this case, the sender of the question, i.e., the questioner, may be different from the user who was trained in the artificial intelligence unit 20 via the information learning device 10. With this configuration, the questioner can obtain answers from the artificial intelligence unit 20 to questions that would otherwise have been asked of the user. Therefore, the user does not need to answer the same question multiple times, reducing their burden. Also, the questioner can conveniently obtain answers at any time without worrying about the user's convenience or time.

[0057] The AI ​​control unit 15 may input the questioner's information into the artificial intelligence unit 20 and send an instruction to output an answer tailored to the questioner. With this configuration, a more appropriate answer can be presented to the questioner.

[0058] ● Processing Flow Here, using Figure 2, we will explain an example of a processing flow for running a learning process on arbitrary data.

[0059] The user inputs and transmits data via the user terminal 30 (step S101). The information learning device 10 acquires the data using the acquisition unit 12 (step S102).

[0060] Next, the information learning device 10 performs the extraction process of the main text data using the extraction unit 14 (step S103). At this time, the extraction process by the extraction unit 14 may refer to the company's internal information database DB2. In addition, in step S103, the AI ​​control unit 15 may send an extraction instruction to the artificial intelligence unit 20 to perform the extraction process.

[0061] Next, the information learning device 10 transmits a learning instruction for the text data extracted by the extraction unit 14 to the artificial intelligence unit 20 via the communication processing unit 16 using the AI ​​control unit 15 (step S104). The artificial intelligence unit 20 then performs a learning process based on the text data using the trained model for responses in response to the learning instruction (step S105).

[0062] As described above, the information learning device according to the present invention makes it possible to efficiently train artificial intelligence on specific electronic documents. In the example described above, we explained how to train the artificial intelligence unit 20 using guidance provided via email or other means within the company. However, the information learning device according to the present invention is not limited to the accumulation of such guidance content, but can also be applied to knowledge sharing within the company.

[0063] Although the present invention has been described above using embodiments, the technical scope of the present invention is not limited to the scope described in the above embodiments, and various modifications and changes are possible within the scope of its gist. [Explanation of Symbols]

[0064] 1. Information Learning System 10 Information Learning Device 11 Display Control Unit 12 Acquisition Department 13 Information storage section 14 Extraction part 15 AI control section 16 Communication Processing Unit 20 Artificial Intelligence Department 30 User terminals DB1 Attribute Information Database DB2 Internal Information Database

Claims

1. An information learning device for learning emails sent to a predetermined email address, An acquisition unit that acquires the email, which includes the email address in the recipient information of the email; An extraction unit extracts body data from the email acquired by the acquisition unit, An AI control unit transmits a learning instruction to the artificial intelligence unit to perform a learning process on a trained model owned by the artificial intelligence unit using the text data, Equipped with, The acquisition unit acquires the email address entered in the FROM field of the email as the sender information of the email, The AI ​​control unit inputs the source information and the text data in association with each other and instructs the artificial intelligence unit to learn the content for each source information. Information learning device.

2. An information learning device for learning emails sent to a predetermined email address, An acquisition unit that acquires the email, which includes the email address in the recipient information of the email; An extraction unit extracts body data from the email acquired by the acquisition unit, An AI control unit transmits a learning instruction to the artificial intelligence unit to perform a learning process on a trained model owned by the artificial intelligence unit using the text data, Equipped with, The recipient information is the email address entered in the TO, CC, and BCC fields of the email. The AI ​​control unit inputs the destination information and the text data in association with each other into the artificial intelligence unit, and instructs the unit to learn the content for each piece of destination information. Information learning device.

3. The acquisition unit is associated with a predetermined account information in a chat tool application and acquires messages that can be viewed with said account information. The AI ​​control unit transmits a learning instruction to execute a learning process on the trained model using the message. The information learning device according to claim 1 or 2.

4. An information learning method for learning emails sent to a predetermined email address, A retrieval process to obtain the email containing the email address in the recipient information of the email, An extraction process for extracting body data from the email obtained by the acquisition process, AI control processing that transmits a learning instruction to the artificial intelligence unit to perform learning processing on a trained model held by the artificial intelligence unit using the text data, The computer executes this, In the acquisition process described above, the email address entered in the FROM field of the email is acquired as the sender information of the email. In the AI ​​control process, the source information and the text data are associated and input to the artificial intelligence unit, and the unit is instructed to learn the content for each source information. Information learning methods.

5. An information learning method for learning emails sent to a predetermined email address, A retrieval process to obtain the email containing the email address in the recipient information of the email, An extraction process for extracting body data from the email obtained by the acquisition process, AI control processing that transmits a learning instruction to the artificial intelligence unit to perform learning processing on a trained model held by the artificial intelligence unit using the text data, The computer executes this, The recipient information is the email address entered in the TO, CC, and BCC fields of the email. In the AI ​​control process, the destination information and the text data are associated and input to the artificial intelligence unit, and the AI ​​is instructed to learn the content for each piece of destination information. Information learning methods.

6. An information learning program that learns emails sent to a predetermined email address, A command to obtain the email containing the email address in the recipient information of the email, An extraction command for extracting body data from the email acquired by the acquisition command, An AI control command that transmits a learning instruction to the artificial intelligence unit to perform a learning process on a trained model held by the artificial intelligence unit using the text data, Have the computer run it, The acquisition command acquires the email address entered in the FROM field of the email as the sender information of the email, The AI ​​control command instructs the artificial intelligence unit to input the source information and the text data in association, and to learn the content for each piece of source information. Information learning program.

7. An information learning program that learns emails sent to a predetermined email address, A command to obtain the email containing the email address in the recipient information of the email, An extraction command for extracting body data from the email acquired by the acquisition command, An AI control command that transmits a learning instruction to the artificial intelligence unit to perform a learning process on a trained model held by the artificial intelligence unit using the text data, Have the computer run it, The recipient information is the email address entered in the TO, CC, and BCC fields of the email. The AI ​​control command instructs the artificial intelligence unit to input the destination information and the text data in association, and to learn the content for each piece of destination information. Information learning program.

Images