JP7686219B1 - Support device, support method, and support program - Google Patents

Abstract

A learning support system is provided that is adapted to a user's learning situation.
The support device 100 inputs information related to the study of a user who uses an electronic book for study into a large-scale language model, and generates information to support the user's study according to the user's study situation. The support device 100 outputs the generated information to support the user's study.
[Selected figure] Figure 2

Description

Article 30, paragraph 2 of the Patent Act applies. Website address https://admin.bookns.jp/ Published on December 25, 2023 Publisher SIE Co., Ltd.

The present invention relates to an assistance device, an assistance method, and an assistance program.

Learning support using Internet technology may be provided to learners (hereinafter, simply referred to as "users") studying languages, programming, qualification exams, etc. For example, there is known a conventional technology that identifies the lesson that a subject is to take from among multiple lessons that make up a course related to an exam, and outputs information indicating the learning effect of that lesson to the subject's terminal (see, for example, Patent Document 1).

JP 2020-166481 A

However, conventional technologies have problems in supporting learning according to the user's learning situation. For example, conventional technologies only show information indicating learning effectiveness based on the user's rate of correct answers to questions in assignments given after a lesson, so there are problems in providing learning support according to the learning situation of the user.

Therefore, in order to solve the above problems and achieve the objective, the support device of the present invention is characterized by having a generation unit that inputs information related to the learning of a user who uses an e-book for learning into a large-scale language model and generates information to support the learning of the user according to the user's learning situation, and an output unit that outputs the generated information to support the learning of the user.

The present invention has the effect of enabling learning support according to the user's learning situation.

FIG. 1 is a diagram showing an overview of learning support according to an embodiment. FIG. 2 is a diagram illustrating an example of the configuration of the support device according to the first embodiment. FIG. 3 is a table diagram illustrating an example of teaching material information according to the first embodiment. FIG. 4 is a table illustrating an example of user learning information according to the first embodiment. FIG. 5 is a diagram illustrating an example of a reception screen according to the first embodiment. FIG. 6 is a diagram showing a series of flows of the generation process and output process according to the first embodiment. FIG. 7 is a diagram illustrating an example of a prompt according to the first example. FIG. 8 is a diagram illustrating an example of a prompt according to the first example. FIG. 9 is a diagram showing an example of learning support information output in a chat format according to the first example. FIG. 10 is a diagram illustrating an example of a prompt according to the second example. FIG. 11 is a diagram showing an example of learning support information output in a chat format according to the second example. FIG. 12 is a diagram showing an example of a prompt used for coaching optimization according to the second example. FIG. 13 is a diagram illustrating an example of a prompt according to the third example. FIG. 14 is a diagram showing an example of learning support information output in a chat format according to the third example. FIG. 15 is a flowchart of the process of the support device according to the first embodiment. FIG. 16 is a diagram illustrating an example of the configuration of the support device according to the second embodiment. FIG. 17 is a table diagram illustrating an example of result information according to the second embodiment. FIG. 18 is a diagram illustrating an example of a process of the support device according to the second embodiment. FIG. 19 is a diagram showing an example of a hacking practice screen according to the second embodiment. FIG. 20 is a flowchart of the process of the support device according to the second embodiment. FIG. 21 is a diagram illustrating an example of a hardware configuration of a computer that realizes the support device according to the embodiment.

Below, the form for carrying out the present invention (hereinafter, "embodiments") will be described with reference to the drawings. Note that each embodiment is not limited to the contents described below.

<Introduction>
(background)
With the recent development of Internet technology, users who study languages, programming, qualification exams, etc. can study online. In online learning, various learning support may be provided according to the user's learning situation. For example, a reference technology is known that outputs the learning effect of a course to a subject based on the user's working status on the assignments of multiple lessons that make up the course related to an exam.

However, the above-mentioned reference technology only shows information indicating the learning effect based on the user's correct answer rate for course assignments. Therefore, there are problems with providing learning support according to the learning situation of the user.

(Processing by the Support Device 100)
Therefore, the support device 100 according to the present embodiment provides appropriate learning support, such as teaching, coaching, setting questions suited to the user, and follow-up, according to the user's learning situation.

Here, an overview of the processing performed by the support device 100 will be described. FIG. 1 is a diagram showing an overview of learning support according to an embodiment. The support device 100 shown in FIG. 1 is an example of a computer that provides technology for realizing the information processing described below.

In this embodiment, a situation will be described in which the support device 100 provides a specific learning support to a user who uses an electronic book for learning. The "electronic book" mentioned above refers to a specific learning material provided as digital content, and includes, for example, the digitized book itself, and still images or videos that explain the book.

The support device 100 inputs information about the learning situation of the user 10 who is studying using an electronic book ("Learning situation of the user 10" shown in (1-1) of FIG. 1) into a large-scale language model as prior knowledge ((1-2) of FIG. 1). Then, based on the learning situation of the user 10 input as prior knowledge, the support device 10 teaches and coaches the user 10, answers questions from the user 10, asks questions, etc.

For example, the support device 100 creates a study plan for the user 10 to study efficiently, manages the progress of the study based on the study plan, and gives encouragement or reprimand to the user 10, depending on the learning situation of the user 10 ("Coaching" shown in (2-1) of FIG. 1). Based on the learning situation of the user 10, the support device 100 also asks practice questions according to the academic level and learning progress of the user 10, and follows up on the answers to the practice questions ("Teaching" shown in (2-2) of FIG. 1). On the other hand, the support device 100 outputs answers to "Questions ((2-3) of FIG. 1)" input by the user 10, depending on the learning situation of the user 10 and the electronic books used by the user 10 ("Answer" shown in (2-4) of FIG. 1).

The support device 100 can realize the above-mentioned "coaching (FIG. 1 (2-1))," "teaching (FIG. 1 (2-2))," and "answering (FIG. 1 (2-4))" in a natural language dialogue format (chat format), for example, as shown in FIG. 1 (3).

In this way, the support device 100 according to this embodiment has the effect of enabling support for learning according to the user's learning situation by providing teaching, coaching, answering questions from the user, etc., according to the user's learning situation.

First Embodiment
From here, a first embodiment realized by the support device 100 will be described. In the first embodiment, the support device 100 generates coaching to support the user's learning progress, teaching to present practice questions generated according to the user, answers to questions from the user, and the like, in accordance with the user's learning situation, and outputs them to the user.

(Support device 100)
First, the support device 100 according to the first embodiment will be described in detail. Fig. 2 is a diagram showing an example of the configuration of the support device 100 according to the first embodiment. As shown in Fig. 2, the support device 100 according to the first embodiment includes a communication unit 110, a storage unit 120, and a control unit 130.

Although not shown in FIG. 2, the support device 100 may include an input unit such as a keyboard or a mouse for receiving input of operations by an administrator or the like. The support device 100 may also include a display unit such as a display for displaying to the administrator the teaching material information, user learning information, etc. stored in the storage unit 120.

(Communication unit 110)
The communication unit 110 performs data communication related to the input of information relating to the user's learning, including information on the user's learning status when studying using an e-book (hereinafter, this may be referred to as "user learning information"), natural language conversations related to teaching and coaching input by the user, answers to practice problems, questions, etc., and output of generated information to support the user's learning (hereinafter, this may be referred to as "learning support information") to a terminal device, etc.

The communication unit 110 is realized, for example, by a NIC (Network Interface Card) or a network interface controller. The communication unit 110 is connected to a network (e.g., the Internet) by wire or wirelessly. The communication unit 110 transmits and receives information to and from external devices via the network.

The communication unit 110 may transmit and receive information using any communication standard or technology, such as Wi-Fi (registered trademark), Bluetooth (registered trademark), SIM (Subscriber Identity Module), or LPWA (Low Power Wide Area).

(Storage unit 120)
The storage unit 120 stores data and programs used for various processes by the control unit 130, and various data acquired by the operation of the control unit 130. The storage unit 120 is realized by, for example, a semiconductor memory element such as a random access memory (RAM) or a flash memory, or a storage device such as a hard disk or an optical disk. As shown in FIG. 2, the storage unit 120 has a teaching material information DB 121, a user learning information DB 122, and a generative model DB 123.

(Educational material information DB121)
The teaching material information DB 121 is a database that stores teaching materials such as e-books used by users for studying. Specifically, the teaching material information DB 121 stores, as teaching material information, the titles of e-books and the like used by users for studying (titles), link information (links) for accessing e-books stored in a predetermined storage device (links), information (order) indicating the display order when displaying a list of teaching materials (order), information (number of contents) indicating the number of contents of a predetermined granularity in the e-book, information (tasks) regarding practice problems associated with the e-books, information (types) indicating the publication type of the teaching materials, and the like.

Here, an example of teaching material information stored by the teaching material information DB 121 will be described. FIG. 3 is a table diagram showing an example of teaching material information according to the first embodiment. The teaching material information DB 121 stores items such as "title", "link", "order", "number of contents", "assignment", and "type" and information related to the items in a table format or the like, in association with "No.", which is information that identifies individual teaching material information.

For example, as shown in FIG. 3, the teaching material information DB 121 stores a title "A", a link "B", a sequence "C", a number of contents "D", an assignment "E", and a type "F", each of which is identified by a No. "1", in association with each other. Note that the above alphabets "A to F" are information showing a legend for each item of individual teaching material information stored in the teaching material information DB 121, and the information actually stored is not particularly limited.

In addition, using the teaching material information stored in the teaching material information DB 121, the support device 100 can display to the user, administrator, etc., in a list format including items such as "type," "thumbnail," "title," "order," "chapter," "section," and "task," as shown in FIG. 3 (1).

The above "title" is information that identifies the e-book used by the user for study, and includes, for example, the title, volume number, issue number, and other information of the e-book expressed in natural language such as text, numbers, and symbols. "Link" is information for accessing the storage medium that stores the e-book itself or thumbnail information of the e-book, and includes, for example, information such as a Uniform Resource Locator (URL) associated with each e-book. "Order" is information used to determine the display order when displaying a list of e-books that are teaching materials.

"Number of contents" is the number of contents organized at a specific granularity contained in an e-book, and includes, for example, the "number of chapters (chapters)" and the "number of sections (sections)". "Assignments" correspond to "chapters" and "sections" in an e-book, and are information about assignments such as practice questions to be given to users, and include information such as the content and correct answers of the practice questions, the number of practice questions, etc. "Type" is a status that indicates the public status of an e-book, and includes, for example, "public" that can be used by all registered users, and "limited" that can be used only by specific users, etc.

(User learning information DB 122)
The user learning information DB 122 is a database that stores information (user learning information) related to a user's learning, including information such as the learning situation of a user who uses an electronic book to learn. Specifically, the user learning information DB 122 stores information for identifying a user (user identification information), learning materials used by the user for learning (used learning materials), a learning history of the user (learning history), a learning goal of the user (learning objective), a conversation history between the user and the support device 100 (conversation history), and the like as the user learning information.

Here, an example of user learning information stored by the user learning information DB 122 will be described. FIG. 4 is a table diagram showing an example of user learning information according to the first embodiment. The user learning information DB 122 stores items such as "user identification information," "teaching materials used," "learning history," "learning goals," and "conversation history" and information related to the items in a table format or the like, in association with "No.", which is information that identifies individual user learning information.

For example, as shown in FIG. 4, the user learning information DB122 stores the user identification information "User 10" identified by No. "1", the teaching material used "G", the learning history "H", the learning goal "I", and the conversation history "J" in association with each other. Note that the above alphabets "G to J" are information showing a legend for each item of individual user learning information stored by the user learning information DB122, and the information actually stored is not particularly limited.

In addition, using the user learning information stored in the user learning information DB 122, the support device 100 can display to the user, administrator, etc., in a list format including items such as "chapter," "section," "task," "type," "correct answer rate," and "result," as shown in (1) of FIG. 4.

The above "user identification information" refers to information that identifies an individual user who uses an e-book for study, and includes, for example, user-specific identification information such as the user's name or nickname, telephone number, membership number for a service, etc. The user study information DB122 can store user identification information while protecting personal information by deleting or replacing information that may identify an individual user with obscured characters, etc., based on known technology.

"Teaching materials used" is information for identifying the teaching materials used by the user for studying, and includes, for example, information identifying the "title" of teaching material information stored in the teaching material information DB 121. "Study history" is historical information about the user's study using an e-book, and includes, for example, information about study dates, study time, and study progress. "Study goals" is information about the study objectives set by the user studying using an e-book, and includes, for example, goals such as "complete study of content XX by XX month" or "study XX hours per week." "Conversation history" is historical information (past conversation data) of conversations in natural language between the user studying and the support device 100.

(Generative model DB 123)
The generation model DB 123 is a database that stores a predetermined generation model used for generating learning support information for a user by the generation unit 132 described later. For example, the generation model DB 123 can store a large-scale language model or the like as the generation model. Specifically, the support device 100 can use, as the large-scale language model, "ChatGPT (registered trademark)", which is a large-scale language model having general-purpose knowledge (for example, see Reference 1).

(Reference 1): ChatGPT (OpenAI), <URL: https://openai.com/chatgpt>, <Retrieved on September 3, 2020>

(Control unit 130)
Here, the explanation will be continued by returning to Fig. 2. The control unit 130 is realized by a processor, a micro processing unit (MPU), a central processing unit (CPU), or the like executing various programs stored in the storage unit 120 using the RAM as a working area.

The control unit 130 is realized by an integrated circuit such as an ASIC (Application Specific Integrated Circuit) or an FPGA (Field Programmable Gate Array). As shown in FIG. 2, the control unit 130 has a reception unit 131, a generation unit 132, and an output unit 133.

(Reception Unit 131)
The receiving unit 131 receives information such as conversations related to teaching or coaching, answers to practice questions, questions, etc., input by the user in a chat format or the like. The receiving unit 131 also receives information used by the generating unit 132 (described later) to generate practice questions.

Now, with reference to FIG. 5, an example of a reception screen for information used to generate practice questions and answers to the generated practice questions by the reception unit 131 will be described. FIG. 5 is a diagram showing an example of a reception screen according to the first embodiment.

As shown in FIG. 5, the reception unit 131 receives conditions for generating practice questions from an administrator or the like via a reception screen for conditions for generating practice questions shown in FIG. 5 (1). For example, the reception unit 131 receives conditions for which content of the teaching materials, such as "teaching material," "chapter," or "section," a practice question is to be generated for (FIG. 5 (1-1)). The reception unit 131 can also display a preview screen of the content identified based on the conditions shown in FIG. 5 (1-1) (FIG. 5 (1-2)). The reception unit 131 further receives conditions such as "number of options," "number of questions to be created," etc. (FIG. 5 (1-3)).

The reception unit 131 can then receive answers from the user to the practice questions posed via the practice question screen shown in FIG. 5 (2). Note that the reception unit 131 can also receive answers to practice questions in a chat format, in addition to the practice question screen (FIG. 5 (2)).

(Generation unit 132)
The generating unit 132 inputs user learning information of a user who uses an electronic book for learning into the large-scale language model, and generates learning support information for the user according to the user's learning situation. Specifically, the generating unit 132 inputs user learning information including at least one of an electronic book used by the user, the user's learning history, the user's learning goal, and past conversation data into the large-scale language model, and generates learning support information for the user.

Specifically, the generation unit 132 generates, as the user's learning support information, messages to support the user's learning (hereinafter, "learning support messages"), answers to questions, practice questions to be presented to the user, and information to follow up on the answers to the practice questions. The generation process by the generation unit 132 will be described in the following section using specific prompt examples.

(Output unit 133)
The output unit 133 outputs the generated learning support information for the user. For example, the output unit 133 outputs the generated learning support information for the user to the user in a chat format. Note that the output process by the output unit 133 will be described in the following section by taking as an example the output of information generated using a specific prompt example.

(An example of the overall process flow)
From here, an example of generation and output of learning support information by the support device 100 will be described. First, a flow of generation and output of learning support information by the support device 100 will be described with reference to Fig. 6. Fig. 6 is a diagram showing a series of flows of generation processing and output processing according to the first embodiment. Fig. 6 shows a terminal device 200 operated by a user, and the support device 100 that performs generation processing and output processing of the user's learning support information.

The terminal device 200 transmits input information related to learning, such as "coaching/teaching consultation" and "answers to practice questions" received from the user, to the support device 100 ((1) in Figure 6).

The support device 100 performs a process of generating learning support information for a user in response to input information related to learning received from the terminal device 200. First, the support device 100 (generation unit) generates a prompt including an instruction to assign a predetermined role to the large-scale language model, an instruction to input user learning information, and an instruction to generate learning support information for the user ((2-1) in FIG. 6). Next, the support device 100 (generation unit) inputs the generated prompt to the large-scale language model ((2-2) in FIG. 6). Then, the support device 100 (generation unit) generates learning support information for the user based on the generated prompt.

Specifically, the support device 100 (generation unit) inputs the prompt into the large-scale language model and sets a predetermined role, such as "AI instructor who supports the user." In addition, the support device 100 (generation unit) causes the large-scale language model to acquire user learning information from the user learning information DB 122 ((3-1) in FIG. 6).

Then, the support device 100 (generation unit) uses the acquired user learning information to generate learning support information in the large-scale language model ((3-2) in FIG. 6). For example, the support device 100 (generation unit) causes the large-scale language model to generate "learning support messages ((3-3) in FIG. 6)" and "answers to questions ((3-4) in FIG. 6)" as learning support information related to coaching. The support device 100 (generation unit) also causes the large-scale language model to generate "practice questions ((3-5) in FIG. 6)" and "answer grading and follow-up ((3-6) in FIG. 6)" as learning support information related to teaching.

The support device 100 (output unit) outputs the generated learning support information to the terminal device 200 ((4) in FIG. 6). As a result, the terminal device 200 can display information such as "learning support messages," "answers to questions from the user," and "practice questions and answer information" to the user as learning support information ((5) in FIG. 6).

From here, an example of the generation and output process of learning support information by the support device 100 will be explained using more specific prompts and output screens as examples, with reference to Figures 7 to 14.

(First Example: Generation and Output of Learning Support Message)
First, an example will be described in which the support device 100 generates a "learning support message" as learning support information and outputs it to the user. In the first example, the support device 100 generates a "learning support message" that is a message for supporting the user to effectively progress in learning according to the user's learning situation, and outputs it to the user.

In specific processing, the support device 100 (generation unit) inputs a command to generate a user learning support message that includes at least one of words of encouragement to the user according to the user's learning progress, a summary of the user's learning content, and a message to confirm whether the user has any questions into the large-scale language model, and generates the user's learning support message as user learning support information.

Here, examples of input prompts and output examples in the process according to the first example will be described with reference to Figs. 7 and 8. Figs. 7 and 8 are diagrams showing examples of prompts according to the first example.

Figure 7 shows an example of a prompt (Figure 7 (1)) for causing a large-scale language model to generate a learning support message according to the user's learning situation, and an example of a learning support message (Figure 7 (2)) that is generated and output based on the prompt. The prompt shown in Figure 7 (1) includes "#Background information (Figure 7 (1-1))," "#Status (Figure 7 (1-2))," and "#Command (Figure 7 (1-3))."

"#Background information" is a command for assigning a role desired by an administrator or the like to a large-scale language model. For example, by using the command shown in (1-1) of FIG. 7, the assistance device 100 (generation unit) can assign the role of "being an AI instructor who supports users who are learning," "having a role to increase the user's motivation for continued learning," and "generating information according to the user's learning status (#status) described in the prompt (#command)" to the large-scale language model.

"#Status" is information expressing the user's learning status in natural language, and is written based on the information stored in the user learning information DB 122. For example, the information shown in (1-2) of FIG. 7 means that "User A" has "studied up to topic <1>."

The "# command" is a command to cause the large-scale language model to execute a specified process. For example, the command shown in (1-3) of FIG. 7 allows the support device 100 (generation unit) to cause the large-scale language model to execute the following: "Generate a message to encourage the user to continue studying similar to the previous day, depending on the user's study status," "Generate a message to encourage the user," "Summarize the content of the previous day's study," "Generate a message to check whether the user has any questions," etc.

Then, the support device 100 (output unit) outputs the learning support message generated based on the prompt shown in FIG. 7 (1), as shown in FIG. 7 (2).

Next, an example of generating a learning plan in a large-scale language model according to the user's learning situation will be described with reference to FIG. 8. FIG. 8 shows an example of a prompt ((1) in FIG. 8) for causing a large-scale language model to generate a learning plan in accordance with the user's learning situation, and an example of a learning plan generated and output based on the prompt ((2) in FIG. 8).

The prompt shown in FIG. 8 (1) includes "#Background information", "#Command (FIG. 8 (1-1))", "#Constraint (FIG. 8 (1-2))", "#Functions that can be used for learning (FIG. 8 (1-3))", "#Individual task data (FIG. 8 (1-4))", and "#Learning progress data (FIG. 8 (1-5))". Note that the "#Background information" is processed in the same way as in FIG. 7, based on the command in FIG. 8, so a description is omitted.

The "# command" is a command to cause the large-scale language model to execute a specified process, as in the example of FIG. 7. For example, the command shown in FIG. 8 (1-1) can cause the support device 100 (generation unit) to cause the large-scale language model to execute operations such as "comparing the user's personal task data with learning progress data while strictly adhering to set constraint conditions, and generating advice for the user's learning progress."

"#Constraint" is an instruction for specifying (restricting) the method of executing processing by a large-scale language model. For example, the instruction shown in (1-2) of FIG. 8 allows the support device 100 (generation unit) to impose constraints on the large-scale language model when generating a study plan, such as "generating a study plan using only functions that can be used for study," "limiting the number of characters," "modifying the study schedule," and "generating sentences to communicate the study plan to the user."

"#Functions usable for learning" is information indicating functions related to learning that can be used by a user when studying using the services realized by the support device 100, and is used as a constraint condition when the large-scale language model generates a study plan. For example, as shown in (1-3) of FIG. 8, "#Functions usable for learning" includes information identifying various functions to be used for learning, such as "e-book/video viewing function," "problem practice function," "question function," "interpersonal coaching function," and "remote lab function."

"#PersonalAssignmentData" includes information related to the user's learning goals. The information included in "#PersonalAssignmentData" may be user learning information stored in the user learning information DB 122. For example, as shown in (1-4) of FIG. 8, "#PersonalAssignmentData" includes information such as "study start date," "study completion deadline," "study goals using e-books (e-book goals), learning goals using videos (video goals), and learning goals using questions (question goals)."

"#Study progress data" includes information related to the user's learning progress. The information included in "#Study progress data" may be user learning information stored in the user learning information DB 122. For example, as shown in (1-5) of FIG. 8, "#Study progress data" includes information such as "current date," "study progress using e-books (e-book progress), learning progress using videos (video progress), and learning progress using questions (question progress)."

Then, the support device 100 (output unit) outputs a learning support message generated based on the prompt shown in FIG. 8 (1), as shown in FIG. 8 (2). For example, the message shown in FIG. 8 (2) includes a "current situation analysis," "necessary actions," "short-term goals," "suggestions for support functions," "suggestions for reviewing the schedule," etc. (FIG. 8 (2-1)), based on a comparison between "#individual task data" and "#learning progress data," as well as a message to the user (FIG. 8 (2-2)).

For example, "Current situation analysis" includes messages such as "As of XX/XX/XX, there are XX months left until the deadline. Progress in learning using e-books is XX%, progress in learning using videos is XX%, and progress in learning using questions is XX%." "Necessary actions" include messages such as "XX% is remaining until you reach your goal progress. You need to advance by XX% each day. As a recommended method, make full use of the e-book viewing function and set a goal of reading at least XX chapters each day."

"Short-term goals" include "Progress in learning using e-books to XX%, "Progress in learning using videos to XX%, "Progress in learning using questions to XX%, etc. " "Support function suggestions" include "If there's anything you don't understand, use the question function to clarify it. Furthermore, if you are faced with an individual challenge, it may be a good idea to consider using the interpersonal coaching function to receive direct advice." "Schedule review suggestions" include "Considering the current pace and estimated pace of progress, it is highly likely that it will be difficult to achieve the deadline, especially with regard to the e-book portion. Therefore, please consider postponing the deadline to XX month."

The message to the user (Figure 8 (2-2)) includes, for example, "Please proceed with your daily studies according to this plan, regularly evaluate your progress, and make adjustments as necessary."

Here, the "example of output in chat format" described in Figures 7 and 8 will be described with reference to Figure 9. Figure 9 is a diagram showing an example of output in chat format of learning support information relating to the first example.

Figure 9 shows a chat-style screen for outputting learning support messages and learning plans generated by a large-scale language model. For example, the support device 100 (output unit) can display to the user learning support messages and learning plans generated by a large-scale language model, as shown in (1) of Figure 9.

The support device 100 (reception unit) can also receive replies, questions, etc. to the learning support message and learning plan input by the user ((2) in FIG. 9). The support device 100 (generation unit) can then generate further learning support messages and learning plans in response to the received replies, questions, etc.

(Second example: Answer to a question from a user)
Next, an example will be described in which the support device 100 generates an "answer to a question from a user" as learning support information and outputs it to the user. In a second example, the support device 100 generates an answer to a question in accordance with the user's learning situation and a question from the user, and outputs it to the user.

In specific processing, the support device 100 (reception unit) receives a question input from a user in the form of a chat. The support device 100 (generation unit) then inputs the received question and an instruction to generate an answer to the question into a large-scale language model, and generates the answer to the question as learning support information for the user.

Here, an example of a prompt to be input and an example of an output in the process according to the second example will be described with reference to FIG. 10. FIG. 10 is a diagram showing an example of a prompt according to the second example. FIG. 10 shows an example of a plurality of prompts ((1) and (2) in FIG. 10) for causing a large-scale language model to generate an answer according to the user's learning situation and a question from the user, and an example of an answer that is generated and output based on the prompts ((3) in FIG. 10).

The first prompt shown in FIG. 10 (1) includes "#background information", "#command (FIG. 10 (1-1))", and "#constraint (FIG. 10 (1-2))". Note that the "#background information" is processed in the same way as in FIG. 7 based on the command described in FIG. 10, so a description is omitted.

The "# command" is a command to cause the large-scale language model to execute a specified process, as in the example of FIG. 7. For example, the command shown in (1-1) of FIG. 10 allows the support device 100 (generation unit) to cause the large-scale language model to execute a command such as "when a user asks a question about a specified content (teaching material/theme), generate an answer to the question in strict accordance with the set constraints."

As with the example in FIG. 8, "#Constraint" is an instruction for specifying (restricting) the method of executing processing by the large-scale language model. For example, the instruction shown in FIG. 10 (1-2) allows the assistance device 100 (generation unit) to impose constraints on the large-scale language model when generating an answer, such as "character limit" and "if an answer is not possible, suggest an alternative learning function."

The second prompt shown in FIG. 10 (2) includes a "# command (FIG. 10 (2-1))." For example, the command shown in FIG. 10 (2-1) causes the large-scale language model to "generate a message summarizing the learning topic and asking whether there are any questions related to the learning topic the first time," and "generate a message based on the message entered by the user and the message generated immediately before from the second time onwards."

Then, the support device 100 (output unit) outputs the initial message generated based on the prompts shown in (1) and (2) of FIG. 10 and the answer to the user's question, as shown in (3) of FIG. 10. For example, the support device 100 (output unit) outputs text such as "... (summary of the study topic)... Do you have any questions? ((3-1) of FIG. 10)" as the initial message.

Next, the user inputs a question such as "I completely forgot about XX. Can you explain it to me? ((3-2) in FIG. 10)" in response to the question input by the user. The support device 100 (output unit) then outputs a message such as "... (detailed answer)... ((3-3) in FIG. 10)" in response to the question input by the user.

Here, the "example of chat-format output" described in FIG. 10 will be described with reference to FIG. 11. FIG. 11 is a diagram showing an example of chat-format output of learning support information relating to the second example. FIG. 11 shows a chat-format screen for outputting an initial message generated by a large-scale language model, answers to questions, etc.

For example, the support device 100 (output unit) can display to the user an initial message generated by a large-scale language model as shown in FIG. 11 (1). The support device 100 (reception unit) can also receive questions, etc. input by the user (FIG. 11 (2)). The support device 100 (generation unit) can then display to the user an answer generated in response to the received question, etc. (FIG. 11 (3)). This allows the support device 100 to output an answer in an interactive format to a question from the user.

Furthermore, the support device 100 can perform coaching using information included in the teaching materials. Here, an example of optimizing coaching using information included in the teaching materials will be described with reference to FIG. 12. FIG. 12 is a diagram showing an example of a prompt used for coaching optimization in the second example.

First, the support device 100 (reception unit) receives a question from a user. Next, the support device 100 (generation unit) searches for text data related to the content of the user's question based on vector data, etc. Next, the support device 100 (generation unit) inputs the content of the user's question and the searched text data into a large-scale language model using a first prompt shown in FIG. 12 (1) and a second prompt shown in FIG. 12 (2).

The support device 100 (generation unit) searches for different text data based on the generation result output from the large-scale language model and the question from the user. The support device 100 (generation unit) then inputs the generation result output from the large-scale language model, the question from the user, and the searched different text data into the large-scale language model again using the first prompt shown in (1) of FIG. 12 and the third prompt shown in (3) of FIG. 12. This process is repeated until the number of times, generation accuracy, etc., meets predetermined conditions.

Then, the support device 100 (output unit) outputs the generated message to the user. Based on the above process, the support device 100 can provide the user with optimized coaching using the information contained in the teaching materials.

(Third example: Generation and output of exercise questions and answer information)
Next, an example will be described in which the support device 100 generates "practice questions" as learning support information and outputs them to the user. In a third example, the support device 100 generates practice questions according to the user's learning situation and outputs them to the user.

Furthermore, the support device 100 generates a grading result and commentary information based on the grading result for the user's answer to the practice questions posed, and outputs the result to the user. Note that the "commentary information" here includes information such as whether the user's answer was correct or incorrect, a detailed explanation of the content of the correct answer, and transition information to educational data related to the correct answer.

Specifically, the support device 100 (generation unit) inputs commands to generate practice questions based on constraints such as the number of questions, options, difficulty level, not generating the same question, character limit, etc., related to the format of the practice questions (prescribed tasks) that correspond to the user's learning situation into a large-scale language model, and generates practice questions as learning support information for the user.

Furthermore, the support device 100 (reception unit) receives answers to the practice questions from the user. Next, the support device 100 (generation unit) inputs the received answers and a command to generate a marking result and commentary information for the answers into a large-scale language model, and generates the marking result and commentary information as learning support information for the user.

Here, an example of a prompt to be input and an example of an output in the process according to the third example will be described. FIG. 13 is a diagram showing an example of a prompt according to the third example. FIG. 13 shows an example of a plurality of prompts for causing a large-scale language model to generate practice problems according to the user's learning situation ((1) and (2) in FIG. 13), practice problems that are generated and output based on the prompts, and an example of the results of scoring the practice problems and explanation information ((3) in FIG. 13).

The first prompt shown in FIG. 13 (1) includes "#background information", "#command (FIG. 13 (1-1))", and "#constraint (FIG. 13 (1-2))". Note that the "#background information" is processed in the same way as in FIG. 7 based on the command described in FIG. 13, so a description is omitted.

The "# command" is a command to cause the large-scale language model to execute a specified process, as in the example of FIG. 7. For example, the command shown in FIG. 13 (1-1) can cause the assistance device 100 (generation unit) to cause the large-scale language model to execute a command such as "generating practice questions of a specified format, such as multiple-choice questions, according to the user's learning situation, while strictly adhering to the set constraints."

Similar to the example in FIG. 8, "#Constraint" is an instruction for specifying (limiting) the method of executing processing by the large-scale language model. For example, the instruction shown in FIG. 13 (1-2) allows the support device 100 (generation unit) to impose constraints on the large-scale language model when generating practice questions, such as "limiting the number of questions," "limiting answer options," "adjusting difficulty," "asking different questions," and "limiting the number of characters."

The second prompt shown in FIG. 13(2) includes the "# command (FIG. 13(2-1))." Note that the second prompt shown in FIG. 13(2) is similar to the prompt shown in FIG. 10(2-1), so a description of it will be omitted.

Then, the support device 100 (output unit) outputs the practice questions generated based on the prompts shown in (1) and (2) of Figure 13, as well as the scoring results and explanation information for the practice questions, as shown in (3) of Figure 13.

For example, the support device 100 (output unit) outputs text such as "... (summary of the learning theme)... Now, let's try solving the following problem" or "Question: What is the role of XX?... (omitted)... Which do you think is the correct answer?" as messages to be displayed when asking practice questions ((3-1) in Figure 13).

The user inputs an answer such as "I think 'A' is the correct answer" to (3-1) in FIG. 13 ((3-2) in FIG. 13). Then, the support device 100 (output unit) outputs text such as "That's right! ... (explanation) ..." or "Let's move on to the next question. This one is a little more difficult ..." in response to the answer from the user ((3-3) in FIG. 13).

Here, the "example of chat-format output" explained in FIG. 13 will be explained with reference to FIG. 14. FIG. 14 is a diagram showing an example of chat-format output of learning support information relating to the third example. FIG. 14 shows a chat-format screen for outputting, in natural language, practice questions generated by a large-scale language model, the grading results of the answers to the practice questions, explanatory information, and the like.

For example, the support device 100 (output unit) can display a message to the user regarding the setting of practice questions generated by a large-scale language model, as shown in FIG. 14 (1). In addition, the support device 100 (reception unit) can receive answers to the practice questions input by the user (FIG. 14 (2)).

The support device 100 (output unit) can then display to the user the scoring results and explanation information generated according to the accepted answers, etc. (FIG. 14 (3)). Furthermore, the support device 100 (output unit) can present as the next question a practice question that has been generated with its difficulty adjusted according to the user's answer (FIG. 14 (3)).

The support device 100 can receive answers from the user and output the marking results and explanation information generated according to the received answers, as shown in (4) and (5) of FIG. 14. The support device 100 also supports the user in effectively carrying out problem exercises and checking explanations by repeatedly executing the above process.

(Processing procedure according to the first embodiment)
From here, a description will be given of a series of processing procedures realized by the support device 100 according to the first embodiment. Fig. 15 is a flowchart of the processing of the support device 100 according to the first embodiment.

The support device 100 waits for processing until the conditions for starting generation of learning support information are met (No in S101). Then, when the conditions for starting generation of learning support information are met, the support device 100 starts processing (Yes in S101). Note that the generation start conditions here are conditions that can be set arbitrarily, and include, for example, input from a user, the passage of a predetermined time, the arrival of a set schedule, instructions from an administrator, etc.

The generation unit 132 inputs the user learning information into the large-scale language model to generate learning support information according to the user's learning situation (S102). Next, the output unit 133 outputs the learning support information according to the user's learning situation (S103).

If the processing termination condition is not met (No in S104), the support device 100 returns to the previous step and continues the processing. On the other hand, if the processing termination condition is met (Yes in S104), the support device 100 ends the process. Note that the processing termination condition here is a condition that can be set arbitrarily, and includes, for example, input from the user, the passage of a specified time, the arrival of a set schedule, instructions from an administrator, etc.

(effect)
Hereinafter, a description will be given of the effects of the support device 100 according to the first embodiment. Conventionally, when learning online, there is a problem in providing learning support according to the learning situation of a user who is learning.

The generation unit 132 of the support device 100 according to the first embodiment inputs user learning information that uses e-books for learning into a large-scale language model, and generates learning support information for the user according to the user's learning situation. The output unit 133 of the support device 100 outputs the generated learning support information for the user.

Therefore, the support device 100 has the effect of enabling appropriate learning support according to the user's learning situation. Furthermore, the support device 100 achieves the respective predetermined effects by executing the processes described below.

The generation unit 132 inputs a command to generate a user learning support message including at least one of words of encouragement to the user according to the user's learning progress, a summary of the user's learning content, and a message to confirm whether the user has any questions into the large-scale language model, and generates the user's learning support message as user learning support information.

By the above-mentioned process, for example, when the user's learning progress is delayed, the support device 100 can display words of encouragement (words of support) such as "You're behind in your learning. Is there something wrong? It's tough, but let's do our best!" to the user. In addition, the support device 100 can display a "summary of what you've learned up to the last time" to support the user's learning, and then display a message such as "Do you have any questions about what you've learned up to the last time?" to resolve any points the user is unsure of and deepen the user's understanding of the learning content. As a result, the support device 100 has the effect of being able to support the user in making efficient/effective progress in their learning according to the user's learning.

The reception unit 131 receives questions input from users in the form of chat. The generation unit 132 inputs the received questions and instructions for generating answers to the questions into a large-scale language model, and generates answers to the questions as learning support information for the user.

By performing the above-described process, when a user asks a question such as "What is XX?", the support device 100 can display an explanation about "XX" to the user as a response to the question. As a result, the user can appropriately carry out independent learning based on the response to the user's question returned from the support device 100. Therefore, the support device 100 has the effect of enabling the user to independently learn more efficiently and effectively than before.

The generation unit 132 inputs a command to generate practice questions (predetermined tasks) based on constraints regarding the format of the practice questions (predetermined tasks) according to the user's learning situation into the large-scale language model, and generates the practice questions (predetermined tasks) as learning support information for the user.

By the above-mentioned process, the support device 100 can generate appropriate practice questions according to the user's learning progress and the electronic book the user uses for learning, and provide them to the user. As a result, the user can carry out appropriate independent learning, such as carrying out problem practice, based on the practice questions according to the user's learning progress provided by the support device 100. Therefore, the support device 100 has the effect of enabling the user to carry out independent learning more efficiently/effectively than before.

The reception unit 131 receives answers to practice questions (predetermined tasks) from a user. The generation unit 132 inputs the received answers and a command to generate a marking result and commentary information for the answers into a large-scale language model, and generates the marking result and commentary information as learning support information for the user.

By the above-mentioned process, the support device 100 can provide a grading result such as correct or incorrect for the user's answer to the practice problem generated by the support device 100. In addition, the support device 100 can provide a chapter or section of an e-book related to the user's answer or grading result, or an individual description, or explanatory information such as other explanatory videos, in association with the user's answer.

As a result, the user can appropriately conduct self-study, such as reviewing the results of their own practice problems and revising content that they have a low level of understanding, based on the scoring results and commentary information provided by the support device 100. Therefore, the support device 100 has the effect of enabling the user to conduct self-study more efficiently and effectively than before.

The output unit 133 outputs the generated learning support information for the user to the user in a chat format. Through the above-described processing, the support device 100 realizes natural exchanges in a conversational format using natural language for coaching and teaching the user, accepting and answering questions from the user, etc. As a result, the support device 100 has the effect of making it possible to provide an environment in which the user can naturally consult with or ask questions of the support device 100.

The generation unit 132 inputs information about the user's learning, including at least one of the user's learning history, the user's learning goals, and past conversation data, into a large-scale language model, and generates learning support information according to the user's learning situation using the tasks.

By the above-mentioned processing, for example, the support device 100 can teach and coach the user and answer questions according to the user's learning situation and learning goals. As a result, the user can receive educational support from the support device 100 according to his/her own learning situation, and can resolve unclear points and deepen understanding of the learning content more appropriately than before. Therefore, the support device 100 has the effect of being able to support the user to progress in learning efficiently/effectively.

Second Embodiment
Hereinafter, a second embodiment will be described in which the support device 100 further supports the implementation of a hacking exercise in addition to teaching and coaching the user described in the first embodiment. The support device 100 according to the second embodiment performs team division for the hacking exercise, construction of an implementation environment for the hacking exercise, generation and output of feedback information for the results of the hacking exercise, etc.

The hacking exercise mentioned above involves users being divided into an "attacking team" and a "defensive team," with the attacking team carrying out virtual hacking (cracking, etc.) attacks and the defensive team defending against the attacking team's virtual hacking attacks.

(Support device 100)
Next, the support device 100 according to the second embodiment will be described in detail. FIG. 16 is a diagram showing an example of the configuration of the support device 100 according to the second embodiment. As shown in FIG. 16, the support device 100 according to the second embodiment has a communication unit 110, a storage unit 120, and a control unit 130, similar to the first embodiment. Note that the support device 100 according to the second embodiment and the support device 100 according to the first embodiment have common functions. Therefore, in the following items, the description of the functions common to the first embodiment and the second embodiment will be omitted as appropriate.

(Storage unit 120)
The storage unit 120 has a teaching material information DB 121 , a user learning information DB 122 , a generative model DB 123 , and further a result information DB 124 .

(Result information DB 124)
The result information DB 124 is a database that stores information (result information) on the results of the hacking exercise performed by each team. Specifically, the result information DB 124 stores information (teams) that identify the teams that participated in the hacking exercise and the exercise results of the hacking exercise as result information.

Here, an example of result information stored by the result information DB 124 will be described. FIG. 17 is a table diagram showing an example of result information according to the second embodiment. As shown in FIG. 17, the result information DB 124 stores the items "Team" and "Exercise Results" and information related to the items in a table format or the like, in association with "No.", which is information that identifies individual result information.

For example, as shown in FIG. 17, the result information DB 124 stores the team "Attack" identified by the No. "1" in association with the practice result "K". Note that the letters "K, L" included in the table diagram shown in FIG. 17 are information indicating a legend for each item of individual result information stored in the result information DB 124, and the information actually stored is not particularly limited.

The above "team" refers to the team that carries out the hacking exercise, and includes information that identifies the "attacking team" that attempts hacking and the "defensive team" that prevents hacking. "Exercise results" is information about the results of the hacking exercise, and includes, for example, "score, number of abnormalities that occurred, final number of abnormalities, time during which abnormalities occurred" for the attacking team, and "score, number of abnormalities resolved, final number of normal conditions, time required to resolve abnormalities, normal operation time" for the defensive team.

(Control unit 130)
16, the control unit 130 includes a reception unit 131, a generation unit 132, an output unit 133, and further includes a team design unit 134 and a construction unit 135.

(Generation unit 132)
The generation unit 132 inputs the results of the exercises performed using the constructed execution environment and an instruction to generate feedback information on the results of the exercises into the large-scale language model, and generates the feedback information on the results of the exercises as learning support information for the user. For example, the generation unit 132 inputs a prompt including "#background information", "#command", "#exercise result", etc. into the large-scale language model to generate the feedback information on the results of the exercises.

The above "#background information" includes information that defines the role, such as "You are an AI instructor who will grade the hacking exercise. #Follow the instructions to generate feedback information regarding the results of the hacking exercise." The "#instructions" include instructions such as "#Grade using the number of abnormalities that occurred, the final number of abnormalities, and the time it took for the abnormalities to occur for the attacking team included in the exercise results," and "#Grade using the number of abnormalities resolved, the final number of normal conditions, the time it took to resolve the abnormalities, and the normal operation time for the defending team included in the exercise results." Note that the above prompt may further include "#constraint conditions."

(Output unit 133)
The output unit 133 outputs feedback information regarding the generated exercise results. For example, the output unit 133 outputs feedback information including exercise results such as "score, number of abnormalities, final number of abnormalities, and time during which abnormalities occurred" for the attacking team and "score, number of abnormalities resolved, final number of normal situations, time during which abnormalities were resolved, and normal operation time" for the defending team to users who participated in the hacking exercise.

(Team design unit 134)
The team design unit 134 assigns users to a first team, which is an offensive team, or a second team, which is a defensive team, according to the user's learning status, and designs the teams. For example, the team design unit 134 assigns users whose learning status progress is included within the top XX as a "defensive team" and users whose learning status progress is included within the bottom XX as an "offensive team" according to the user's learning status based on the user's learning information stored in the user learning information DB 122, and designs the teams.

(Construction Unit 135)
The construction unit 135 constructs an execution environment for executing a hacking exercise between the designed first team (attacking team) and second team (defensive team) according to the user's learning situation. Specifically, the construction unit 135 executes registered scripts based on an RPA (Robotic Process Automation) tool or the like according to the user's learning situation based on the user's learning information stored in the user learning information DB 122, thereby automatically and periodically executing operations such as email, chat, file sharing, and web site browsing.

For example, in the case of a hacking exercise by multiple users whose learning progress is in the top XX, the construction unit 135 automatically executes operations such as "Website browsing." On the other hand, in the case of a hacking exercise by multiple users whose learning progress is in the bottom XX, the construction unit 135 automatically executes operations such as "email, chat, file sharing."

(An example of a hacking exercise screen)
From here, an example of team division, environment construction, and generation and output of feedback information regarding the results of a hacking exercise realized by the support device 100 according to the second embodiment will be described with reference to Fig. 18. Fig. 18 is a diagram showing an example of processing of the support device 100 according to the second embodiment.

First, the support device 100 starts preparations to conduct a hacking exercise. If the conditions for starting the hacking exercise are not met, the support device 100 waits until the exercise starts.

The support device 100 (team design unit) assigns users to "offensive teams" or "defensive teams" according to the users' learning status, and designs "offensive teams" and "defensive teams" ((1) in FIG. 18). Next, the support device 100 (construction unit) constructs a hacking practice environment in which email, chat, file sharing, web site browsing, etc. are automatically performed according to the users' learning status ((2) in FIG. 18).

Then, users assigned to the "attack team" or "defense team" participate in a hacking exercise conducted by the support device 100 ((3-1) and (3-2) in FIG. 18).

When a hacking exercise is carried out, the support device 100 (generation unit) inputs the results of the hacking exercise (exercise results) into a large-scale language model ((4-1) in FIG. 18) and generates feedback information about the results of the hacking exercise ((4-2) in FIG. 18).

The support device 100 (output unit) outputs the generated feedback information to the terminal device 200 (FIG. 18 (5-1)). As a result, the terminal device 200 can display information such as a "summary of the exercise results" and "explanatory information about the results of the hacking exercise" to the user as learning support information (FIG. 18 (5-2)).

Here, an example of feedback information output by the support device 100 (output unit) will be described with reference to FIG. 19. FIG. 19 is a diagram showing an example of a hacking practice screen according to the second embodiment. FIG. 19 shows a practice environment for a hacking practice ((1) in FIG. 19) and a practice result for the hacking practice ((2) in FIG. 19).

The training environment shown in FIG. 19 (1) is an environment in which emails are automatically sent to specified destinations. The attacking team performs hacking on the environment in which emails are automatically sent. Note that the environment in which emails are automatically sent shown in FIG. 19 (1) is merely an example, and other environments in which operations related to chat, file sharing, web site browsing, etc. are automatically performed may also be used.

In the area shown in FIG. 19(2), as the results of the hacking exercise, information such as "score, number of abnormalities, final number of abnormalities, and time during which abnormalities occurred" is displayed for the offensive team, and "score, number of abnormalities resolved, final number of normal conditions, time during which abnormalities were resolved, and normal operation time" is displayed for the defensive team. Note that the results of the hacking exercise shown in FIG. 19(2) may be displayed in real time while the hacking exercise is being carried out, or may be displayed as the overall results after the exercise is completed.

(Processing procedure according to the second embodiment)
From here, a description will be given of a series of processing procedures realized by the support device 100 according to the second embodiment. Fig. 20 is a flowchart of the processing of the support device 100 according to the second embodiment.

The support device 100 waits until the hacking exercise starts (No in S201). Then, when the hacking exercise starts (Yes in S201), the team design unit 134 divides the users into teams according to their learning status (S202). Next, the construction unit 135 constructs an execution environment for the hacking exercise according to the users' learning status (S203).

If the exercise has not ended (No in S204), the support device 100 waits to proceed to the next step. On the other hand, if the exercise has ended (Yes in S204), the generation unit 132 generates feedback information (S205). Next, the output unit 133 outputs the feedback information (S206). Then, the support device 100 ends the step.

(effect)
The team design unit 134 of the support device 100 according to the second embodiment assigns a user to a first team or a second team according to the user's learning status, and designs the teams. The construction unit 135 of the support device 100 constructs an execution environment for executing a hacking exercise with the designed first team and second team according to the user's learning status. Then, the generation unit 132 inputs the result of the exercise performed using the constructed execution environment and an instruction to generate feedback information regarding the result of the exercise into the large-scale language model, and generates the feedback information regarding the result of the exercise as learning support information for the user.

Therefore, the support device 100 enables appropriate hacking exercises to be carried out according to each user's level of understanding, skill level, learning progress, etc.

Specifically, the support device 100 enables appropriate team design, taking into consideration the user's technical and knowledge levels so as not to imbalance the strength balance of each team during the hacking exercise. As a result, the support device 100 realizes more effective hacking exercises than ever before.

In addition, when conducting a hacking exercise, the environment in which the hacking exercise is carried out is taken into consideration. For example, if an execution environment is constructed that is too high or too low compared to the user's technical or knowledge level, the user will not be able to carry out an appropriate hacking exercise, and therefore an effective hacking exercise for improving the user's capabilities cannot be carried out. Therefore, the support device 100 enables appropriate team design while taking into consideration that the balance of capabilities of each team is not biased when constructing an execution environment for a hacking exercise according to the user's technical or knowledge level. As a result, the support device 100 realizes the implementation of a hacking exercise that is more effective than before.

The support device 100 also generates feedback information including scoring results, summaries, commentary information, etc., for the results of the hacking exercise based on a large-scale language model. As a result, the support device 100 enables the generation of feedback information in real time and feedback information with a larger amount of information than before. As a result, the support device 100 has the effect of enabling more effective hacking exercises than before. Note that the above "large amount of information" means, for example, "scores for each team," "good and bad points for each team," "points for improvement," "information collected from related e-books or external sources," etc.

<Modification>
Below, modified examples realized by the assistance device 100 according to this embodiment will be described.

Alternative Forms of Generating and Marking Exercises
The assistance device 100 according to the first embodiment described above has been described as generating practice questions and scoring the practice questions based on a large-scale language model, but is not limited to this.

For example, the support device 100 presents the user with practice questions input by an administrator or the like. Next, the support device 100 accepts answers to the practice questions from the user. The support device 100 then compares the accepted answers with correct answers registered in advance and performs grading.

By the above process, the support device 100 presents the user with a practice problem for remotely setting up an actual device such as a communication device, and the user sets up the communication device based on remote operation for the practice problem. The support device 100 can then compare the setting state of the communication device set by the user based on remote operation with the setting state of the communication device that has been registered in advance as the correct answer, and perform grading.

(Data, etc.)
The electronic books, coaching, teaching, practice questions (predetermined tasks), learning support messages, etc., as well as the names of the functional parts, steps, processes, and names of steps or processes of the support device 100 used in the description of the above embodiments are merely examples and can be changed as desired.

For example, the teaching material information DB 121 has been described as storing items such as "title", "link", "order", "number of contents", "assignment", and "type" and information related to the items in a table format or the like in association with "No", which is information identifying individual teaching material information, but the items, contents, storage format, etc. stored are not limited. The user learning information DB 122 has been described as storing items such as "user identification information", "teaching materials used", "learning history", "learning goal", and "conversation history" and information related to the items in a table format or the like in association with "No", which is information identifying individual user learning information, but the items, contents, storage format, etc. stored are not limited. The result information DB 124 has been described as storing items such as "team" and "exercise result" and information related to the items in a table format or the like in association with "No", which is information identifying individual result information, but the items, contents, storage format, etc. stored are not limited.

(About the use of generative models)
In the present embodiment, the model (large-scale language model) used by the assistance device 100 is stored in the generative model DB 123 of the storage unit 120, but the present invention is not limited to this. For example, the assistance device 100 can access an external information processing device (such as a server) and use a predetermined model.

(Flowcharts, etc.)
The steps in the flowcharts and the like may be interchanged and executed as long as there is no contradiction, and some steps may not be executed. In addition, conjunctions such as "next,""continue,""further,""at this time," and "on this occasion" in the explanation of the flowcharts do not limit the order or timing of execution of the processes in the flowcharts.

(Systems, etc.)
Of the processes described in the above embodiments and modifications, all or part of the processes described as being performed automatically can be performed manually, or all or part of the processes described as being performed manually can be performed automatically by a known method. In addition, the information including the processing procedures, specific names, various data and parameters shown in the above documents and drawings can be changed arbitrarily unless otherwise specified. For example, the various information shown in each drawing is not limited to the illustrated information.

In addition, each component of each device shown in the figure is a functional concept, and does not necessarily have to be physically configured as shown in the figure. In other words, the specific form of distribution and integration of each device is not limited to that shown in the figure, and all or part of them can be functionally or physically distributed and integrated in any unit depending on various loads, usage conditions, etc.

The above-mentioned components include those that can be easily imagined by a person skilled in the art, those that are substantially the same, and those that are within the scope of what is called equivalent. Furthermore, the above-mentioned embodiments and variations can be combined as appropriate to the extent that they do not cause inconsistencies in the processing content.

The aforementioned "section, module, unit" can be read as "means" or "circuit," etc. For example, a control unit can be read as a control means or a control circuit.

Although some of the embodiments have been described in detail above with reference to the drawings, these are merely examples, and it is possible to implement the present embodiments in other forms with various modifications and improvements based on the knowledge of those skilled in the art, including the forms described in the disclosure section of the invention.

<Hardware Configuration>
The devices included in the support device 100 according to this embodiment are realized, for example, by a computer 1000 having a configuration as shown in Fig. 21. Fig. 21 is a diagram showing an example of a hardware configuration of a computer that realizes the support device 100 according to the embodiment.

The computer 1000 has a configuration in which a CPU 1100, a memory 1200, an auxiliary storage device 1300, an input interface 1400, an output interface 1500, and a communication interface 1600 are connected by a bus 1700.

The CPU 1100 operates based on a program stored in the memory 1200 or the auxiliary storage device 1300, and controls each functional unit. The memory 1200 is composed of, for example, a RAM (Random Access Memory) or a ROM (Read Only Memory), and stores a boot program executed by the CPU 1100 when the computer 1000 starts up, programs that depend on the hardware of the computer 1000, etc.

For example, when the computer 1000 functions as the support device 100 according to this embodiment, the CPU 1100 of the computer 1000 can realize the functions of the control unit 130 by executing a program loaded onto the memory 1200.

The auxiliary storage device 1300 stores programs executed by the CPU 1100 and data used by the programs. The CPU 1100 controls an input device 1410 such as a keyboard or mouse via an input interface 1400. The CPU 1100 also acquires data from the input device 1410 via the input interface 1400.

The CPU 1100 controls an output device 1510 such as a display or a printer via the output interface 1500. The CPU 1100 also outputs generated data, etc. to the output device 1510 via the output interface 1500.

The communication interface 1600 receives data from other devices via a specified network NW and sends it to the CPU 1100, and transmits data generated by the CPU 1100 to other devices via the specified network NW.

100 Support device 110 Communication unit 120 Storage unit 121 Teaching material information DB
122 User learning information DB
123 Generative Model DB
124 Results information DB
130 Control unit 131 Reception unit 132 Generation unit 133 Output unit 134 Team design unit 135 Construction unit

Claims (18)

a generation unit that inputs learning information of a user who uses an electronic book for learning, the learning information including at least one of learning materials used for learning, learning history, learning goals, and conversation history, into a large-scale language model, and generates information to support the learning of the user , the information including at least one of messages to support the learning of the user according to the learning situation of the user, answers to questions from the user, practice questions to be given to the user, and information to follow up on the answers to the practice questions;
an output unit that outputs the generated information for supporting the user's learning,
The generation unit is
a command to generate a message to support the user's learning, the message including at least one of words of encouragement for the user according to the progress of the user's learning, a learning summary showing a summary of the learning content by the user up to the last time in natural language, and a message to confirm whether the user has any questions about the learning, is input to the large-scale language model, and the message to support the user's learning is generated as information to support the user's learning;
16. An assistance device comprising: a generation unit that inputs learning information of a user who uses an electronic book for learning, the learning information including at least one of learning materials used for learning, learning history, learning goals, and conversation history, into a large-scale language model, and generates information to support the learning of the user , the information including at least one of messages to support the learning of the user according to the learning situation of the user, answers to questions from the user, practice questions to be given to the user, and information to follow up on the answers to the practice questions;
an output unit that outputs the generated information for supporting the user's learning,
The generation unit is
inputting into the large-scale language model a command to generate a predetermined task based on constraints on a question format of the predetermined task, including at least one of the number of practice questions corresponding to the learning status of the user, options, difficulty level, restrictions on question generation, and restrictions on the number of characters, and generating the predetermined task as information to support the learning of the user;
16. An assistance device comprising: a generation unit that inputs learning information of a user who uses an electronic book for learning, the learning information including at least one of learning materials used for learning, learning history, learning goals, and conversation history, into a large-scale language model, and generates information to support the learning of the user , the information including at least one of messages to support the learning of the user according to the learning situation of the user, answers to questions from the user, practice questions to be given to the user, and information to follow up on the answers to the practice questions;
an output unit that outputs the generated information for supporting the user's learning;
a team design unit that assigns the user to a first team or a second team according to a learning situation of the user and designs the team;
A construction unit that constructs an execution environment for executing the hacking exercise between the first team and the second team according to a learning situation of the user;
An assistance device comprising: a generation unit that inputs learning information of a user who uses an electronic book for learning, the learning information including at least one of learning materials used for learning, learning history, learning goals, and conversation history, into a large-scale language model, and generates information to support the learning of the user , the information including at least one of messages to support the learning of the user according to the learning situation of the user, answers to questions from the user, practice questions to be given to the user, and information to follow up on the answers to the practice questions;
an output unit that outputs the generated information for supporting the user's learning,
The generation unit is
inputting information about the user's learning, including at least one of an electronic book used by the user, the user's learning history, the user's learning goal, and past conversation data, into the large-scale language model to generate information to support the user's learning;
16. An assistance device comprising: The method further includes a reception unit that receives input of a question about learning from the user in a chat format,
The generation unit is
inputting the received question and an instruction to generate an answer to the question into the large-scale language model, and generating the answer to the question as information to assist the user in learning;
5. The support device according to claim 1, wherein the support device is a computer. The method further includes a reception unit that receives an answer to the predetermined task from the user,
The generation unit is
inputting the accepted answer and a command to generate a marking result and commentary information for the answer into the large-scale language model, thereby generating the marking result and the commentary information as information to support the user's learning;
3. The support device according to claim 2 . a team design unit that assigns the user to a first team or a second team according to a learning situation of the user and designs the team;
A construction unit that constructs an execution environment for executing the hacking exercise between the first team and the second team according to the learning situation of the user,
7. The support device according to claim 1 , 2, 3 or 6. The generation unit is
inputting the results of the hacking exercise carried out using the constructed execution environment and an instruction to generate feedback information regarding the results of the hacking exercise into the large-scale language model, thereby generating feedback information regarding the results of the hacking exercise as information to support the user's learning;
8. The support device according to claim 7 . The output unit is
outputting the generated information for supporting the user's learning to the user in a chat format;
7. The support device according to claim 1 , 2, 3, 4 or 6. The generation unit is
inputting information about the user's learning, including at least one of an electronic book used by the user, the user's learning history, the user's learning goal, and past conversation data, into the large-scale language model to generate information to support the user's learning;
7. The support device according to claim 1, wherein the support device is a computer. A support method to be executed by a support device,
a generation process of inputting learning information of a user who uses an electronic book for learning, including at least one of learning materials used for learning by the user who uses the electronic book for learning, a learning history, a learning goal, and a conversation history , into a large-scale language model to generate information to support the learning of the user, including at least one of messages to support the learning of the user according to the learning situation of the user, answers to questions from the user, practice questions to be given to the user, and information to follow up on the answers to the practice questions ;
and an output step of outputting the generated information for supporting the user's learning,
The generating step includes:
a command to generate a message to support the user's learning, the message including at least one of words of encouragement for the user according to the progress of the user's learning, a learning summary showing a summary of the learning content by the user up to the last time in natural language, and a message to confirm whether the user has any questions about the learning, is input to the large-scale language model, and the message to support the user's learning is generated as information to support the user's learning;
A support method comprising : A support method to be executed by a support device,
a generation process of inputting learning information of a user who uses an electronic book for learning, including at least one of learning materials used for learning by the user who uses the electronic book for learning, a learning history, a learning goal, and a conversation history , into a large-scale language model to generate information to support the learning of the user, including at least one of messages to support the learning of the user according to the learning situation of the user, answers to questions from the user, practice questions to be given to the user, and information to follow up on the answers to the practice questions ;
and an output step of outputting the generated information for supporting the user's learning,
The generating step includes:
inputting into the large-scale language model a command to generate a predetermined task based on constraints on a question format of the predetermined task, including at least one of the number of practice questions corresponding to the learning status of the user, options, difficulty level, restrictions on question generation, and restrictions on the number of characters, and generating the predetermined task as information to support the learning of the user;
A support method comprising : A support method to be executed by a support device,
a generation process of inputting learning information of a user who uses an electronic book for learning, including at least one of learning materials used for learning by the user who uses the electronic book for learning, a learning history, a learning goal, and a conversation history , into a large-scale language model to generate information to support the learning of the user, including at least one of messages to support the learning of the user according to the learning situation of the user, answers to questions from the user, practice questions to be given to the user, and information to follow up on the answers to the practice questions ;
an output step of outputting the generated information for supporting the user's learning;
a team design process of assigning the user to a first team or a second team according to a learning situation of the user, and designing a team;
A construction process of constructing an execution environment for executing the hacking exercise between the first team and the second team according to the learning situation of the user;
Assistance method comprising: A support method to be executed by a support device,
a generation process of inputting learning information of a user who uses an electronic book for learning, including at least one of learning materials used for learning by the user who uses the electronic book for learning, a learning history, a learning goal, and a conversation history , into a large-scale language model to generate information to support the learning of the user, including at least one of messages to support the learning of the user according to the learning situation of the user, answers to questions from the user, practice questions to be given to the user, and information to follow up on the answers to the practice questions ;
and an output step of outputting the generated information for supporting the user's learning,
The generating step includes:
inputting information about the user's learning, including at least one of an electronic book used by the user, the user's learning history, the user's learning goal, and past conversation data, into the large-scale language model to generate information to support the user's learning;
A support method comprising : a generation step of inputting learning information of a user who uses an electronic book for learning, including at least one of learning materials used for learning by the user who uses the electronic book for learning, a learning history, a learning goal, and a conversation history , into a large-scale language model to generate information to support the learning of the user , including at least one of a message to support the learning of the user according to the learning situation of the user, an answer to a question from the user, a practice question to be given to the user, and information to follow up on the answer to the practice question;
and an output step of outputting the generated information for supporting the user's learning ,
The generating step includes:
a command to generate a message to support the user's learning, the message including at least one of words of encouragement for the user according to the progress of the user's learning, a learning summary showing a summary of the learning content by the user up to the last time in natural language, and a message to confirm whether the user has any questions about the learning, is input to the large-scale language model, and the message to support the user's learning is generated as information to support the user's learning;
Support programs. a generation step of inputting learning information of a user who uses an electronic book for learning, including at least one of learning materials used for learning by the user who uses the electronic book for learning, a learning history, a learning goal, and a conversation history , into a large-scale language model to generate information to support the learning of the user , including at least one of a message to support the learning of the user according to the learning situation of the user, an answer to a question from the user, a practice question to be given to the user, and information to follow up on the answer to the practice question;
and an output step of outputting the generated information for supporting the user's learning ,
The generating step includes:
inputting into the large-scale language model a command to generate a predetermined task based on constraints on a question format of the predetermined task, including at least one of the number of practice questions corresponding to the learning status of the user, options, difficulty level, restrictions on question generation, and restrictions on the number of characters, and generating the predetermined task as information to support the learning of the user;
Support programs. a generation step of inputting learning information of a user who uses an electronic book for learning, including at least one of learning materials used for learning by the user who uses the electronic book for learning, a learning history, a learning goal, and a conversation history , into a large-scale language model to generate information to support the learning of the user , including at least one of a message to support the learning of the user according to the learning situation of the user, an answer to a question from the user, a practice question to be given to the user, and information to follow up on the answer to the practice question;
an output step of outputting the generated information for supporting the user's learning;
a team design step of assigning the user to a first team or a second team according to a learning situation of the user, and designing a team;
A construction step of constructing an execution environment for executing the hacking exercise between the first team and the second team according to a learning situation of the user;
A support program characterized by causing a computer to execute the above. a generation step of inputting learning information of a user who uses an electronic book for learning, including at least one of learning materials used for learning by the user who uses the electronic book for learning, a learning history, a learning goal, and a conversation history , into a large-scale language model to generate information to support the learning of the user , including at least one of a message to support the learning of the user according to the learning situation of the user, an answer to a question from the user, a practice question to be given to the user, and information to follow up on the answer to the practice question;
and an output step of outputting the generated information for supporting the user's learning ,
The generating step includes:
inputting information about the user's learning, including at least one of an electronic book used by the user, the user's learning history, the user's learning goal, and past conversation data, into the large-scale language model to generate information to support the user's learning;
Support programs.

Images