Information processing system for learning support using generative AI

Abstract

To provide a learning support system using a generative AI that provides a user with a timely answer that is accurate to a certain degree or more according to a learning history of the user.SOLUTION: In an information processing system for learning support using a generative AI, one or more information processing devices includes: learning data provision means for providing a user with learning data associated with user information; input data acquisition means for acquiring a learning history of a user identified from the learning data and input data from the user associated with the learning history; prompt acquisition means for acquiring a prompt including an instruction to generate an answer according to the learning history based on data including the input data; and answer data generation means that allows the generative AI to generate answer data according to the learning history based on the prompt. The generative AI refers to data, including the learning data and related data associated with the learning data.SELECTED DRAWING: Figure 1

Description

Technical Field

[0001] The present invention relates to an information processing system for learning support using generative AI.

Background Art

[0002] With the development of communication technology, systems for conducting lectures online or providing learning have emerged. However, especially in lectures related to specialized fields, due to the complexity of the content, it is required that two-way communication between the user and the teacher be possible, for example, that questions and answers can be conducted online.

[0003] As a method for a user to ask a teacher questions online, for example, a student terminal having a question sending means for issuing a question to a teacher terminal and sending an email via a LAN, and an answer receiving means for receiving an answer from the teacher terminal, a teacher terminal having a question receiving means for receiving a question from the student terminal, and an answer sending means for sending an answer to the student terminal via a LAN (see Patent Document 1), and a learning support system that selects an answerer suitable for answering a learner's question by referring to an answerer database 10, and performs a process of acquiring and transmitting the learning history of the learner from a learning history database 11 together with the question content to the selected answerer (see Patent Document 2) exist.

Prior Art Documents

Patent Documents

[0004]

Patent Document 1

Patent Document 2

Disclosure of the Invention

Problems to be Solved by the Invention

[0005] The learning support system disclosed in Patent Document 1 requires teachers to respond to questions individually, which can make timely responses difficult. The learning support system disclosed in Patent Document 2 requires respondents to answer questions that are not stored in the question and answer database, which contains past question and answer data, and may not be able to provide timely responses. Furthermore, given the current severe labor shortage, securing teachers who can appropriately answer specialized questions is difficult. Additionally, there's a possibility that answers may include knowledge and information that the user hasn't studied. Furthermore, Patent Documents 1 and 2 do not address inquiries other than questions from users, such as the evaluation of assignments created as part of learning.

[0006] In summary, the primary objective of the present invention is to provide a learning support information processing system that uses generative AI capable of providing users with timely responses to data they input, with a certain level of accuracy corresponding to the user's learning history (course completion status).

[0007] Another primary objective of the present invention is to provide a learning support information processing system that uses a generative AI capable of providing users with timely recommendation information that has a certain level of accuracy based on the user's learning history and predetermined task answer indicator evaluations, etc., in response to the user's answers to tasks submitted by the user.

[0008] Another primary objective of the present invention is to provide a learning support information processing system that uses a generative AI capable of providing users with timely question answers that are appropriate to the user's learning history and have a certain level of accuracy.

[0009] Another main objective of the present invention is to provide a learning support information processing system that uses a generative AI capable of providing users with a timely summary of a question, an answer to the question, or hints related to the question, as well as additional question-related information related to the answer to the question, etc., in accordance with the user's learning history.

[0010] Another main objective of the present invention is to provide a learning support information processing system that uses a generative AI capable of providing the user with the results of a source code review and additional review-related information in a timely manner, based on the user's learning history and predetermined rules, for source code input by the user based on predetermined program specifications.

[0011] Another primary objective of the present invention is to provide a learning support information processing system that uses a generative AI capable of providing users with timely responses based on information contained in images, in response to user input data including images. [Means for solving the problem]

[0012] To achieve the above objective, the present invention includes the following embodiments.

[0013] (1) A learning support information processing system using a generative AI, comprising one or more information processing devices, the information processing device comprising: learning data provision means for providing learning data associated with user information to a user; input data acquisition means for acquiring the user's learning history identified from the learning data and input data from the user associated with the learning history; prompt acquisition means for acquiring prompts including instructions for generating answers corresponding to the learning history based on data including the input data; and answer data generation means for causing a generative AI to generate answer data corresponding to the learning history based on the prompt, wherein the generative AI refers to data including the learning data and related data associated with the learning data.

[0014] (2) A learning support information processing system using a generative AI, wherein the configuration described in (1) above, the generative AI refers to data including task data linked to the learning data, and correct answer data and task answer evaluation index data linked to the task data as the related data, the learning data providing means further provides the task data to the user, the input data acquisition means acquires task answer data from the user corresponding to the task data as input data, the prompt acquisition means acquires the prompt including an instruction to generate the answer including recommendation information based on the correct answer data and task answer evaluation index data for the task answer data, and the answer data generation means causes the generative AI to generate the answer data based on the prompt.

[0015] (3) A learning support information processing system using a generative AI, wherein the generative AI is configured to refer to data including question-answer format data linked to the learning data as the related data, the input data acquisition means acquires question data from the user as the input data, the prompt acquisition means acquires a prompt including an instruction to generate an answer including a question-answer based on the question-answer format data for the question data, and the answer data generation means causes the generative AI to generate the answer data based on the prompt.

[0016] (4) A learning support information processing system using a generative AI, wherein the configuration described in (1) above, wherein the input data acquisition means acquires question data from a user as the input data, the prompt acquisition means acquires the prompt from the question data, which includes a summary of the question, an answer to the question or hint information for the question, and an instruction to generate the answer, which includes additional question-related information related to the answer or hint information, and the answer data generation means causes the generative AI to generate the answer data based on the prompt.

[0017] (5) In the configuration described in (1) above, the generative AI is configured to refer to data including program specification data linked to the learning data, source code linked to the program specification data, and pass review data that defines the rules that the source code should satisfy, as the related data. An information processing system for learning support using a generative AI, wherein the input data acquisition means acquires source code entered by the user based on predetermined program specification data as the input data, the prompt acquisition means determines whether the source code entered by the user satisfies the rules defined in the passing review data and, if the rules are met, acquires the prompt which includes an instruction to generate the above answer including additional review-related information, and the answer data generation means causes the generative AI to generate the above answer data based on the prompt.

[0018] (6) A learning support information processing system using a generative AI, wherein the configuration is as described in any one of (1) to (5) above, the input data acquisition means acquires the input data including image data, the prompt acquisition means acquires the prompt including an instruction to generate the answer including an answer based on information extracted from the image data, and the answer data generation means causes the generative AI to generate the answer data based on the prompt.

[0019] (7) A learning support information processing device using a generative AI, comprising: a learning data provision means that provides learning data linked to user information to a user; an input data acquisition means that acquires the user's learning history identified from the learning data and input data from the user linked to the learning history; a prompt acquisition means that acquires a prompt that includes an instruction to generate an answer corresponding to the learning history based on data including the input data; and an answer data generation means that causes a generative AI to generate answer data corresponding to the learning history based on the prompt, wherein the generative AI refers to data including the learning data and related data linked to the learning data.

[0020] (8) In the configuration described in (7) above, the generative AI refers to, as the related data, data including problem data associated with the learning data, as well as correct answer data and problem-solving evaluation index data associated with the problem data. The learning data providing means further provides the problem data to the user, the input data acquisition means acquires, as the input data, problem-solving data from the user corresponding to the problem data, the prompt acquisition means acquires the prompt including a generation instruction of the answer including recommendation information based on the correct answer data and the problem-solving evaluation index data for the problem-solving data, and the answer data generation means causes the generative AI to generate the answer data based on the prompt. An information processing apparatus for learning support using a generative AI.

[0021] (9) In the configuration described in (7) above, the generative AI refers to, as the related data, data including question-and-answer format data associated with the learning data. The input data acquisition means acquires, as the input data, question data from the user, the prompt acquisition means acquires the prompt including a generation instruction of the answer including a question-and-answer based on the question-and-answer format data for the question data, and the answer data generation means causes the generative AI to generate the answer data based on the prompt. An information processing apparatus for learning support using a generative AI.

[0022] (10) In the configuration described in (7) above, the input data acquisition means acquires, as the input data, question data from the user, the prompt acquisition means acquires the prompt including a generation instruction of the answer including a summary of the question, an answer or hint information for the question, and question-related additional information related to the answer or hint information for the question from the question data, and the answer data generation means causes the generative AI to generate the answer data based on the prompt. An information processing apparatus for learning support using a generative AI.

[0023] (11) In the configuration described in (7) above, the generation-based AI refers to data including program specification data associated with the learning data, source code associated with the program specification data, and pass review data that defines the rules to be satisfied by the source code, as the associated data, An information processing apparatus for learning support using a generation-based AI, wherein the input data acquisition means acquires source code input by a user based on predetermined program specification data as the input data, the prompt acquisition means determines whether the source code input by the user satisfies the rules defined by the pass review data, and when the rules are satisfied, acquires the prompt including an instruction to generate the response including review-related additional information, and the response data generation means causes the generation-based AI to generate the response data based on the prompt.

[0024] (12) In the configuration according to any one of (7) to (11) above, the input data acquisition means acquires the input data including image data, the prompt acquisition means acquires the prompt including an instruction to generate the response including a response based on information extracted from the image data, and the response data generation means causes the generation-based AI to generate the response data based on the prompt. An information processing apparatus for learning support using a generation-based AI.

[0025] (13) A learning support information processing program using a generation-based AI, which causes one or more information processing apparatuses to execute steps of providing learning data associated with user information to a user, acquiring a learning history of the user specified from the learning data and input data from the user associated with the learning history, acquiring a prompt including an instruction to generate a response corresponding to the learning history based on data including the input data, and causing the generation-based AI to generate response data corresponding to the learning history based on the prompt, wherein the generation-based AI refers to data including the learning data and associated data associated with the learning data.

[0026] (14) A learning support information processing program using a generative AI, wherein the generative AI refers to the data including the task data linked to the learning data, and the correct answer data and task answer evaluation index data linked to the task data as the related data, and the information processing device is instructed to provide the task data to the user along with the learning data, and when it obtains the task answer data from the user corresponding to the task data as input data, it executes a process to obtain the prompt including an instruction to generate the answer, which includes recommendation information based on the correct answer data and task answer evaluation index data for the task answer data.

[0027] (15) A learning support information processing program using a generative AI, wherein the generative AI is configured to refer to data including question-answer format data linked to the learning data as the related data, and the information processing device is made to obtain question data from the user as input data and to obtain a prompt including an instruction to generate the answer including a question-answer based on the question-answer format data for the question data.

[0028] (16) A learning support information processing program using a generative AI, wherein the information processing device has the configuration described in (13) above, and the information processing device has the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device the information processing device

[0029] (17) In the configuration described in (13) above, the generative AI is configured to refer to data including program specification data linked to the learning data, source code linked to the program specification data, and passing review data that defines the rules that the source code should satisfy, as the related data, and causes the information processing device to execute a process to obtain source code entered by the user based on predetermined program specification data as input data, to determine whether the source code entered by the user satisfies the rules defined in the passing review data, and if the rules are satisfied, to obtain the prompt including an instruction to generate the above answer including additional review-related information, using a generative AI.

[0030] (18) A learning support information processing program using a generative AI, wherein the information processing device has the configuration described in any one of (13) to (17) above, and causes the information processing device to execute a process to acquire the input data including image data and to acquire the prompt including an instruction to generate the answer including an answer based on information extracted from the image data.

[0031] (19) A learning support information processing method using a generative AI, comprising the steps of: providing a user with learning data associated with user information; obtaining the user's learning history identified from the learning data and input data from the user associated with the learning history; obtaining a prompt including an instruction to generate an answer corresponding to the learning history based on data including the input data; and causing a generative AI to generate answer data corresponding to the learning history based on the prompt, wherein the generative AI refers to data including the learning data and related data associated with the learning data.

[0032] (20) A learning support information processing method using a generative AI, wherein the generative AI is configured to refer to the related data, which includes task data linked to the learning data, and correct answer data and task answer evaluation index data linked to the task data, and provides the user with the learning data along with the task data, and obtains task answer data from the user corresponding to the task data as input data, and obtains the prompt which includes an instruction to generate the answer which includes recommendation information based on the correct answer data and task answer evaluation index data for the task answer data.

[0033] (21) A learning support information processing method using a generative AI, wherein the generative AI is configured to refer to data including question answer format data linked to the learning data as the related data, and the information processing device obtains question data from the user as input data and obtains a prompt including an instruction to generate the answer including a question answer based on the question answer format data for the question data.

[0034] (22) A learning support information processing method using a generative AI, wherein the configuration described in (19) above, the method obtains question data from a user as input data, and obtains a prompt from the question data that includes a summary of the question, an answer or hint information for the question, and an instruction to generate the answer that includes additional question-related information relating to the answer or hint information.

[0035] (23) A learning support information processing method using a generative AI, wherein the generative AI refers to data including program specification data linked to the learning data, source code linked to the program specification data, and passing review data that defines the rules that the source code should satisfy, as the related data, and obtains source code entered by the user based on predetermined program specification data as input data, determines whether the source code entered by the user satisfies the rules defined in the passing review data, and obtains the prompt including an instruction to generate the above answer including additional review-related information if the rules are satisfied.

[0036] (24) A learning support information processing method using a generative AI, wherein the configuration is as described in any one of (19) to (23) above, and the method acquires the above input data including image data, and acquires the above prompt including an instruction to generate the above answer which includes information extracted from the above image data. [Effects of the Invention]

[0037] According to the present invention, it is possible to provide a learning support information processing system that uses a generative AI to provide users with timely answers that have a certain level of accuracy or higher, in response to data input by the user, according to the user's learning history.

[0038] Furthermore, according to the present invention, it is possible to provide a learning support information processing system that uses a generative AI to provide users with timely recommendation information that has a certain level of accuracy based on the user's learning history and predetermined task answer indicator evaluations, etc., in response to the user's answers to tasks submitted by the user.

[0039] Furthermore, according to the present invention, it is possible to provide a learning support information processing system that uses a generative AI to provide users with timely question answers that are appropriate to the user's learning history and have a certain level of accuracy.

[0040] Furthermore, according to the present invention, it is possible to provide a learning support information processing system that uses a generative AI to provide the user with a summary of the question, the answer to the question, or hint information for the question, as well as additional question-related information related to the answer to the question, etc., in a timely manner, in accordance with the user's learning history, in response to the user's questions.

[0041] Furthermore, according to the present invention, it is possible to provide a learning support information processing system that uses a generative AI to provide the user with source code input by the user based on predetermined program specifications, in accordance with the user's learning history and based on predetermined rules, and to provide the user with source code review and additional review-related information in a timely manner if the predetermined rules are met.

[0042] Furthermore, according to the present invention, it is possible to provide a learning support information processing system that uses a generative AI to provide users with timely responses based on information contained in images, in response to user input data including images. [Brief explanation of the drawing]

[0043] [Figure 1] This figure shows the overall schematic configuration of the information processing system according to the first embodiment. [Figure 2] This is a functional block diagram of the user terminal related to the said embodiment. [Figure 3] This is a functional block diagram of the server according to the same embodiment. [Figure 4] In relation to the same embodiment, Figure 4(a) is a table showing an example of checkpoints for evaluating problem answers included in the problem answer evaluation index data, and Figure 4(b) is a table showing an example of assumed problem answer evaluation data. [Figure 5] This is a flowchart illustrating the overall flow of information processing related to the said embodiment. [Figure 6] This figure shows the display screen of the user terminal when inputting problem-solving information, relating to the same embodiment. [Figure 7] This figure shows an example of a prompt display screen related to the same embodiment. [Figure 8] This figure shows an example of a display screen for response information shown on a user terminal, relating to the same embodiment. [Figure 9] This is a flowchart illustrating the overall flow of information processing in the second embodiment. [Figure 10] This figure shows the display screen of a user terminal, including a question form, relating to the same embodiment. [Figure 11] This figure shows an example of a prompt display screen related to the same embodiment. [Figure 12] This figure shows an example of a display screen for response information shown on a user terminal, relating to the same embodiment. [Figure 13] This figure relates to a third embodiment and shows the display screen of a user terminal including a question form. [Figure 14] This figure shows the display screen of a user terminal in which question information has been entered into a question form, relating to the same embodiment. [Figure 15] This figure shows an example of a display screen for response information shown on a user terminal, relating to the same embodiment. [Figure 16] This figure shows a display screen (1) of a predetermined program specification displayed on a user terminal, relating to the fourth embodiment. [Figure 17] This figure shows a display screen (2) of a predetermined program specification displayed on a user terminal, relating to the same embodiment. [Figure 18] This figure shows an example of a display screen for response information (when the predetermined rules are not met) that is displayed on the user terminal in relation to the same embodiment. [Figure 19] This figure shows an example of a display screen for response information (when the predetermined rules are met) that is displayed on the user terminal in relation to the same embodiment. [Modes for carrying out the invention]

[0044] One embodiment of the present invention will be described below with reference to the figures.

[0045] 1. First Embodiment <Overview of the Information Processing System> The learning support information processing system 1 (hereinafter referred to as "System 1") using a generative AI according to this embodiment is a system that supports user learning, and as shown in Figure 1, a user terminal 20, a server 30, and a generative AI 40 are connected via a communication network (hereinafter referred to as "Network 10"). There may be multiple user terminals 20 and servers 30.

[0046] Network 10 may be a wired communication method, a wireless communication method, or a combination of both.

[0047] <User terminal 20> The user terminal 20 is, for example, an information processing device such as a PC, tablet, or smartphone. The user terminal 20 comprises a control unit 210, a communication unit 220, an input unit 230, an output unit 240, and a storage unit 250, which are connected to each other via signal lines 260 (see Figure 2).

[0048] The control unit 210 includes, for example, one or more processors, including a CPU (Central Processing Unit) or an MPU (Micro Processing Unit). The control unit 210 has the function of controlling, processing, and managing the operation of the entire user terminal 20 by referring to the program stored in the memory unit 250.

[0049] The control unit 210 has the following functions: - A function that requests access to learning data from server 30, and based on the learning data including learning content data and learning segment data obtained from server 30 in response to the request, as well as task data, displays learning information, such as a learning video, on the output unit 240 to allow the user to perform the learning, and also displays task information on the output unit 240 at predetermined timings (for example, learning segments of learning content identified by the learning segment data). • A function to acquire learning logs, such as learning history, generated by the user's learning process. This learning history corresponds to the learning segment data received from server 30, which allows for the identification of the learning content the user has studied and its progress. • A function to acquire task answer data entered by the user via the input unit 230 and link said task answer data to the learning history. Also, a function to acquire other input data. • A function to transmit data to an external device, including a server 30, via the communication unit 220, and to receive data from said external device. • A function that converts data to be sent and received into a specified format. - A function that performs processing to display data input via the input unit 230 and data received from the external device on the output unit 240. - A function that, when inputting or outputting data to or from the user terminal 20, executes a process to display a user-operable graphical user interface (GUI) screen on the output unit 240. - A function to store data acquired and generated by the user terminal 20 and data transmitted and received via the communication unit 220 in the storage unit 250.

[0050] The communication unit 220 is a communication means that connects the user terminal 20 to the network 10 to perform communication. Communication may be performed via wired or wireless connection. The communication protocol is also not particularly limited. The data received by the communication unit 220 is sent to the control unit 210 via the signal line 260.

[0051] The input unit 230 may be installed internally in the user terminal 20 or it may be externally attached. The input unit 230 may be, for example, a mouse, buttons, a keyboard, a scanner, a microphone capable of voice input, or a touch panel integrated with the output unit 240. The input unit 230 has the function of accepting data input operations from the user.

[0052] The output unit 240 may be installed internally in the user terminal 20 or it may be externally attached. The output unit 240 may be, for example, a display (including the touch panel described above), a speaker, a printer, a head-mounted display, etc. The output unit 240 has functions for displaying data on a screen, outputting audio, printing data, etc.

[0053] The memory unit 250 is a storage device such as RAM, ROM, or storage, and may be installed internally in the user terminal 20 or be externally connected. The storage unit 250 stores data transmitted and received by the user terminal 20, programs referenced to execute each process, and data acquired and generated by each process. This includes data that is temporarily stored.

[0054] The user terminal 20 configured in this way requests access to the server 30, and in response, the server 30 obtains learning data, which the terminal then uses to have the user perform learning. At predetermined times, the terminal provides the user with task information, and when it receives the user's input operation for task answer information, it links the user's learning history to the obtained task answer data and sends it to the server 30. The user terminal 20 also outputs and displays answer information based on the answer data obtained from the server 30 on the output unit 240.

[0055] <Server 30> Server 30 is an information processing device. Server 30 comprises a control unit 310, a communication unit 320, and a storage unit 330, which are connected to each other via signal lines 340 (see Figure 3). Furthermore, as described above, Server 30 is connected to the generation system AI 40 via the network 10.

[0056] The control unit 310 includes, for example, one or more processors, including a CPU or MPU. The control unit 310 has the function of controlling, processing, and managing the operation of the entire server 30 by referring to a program stored in the storage unit 330. Furthermore, the control unit 310 stores the data transmitted and received via the communication unit 320 in the storage unit 330.

[0057] Furthermore, the control unit 310 also functions as a user terminal identification unit 312, a learning data control unit 314, a prompt creation unit 316, and an answer generation control unit 318.

[0058] The user terminal identification unit 312, upon receiving an access request from the user terminal 20, has the function of identifying the user terminal 20 to which learning data and assignment data will be transmitted, based on the user terminal information stored in the user DB 332 of the storage unit 330. The user terminal identification unit 312 also has the function of storing the learning log (learning history) obtained from the user terminal 20 in the user DB 332 of the storage unit 330.

[0059] The learning data control unit 314 has the function of identifying the learning data and task data to be transmitted to the identified user terminal 20 by referring to the learning data and task data stored in the learning DB 334 of the storage unit 330.

[0060] The prompt creation unit 316, upon acquiring the above-mentioned problem answer data from the user terminal 20, has the function of creating a prompt by referring to the problem answer data, the learning segment data, problem data, correct answer data, and problem answer evaluation index data stored in the learning DB 334 of the storage unit 330, and the instruction condition data stored in the instruction condition DB 336. The prompt creation unit 316 may also be configured to select predetermined instruction conditions from the instruction condition data according to the type of learning data corresponding to the above-mentioned problem answer data (e.g., learning subject, etc.).

[0061] The response generation control unit 318 has the function of transmitting a prompt created by the prompt creation unit 316 to the generation system AI 40 via the network 10, and causing the generation system AI 40 to generate response data based on that prompt.

[0062] Furthermore, the control unit 310 has functions such as transmitting learning data to a predetermined user terminal 20 via the communication unit 320, receiving various data including learning logs and task answer data from the user terminal 20, receiving answer data generated by the generation AI 40, transmitting and receiving other predetermined data, and converting the transmitted and received data into a predetermined format.

[0063] The communication unit 320 is a communication means that connects the server 30 to the network 10 and performs communication. Communication may be performed via wired or wireless connection. The communication protocol is also not particularly limited. The data received by the communication unit 320 is sent to the control unit 310 via the signal line 340.

[0064] The memory unit 330 is a storage device such as RAM, ROM, or storage, and may be installed internally in the server 30 or connected to the server 30 via the network 10. The memory unit 330 stores data transmitted and received by the server 30, programs referenced to execute each process, data generated by each process, etc. This data includes temporarily stored data. Furthermore, the memory unit 330 includes a user DB 332, a learning DB 334, an instruction condition DB 336, and an answer DB 338.

[0065] The user database 332 stores information for identifying the user terminal 20 (such as the user ID assigned to each user), information on available learning data, and learning logs obtained from the user terminal 20.

[0066] The learning DB334 stores the following information. • Learning content data, learning segment data linked to the learning content (e.g., segment like Chapter 1, Part 1), and assignment data linked to the learning data. • Correct answer data linked to the above problem data. The correct answer data is model answer information corresponding to the problem. • Problem-solving evaluation index data linked to the above problem data. The problem-solving evaluation index data includes evaluation manual data that includes perspectives, checkpoints, and methods for evaluating problem-solving responses from users, and assumed problem-solving evaluation data that includes expected problem-solving responses from users and assumed evaluations of those responses. An example of the data (table) included in the problem-solving evaluation index data is shown in Figure 4.

[0067] The instruction condition DB336 stores predetermined instruction condition data that is referenced when creating prompts. Examples of such predetermined instruction conditions include the following: • Generate answers based on the learning history of the user terminal 20 and the corresponding learning content. The system does not include the correct answer to the problem itself, but generates a response that includes recommendation information referencing the correct answer data and problem-solving evaluation index data. • Outputs the response in the specified output format. The recommendation information mentioned above includes hints and recommendations for users' answers to the challenges, and is intended to support users in arriving at the model answer.

[0068] The response DB338 stores the above response data received from the generation system AI40.

[0069] When the server 30 configured in this way receives an access request from a user terminal 20, it identifies the user terminal 20 that made the access request and sends predetermined learning data and assignment data to that user terminal 20. Furthermore, when the server 30 obtains task answer data from the user terminal 20, it creates a predetermined prompt by referring to the task answer data, learning segment data, task data, correct answer data, task answer evaluation index data, and instruction condition data, and sends this to the generation system AI 40, causing the generation system AI 40 to generate response data based on the prompt, and receives the response data generated by the generation system AI 40.

[0070] <Generation AI40> The generative AI 40 is a type of artificial intelligence technology and is an information processing device that has the ability to learn from large amounts of data and generate text and other data. In this embodiment, the generative AI 40 has been further trained on data including the above-mentioned training data, task data, correct answer data, and task answer evaluation index data in order to refer to these data. Additional training means, for example, transfer learning and fine tuning. Note that the above-mentioned data may include other data. Furthermore, the generation system AI40 may refer to the above-mentioned correct answer data by performing a so-called RAG (Retrieval-Augmented Generation) that searches an external database containing the above-mentioned correct answer data. The same applies to other embodiments.

[0071] When the generation AI 40 receives the above prompt from the server 30, it generates response data based on the prompt, which includes the above recommendation information that corresponds to the user's learning history and refers to the above correct answer data and the above task answer evaluation index data, and sends the generated response data to the server 30.

[0072] <Information Processing Methods> The following describes the information processing method using System 1 (see Figure 5).

[0073] First, the user terminal 20 accesses the server 30 via a predetermined application (not shown) and requests the server 30 to access the training data (step S102).

[0074] When the control unit 310 of the server 30 receives an access request from the user terminal 20 (step S202), it identifies the user terminal 20 to which learning data and task data will be transmitted based on the user terminal information stored in the user DB 332 of the storage unit 330 (step S204). The control unit 310 then refers to the learning DB 334 of the storage unit 330 to identify the learning data and task data to be transmitted to the identified user terminal 20 (step S206), and transmits them to the user terminal 20 (step S208).

[0075] The control unit 210 of the user terminal 20 receives learning data and task data from the server 30 (step S104), and based on the learning content data included in the received learning data, displays learning information, such as a learning video, on the output unit 240 to allow the user to perform the learning (step S106). During the execution of the learning, the learning video is processed to be streamed. Furthermore, during the learning process, the control unit 210 may acquire the learning log (learning history) from the user terminal 20 each time, temporarily store the acquired learning history in the storage unit 250, and send it to the server 30 as appropriate. This learning history corresponds to the learning segment data received from the server 30, thereby identifying the learning content the user has studied and its progress.

[0076] The control unit 210 of the user terminal 20 displays task information on the output unit 240 based on task data received from the server 30 during the execution of learning or at a learning segment based on received learning segment data (step S108). In this embodiment, the task is to create source code based on predetermined conditions. That is, the control unit 210 causes the output unit 240 to display the corresponding task information (creation of source code) at a learning segment of a predetermined learning content (see Figure 6). An example in which task information is displayed on the output unit 240 at the end of Chapter 1, Part 2 will be described below.

[0077] When a user operates the input unit 230 of the user terminal 20 to input problem-solving information corresponding to the above problem into the input form 6 of Figure 6 and clicks the submit button, the control unit 210 of the user terminal 20 associates the learning history (Chapter 1, Part 2) with the problem-solving data obtained in this way and sends it to the server 30 (Step S112).

[0078] When the control unit 310 of the server 30 receives the above-mentioned problem answer data linked to the learning history from the user terminal 20 (step S210), it creates a prompt for the generation system AI 40 to generate answer data by referring to the problem answer data, the learning segment data, problem data, correct answer data and problem answer evaluation index data stored in the learning DB 334 of the storage unit 330, and the instruction condition data stored in the instruction condition DB 336 (step S212, see Figure 7). The above prompts are created based on instruction conditions such as, for example, "generate an answer within the range of learning content corresponding to the learning history (learning segment) of the user terminal 20 to be acquired," "generate an answer that does not include the correct answer to the task itself, but includes recommendation information that refers to the correct answer data and the task answer evaluation index data," and "output the answer in a predetermined format."

[0079] The control unit 310 of the server 30 sends the created prompt to the generation system AI 40 (step S214).

[0080] When the generation AI 40 receives a prompt from the server 30 (step S302), it analyzes the prompt (step S304) and generates response data based on the analysis results (step S306).

[0081] The generation AI 40 sends the generated response data to the server 30 (step S308). The generated response data is generated within the scope of the user terminal 20's learning history and corresponding learning content, and does not include the correct answer to the task itself, but it does include recommendation information that refers to the above correct answer data and the above task answer evaluation index data, and is output in a predetermined format.

[0082] The control unit 310 of the server 30 receives response data from the generation system AI 40 (step S216), converts the received response data into a predetermined format (step S218), and transmits it to the user terminal 20 (step S220).

[0083] The control unit 210 of the user terminal 20 receives response data from the server 30 (step S114), and based on the received response data, outputs and displays the response information on the output unit 240 (step S116, see Figure 8).

[0084] As described above, according to this embodiment, in response to a user's answer to a problem, the user terminal 20 provides an answer within the scope of its learning history and corresponding learning content, making it possible to provide the user with an answer that matches their learning level. Furthermore, Server 30 causes the generative AI 40, which has been further trained to refer to data including the above-mentioned learning data, task data, correct answer data, and task answer evaluation index data, to generate answer data. As a result, even for specialized tasks, it is possible to provide users with appropriate answers that analyze the user's task answers in a timely manner while maintaining a certain level of accuracy and consistency. This makes it possible for System 1 to provide highly convenient learning support to many users at any time. Furthermore, while the above responses do not include the correct answer to the problem itself, they do include recommendation information, such as hints and recommendations for user-submitted solutions to the problem, which helps users arrive at the model answer. This provides users with an opportunity to think for themselves and supports their growth.

[0085] 2. Second Embodiment The same reference numerals are used for components that are identical or similar to those in the first embodiment, and detailed descriptions of parts other than the differences are omitted.

[0086] In this embodiment, the user terminal 20 receives question form data from the server 30. Note that sending task data to the user terminal 20 is optional.

[0087] <User terminal 20> When the control unit 210 of the user terminal 20 receives the above-mentioned question form data, it displays the question form on the output unit 240 at a predetermined timing, and the input unit 230 of the user terminal 20 is configured to accept the input of question information from the user using the question form. Furthermore, the control unit 210 associates the learning history with the question data obtained from the input of question information from the user, and transmits this to the server 30 via the communication unit 320.

[0088] <Server 30> The control unit 310 (user terminal identification unit 312) of the server 30 has the function of identifying the user terminal 20 to which learning data and question form data will be sent, based on the user terminal information stored in the user DB 332 of the storage unit 330, when it receives an access request from the user terminal 20.

[0089] The prompt creation unit 316 has the function of creating a prompt by referring to the question data, learning segmentation data, question answer format data, and instruction condition data stored in the instruction condition DB 336 after obtaining question data from the user terminal 20.

[0090] The learning DB 334 of the memory unit 330 further stores question-answer format data for generating answers based on predetermined conditions. The question-answer format data includes, for example, information to improve the accuracy of the answers, and is data that links the type of learning data (learning subject), keyword conditions included in the question, and information that should be included in the answer. That is, for example, it is data that includes information such as, "Learning subject: Creating a calculator application for programming; Keyword conditions: Includes "error", "calculation", and "output"; Answer: Generate an answer that takes into account the type of operator, the type of code, and the order in which the code is written for performing arithmetic operations."

[0091] The instruction condition DB336 includes instructions to generate an answer that includes a question and answer that references question and answer format data, as predetermined instruction condition data. In this embodiment, the system is configured to also include instructions to generate an answer that includes additional information related to the question (referred to simply as "additional information" in this embodiment), in addition to the question and answer.

[0092] When the server 30 configured in this way receives an access request from a user terminal 20, it identifies the user terminal 20 that made the access request and sends predetermined learning data and question form data to that user terminal 20. Furthermore, when the server 30 obtains question data from the user terminal 20, it creates a predetermined prompt by referring to the question data, learning segmentation data, question answer format data, and instruction condition data, sends this to the generating AI 40, and receives the answer data generated by the generating AI 40 based on the prompt.

[0093] <Generation AI40> In this embodiment, the generation system AI 40 is trained on additional data including the above-mentioned question answer format data in order to refer to the data including this data. Note that the above data may include other data.

[0094] When the generation AI 40 receives the above prompt from the server 30, it generates response data that includes question and answer and additional information corresponding to the user's learning history and referring to the above question and answer format data, based on the prompt, and sends the generated response data to the server 30.

[0095] <Information Processing Methods> The following describes the information processing method using System 1 (see Figure 9). Note that the same processing as in the first embodiment will be omitted from the explanation as appropriate.

[0096] First, the user terminal 20 accesses the server 30 and requests access to the training data from the server 30 (step S402).

[0097] When the control unit 310 of the server 30 receives an access request from the user terminal 20 (step S402), it identifies the user terminal 20 to which learning data and question form data will be sent based on the user terminal information stored in the user DB 332 of the storage unit 330 (step S404). The control unit 310 then refers to the learning DB 334 of the storage unit 330 to identify the learning data and question form data to be sent to the identified user terminal 20 (step S506), and sends them to the user terminal 20 (step S508).

[0098] The control unit 210 of the user terminal 20 receives learning data and question form data from the server 30 (step S404), and based on the learning content data included in the received learning data, displays a learning video on the output unit 240 to allow the user to perform the learning (step S406).

[0099] The control unit 210 of the user terminal 20 displays a question form on the output unit 240 based on the question form data received from the server 30 during the execution of learning or at a learning segment based on the received learning segment data (step S408). In this embodiment, the control unit 210 displays a question form on the output unit 240 at a predetermined learning section of the learning content (e.g., Chapter 1, Part 2) (see Figure 10). Below, an example in which a question form (input form) is displayed on the output unit 240 at the end of Chapter 1, Part 2 will be described.

[0100] When a user operates the input unit 230 of the user terminal 20 to enter question information into the input form in Figure 10 and clicks the submit button, the control unit 210 of the user terminal 20 associates the acquired question data with the learning history (Chapter 1, Part 2) and sends it to the server 30 (Step S412).

[0101] When the control unit 310 of the server 30 receives the above-mentioned question data linked to the learning history from the user terminal 20 (step S510), it creates a prompt for the generation system AI 40 to generate answer data by referring to the question data, learning segmentation data, question answer format data, and instruction condition data stored in the instruction condition DB 336 (step S512, see Figure 11). The above prompts are created based on instruction condition data such as, for example, "generate answers to questions within the range of the learning history of the user terminal 20 to be acquired and the corresponding learning data," "generate answers to questions that refer to question answer format data," "include additional information related to the question in the answer to the question," and "output the answer in a predetermined output format."

[0102] The control unit 310 of the server 30 sends the created prompt to the generation system AI 40 (step S514).

[0103] When the generation AI 40 receives the above prompt from the server 30 (step S602), it analyzes the prompt (step S604) and generates response data based on the analysis results (step S606).

[0104] The generation AI 40 sends the generated response data to the server 30 (step S608). The generated response data is generated within the scope of the user terminal 20's learning history and corresponding learning content, and includes answers to questions and additional information that refer to the above-mentioned question-answer format data, and is output in a predetermined format.

[0105] The control unit 310 of the server 30 receives response data from the generation system AI 40 (step S516), converts the received response data into a predetermined format (step S518), and transmits it to the user terminal 20 (step S520).

[0106] The control unit 210 of the user terminal 20 receives response data from the server 30 (step S114), and based on the received response data, outputs and displays the response information on the output unit 240 (step S416, see Figure 12).

[0107] As described above, according to this embodiment, in response to a question from the user, an answer is provided within the scope of the learning history of the user terminal 20 and the corresponding learning content, making it possible to provide the user with an answer that is appropriate to the user's learning level. Furthermore, the server 30 causes the generative AI 40, which has been further trained with data including the above-mentioned learning data and question-answer format data, to generate answer data. As a result, even for specialized questions, the system 1 can provide users with appropriate answers in a timely manner while maintaining a certain level of accuracy and consistency. This enables the system 1 to provide highly convenient learning support to many users at any time. Furthermore, since the above answers include additional information related to the user's question, it is possible to give the user an opportunity to think for themselves and support their growth.

[0108] 3. Third Embodiment The same reference numerals are used for components that are identical or similar to those in the first and second embodiments, and detailed descriptions of components other than the differences are omitted.

[0109] In this embodiment, as in the second embodiment, question form data is sent from the server 30 to the user terminal 20. The question form displayed on the output unit 240 of the user terminal 20 allows the user to input "What you want to solve," "Prerequisites," "What you have tried / what you have understood so far," and "Supplementary information" for each item (see Figure 13).

[0110] <Server 30> The server's prompt creation unit 316 has the function of creating a prompt by referring to the question data, learning segmentation data, question answer format data, and instruction condition data stored in the instruction condition DB 336 after obtaining question data from the user terminal 20. The prompt includes an instruction to generate answer data from the above question, which includes "summary of the question," "answer to the question or hint information for the question," and "question-related additional information related to the answer or hint information for the question." The above-mentioned "answer to the question or hint information for the question" refers to either a direct answer to the question or hint information that guides the user to an answer, but is not a direct answer to the question. For example, one of these is selected by a program stored in the memory unit 330, based on the type of learning data (e.g., learning subject). In this embodiment, the above-mentioned answer data is configured to include a direct answer to the question.

[0111] Furthermore, if the above answer data includes hint information for the above question, the prompt generation unit 316 may refer to the hint information data stored in the learning DB 334 of the storage unit 330 in order to generate a prompt to obtain the hint information. The hint information data includes, for example, information on the thought steps necessary to arrive at an answer to the user's question, and includes information such as, "Answer generation conditions: Learning about Eclipse; Question includes the keywords "screen," "arrow," "panel," and "layout"; Hint information: Includes an explanation of the Perspective function."

[0112] When the server 30 configured in this way obtains question data from the user terminal 20, it creates a predetermined prompt by referring to the question data, learning segmentation data, and instruction condition data, sends this to the generating AI 40, and receives the answer data generated by the generating AI 40 based on the prompt.

[0113] <Generation AI40> When the generation AI 40 receives the above prompt from the server 30, it generates response data based on the prompt, corresponding to the user's learning history, and including "summary of the question," "answer to the question or hint information for the question," and "question-related additional information related to the answer to the question or hint information," and sends the generated response data to the server 30.

[0114] <Information Processing Methods> The following describes the information processing method using System 1. Since the overall flow of information processing in this embodiment is the same as in the second embodiment, the explanation will be given using Figure 9. Furthermore, explanations of processes similar to those in the first and second embodiments will be omitted as appropriate.

[0115] First, the user terminal 20 accesses the server 30 and requests access to the training data from the server 30 (step S402).

[0116] When the control unit 310 of the server 30 receives an access request from the user terminal 20 (step S502), it identifies the user terminal 20 to which learning data and question form data will be sent based on the user terminal information stored in the user DB 332 of the storage unit 330 (step S504). The control unit 310 then refers to the learning DB 334 of the storage unit 330 to identify the learning data and question form data to be sent to the identified user terminal 20 (step S506), and sends them to the user terminal 20 (step S508).

[0117] The control unit 210 of the user terminal 20 receives learning data and question form data from the server 30 (step S304), and based on the learning content data included in the received learning data, displays a learning video on the output unit 240 to allow the user to perform the learning (step S406).

[0118] The control unit 210 of the user terminal 20 displays a question form on the output unit 240 based on the question form data received from the server 30 during the execution of learning or at a learning segment based on the received learning segment data (step S108). In this embodiment, the control unit 210 displays a question form on the output unit 240 at a predetermined learning section of learning content (e.g., Chapter 1, Part 3) (see Figure 13). Below, an example in which a question form (input form) is displayed on the output unit 240 at the end of Chapter 1, Part 3 will be described.

[0119] When a user operates the input unit 230 of the user terminal 20 to enter question information into the input form and clicks the submit button (see Figure 14), the control unit 210 of the user terminal 20 associates the acquired question data with the learning history (Chapter 1, Part 3) and sends it to the server 30 (Step S412). In this embodiment, the input form allows input of "What you want to solve," "Prerequisites," "What you have tried / what you have understood so far," and "Supplementary information" for each item. This enables the generation AI 40 to generate question summaries and the like more accurately.

[0120] When the control unit 310 of the server 30 receives the above-mentioned question data, to which the learning history is linked, from the user terminal 20 (step S510), it creates a prompt to cause the generation system AI 40 to generate answer data by referring to the question data, learning segmentation data, and instruction condition data stored in the instruction condition DB 336. The above prompts are created based on instruction condition data such as, for example, "generate answers to questions within the range of the learning history of the user terminal 20 to be acquired and the corresponding learning data," "generate a summary of the question, the answer to the question, and additional information related to the answer to the question from the question," and "output the answer to the question in a predetermined output format."

[0121] The control unit 310 of the server 30 sends the created prompt to the generation system AI 40 (step S514).

[0122] When the generation AI 40 receives the above prompt from the server 30 (step S602), it analyzes the prompt (step S604) and generates response data based on the analysis results (step S606).

[0123] The generation AI 40 sends the generated response data to the server 30 (step S608). The generated response data is generated within the scope of the user terminal 20's learning history and corresponding learning content, and includes a summary of the question, the answer to the question, and additional question-related information related to the answer to the question, and is output in a predetermined format.

[0124] The control unit 310 of the server 30 receives response data from the generation system AI 40 (step S516), converts the received response data into a predetermined format (step S518), and transmits it to the user terminal 20 (step S520).

[0125] The control unit 210 of the user terminal 20 receives response data from the server 30 (step S414), and based on the received response data, outputs and displays the response information on the output unit 240 (step S416, see Figure 15).

[0126] As described above, according to this embodiment, in response to a question from the user, an answer is provided within the scope of the learning history of the user terminal 20 and the corresponding learning content, making it possible to provide the user with an answer that is appropriate to the user's learning level. Furthermore, the server 30 causes the generative AI 40, which has been further trained with data including the aforementioned learning data, to generate response data. As a result, even for specialized questions, it can provide users with appropriate answers in a timely manner while maintaining a certain level of accuracy and consistency. This enables the system 1 to provide highly convenient learning support to many users at any time. Furthermore, since the above response includes a summary of the question, the answer to the question or hints related to the question, and additional information related to the answer to the question, it is possible to provide answers that are easy for users to understand and to give users an opportunity to think for themselves, thereby supporting user growth.

[0127] 4. Fourth Embodiment The same reference numerals are used for components that are identical or similar to those in the first embodiment, and detailed descriptions of parts other than the differences are omitted.

[0128] In this embodiment, the user terminal 20 receives data from the server 30 that allows it to input source code associated with a predetermined program specification as the problem data, and the input unit 230 of the user terminal 20 receives source code input from the user based on the predetermined program specification. Hereinafter, the source code entered by the user will be referred to as "user-input source code".

[0129] <Server 30> The prompt creation unit 316 has the function of creating a prompt by obtaining user input source code from the user terminal 20, referring to the user input source code, learning segmentation data, program specification data associated with the learning data, source code associated with the program specification data, and pass review data that defines the rules that the source code should satisfy, as well as instruction condition data stored in the instruction condition DB 336.

[0130] The learning DB 334 of the memory unit 330 stores program specification data associated with the learning data, source code associated with the program specification data, and acceptance review data that defines the rules that the source code must satisfy. The acceptance review data is evaluation data that includes, for example, the following aspects: • Is the user-input source code executable (does the program work)? Does the user-input source code meet the requirements of the corresponding program specification? • Is the user-input source code consistent? Does the user-input source code conform to the specified style?

[0131] The instruction condition DB336 stores instructions as predetermined instruction condition data, which determine whether the user-input source code satisfies the rules defined in the acceptance review data, and if the rules are met, generate a response that includes related program specification data and additional information (review-related additional information) about the source code associated with it.

[0132] When the server 30 configured in this way obtains user input source code from the user terminal 20, it creates a predetermined prompt by referring to the user input source code, learning segmentation data, and instruction condition data, sends this to the generation system AI 40, and receives the response data generated by the generation system AI 40 based on the prompt.

[0133] <Generation AI40> When the generation AI 40 receives the above prompt from the server 30, it generates response data based on the prompt, which includes a determination result of whether the user input source code satisfies the rules defined in the above-mentioned passing review data, according to the user's learning history, and if the rules are met, it includes the above-mentioned additional information, and sends the generated response data to the server 30.

[0134] <Information Processing Methods> The following describes the information processing method using System 1. Since the overall flow of information processing in this embodiment is the same as in the first embodiment, we will use Figure 5 for explanation. Furthermore, explanations of processes similar to those in the first embodiment will be omitted as appropriate.

[0135] First, the user terminal 20 accesses the server 30 and requests access to the training data from the server 30 (step S102).

[0136] When the control unit 310 of the server 30 receives an access request from a user (step S202), it identifies a user terminal 20 to which learning data and task data should be transmitted based on the user terminal information stored in the user DB 332 of the storage unit 330 (step S204). The control unit 310 then refers to the learning DB 334 of the storage unit 330 to identify the learning data and task data to be transmitted to the identified user terminal 20 (step S206), and transmits them to the user terminal 20 (step S208).

[0137] The control unit 210 of the user terminal 20 receives learning data and task data from the server 30 (step S104), and based on the learning content data included in the received learning data, displays a learning video on the output unit 240 to allow the user to perform the learning (step S106).

[0138] The control unit 210 of the user terminal 20 displays task information on the output unit 240 based on task data received from the server 30 during the execution of learning or at a learning segment based on received learning segment data (step S108). In this embodiment, the task is to create source code (source code for creating a web application with predetermined functions) based on predetermined program specifications, and the control unit 210 displays a source code creation screen (task information) corresponding to the learning segment of predetermined learning content on the output unit 240 (see Figures 16 and 17). An example in which task information is displayed on the output unit 240 at the end of Chapter 1, Part 4 will be described below.

[0139] When a user operates the input unit 230 of the user terminal 20 to input source code based on the above task (predetermined program specifications) into an input form (not shown) and clicks the submit button, the control unit 210 of the user terminal 20 associates the input source code, i.e., the user-input source code, with the learning history (Chapter 1, Part 4) and sends it to the server 30 (Step S112). Note that the user-input source code may also be input by uploading an electronic file.

[0140] When the control unit 310 of the server 30 receives the user input source code associated with the learning history from the user terminal 20 (step S210), it refers to the user input source code, learning segment data, program specification data associated with the learning data, the source code associated with the program specification data, the pass review data that defines the rules that the source code should satisfy, and the instruction condition data stored in the instruction condition DB 336 to create a prompt for the generation system AI 40 to generate response data. The above prompts are created based on instructional conditions such as, for example, "generate an answer within the scope of the learning history of the user terminal 20 to be acquired and the corresponding learning content," and "determine whether the input source code satisfies the rules that the source code associated with the relevant program specification must satisfy, and generate additional information related to the program specification and its associated source code only if the rules are satisfied, and if the rules are not satisfied, generate an answer including the parts that are not satisfied, the reasons for the non-satisfaction, and suggested corrections."

[0141] The control unit 310 of the server 30 sends the created prompt to the generation system AI 40 (step S214).

[0142] When the generation AI 40 receives a prompt from the server 30 (step S302), it analyzes the prompt (step S304) and generates response data based on the analysis results (step S306).

[0143] The generation AI 40 sends the generated response data to the server 30 (step S308). The generated response data is generated within the scope of the user terminal 20's learning history and corresponding learning content. If the user input source code satisfies the rules that the source code associated with the relevant program specification must satisfy, it includes the program specification and additional information related to the source code associated with it (review additional information). If the above rules are not satisfied, it includes the parts that are not satisfied, the reasons for the non-satisfaction, and suggested corrections, and is output in a predetermined format.

[0144] The control unit 310 of the server 30 receives response data from the generation system AI 40 (step S216), converts the received response data into a predetermined format (step S218), and transmits it to the user terminal 20 (step S220).

[0145] The control unit 210 of the user terminal 20 receives response data from the server 30 (step S114), and based on the received response data, outputs and displays the response information on the output unit 240 (step S116, see Figures 18 and 19). In other words, Figure 18 shows the response information displayed in the output unit 240 when the user-input source code does not satisfy the rules that the source code associated with the corresponding program specification must satisfy. Furthermore, in this embodiment, the user-input source code can be entered multiple times, and Figure 19 shows the response information displayed in the output unit 240 when the re-entered user-input source code satisfies the above rules.

[0146] As described above, according to this embodiment, since the user input source code is provided with answers within the scope of the learning history of the user terminal 20 and the corresponding learning content, it becomes possible to provide the user with answers that are appropriate to the user's learning level. Furthermore, Server 30 causes Generative AI 40, which has been further trained with data including program specification data linked to the learning data, source code linked to the program specification data, and passing review data that defines the rules that the source code should satisfy, to generate response data. As a result, it can provide users with timely and appropriate reviews (responses) based on predetermined rules, while maintaining a certain level of accuracy and consistency for the user-input source code. This enables System 1 to provide highly convenient learning support to many users at any time. Furthermore, the above response includes, if the user-input source code satisfies the rules that source code associated with the relevant program specification should meet, additional information related to the program specification and its associated source code (review supplementary information). If the above rules are not met, it includes the parts that are not met, the reasons for the non-compliance, and suggested corrections. This supports users in finding source code that satisfies the above rules themselves, and also allows for the provision of advanced information according to the user's level of achievement (after understanding the source code rules), thereby appropriately supporting the user's growth.

[0147] 5. Modified Examples of the First to Fourth Embodiments At least one of the above problem answer information, question information, and user input source code may be configured to be input as image data. In this case, the prompt generation unit 316 of the server 30 is configured to analyze the character information and image information contained in the image data and generate a prompt that includes an instruction to generate a response based on the results of the analysis. Alternatively, for example, a trained AI trained using image analysis data may be stored in the storage unit 330 of the server 30, and an image recognition unit (not shown) provided in the server 30 may use the trained AI to analyze the character information and image information contained in the image data, and the prompt creation unit 316 may be configured to generate a prompt that includes an instruction to generate a response based on the analysis results.

[0148] Furthermore, for example, before creating a prompt, the control unit 310 of the server 30 may be configured to have the generation AI 40 or other generation AI perform an analysis of the image data, and the prompt creation unit 316 may generate a prompt based on the results of that analysis.

[0149] 6. Other variations The timing of displaying each input form on the output unit 240 of the user terminal 20, and the timing of inputting the problem answer information, question information, and user input source code into each input form, can be changed as appropriate.

[0150] Furthermore, the function of the server 30's control unit 310 as a prompt generation unit 316 can be delegated to another server (not shown) via an API (Application Programming Interface). Furthermore, the generation AI 40 may be configured to generate other response data by referring to previously generated response data (history).

[0151] Furthermore, the prompt generation unit 316 of the server 30 may be configured to repeatedly generate prompts. That is, when the prompt generation unit 316 creates the first prompt, it causes the generation system AI 40 to create the first response data based on that first prompt. Furthermore, when the server 30 receives the initial response data, the prompt creation unit 316 creates a prompt (first improvement prompt) to check whether there are any improvements to the initial response data and to create a response data that takes those improvements into account, and causes the generation system AI 40 to generate the first improved response data based on the first improvement prompt. When server 30 receives the first improvement response data, prompt creation unit 316 creates a prompt (second improvement prompt) to confirm whether there are any improvements to the first improvement response data and to create response data based on those improvements, and causes generation system AI 40 to generate the first improvement response data based on the second improvement prompt. In this case, the number of times the prompt creation unit 316 creates improvement prompts is controlled based on a program stored in the storage unit 330, and when the prompt creation unit 316 creates an improvement prompt, it is configured to refer to improvement prompt data stored in the instruction condition DB of the storage unit 330.

[0152] Furthermore, the program in this embodiment is a program that causes the user terminal 20 and the server 30 to execute the above-mentioned process. Furthermore, the information processing device of this embodiment is an information processing device that performs the above-mentioned processing.

[0153] It should be noted that the present invention is not limited to the embodiments described above, and various modifications are possible without departing from the spirit of the invention. [Explanation of symbols]

[0154] 1. A learning support information processing system (system) using generative AI. 10. Network (communication network) 20 ... User terminal 30 ... Server 40 …Generative AI

Claims

1. Equipped with one or more information processing devices, The above information processing device is A means of providing learning data that provides users with learning data linked to user information, An input data acquisition means for acquiring the user's learning history identified from the above learning data and the user's input data linked to said learning history, A prompt acquisition means that acquires a prompt including an instruction to generate a response corresponding to the learning history based on the data including the above input data, The system includes a means for generating response data that causes a generative AI to generate response data corresponding to the learning history based on the above prompt, The above-mentioned generative AI refers to data including the above-mentioned training data, as well as related data including task data associated with the training data, correct answer data associated with the task data, and task answer evaluation index data. The above learning data provision means further provides the above task data to the user, The above input data acquisition means acquires problem answer data from users corresponding to the above problem data as the above input data, The prompt acquisition means acquires the prompt, which includes an instruction to generate the above answer, which includes recommendation information that does not include the correct answer to the problem itself, based on the above correct answer data and the above problem answer evaluation index data for the above problem answer data, and The above-mentioned response data generation means is a learning support information processing system that uses a generative AI to cause the above-mentioned generative AI to generate the above-mentioned response data based on the above-mentioned prompt.

2. The above-mentioned generative AI is configured to refer to data that includes question-answer format data linked to the above-mentioned training data as related data. The above input data acquisition means acquires question data from the user as the above input data, The prompt acquisition means acquires the prompt which includes an instruction to generate the answer, which includes a question answer based on the question answer format data for the question data, The learning support information processing system using the generation AI according to claim 1, wherein the above-mentioned response data generation means causes the generation AI to generate the above-mentioned response data based on the above-mentioned prompt.

3. The above input data acquisition means acquires question data from the user as the above input data, The prompt acquisition means acquires the prompt from the question data, which includes a summary of the question, an answer or hint information for the question, and an instruction to generate the answer, which includes additional question-related information relating to the answer or hint information. The learning support information processing system using the generation AI according to claim 1, wherein the above-mentioned response data generation means causes the generation AI to generate the above-mentioned response data based on the above-mentioned prompt.

4. The above-mentioned generative AI is configured to refer to data including program specification data linked to the above-mentioned training data, source code linked to said program specification data, and pass / fail review data that defines the rules that said source code should satisfy, as the above-mentioned related data. The above input data acquisition means acquires source code entered by the user based on predetermined program specification data as the above input data, The prompt acquisition means acquires a prompt that includes an instruction to determine whether the source code entered by the user satisfies the rules defined in the approved review data, and if the rules are met, to generate the above response which includes additional review-related information. The learning support information processing system using the generation AI according to claim 1, wherein the above-mentioned response data generation means causes the generation AI to generate the above-mentioned response data based on the above-mentioned prompt.

5. The above input data acquisition means acquires the above input data including image data, The prompt acquisition means acquires the prompt which includes an instruction to generate the response, which includes a response based on information extracted from the image data. The learning support information processing system using the generation AI according to claim 1, wherein the above-mentioned response data generation means causes the generation AI to generate the above-mentioned response data based on the above-mentioned prompt.

Images