System

Abstract

A system is provided.SOLUTION: A system comprising: means for automatically generating test questions for a particular subject; means for acquiring information based on requested criteria; means for invoking a generative AI model to generate new test questions from the acquired information; and means for formatting and outputting the generated test questions and commentary.SELECTED DRAWING: Figure 1

Description

[Technical Field]

[0001] The technology of the present disclosure relates to a system. [Background technology]

[0002] Patent document 1 discloses a persona chatbot control method performed by at least one processor, the method including the steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to a description of the chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance in response to the user utterance. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Japanese Patent Publication No. 2022-180282 Summary of the Invention [Problem to be solved by the invention]

[0004] Conventional test question creation not only requires a great deal of effort from teachers and instructors, but also makes it difficult to adjust the difficulty of the questions, which can have a negative impact on student grades. Furthermore, even if past questions are used as reference when creating questions, there are limitations to extracting appropriate questions or generating new questions. Therefore, there is a need for efficient and effective test question generation. [Means for solving the problem]

[0005] To solve the above problems, the present invention provides the following means: a system including a generation means for automatically generating test questions for specific subjects, an acquisition means for acquiring data based on requested conditions, a calling means for calling a generative AI model that generates new test questions from the acquired data, and an output means for formatting and outputting the generated test questions and explanations. The system also includes a display means for displaying the generated test questions and explanations on a user interface, a communication means for requesting and responding to test question generation between a server and a terminal, an input means for a user to input conditions for test question generation, and a transmission means for transmitting requests to the server based on the input conditions, thereby improving the efficiency and quality of test question creation and significantly reducing the burden on teachers and instructors.

[0006] A "specific subject" is an academic field or subject for which test questions are generated.

[0007] "Exam item" means a question or task used to assess a learner's understanding or knowledge.

[0008] "Generation means" refers to the method or technology for automatically creating test questions for a specific subject.

[0009] "Acquisition means" refers to the methods and techniques used to collect the required data based on the requested conditions.

[0010] A "generative AI model" is a model that uses artificial intelligence technology to generate new test questions and explanations.

[0011] "Invocation means" refers to the method or technology used to run a generative AI model to generate new test questions.

[0012] "Output means" refers to the method or technology for outputting the generated test questions and explanations in an appropriate format.

[0013] "User interface" refers to an interface that includes a screen and input devices through which a user interacts with a system.

[0014] "Display means" refers to the method or technology used to display test questions and explanations on a user interface.

[0015] "Communication means" refers to the method or technology for sending and receiving data between a server and a terminal.

[0016] "Input means" refers to the method or technique by which a user inputs test question generation conditions into the system.

[0017] "Transmission means" refers to the method or technology for transmitting the conditions entered by the user to the server.

[0018] A "request" is an operation or message that a user issues to the system requesting that it generate test questions under specific conditions.

[0019] A "database" is a recording medium that stores information such as past exam questions and study materials. [Brief explanation of the drawings]

[0020] [Figure 1] 1 is a conceptual diagram showing an example of the configuration of a data processing system according to a first embodiment. [Figure 2] 1 is a conceptual diagram showing an example of main functions of a data processing device and a smart device according to a first embodiment. [Figure 3] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a second embodiment. [Figure 4] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and smart glasses according to a second embodiment. [Figure 5] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a third embodiment. [Figure 6]FIG. 11 is a conceptual diagram showing an example of main functions of a data processing device and a headset-type terminal according to a third embodiment. [Figure 7] FIG. 10 is a conceptual diagram showing an example of the configuration of a data processing system according to a fourth embodiment. [Figure 8] FIG. 10 is a conceptual diagram showing an example of main functions of a data processing device and a robot according to a fourth embodiment. [Figure 9] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 10] 1 shows an emotion map onto which multiple emotions are mapped. [Figure 11] FIG. 3 is a sequence diagram showing a processing flow of the data processing system according to the first embodiment. [Figure 12] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 1. [Figure 13] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system according to the second embodiment when an emotion engine is combined. [Figure 14] FIG. 10 is a sequence diagram showing the flow of processing in the data processing system in Application Example 2 when an emotion engine is combined. DETAILED DESCRIPTION OF THE INVENTION

[0021] An example of an embodiment of a system according to the technology of the present disclosure will be described below with reference to the accompanying drawings.

[0022] First, the terms used in the following description will be explained.

[0023] In the following embodiments, a coded processor (hereinafter simply referred to as a "processor") may be a single arithmetic device or a combination of multiple arithmetic devices. Furthermore, a processor may be a single type of arithmetic device or a combination of multiple types of arithmetic devices. Examples of arithmetic devices include a CPU (Central Processing Unit), a GPU (Graphics Processing Unit), a GPGPU (General-Purpose computing on Graphics Processing Units), and an APU (Accelerated Processing Unit).

[0024] In the following embodiments, a coded RAM (Random Access Memory) is a memory in which information is temporarily stored and is used as a working memory by a processor.

[0025] In the following embodiments, the coded storage is one or more non-volatile storage devices that store various programs, various parameters, etc. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), and magnetic tapes.

[0026] In the following embodiments, a communication I / F (Interface) with a symbol is an interface including a communication processor, an antenna, etc. The communication I / F controls communication between multiple computers. Examples of communication standards applied to the communication I / F include wireless communication standards including 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), Bluetooth (registered trademark), etc.

[0027] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." In other words, "A and / or B" means that it may be only A, only B, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" is also applied when three or more things are expressed connected by "and / or."

[0028] [First embodiment]

[0029] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0030] 1, a data processing system 10 includes a data processing device 12 and a smart device 14. An example of the data processing device 12 is a server.

[0031] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0032] The smart device 14 includes a computer 36, a reception device 38, an output device 40, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The reception device 38, the output device 40, and the camera 42 are also connected to the bus 52.

[0033] The reception device 38 includes a touch panel 38A, a microphone 38B, and the like, and receives user input. The touch panel 38A detects contact with an indicator (for example, a pen or a finger) to receive user input by the touch of the indicator. The microphone 38B detects the user's voice to receive user input by voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the data indicating the user input.

[0034] The output device 40 includes a display 40A and a speaker 40B, and presents data to the user 20 by outputting the data in a form of expression that the user 20 can perceive (for example, audio and / or text). The display 40A displays visible information such as text and images in accordance with instructions from the processor 46. The speaker 40B outputs audio in accordance with instructions from the processor 46. The camera 42 is a compact digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0035] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 control the exchange of various information between the processor 46 and the processor 28 via the network 54.

[0036] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0037] 2, in the data processing device 12, a specific process is performed by the processor 28. A specific processing program 56 is stored in the storage 32. The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific process is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0038] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0039] In the smart device 14, the processor 46 performs the reception output process. The storage 50 stores a reception output program 60. The reception output program 60 is used in conjunction with the specific processing program 56 by the data processing system 10. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0040] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0041] This invention relates to a system that uses a generative AI model to create and explain test questions. The purpose of this system is to reduce the burden on question creators in educational institutions and qualification exams, and to provide high-quality test questions with appropriate difficulty.

[0042] System Overview

[0043] server

[0044] The server receives test question creation requests from users, acquires the appropriate data, and generates new test questions and explanations using a generative AI model. Specifically, it performs the following processes:

[0045] 1. Accepting user input

[0046] The server receives requests from users through API endpoints.

[0047] The request data includes detailed conditions such as test subjects, question format, difficulty level, and number of questions.

[0048] Analyze the received request data and extract each parameter.

[0049] 2. Referencing the database

[0050] The server retrieves appropriate question data from a past exam question database based on the specified exam subject.

[0051] A database query is used to extract data that matches the input criteria and perform additional filtering.

[0052] 3. Invoking the generative AI model

[0053] The server preprocesses the acquired data and converts it into a format that can be input into the generative AI model.

[0054] New test questions and explanations are generated using a generative AI model.

[0055] 4. Configuring the Output

[0056] The generated test questions and explanations are formatted and returned to the user.

[0057] Terminal

[0058] The terminal is responsible for the interface with the user, accepting input, sending requests to the server, and displaying the results. Specifically, it performs the following processes:

[0059] 1. Submit a request

[0060] The terminal receives the user's input and sends the conditions for creating test questions to the server as a request.

[0061] 2. Receiving a response

[0062] The test questions and explanations data returned from the server are received and displayed on the user interface.

[0063] User

[0064] The user is the entity that inputs the conditions for creating test questions and uses the system to obtain original test questions. Specifically, the user performs the following operations:

[0065] 1. Enter the conditions

[0066] The user inputs the conditions for creating test questions through the terminal and sends a request.

[0067] 2. Check the results

[0068] Check the test questions and explanations sent from the server and make a request again if necessary.

[0069] Specific examples

[0070] For example, if a user wants to create five multiple-choice questions for intermediate level math, the process is as follows:

[0071] 1. Users

[0072] The user enters "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions, and then sends a request from the device.

[0073] 2. Terminal

[0074] The terminal transmits the conditions entered by the user to the server.

[0075] 3. Server

[0076] The server analyzes the received conditions and retrieves the appropriate data from the past exam question database.

[0077] A generative AI model is called and new test questions and their explanations are generated based on the acquired data.

[0078] The generated test questions and explanations are formatted and sent back to the terminal.

[0079] 4. Terminal

[0080] The terminal receives the test questions and explanations returned from the server and displays them to the user.

[0081] 5. Users

[0082] The user checks the test questions and explanations displayed on the terminal and makes a request again if necessary.

[0083] This system allows teachers and instructors to efficiently create high-quality test questions, contributing to improving students' academic abilities. In addition, users can easily customize test questions, making it extremely flexible and fast to use in educational settings.

[0084] The processing flow will be explained below.

[0085] Program processing flow

[0086] server

[0087] Step 1:

[0088] The API endpoint receives requests from users. The request data includes detailed conditions such as test subject, question format, difficulty level, and number of questions.

[0089] Step 2:

[0090] Parse the received request data and extract each parameter (exam subject, question format, difficulty level, number of questions).

[0091] Step 3:

[0092] Access the appropriate past exam question database based on the exam subject. Use a database query to extract past exam question data that matches the input criteria.

[0093] Step 4:

[0094] The extracted data is preprocessed and converted into a format that can be input to the generative AI model, for example by tokenizing the text data and adding necessary metadata.

[0095] Step 5:

[0096] A generative AI model is then invoked to generate new test questions and explanations using the preprocessed data as input, with each question containing four options and a correct answer.

[0097] Step 6:

[0098] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[0099] Step 7:

[0100] The formatted data is sent to the terminal as an API response.

[0101] Terminal

[0102] Step 1:

[0103] Through the user interface, the user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[0104] Step 2:

[0105] The entered conditions are combined into a single JSON request data, for example, { "Subject": "Mathematics", "Format": "Multiple Choice", "Difficulty": "Intermediate", "Number of Questions": 5}.

[0106] Step 3:

[0107] Send the request data to the server using an HTTP request (POST method).

[0108] Step 4:

[0109] Receive the response data (test questions and explanations) sent back from the server as an HTTP response (200 OK).

[0110] Step 5:

[0111] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[0112] Step 6:

[0113] The formatted test questions and explanations are displayed in the user interface, and are output to the screen so that the user can check them.

[0114] User

[0115] Step 1:

[0116] Operate the terminal's user interface and enter the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[0117] Step 2:

[0118] Once you have completed the input, click the send button on your device to send the request.

[0119] Step 3:

[0120] Check that the response from the server is displayed on the terminal. The test questions and explanations are displayed on the screen so that the user can check them.

[0121] Step 4:

[0122] To revise the test questions and explanations as needed, enter the criteria again and submit a new request.

[0123] Example 1

[0124] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0125] Conventional test question creation not only requires a significant amount of time and effort from the question creator, but also has the problem of difficulty in maintaining appropriate level of difficulty and quality. Furthermore, efficiently utilizing a database of past questions to generate new test questions is also a technical challenge. There is a need to solve these problems and improve the efficiency and quality of test question creation in educational institutions and qualification examinations.

[0126] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0127] In this invention, the server includes input means for accepting a test question creation request from a user, analysis means for analyzing the received request data and extracting conditions such as test subject, question format, difficulty level, and number of questions, acquisition means for acquiring appropriate data from a past question database based on the specified test subject, preprocessing means for preprocessing the acquired data and converting it into a format to be input to the generative AI model, generation means for generating new test questions and explanations using the generative AI model, output means for formatting and outputting the generated test questions and explanations, and communication means for returning the output data to the user. This makes it possible to automatically generate test questions that meet the conditions specified by the user and quickly provide high-quality, validated questions and explanations.

[0128] A "user" is an entity that uses the system to make a request to create test questions.

[0129] A "server" is a computer system that receives requests from users, generates test questions and explanations, and returns them to the users.

[0130] A "terminal" is a device that a user uses to input test question creation conditions and check the results.

[0131] "Input means" is a function that accepts a test question creation request from a user.

[0132] The "analysis means" is a function that analyzes the request data received from the user and extracts conditions such as test subjects, question format, difficulty level, and number of questions.

[0133] "Acquisition means" is a function that acquires data from a past exam question database based on a specified exam subject.

[0134] "Preprocessing means" is a function that converts acquired data into a format that can be input into a generative AI model.

[0135] "Generation means" is a function that generates new test questions and explanations using a generative AI model.

[0136] "Output means" is a function that formats the generated test questions and explanations and provides them to the user.

[0137] "Communication means" refers to the function that sends and receives requests for test question generation between the server and the terminal.

[0138] A "database" is a collection of information in which past exam questions are stored.

[0139] A "generative AI model" is an artificial intelligence model used to automatically generate test questions.

[0140] A "prompt sentence" is an instruction sentence that causes a generative AI model to generate test questions.

[0141] The "user interface" is a display screen for displaying the generated test questions and explanations.

[0142] This invention relates to a system that leaves the creation of test questions and explanations to a generative AI model. Specifically, it shows the process of automatically generating test questions through interactions between a server, a terminal, and a user.

[0143] Server Operation

[0144] The server is a high-performance computer system that has the following functions:

[0145] 1. Input method:

[0146] The server accepts test question creation requests from users via an API endpoint. This input is sent in JSON format and includes the test subject, question format, difficulty level, number of questions, etc.

[0147] 2. Analysis method:

[0148] The server analyzes the received JSON data and extracts each parameter (exam subject, question format, difficulty level, number of questions). This analysis process is implemented using programming languages ​​such as Python and JavaScript.

[0149] 3. Acquisition method:

[0150] The server connects to a database of past exam questions and retrieves the appropriate data based on the selected exam subject. This database stores past exam questions and related information.

[0151] 4. Pretreatment methods:

[0152] The acquired data is converted into a format that can be input to the generative AI model, specifically by pre-processing the data to convert it into text-based prompts.

[0153] 5. Generation means:

[0154] Generate new test questions and their explanations using a generative AI model (e.g., OpenAI's GPT-4), using the following prompt:

[0155] "Please create a test based on the following criteria:

[0156] Subject: Mathematics

[0157] Question format: 4 choices

[0158] Difficulty: Intermediate

[0159] Number of questions: 5"

[0160] 6. Output Method:

[0161] The generated test questions and explanations are formatted and returned to the user in JSON format, which includes the question text, options, correct answers, explanations, etc.

[0162] 7. Means of communication:

[0163] The server sends the generated test questions and explanations to the user's device using the HTTP protocol.

[0164] Device behavior

[0165] The terminal provides an interface for users to create test questions and has the following functions:

[0166] 1. Input method:

[0167] The terminal accepts the conditions for creating test questions as input from the user, which is usually provided in the form of a form.

[0168] 2. Means of transmission:

[0169] The conditions entered by the user are converted into JSON format and sent to the server using an HTTP POST request.

[0170] 3. Receiving means:

[0171] The JSON data returned from the server is received and the test questions and explanations are displayed in the user interface, which can be provided as a browser-based UI or a dedicated application.

[0172] User Actions

[0173] 1. Enter the conditions:

[0174] The user inputs the conditions for creating test questions through a terminal, including the subject, question format, difficulty level, and number of questions.

[0175] 2. Check the results:

[0176] The user can check the test questions and explanations returned by the server and make a request again if necessary. For example, if the difficulty or quality of the questions does not meet their expectations, they can change the conditions and make a request again.

[0177] Specific examples

[0178] For example, to create five multiple-choice questions for intermediate level math, the user enters the following conditions into the input form on the device:

[0179] Subject: Mathematics

[0180] Question format: 4 choices

[0181] Difficulty: Intermediate

[0182] Number of questions: 5

[0183] The device converts this data into JSON format and sends it to the server. The server analyzes the input conditions and retrieves the appropriate question data from the database. Based on this, it invokes the generative AI model to generate new test questions and explanations, which are then formatted and sent back to the device. The device then displays the received data on the user interface so that the user can view it.

[0184] This system will improve the efficiency and quality of test question creation, and can be used flexibly in educational institutions and qualification exams.

[0185] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0186] Step 1:

[0187] Accepting user input

[0188] Specific explanation

[0189] The user enters the conditions for creating test questions into the input form on the terminal. These include the test subject (e.g., "Mathematics"), question format (e.g., "Multiple Choice"), difficulty level (e.g., "Intermediate"), and number of questions (e.g., "5"). When the user clicks the "Submit" button, these conditions are converted into JSON format.

[0190] input

[0191] Exam subjects, question format, difficulty level, number of questions

[0192] output

[0193] Request data in JSON format

[0194] Step 2:

[0195] Submitting a Request

[0196] Specific explanation

[0197] The terminal converts the conditions entered by the user into JSON format and sends it to the server as an HTTP POST request.

[0198] input

[0199] Request data in JSON format

[0200] output

[0201] HTTP POST request

[0202] Step 3:

[0203] Accepting a request

[0204] Specific explanation

[0205] The server receives requests from users via the API endpoint and parses the request data, extracting conditions such as test subject, question format, difficulty level, and number of questions from the JSON-formatted request data.

[0206] input

[0207] HTTP POST request

[0208] output

[0209] Extracted data on exam subjects, question format, difficulty level, and number of questions

[0210] Step 4:

[0211] Database Reference

[0212] Specific explanation

[0213] The server retrieves appropriate question data from the past exam database based on the extracted conditions. For example, it uses an SQL query to search for "intermediate" level questions in the "Mathematics" subject.

[0214] sql

[0215] SELECT FROM question_bank WHERE subject = 'Mathematics' AND difficulty = 'Intermediate' LIMIT 10;

[0216] input

[0217] Exam subjects and difficulty level

[0218] output

[0219] Appropriate problem data

[0220] Step 5:

[0221] Data Preprocessing

[0222] Specific explanation

[0223] The server converts the acquired data into a format that can be input to the generative AI model. For example, it formats the acquired problem data into a list format.

[0224] text

[0225] {"question": "What is the solution to x^2 + 2x + 1 = 0?", "choices": ["1", "-1", "0", "2"], "answer": "1"}

[0226] input

[0227] Appropriate problem data

[0228] output

[0229] Data formats that can be input to generative AI models

[0230] Step 6:

[0231] Invoking a generative AI model

[0232] Specific explanation

[0233] The server uses the converted data to send prompts to a generative AI model (e.g., GPT-4) to generate new test questions and explanations.

[0234] text

[0235] "Please create a test based on the following criteria:

[0236] Subject: Mathematics

[0237] Question format: 4 choices

[0238] Difficulty: Intermediate

[0239] Number of questions: 5"

[0240] input

[0241] Prompt statement

[0242] output

[0243] New exam questions and explanations

[0244] Step 7:

[0245] Configuring the Output

[0246] Specific explanation

[0247] The generated test questions and explanations are formatted and converted into JSON format for sending back to the user. For example, formatting it as follows:

[0248] json

[0249] {

[0250] "questions": [

[0251] {"question": "What is the solution to x^2 + 2x + 1 = 0?", "choices": ["1", "-1", "0", "2"], "answer": "1", "explanation": "Using the solution formula..."}

[0252] ]

[0253] }

[0254] input

[0255] New exam questions and explanations

[0256] output

[0257] JSON format data

[0258] Step 8:

[0259] Sending a response

[0260] Specific explanation

[0261] The server sends the formatted JSON data to the device as an HTTP response.

[0262] input

[0263] JSON format data

[0264] output

[0265] HTTP response

[0266] Step 9:

[0267] Receiving a response

[0268] Specific explanation

[0269] The terminal receives the response data from the server, analyzes it, and displays the test questions and explanations on the user interface.

[0270] input

[0271] HTTP response

[0272] output

[0273] Test questions and explanations displayed on the user interface

[0274] Step 10:

[0275] Checking the results

[0276] Specific explanation

[0277] The user can check the test questions and explanations displayed on the terminal and make a re-request if necessary.

[0278] input

[0279] Test questions and explanations displayed on the user interface

[0280] output

[0281] User confirmation / request

[0282] (Application example 1)

[0283] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0284] Conventional test question creation systems require test creators to expend a great deal of time and effort, making it difficult to quickly generate high-quality test questions. Furthermore, it is difficult to automatically generate new test questions by referencing past questions, which often results in inconsistent question quality and difficulty. Furthermore, there was a lack of a mechanism for instantly providing users with the generated test questions and explanations, making it difficult to provide appropriate learning support. There was a need to provide a system that could solve these issues and quickly and efficiently generate and provide test questions to users, particularly in online education and qualification exams.

[0285] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0286] In this invention, the server includes a generation means for automatically generating test questions for a specific domain, an acquisition means for acquiring information based on requested conditions, a call means for calling a generative AI model for generating new test questions from the acquired information, an output means for formatting and outputting the generated test questions and explanations, a prompt generation means for receiving input from a user via a communication interface and creating prompt sentences for generating test questions that match the conditions, and a display means for displaying the generated test questions and explanations on a user interface. This enables users to easily and quickly generate and check high-quality test questions and explanations.

[0287] A "generation means" is a device or system that has the functionality to automatically generate test questions for a particular domain.

[0288] An "acquisition means" is a device or system for acquiring the necessary information or data based on a request.

[0289] A "calling means" is a device or system that has the function of calling a generative AI model and generating new test questions based on the acquired information.

[0290] The "output means" is a device or system for formatting the generated test questions and explanations into an appropriate format and providing them to an external party.

[0291] A "communications interface" is a device or system that allows data to be sent and received between a server and a user's terminal.

[0292] A "prompt generation means" is a device or system that has the function of creating prompt sentences to pose appropriate questions to the generative AI model based on input conditions from the user.

[0293] The "display means" is a device or system for displaying the generated test questions and explanations on a user interface.

[0294] This invention relates to a system that uses a generative AI model to create and explain test questions. The purpose of this system is to reduce the burden on question creators in educational institutions, qualification exams, and other settings, and to provide high-quality test questions. Here, we will explain a specific embodiment of the system that realizes this application example.

[0295] server

[0296] The server includes the following means:

[0297] 1. Generation means

[0298] This function automatically generates test questions for specific areas. Based on the user's requested tasks, it retrieves information from a database of previous questions and generates new questions.

[0299] 2. Acquisition method

[0300] This function retrieves appropriate information from the past exam question database based on user requests, allowing for optimal data collection.

[0301] 3. Calling means

[0302] This function calls a generative AI model based on the data collected by the acquisition means and generates new test questions. This AI model is usually a generative AI model such as OpenAI.

[0303] 4. Output Method

[0304] This function formats the generated test questions and explanations appropriately and outputs them to the user. This ensures that the generated content is presented to the user in a format that is easy for them to understand.

[0305] 5. Prompt Generation Methods

[0306] This function creates prompt sentences that ask appropriate questions to the generative AI model based on input conditions from the user, thereby eliciting appropriate responses from the AI ​​model.

[0307] 6. Display means

[0308] This function displays the generated test questions and explanations on the user interface.

[0309] Terminal

[0310] The terminal includes the following means:

[0311] 1. Communication Interface

[0312] This function sends and receives data between the server and the user, allowing for smooth requests and responses for test questions.

[0313] User

[0314] The user does the following:

[0315] 1. Input Method

[0316] The conditions for generating test questions are entered into the terminal and a request is sent.

[0317] 2. Transmission Method

[0318] A request is sent to the server based on the entered conditions.

[0319] 3. Display means

[0320] Check the generated test questions and explanations on the device interface.

[0321] For example, if a user wants to create five multiple-choice questions for the intermediate level of "Mathematics," he or she inputs the conditions into the terminal and sends a request to the server. The terminal transmits this information through the communication interface, and the server uses the acquisition means to collect appropriate data and creates a prompt sentence using the generation means and prompt generation means.

[0322] Here is an example prompt:

[0323] Create five multiple-choice questions and explanations for intermediate level mathematics.

[0324] The generative AI model generates new test questions and explanations based on this prompt, formats them using an output method, and returns them to the device. The user can then check the generated test questions and explanations on the device and make further requests as necessary. This allows users to create high-quality test questions efficiently and quickly.

[0325] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0326] Step 1:

[0327] The user uses the input means of the terminal to input conditions for generating test questions. Specifically, detailed conditions such as test subjects, question format, difficulty level, number of questions, etc. This input data is sent to the terminal as request data.

[0328] Step 2:

[0329] The terminal receives the request data and sends it to the server via a communication interface. At this time, the request data includes the test subject, question format, difficulty level, number of questions, etc. The terminal converts the request data into the required format and sends it to the server in JSON format, etc.

[0330] Step 3:

[0331] The server receives and analyzes the request data. As a result of the analysis, the test conditions desired by the user (test subject, question format, difficulty level, number of questions) are extracted. Based on the extracted conditions, the server references the past question database and retrieves past question data related to the specified test subject.

[0332] Step 4:

[0333] The retrieval means retrieves appropriate data from the past exam question database, including extracting questions that match specified criteria using a search query, and stores the extracted data in a temporary data store.

[0334] Step 5:

[0335] The server uses the prompt generation means to create a prompt sentence based on the user's input conditions. For example, a sentence such as "Please create five multiple-choice questions and explanations for intermediate level mathematics" is generated. This prompt sentence is prepared to be passed to the generative AI model.

[0336] Step 6:

[0337] The calling means calls the generative AI model based on the prompt sentence. Specifically, it sends the prompt sentence to the generative AI model, and generates new test questions and explanations in response. The generative AI model (e.g., OpenAI) generates appropriate questions and explanations according to the prompt sentence and returns the results to the server.

[0338] Step 7:

[0339] The generated test questions and explanations are formatted by the output means. The formatted data is converted into a format that is easy for the user to understand (e.g., text, PDF, HTML, etc.). The formatted data is then prepared in a format for transmission to the terminal.

[0340] Step 8:

[0341] The server sends the formatted test questions and explanations through a communications interface to the terminal, which receives them and prepares them for display on the user interface.

[0342] Step 9:

[0343] The terminal uses the display means to display the generated test questions and explanations on the user interface, allowing the user to check them on the terminal and make a re-request if necessary.

[0344] If the user changes the conditions and makes a request again, the process is repeated from step 1. This realizes a system that can quickly and efficiently provide high-quality test questions and explanations.

[0345] Furthermore, an emotion engine that estimates the user's emotion may be combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59 and perform identification processing using the user's emotion.

[0346] This invention combines a system that automatically generates test questions for specific subjects with an emotion engine that recognizes user emotions. The purpose of this system is to reduce the burden on question creators in educational institutions and qualification exams, and to improve the quality of users' learning experience.

[0347] System Overview

[0348] server

[0349] The server receives test question creation requests from users, retrieves appropriate data, and generates new test questions and explanations using a generative AI model. It also uses an emotion engine to recognize user emotions and adjust the difficulty of the test questions.

[0350] 1. Accepting user input

[0351] The server receives requests from users through API endpoints. The request data includes detailed conditions such as test subject, question format, difficulty level, number of questions, etc. It also includes emotion recognition data.

[0352] Parse the received request data and emotion data and extract each parameter.

[0353] 2. Referencing the database

[0354] The server retrieves the appropriate question data from the past exam database based on the specified exam subject. It uses a database query to extract data that matches the input criteria and performs additional filtering.

[0355] 3. Invoking the generative AI model

[0356] The server preprocesses the acquired data and converts it into a format that can be input to the generative AI model. It also adjusts the difficulty of the questions based on the emotional data.

[0357] It uses a generative AI model to generate new test questions and explanations, each with four options and a correct answer.

[0358] 4. Configuring the Output

[0359] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[0360] Feedback information including an evaluation by the emotion recognition engine is generated and sent back to the user along with the test questions.

[0361] Terminal

[0362] The device is responsible for interfacing with the user, accepting input, sending requests to the server, and displaying the results. It also contains an emotion engine that recognizes the user's emotions and sends the data to the server.

[0363] 1. Submit a request

[0364] The terminal receives user input, sends the conditions for creating test questions to the server as a request, and also recognizes the user's emotions using an emotion engine and sends the data to the server.

[0365] 2. Receiving a response

[0366] Response data (test questions, explanations, and emotion evaluation results) is received from the server as an HTTP response (200 OK).

[0367] 3. Displaying the results

[0368] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[0369] The formatted test questions, explanations, and emotion evaluation results are displayed in the user interface, and are output to the screen so that the user can check them.

[0370] User

[0371] The user is the entity that inputs the conditions for creating test questions and uses the system to obtain original test questions. In addition, the emotion engine feeds back the user's emotions during learning to the system, improving the quality of the learning experience.

[0372] 1. Enter the conditions

[0373] The user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions) through the terminal.

[0374] 2. Emotional Feedback

[0375] The emotion engine recognizes the user's operating status and facial expressions, and transmits them to the server as emotion data.

[0376] 3. Check the results

[0377] The response from the server is displayed on the terminal. The test questions, explanations, and emotion evaluation results are displayed on the screen for the user to review.

[0378] To revise the test questions and explanations as needed, enter the criteria again and submit a new request.

[0379] Specific examples

[0380] For example, if a user wants to create five multiple-choice questions at an intermediate level in mathematics and wants advice on how to reduce stress levels while studying, the process is as follows:

[0381] 1. Users

[0382] Enter "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions, and send a request from your device. The emotion engine will also capture your learning progress.

[0383] 2. Terminal

[0384] The request data and emotion data are sent to the server.

[0385] 3. Server

[0386] The system analyzes request data, retrieves appropriate data from a database of past questions, and adjusts the difficulty of questions based on emotional data.

[0387] A generative AI model is invoked to generate new test questions and explanations.

[0388] The test questions, explanations, and emotional assessment results are formatted and sent back to the terminal.

[0389] 4. Terminal

[0390] The test questions, explanations, and emotion evaluation results returned from the server are received and displayed on the user interface.

[0391] 5. Users

[0392] Check the test questions, explanations, and emotional assessment results, and receive appropriate advice based on your learning progress.

[0393] If necessary, re-requests will be made to adjust the content and difficulty of the test questions.

[0394] The processing flow will be explained below.

[0395] server

[0396] Step 1:

[0397] The API endpoint receives requests from users. The request data includes test subject, question format, difficulty level, number of questions, and sentiment data.

[0398] Step 2:

[0399] The received request data and emotion data are parsed and each parameter (exam subject, question format, difficulty level, number of questions, emotion data) is extracted.

[0400] Step 3:

[0401] Access the appropriate past exam question database based on the exam subject. Use a database query to extract past exam question data that matches the input criteria.

[0402] Step 4:

[0403] The extracted data is preprocessed and converted into a format that can be input to the generative AI model, for example by tokenizing the text data and adding necessary metadata.

[0404] Step 5:

[0405] The difficulty of the generated test questions is adjusted based on the emotional data. For example, if the user is feeling stressed, the difficulty level is set slightly lower.

[0406] Step 6:

[0407] A generative AI model is then invoked to generate new test questions and explanations using the preprocessed data as input, with each question containing four options and a correct answer.

[0408] Step 7:

[0409] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[0410] Step 8:

[0411] Feedback information including evaluations by the emotion recognition engine is generated and sent back to the terminal along with the test questions.

[0412] Terminal

[0413] Step 1:

[0414] Through the user interface, the user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[0415] Step 2:

[0416] The input conditions are compiled into a single JSON request data. The request also includes the user's emotion data recognized by the emotion engine. For example, the format is { "Subject": "Mathematics", "Format": "Multiple Choice", "Difficulty": "Intermediate", "Number of Questions": 5, "Emotion": "Stress"}.

[0417] Step 3:

[0418] Send the request data to the server using an HTTP request (POST method).

[0419] Step 4:

[0420] Response data (test questions, explanations, and emotion evaluation results) returned from the server is received as an HTTP response (200 OK).

[0421] Step 5:

[0422] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[0423] Step 6:

[0424] The formatted test questions, explanations, and emotion evaluation results are displayed in the user interface, and are output to the screen so that the user can check them.

[0425] User

[0426] Step 1:

[0427] Operate the terminal's user interface and enter the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[0428] Step 2:

[0429] Once you have completed the input, click the send button on your device to send the request.

[0430] Step 3:

[0431] The emotion engine recognizes the user's operating status and facial expressions in real time and generates emotion data, which is included in the request.

[0432] Step 4:

[0433] The request sent from the terminal is processed by the server, and the generated test questions, explanations, and emotion evaluation results are displayed on the terminal.

[0434] Step 5:

[0435] Check the test questions and explanations displayed on the screen and receive appropriate study advice based on the emotional assessment results.

[0436] Step 6:

[0437] If necessary, re-requests will be made to adjust the content and difficulty of the test questions.

[0438] Example 2

[0439] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0440] There is a need to reduce the burden on educational institutions and test question creators and improve the quality of users' learning experience.However, there is no system that can take into account the user's emotional state while studying and adjust the difficulty of questions individually, making it difficult to maintain learning motivation and provide adaptive education.

[0441] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[0442] In this invention, the server includes a generation means for automatically generating test questions for specific subjects, an acquisition means for acquiring data based on requested conditions, a call means for calling a generation AI model that generates new test questions from the acquired data, an output means for formatting and outputting the generated test questions and explanations, an emotion analysis means for recognizing the user's emotions and analyzing the data, a difficulty adjustment means for adjusting the difficulty of the questions based on the emotion data, and a feedback means for returning the generated test questions and explanations to the user and generating feedback including the emotion evaluation results. This reduces the burden on the question creator, improves the quality of the user's learning experience, and enables the difficulty of questions to be adjusted appropriately according to the emotional state of each user.

[0443] A "specific subject" refers to a specific field or academic area that serves as the subject matter for generating test questions.

[0444] A "test question" is material containing questions or tasks used to assess a user's knowledge and understanding.

[0445] "Generation means" refers to the processes and techniques used to automatically create test questions, primarily using generative AI models.

[0446] "Acquisition means" refers to the process or method of searching for and acquiring relevant data or information based on specific conditions.

[0447] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new test questions and explanations.

[0448] "Invocation means" refers to the process or method for running a generative AI model using acquired data to obtain the required output.

[0449] "Output means" refers to the process or method for outputting the generated test questions and explanations in a specified format.

[0450] "Emotion analysis means" refers to techniques and processes for detecting and analyzing a user's emotional state.

[0451] "Difficulty adjustment means" refers to a process or method for appropriately adjusting the difficulty of test questions based on emotional data.

[0452] "Feedback means" refers to a process or technology for returning the generated test questions, explanations, and emotional evaluation results to the user and providing feedback information.

[0453] "User interface" refers to the interface for exchanging information between the system and the user, and displays test questions, explanations, etc.

[0454] "Communication means" refers to the communication technology and protocol used to request and respond to test question generation between the server and the terminal.

[0455] "Input means" refers to a device or interface that allows a user to input conditions for generating test questions.

[0456] "Transmission means" refers to the process or method for sending a request to a server based on the conditions entered by the user.

[0457] MODE FOR CARRYING OUT THE INVENTION

[0458] This invention combines a system for automatically generating test questions for specific subjects with an emotion engine that recognizes user emotions. This system aims to reduce the burden on question creators in educational institutions and qualification exams and improve the quality of users' learning experience.

[0459] server

[0460] A server is a combination of hardware and software that performs multiple functions. In particular, it includes the following elements:

[0461] 1. Generation means

[0462] The server uses a generative AI model to automatically generate test questions for specific subjects. The generative AI model uses artificial intelligence algorithms to generate new questions and their explanations.

[0463] 2. Acquisition method

[0464] The server retrieves data from information sources such as a database of past exam questions based on the conditions requested by the user. To do this, it uses SQL queries to search for and retrieve data that matches the request conditions.

[0465] 3. Calling means

[0466] The server preprocesses the acquired data and converts it into a format that can be input to the generative AI model. It also calls the generative AI model to generate new test questions and explanations. Specifically, it passes the necessary data to the generative AI model as prompts.

[0467] 4. Output Method

[0468] The server formats the generated test questions and explanations in an appropriate format (e.g., JSON or HTML), outputs the data, and temporarily stores it before returning it to the user.

[0469] 5. Emotion analysis method

[0470] The server receives the user's emotional data and uses an emotion engine to analyze it, which detects various emotional states such as the user's stress level and concentration level.

[0471] 6. Difficulty adjustment method

[0472] The server adjusts the difficulty of the questions based on the emotional data, for example, setting the difficulty of the questions lower if the user indicates a high stress level.

[0473] 7. Feedback channels

[0474] The server returns the generated test questions and explanations, as well as the emotional evaluation results, to the user, providing feedback information, so that the user can receive appropriate feedback based on their learning progress and emotional state.

[0475] Terminal

[0476] The device is responsible for the interface with the user, accepting input, sending requests to the server, and displaying the results. It also recognizes the user's emotions and sends the data to the server.

[0477] 1. Submit a request

[0478] The terminal uses an HTTP request to transmit the test question creation conditions and emotion data received from the user to the server.

[0479] 2. Receiving a response

[0480] The terminal receives the response data returned from the server, which includes the test questions, explanations, and emotion evaluation results.

[0481] 3. Displaying the results

[0482] The terminal analyzes the received response data and displays it on the user interface. For example, it formats and displays test questions and explanations using HTML templates.

[0483] User

[0484] To use the system to obtain the original test questions, users must:

[0485] 1. Enter the conditions

[0486] The user inputs the conditions for creating test questions (for example, test subjects, question format, difficulty level, number of questions) through the terminal.

[0487] 2. Emotional Feedback

[0488] During learning, the user's emotional data is captured by the emotion engine through their operating status and facial expressions.

[0489] 3. Check the results

[0490] The user can check the test questions, explanations, and emotion evaluation results returned in response from the server by displaying them on the terminal. If necessary, the user can send a re-request to adjust the content and difficulty of the test questions.

[0491] Specific examples

[0492] For example, if a user wants to create five intermediate-level multiple-choice questions in math and get advice on how to reduce stress levels while studying, the process might go like this:

[0493] 1. Users

[0494] The user enters "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions into the device and sends a request. At this time, the emotion engine captures the user's emotions while studying.

[0495] 2. Terminal

[0496] The request data and emotion data are sent to the server.

[0497] 3. Server

[0498] The system analyzes the request data, retrieves appropriate data from a database of past exam questions, and adjusts the difficulty of the questions based on the emotion data. It then uses a generative AI model to generate new exam questions and explanations. The generated questions, explanations, and emotion evaluation results are formatted and sent back to the device.

[0499] 4. Terminal

[0500] Receives the data returned from the server and displays it in the user interface.

[0501] 5. Users

[0502] Review the test questions, explanations, and emotional evaluation results. If necessary, make a request again to adjust the content and difficulty of the test questions.

[0503] Examples of specific prompts to use

[0504] Create five multiple-choice questions for intermediate level math based on the following criteria, taking into account the user's stress level while studying:

[0505] Exam subject: Mathematics

[0506] Question format: 4 choices

[0507] Difficulty: Intermediate

[0508] Number of questions: 5

[0509] keyword

[0510] Generative AI model, prompt sentence

[0511] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0512] Step 1:

[0513] Accepting user input

[0514] server

[0515] The server receives requests from users through an HTTP API endpoint. Specifically, it receives JSON data (e.g., {"subject": "math", "difficulty": "intermediate", "question_count": 5}) in which the user enters conditions such as test subject, question format, difficulty level, and number of questions. It also receives the user's emotion data at the same time.

[0516] Input: Test question creation conditions (JSON format) and emotion data

[0517] Data processing / calculation: Parse JSON data and extract each parameter (exam subject, question format, difficulty level, number of questions)

[0518] Output: Extracted parameters and emotion data

[0519] Step 2:

[0520] Database Reference

[0521] server

[0522] The server references the past exam question database based on the criteria received from the user, and uses an SQL query to retrieve data that matches the request criteria (e.g. SELECT FROM problem_database WHERE subject = 'math' AND difficulty = 'intermediate').

[0523] Input: Extracted parameters (exam subject, difficulty level, etc.)

[0524] Data processing / calculation: Executing SQL queries and filtering retrieved data

[0525] Output: Filtered data

[0526] Step 3:

[0527] Invoking a generative AI model

[0528] server

[0529] The server preprocesses the acquired data and converts it into a format suitable for input to the generative AI model. It adjusts the difficulty of the questions based on the emotional data (e.g., lowering the difficulty level if the stress level is high). It then calls the generative AI model to generate new test questions and explanations. The generated questions include four options and a correct answer.

[0530] Input: filtered data, sentiment data

[0531] Data processing / calculation: Preprocessing data, converting it to an input format for the generative AI model, and invoking the generative AI model using prompts.

[0532] Output: Generated test questions and explanations (e.g., ["Q1: Question text...", "A: Choice 1", "B: Choice 2", "C: Choice 3", "D: Choice 4", "Answer: B"])

[0533] Step 4:

[0534] Configuring the Output

[0535] server

[0536] The server formats the generated test questions and explanations into an appropriate format (e.g., JSON format), and also generates feedback information including the emotion evaluation results, which is temporarily stored in a database for sending back to the user.

[0537] Input: Generated test questions and explanations, emotional evaluation results

[0538] Data processing / calculation: formatting test questions and explanations, generating feedback information

[0539] Output: Formatted exam questions, explanations and feedback information (JSON format)

[0540] Step 5:

[0541] Send request

[0542] Terminal

[0543] The terminal receives the user's input and sends the request data and emotion data to the server using an HTTP request.

[0544] Input: User-entered test question creation conditions and emotion data

[0545] Data processing / calculation: Converting input data to JSON format, creating and sending HTTP requests

[0546] Output: Request data sent to the server (JSON format)

[0547] Step 6:

[0548] Response reception

[0549] Terminal

[0550] The device receives the response data from the server. After receiving the HTTP status code 200 (OK), it acquires the test questions, explanations, and emotion evaluation results contained in the response body.

[0551] Input: Response data from the server

[0552] Data processing / calculation: Response data analysis (JSON parsing)

[0553] Output: Analyzed test questions, explanations, and emotion evaluation results

[0554] Step 7:

[0555] Displaying the results

[0556] Terminal

[0557] The device analyzes the received response data and displays it on the user interface. It uses HTML templates to format and display test questions, explanations, and emotion evaluation results.

[0558] Input: Analyzed test questions, explanations, and emotion evaluation results

[0559] Data processing / calculation: Rendering using HTML templates

[0560] Output: Formatted test questions, explanations, and emotional evaluation results (displayed on screen)

[0561] Step 8:

[0562] Entering conditions

[0563] User

[0564] The user inputs the conditions for creating test questions (test subjects, question format, difficulty level, number of questions) through the terminal. After the conditions are entered, the user clicks a button to send the request.

[0565] Input: Conditions for creating test questions

[0566] Data processing / calculation: None

[0567] Output: What you type appears on the terminal

[0568] Step 9:

[0569] Emotional feedback

[0570] User

[0571] The emotion engine captures the operation status and facial expressions during learning, and the emotion data is sent to the server.

[0572] Input: User operation status, facial expressions, etc.

[0573] Data processing / calculation: Analysis by emotion engine, data capture

[0574] Output: Captured emotion data

[0575] Step 10:

[0576] Checking the results

[0577] User

[0578] The test questions, explanations, and emotional evaluation results returned from the server are displayed on the device. If necessary, a request is made again to adjust the content and difficulty of the test questions.

[0579] Input: Test questions, explanations, and emotional evaluation results displayed on the terminal

[0580] Data processing / calculation: None

[0581] Output: User confirmation and re-request if necessary

[0582] (Application example 2)

[0583] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart device 14 will be referred to as a "terminal."

[0584] Conventional test question generation systems provide questions uniformly without considering the user's emotions, which reduces the user's learning efficiency. In addition, since feedback based on emotional data is not provided, users tend to feel stressed and may lose motivation to study.

[0585] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[0586] In this invention, the server includes a generation means for automatically generating test questions for specific subjects, an acquisition means for acquiring data based on requested conditions, a call means for calling a generation AI model for generating new test questions from the acquired data, an output means for formatting and outputting the generated test questions and explanations, an emotion recognition means for recognizing a user's emotions, a difficulty adjustment means for adjusting the difficulty of the questions based on the emotion data, and a feedback generation means for generating feedback information based on the emotion data. This makes it possible to provide test questions of an appropriate difficulty level that takes the user's emotions into consideration, improving learning efficiency and maintaining the user's motivation to learn.

[0587] The "means for automatically generating test questions for a specific subject" is a function for automatically generating test questions based on the subjects specified by the user.

[0588] The "acquisition means for acquiring data based on requested conditions" is a function for acquiring necessary data from a database according to conditions input by a user.

[0589] The "calling means for calling the generative AI model" is a function that executes the AI ​​model to generate new test questions using the acquired data.

[0590] The "output means for formatting and outputting the generated test questions and explanations" is a function for formatting the newly generated test questions and their explanations in an appropriate format and outputting them.

[0591] The "emotion recognition means for recognizing the user's emotions" is a function for analyzing the user's facial expressions, actions, etc. to grasp the user's emotional state.

[0592] The "difficulty level adjusting means for adjusting the difficulty level of questions based on emotional data" is a function for providing test questions of a difficulty level appropriate for the user based on the recognized emotional data.

[0593] The "feedback generating means for generating feedback information based on emotional data" is a function for providing appropriate advice and feedback based on the emotional state of the user.

[0594] The "display means for displaying the generated test questions and explanations on the user interface" is a function for displaying the generated test questions and explanations on the screen so that the user can check them.

[0595] "Communication means for requesting and responding to test question generation between the server and the terminal" is a function for sending and receiving requests for test question generation and the generated results between the server and the terminal.

[0596] The "input means for the user to input conditions for generating test questions" is a function for the user to input conditions such as the subject, difficulty level, and format of the test questions that he or she wants to generate.

[0597] The "transmission means for transmitting a request to the server based on the input conditions" is a function for transmitting the conditions input by the user to the server and requesting the generation of test questions.

[0598] The "emotion data transmission means for acquiring emotion data in real time and transmitting it to the server" is a function for collecting the user's emotions in real time and transmitting that data to the server.

[0599] This invention combines a system that automatically generates test questions for specific subjects with an emotion engine that recognizes the user's emotions. This system includes a generation means, an acquisition means, a calling means, an output means, an emotion recognition means, a difficulty level adjustment means, and a feedback generation means. It is particularly characterized by using emotion data to adjust the difficulty of questions and provide appropriate feedback.

[0600] System Overview

[0601] server

[0602] The server has the following features:

[0603] Generation method: Automatically generate test questions for specified subjects.

[0604] Acquisition method: Retrieves the necessary data from the database based on a user request.

[0605] Calling method: Input the acquired data into the generative AI model to generate new test questions.

[0606] Output method: Format the generated test questions and explanations and output them in an appropriate format.

[0607] Emotion recognition means: Recognizes the user's emotions using cameras and sensors and acquires them as data.

[0608] Difficulty adjustment method: Adjust the difficulty of the questions based on emotional data.

[0609] Feedback generation means: Generates feedback information based on emotion data and provides it to the user.

[0610] The hardware used includes servers, databases, and computing resources for running the AI ​​model, while the software includes an emotion recognition engine (EmotionEngine) and a generative AI model (QuestionGenerator).

[0611] Terminal

[0612] The terminal is responsible for:

[0613] Input means: The user inputs the conditions for generating test questions.

[0614] Transmission method: The input conditions are sent to the server as a request, along with the emotion data.

[0615] Display: Displays the generated exam questions, explanations, and feedback.

[0616] For example, smartphones, smart glasses, head-mounted displays, etc. The emotion recognition engine works using the cameras and sensors built into these devices.

[0617] User

[0618] The user does the following:

[0619] Input of conditions: The user inputs the conditions for creating test questions (subject, difficulty level, question format, number) through the terminal.

[0620] Emotion recognition: The learning process is captured by the emotion recognition engine and the emotional data is sent to the server.

[0621] Review feedback: Review the generated exam questions, explanations, and sentiment-based advice.

[0622] Specific examples

[0623] For example, if a user wants to generate five multiple-choice questions for intermediate level math, they can input the following prompts into the generative AI model.

[0624] "Generate five intermediate level multiple choice questions in mathematics. The user's emotion is 'feeling somewhat stressed.' Adjust the difficulty of the questions taking this emotion into account."

[0625] This allows users to receive test questions of an appropriate difficulty level according to their own feelings, which is expected to improve learning efficiency and motivation.

[0626] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0627] Step 1:

[0628] The user inputs the conditions for generating test questions. The user inputs the test subjects, difficulty level, question format, and number of questions using a smartphone or other device. This input data becomes the basis for creating a request to generate test questions.

[0629] Step 2:

[0630] The emotion recognition engine recognizes the user's emotions in real time. It uses the device's built-in camera and sensors to capture the user's facial expressions and behaviors and extract emotional data, which is then used in subsequent processing steps.

[0631] Step 3:

[0632] The input conditions and emotion data are sent to the server. A test question generation request and emotion data are sent to the server using the terminal's transmission means. The input data in this process are the user's condition information acquired in (Step 1) and the emotion information acquired in (Step 2).

[0633] Step 4:

[0634] The server analyzes the request data and retrieves the necessary data. The server analyzes the request data and emotion data and retrieves related data such as past questions from the database based on that. Appropriate questions are extracted based on the request data as input.

[0635] Step 5:

[0636] The server calls the generative AI model to generate new test questions. The acquired data is preprocessed and input to the generative AI model along with the prompt. The model generates new test questions and explanations. The input data is the preprocessed data and the prompt, and the generated output is new test questions and their explanations.

[0637] Step 6:

[0638] Format the generated test questions and explanations and return them to the user. Format the generated data into an appropriate format and send it to the device. Input data is the output data from the generative AI model, which is converted into a format that can be sent to the device.

[0639] Step 7:

[0640] The device displays the test questions, explanations, and emotional feedback on the user interface. The received data is analyzed and displayed in a format that the user can understand. The input data is the response data from the server, and the displayed content is the output.

[0641] Through the above procedure, the user can obtain appropriate test questions and feedback according to his / her own emotions. These processing steps detail the processing flow of a specific system and program based on the present invention.

[0642] The specific processing unit 290 transmits the result of the specific processing to the smart device 14. In the smart device 14, the control unit 46A causes the output device 40 to output the result of the specific processing. The microphone 38B acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 38B to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[0643] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0644] In the above embodiment, an example in which the specific process is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific process may be performed by the smart device 14.

[0645] [Second embodiment]

[0646] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0647] 3, the data processing system 210 includes the data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0648] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0649] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication I / F 44. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, and the camera 42 are also connected to the bus 52.

[0650] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[0651] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[0652] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[0653] Fig. 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Fig. 4, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[0654] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[0655] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0656] In the smart glasses 214, the reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[0657] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the smart glasses 214 will be referred to as the "terminal."

[0658] This invention relates to a system that uses a generative AI model to create and explain test questions. The purpose of this system is to reduce the burden on question creators in educational institutions and qualification exams, and to provide high-quality test questions with appropriate difficulty.

[0659] System Overview

[0660] server

[0661] The server receives test question creation requests from users, acquires the appropriate data, and generates new test questions and explanations using a generative AI model. Specifically, it performs the following processes:

[0662] 1. Accepting user input

[0663] The server receives requests from users through API endpoints.

[0664] The request data includes detailed conditions such as test subjects, question format, difficulty level, and number of questions.

[0665] Analyze the received request data and extract each parameter.

[0666] 2. Referencing the database

[0667] The server retrieves appropriate question data from a past exam question database based on the specified exam subject.

[0668] A database query is used to extract data that matches the input criteria and perform additional filtering.

[0669] 3. Invoking the generative AI model

[0670] The server preprocesses the acquired data and converts it into a format that can be input into the generative AI model.

[0671] New test questions and explanations are generated using a generative AI model.

[0672] 4. Configuring the Output

[0673] The generated test questions and explanations are formatted and returned to the user.

[0674] Terminal

[0675] The terminal is responsible for the interface with the user, accepting input, sending requests to the server, and displaying the results. Specifically, it performs the following processes:

[0676] 1. Submit a request

[0677] The terminal receives the user's input and sends the conditions for creating test questions to the server as a request.

[0678] 2. Receiving a response

[0679] The test questions and explanations data returned from the server are received and displayed on the user interface.

[0680] User

[0681] The user is the entity that inputs the conditions for creating test questions and uses the system to obtain original test questions. Specifically, the user performs the following operations:

[0682] 1. Enter the conditions

[0683] The user inputs the conditions for creating test questions through the terminal and sends a request.

[0684] 2. Check the results

[0685] Check the test questions and explanations sent from the server and make a request again if necessary.

[0686] Specific examples

[0687] For example, if a user wants to create five multiple-choice questions for intermediate level math, the process is as follows:

[0688] 1. Users

[0689] The user enters "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions, and then sends a request from the device.

[0690] 2. Terminal

[0691] The terminal transmits the conditions entered by the user to the server.

[0692] 3. Server

[0693] The server analyzes the received conditions and retrieves the appropriate data from the past exam question database.

[0694] A generative AI model is called and new test questions and their explanations are generated based on the acquired data.

[0695] The generated test questions and explanations are formatted and sent back to the terminal.

[0696] 4. Terminal

[0697] The terminal receives the test questions and explanations returned from the server and displays them to the user.

[0698] 5. Users

[0699] The user checks the test questions and explanations displayed on the terminal and makes a request again if necessary.

[0700] This system allows teachers and instructors to efficiently create high-quality test questions, contributing to improving students' academic abilities. In addition, users can easily customize test questions, making it extremely flexible and fast to use in educational settings.

[0701] The processing flow will be explained below.

[0702] Program processing flow

[0703] server

[0704] Step 1:

[0705] The API endpoint receives requests from users. The request data includes detailed conditions such as test subject, question format, difficulty level, and number of questions.

[0706] Step 2:

[0707] Parse the received request data and extract each parameter (exam subject, question format, difficulty level, number of questions).

[0708] Step 3:

[0709] Access the appropriate past exam question database based on the exam subject. Use a database query to extract past exam question data that matches the input criteria.

[0710] Step 4:

[0711] The extracted data is preprocessed and converted into a format that can be input to the generative AI model, for example by tokenizing the text data and adding necessary metadata.

[0712] Step 5:

[0713] A generative AI model is then invoked to generate new test questions and explanations using the preprocessed data as input, with each question containing four options and a correct answer.

[0714] Step 6:

[0715] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[0716] Step 7:

[0717] The formatted data is sent to the terminal as an API response.

[0718] Terminal

[0719] Step 1:

[0720] Through the user interface, the user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[0721] Step 2:

[0722] The entered conditions are combined into a single JSON request data, for example, { "Subject": "Mathematics", "Format": "Multiple Choice", "Difficulty": "Intermediate", "Number of Questions": 5}.

[0723] Step 3:

[0724] Send the request data to the server using an HTTP request (POST method).

[0725] Step 4:

[0726] Receive the response data (test questions and explanations) sent back from the server as an HTTP response (200 OK).

[0727] Step 5:

[0728] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[0729] Step 6:

[0730] The formatted test questions and explanations are displayed in the user interface, and are output to the screen so that the user can check them.

[0731] User

[0732] Step 1:

[0733] Operate the terminal's user interface and enter the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[0734] Step 2:

[0735] Once you have completed the input, click the send button on your device to send the request.

[0736] Step 3:

[0737] Check that the response from the server is displayed on the terminal. The test questions and explanations are displayed on the screen so that the user can check them.

[0738] Step 4:

[0739] To revise the test questions and explanations as needed, enter the criteria again and submit a new request.

[0740] Example 1

[0741] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0742] Conventional test question creation not only requires a significant amount of time and effort from the question creator, but also has the problem of difficulty in maintaining appropriate level of difficulty and quality. Furthermore, efficiently utilizing a database of past questions to generate new test questions is also a technical challenge. There is a need to solve these problems and improve the efficiency and quality of test question creation in educational institutions and qualification examinations.

[0743] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[0744] In this invention, the server includes input means for accepting a test question creation request from a user, analysis means for analyzing the received request data and extracting conditions such as test subject, question format, difficulty level, and number of questions, acquisition means for acquiring appropriate data from a past question database based on the specified test subject, preprocessing means for preprocessing the acquired data and converting it into a format to be input to the generative AI model, generation means for generating new test questions and explanations using the generative AI model, output means for formatting and outputting the generated test questions and explanations, and communication means for returning the output data to the user. This makes it possible to automatically generate test questions that meet the conditions specified by the user and quickly provide high-quality, validated questions and explanations.

[0745] A "user" is an entity that uses the system to make a request to create test questions.

[0746] A "server" is a computer system that receives requests from users, generates test questions and explanations, and returns them to the users.

[0747] A "terminal" is a device that a user uses to input test question creation conditions and check the results.

[0748] "Input means" is a function that accepts a test question creation request from a user.

[0749] The "analysis means" is a function that analyzes the request data received from the user and extracts conditions such as test subjects, question format, difficulty level, and number of questions.

[0750] "Acquisition means" is a function that acquires data from a past exam question database based on a specified exam subject.

[0751] "Preprocessing means" is a function that converts acquired data into a format that can be input into a generative AI model.

[0752] "Generation means" is a function that generates new test questions and explanations using a generative AI model.

[0753] "Output means" is a function that formats the generated test questions and explanations and provides them to the user.

[0754] "Communication means" refers to the function that sends and receives requests for test question generation between the server and the terminal.

[0755] A "database" is a collection of information in which past exam questions are stored.

[0756] A "generative AI model" is an artificial intelligence model used to automatically generate test questions.

[0757] A "prompt sentence" is an instruction sentence that causes a generative AI model to generate test questions.

[0758] The "user interface" is a display screen for displaying the generated test questions and explanations.

[0759] This invention relates to a system that leaves the creation of test questions and explanations to a generative AI model. Specifically, it shows the process of automatically generating test questions through interactions between a server, a terminal, and a user.

[0760] Server Operation

[0761] The server is a high-performance computer system that has the following functions:

[0762] 1. Input method:

[0763] The server accepts test question creation requests from users via an API endpoint. This input is sent in JSON format and includes the test subject, question format, difficulty level, number of questions, etc.

[0764] 2. Analysis method:

[0765] The server analyzes the received JSON data and extracts each parameter (exam subject, question format, difficulty level, number of questions). This analysis process is implemented using programming languages ​​such as Python and JavaScript.

[0766] 3. Acquisition method:

[0767] The server connects to a database of past exam questions and retrieves the appropriate data based on the selected exam subject. This database stores past exam questions and related information.

[0768] 4. Pretreatment methods:

[0769] The acquired data is converted into a format that can be input to the generative AI model, specifically by pre-processing the data to convert it into text-based prompts.

[0770] 5. Generation means:

[0771] Generate new test questions and their explanations using a generative AI model (e.g., OpenAI's GPT-4), using the following prompt:

[0772] "Please create a test based on the following criteria:

[0773] Subject: Mathematics

[0774] Question format: 4 choices

[0775] Difficulty: Intermediate

[0776] Number of questions: 5"

[0777] 6. Output Method:

[0778] The generated test questions and explanations are formatted and returned to the user in JSON format, which includes the question text, options, correct answers, explanations, etc.

[0779] 7. Means of communication:

[0780] The server sends the generated test questions and explanations to the user's device using the HTTP protocol.

[0781] Device behavior

[0782] The terminal provides an interface for users to create test questions and has the following functions:

[0783] 1. Input method:

[0784] The terminal accepts the conditions for creating test questions as input from the user, which is usually provided in the form of a form.

[0785] 2. Means of transmission:

[0786] The conditions entered by the user are converted into JSON format and sent to the server using an HTTP POST request.

[0787] 3. Receiving means:

[0788] The JSON data returned from the server is received and the test questions and explanations are displayed in the user interface, which can be provided as a browser-based UI or a dedicated application.

[0789] User Actions

[0790] 1. Enter the conditions:

[0791] The user inputs the conditions for creating test questions through a terminal, including the subject, question format, difficulty level, and number of questions.

[0792] 2. Check the results:

[0793] The user can check the test questions and explanations returned by the server and make a request again if necessary. For example, if the difficulty or quality of the questions does not meet their expectations, they can change the conditions and make a request again.

[0794] Specific examples

[0795] For example, to create five multiple-choice questions for intermediate level math, the user enters the following conditions into the input form on the device:

[0796] Subject: Mathematics

[0797] Question format: 4 choices

[0798] Difficulty: Intermediate

[0799] Number of questions: 5

[0800] The device converts this data into JSON format and sends it to the server. The server analyzes the input conditions and retrieves the appropriate question data from the database. Based on this, it invokes the generative AI model to generate new test questions and explanations, which are then formatted and sent back to the device. The device then displays the received data on the user interface so that the user can view it.

[0801] This system will improve the efficiency and quality of test question creation, and can be used flexibly in educational institutions and qualification exams.

[0802] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0803] Step 1:

[0804] Accepting user input

[0805] Specific explanation

[0806] The user enters the conditions for creating test questions into the input form on the terminal. These include the test subject (e.g., "Mathematics"), question format (e.g., "Multiple Choice"), difficulty level (e.g., "Intermediate"), and number of questions (e.g., "5"). When the user clicks the "Submit" button, these conditions are converted into JSON format.

[0807] input

[0808] Exam subjects, question format, difficulty level, number of questions

[0809] output

[0810] Request data in JSON format

[0811] Step 2:

[0812] Submitting a Request

[0813] Specific explanation

[0814] The terminal converts the conditions entered by the user into JSON format and sends it to the server as an HTTP POST request.

[0815] input

[0816] Request data in JSON format

[0817] output

[0818] HTTP POST request

[0819] Step 3:

[0820] Accepting a request

[0821] Specific explanation

[0822] The server receives requests from users via the API endpoint and parses the request data, extracting conditions such as test subject, question format, difficulty level, and number of questions from the JSON-formatted request data.

[0823] input

[0824] HTTP POST request

[0825] output

[0826] Extracted data on exam subjects, question format, difficulty level, and number of questions

[0827] Step 4:

[0828] Database Reference

[0829] Specific explanation

[0830] The server retrieves appropriate question data from the past exam database based on the extracted conditions. For example, it uses an SQL query to search for "intermediate" level questions in the "Mathematics" subject.

[0831] sql

[0832] SELECT FROM question_bank WHERE subject = 'Mathematics' AND difficulty = 'Intermediate' LIMIT 10;

[0833] input

[0834] Exam subjects and difficulty level

[0835] output

[0836] Appropriate problem data

[0837] Step 5:

[0838] Data Preprocessing

[0839] Specific explanation

[0840] The server converts the acquired data into a format that can be input to the generative AI model. For example, it formats the acquired problem data into a list format.

[0841] text

[0842] {"question": "What is the solution to x^2 + 2x + 1 = 0?", "choices": ["1", "-1", "0", "2"], "answer": "1"}

[0843] input

[0844] Appropriate problem data

[0845] output

[0846] Data formats that can be input to generative AI models

[0847] Step 6:

[0848] Invoking a generative AI model

[0849] Specific explanation

[0850] The server uses the converted data to send prompts to a generative AI model (e.g., GPT-4) to generate new test questions and explanations.

[0851] text

[0852] "Please create a test based on the following criteria:

[0853] Subject: Mathematics

[0854] Question format: 4 choices

[0855] Difficulty: Intermediate

[0856] Number of questions: 5"

[0857] input

[0858] Prompt statement

[0859] output

[0860] New exam questions and explanations

[0861] Step 7:

[0862] Configuring the Output

[0863] Specific explanation

[0864] The generated test questions and explanations are formatted and converted into JSON format for sending back to the user. For example, formatting it as follows:

[0865] json

[0866] {

[0867] "questions": [

[0868] {"question": "What is the solution to x^2 + 2x + 1 = 0?", "choices": ["1", "-1", "0", "2"], "answer": "1", "explanation": "Using the solution formula..."}

[0869] ]

[0870] }

[0871] input

[0872] New exam questions and explanations

[0873] output

[0874] JSON format data

[0875] Step 8:

[0876] Sending a response

[0877] Specific explanation

[0878] The server sends the formatted JSON data to the device as an HTTP response.

[0879] input

[0880] JSON format data

[0881] output

[0882] HTTP response

[0883] Step 9:

[0884] Receiving a response

[0885] Specific explanation

[0886] The terminal receives the response data from the server, analyzes it, and displays the test questions and explanations on the user interface.

[0887] input

[0888] HTTP response

[0889] output

[0890] Test questions and explanations displayed on the user interface

[0891] Step 10:

[0892] Checking the results

[0893] Specific explanation

[0894] The user can check the test questions and explanations displayed on the terminal and make a re-request if necessary.

[0895] input

[0896] Test questions and explanations displayed on the user interface

[0897] output

[0898] User confirmation / request

[0899] (Application example 1)

[0900] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[0901] Conventional test question creation systems require test creators to expend a great deal of time and effort, making it difficult to quickly generate high-quality test questions. Furthermore, it is difficult to automatically generate new test questions by referencing past questions, which often results in inconsistent question quality and difficulty. Furthermore, there was a lack of a mechanism for instantly providing users with the generated test questions and explanations, making it difficult to provide appropriate learning support. There was a need to provide a system that could solve these issues and quickly and efficiently generate and provide test questions to users, particularly in online education and qualification exams.

[0902] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[0903] In this invention, the server includes a generation means for automatically generating test questions for a specific domain, an acquisition means for acquiring information based on requested conditions, a call means for calling a generative AI model for generating new test questions from the acquired information, an output means for formatting and outputting the generated test questions and explanations, a prompt generation means for receiving input from a user via a communication interface and creating prompt sentences for generating test questions that match the conditions, and a display means for displaying the generated test questions and explanations on a user interface. This enables users to easily and quickly generate and check high-quality test questions and explanations.

[0904] A "generation means" is a device or system that has the functionality to automatically generate test questions for a particular domain.

[0905] An "acquisition means" is a device or system for acquiring the necessary information or data based on a request.

[0906] A "calling means" is a device or system that has the function of calling a generative AI model and generating new test questions based on the acquired information.

[0907] The "output means" is a device or system for formatting the generated test questions and explanations into an appropriate format and providing them to an external party.

[0908] A "communications interface" is a device or system that allows data to be sent and received between a server and a user's terminal.

[0909] A "prompt generation means" is a device or system that has the function of creating prompt sentences to pose appropriate questions to the generative AI model based on input conditions from the user.

[0910] The "display means" is a device or system for displaying the generated test questions and explanations on a user interface.

[0911] This invention relates to a system that uses a generative AI model to create and explain test questions. The purpose of this system is to reduce the burden on question creators in educational institutions, qualification exams, and other settings, and to provide high-quality test questions. Here, we will explain a specific embodiment of the system that realizes this application example.

[0912] server

[0913] The server includes the following means:

[0914] 1. Generation means

[0915] This function automatically generates test questions for specific areas. Based on the user's requested tasks, it retrieves information from a database of previous questions and generates new questions.

[0916] 2. Acquisition method

[0917] This function retrieves appropriate information from the past exam question database based on user requests, allowing for optimal data collection.

[0918] 3. Calling means

[0919] This function calls a generative AI model based on the data collected by the acquisition means and generates new test questions. This AI model is usually a generative AI model such as OpenAI.

[0920] 4. Output Method

[0921] This function formats the generated test questions and explanations appropriately and outputs them to the user. This ensures that the generated content is presented to the user in a format that is easy for them to understand.

[0922] 5. Prompt Generation Methods

[0923] This function creates prompt sentences that ask appropriate questions to the generative AI model based on input conditions from the user, thereby eliciting appropriate responses from the AI ​​model.

[0924] 6. Display means

[0925] This function displays the generated test questions and explanations on the user interface.

[0926] Terminal

[0927] The terminal includes the following means:

[0928] 1. Communication Interface

[0929] This function sends and receives data between the server and the user, allowing for smooth requests and responses for test questions.

[0930] User

[0931] The user does the following:

[0932] 1. Input Method

[0933] The conditions for generating test questions are entered into the terminal and a request is sent.

[0934] 2. Transmission Method

[0935] A request is sent to the server based on the entered conditions.

[0936] 3. Display means

[0937] Check the generated test questions and explanations on the device interface.

[0938] For example, if a user wants to create five multiple-choice questions for the intermediate level of "Mathematics," he or she inputs the conditions into the terminal and sends a request to the server. The terminal transmits this information through the communication interface, and the server uses the acquisition means to collect appropriate data and creates a prompt sentence using the generation means and prompt generation means.

[0939] Here is an example prompt:

[0940] Create five multiple-choice questions and explanations for intermediate level mathematics.

[0941] The generative AI model generates new test questions and explanations based on this prompt, formats them using an output method, and returns them to the device. The user can then check the generated test questions and explanations on the device and make further requests as necessary. This allows users to create high-quality test questions efficiently and quickly.

[0942] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0943] Step 1:

[0944] The user uses the input means of the terminal to input conditions for generating test questions. Specifically, detailed conditions such as test subjects, question format, difficulty level, number of questions, etc. This input data is sent to the terminal as request data.

[0945] Step 2:

[0946] The terminal receives the request data and sends it to the server via a communication interface. At this time, the request data includes the test subject, question format, difficulty level, number of questions, etc. The terminal converts the request data into the required format and sends it to the server in JSON format, etc.

[0947] Step 3:

[0948] The server receives and analyzes the request data. As a result of the analysis, the test conditions desired by the user (test subject, question format, difficulty level, number of questions) are extracted. Based on the extracted conditions, the server references the past question database and retrieves past question data related to the specified test subject.

[0949] Step 4:

[0950] The retrieval means retrieves appropriate data from the past exam question database, including extracting questions that match specified criteria using a search query, and stores the extracted data in a temporary data store.

[0951] Step 5:

[0952] The server uses the prompt generation means to create a prompt sentence based on the user's input conditions. For example, a sentence such as "Please create five multiple-choice questions and explanations for intermediate level mathematics" is generated. This prompt sentence is prepared to be passed to the generative AI model.

[0953] Step 6:

[0954] The calling means calls the generative AI model based on the prompt sentence. Specifically, it sends the prompt sentence to the generative AI model, and generates new test questions and explanations in response. The generative AI model (e.g., OpenAI) generates appropriate questions and explanations according to the prompt sentence and returns the results to the server.

[0955] Step 7:

[0956] The generated test questions and explanations are formatted by the output means. The formatted data is converted into a format that is easy for the user to understand (e.g., text, PDF, HTML, etc.). The formatted data is then prepared in a format for transmission to the terminal.

[0957] Step 8:

[0958] The server sends the formatted test questions and explanations through a communications interface to the terminal, which receives them and prepares them for display on the user interface.

[0959] Step 9:

[0960] The terminal uses the display means to display the generated test questions and explanations on the user interface, allowing the user to check them on the terminal and make a re-request if necessary.

[0961] If the user changes the conditions and makes a request again, the process is repeated from step 1. This realizes a system that can quickly and efficiently provide high-quality test questions and explanations.

[0962] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[0963] This invention combines a system that automatically generates test questions for specific subjects with an emotion engine that recognizes user emotions. The purpose of this system is to reduce the burden on question creators in educational institutions and qualification exams, and to improve the quality of users' learning experience.

[0964] System Overview

[0965] server

[0966] The server receives test question creation requests from users, retrieves appropriate data, and generates new test questions and explanations using a generative AI model. It also uses an emotion engine to recognize user emotions and adjust the difficulty of the test questions.

[0967] 1. Accepting user input

[0968] The server receives requests from users through API endpoints. The request data includes detailed conditions such as test subject, question format, difficulty level, number of questions, etc. It also includes emotion recognition data.

[0969] Parse the received request data and emotion data and extract each parameter.

[0970] 2. Referencing the database

[0971] The server retrieves the appropriate question data from the past exam database based on the specified exam subject. It uses a database query to extract data that matches the input criteria and performs additional filtering.

[0972] 3. Invoking the generative AI model

[0973] The server preprocesses the acquired data and converts it into a format that can be input to the generative AI model. It also adjusts the difficulty of the questions based on the emotional data.

[0974] It uses a generative AI model to generate new test questions and explanations, each with four options and a correct answer.

[0975] 4. Configuring the Output

[0976] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[0977] Feedback information including an evaluation by the emotion recognition engine is generated and sent back to the user along with the test questions.

[0978] Terminal

[0979] The device is responsible for interfacing with the user, accepting input, sending requests to the server, and displaying the results. It also contains an emotion engine that recognizes the user's emotions and sends the data to the server.

[0980] 1. Submit a request

[0981] The terminal receives user input, sends the conditions for creating test questions to the server as a request, and also recognizes the user's emotions using an emotion engine and sends the data to the server.

[0982] 2. Receiving a response

[0983] Response data (test questions, explanations, and emotion evaluation results) is received from the server as an HTTP response (200 OK).

[0984] 3. Displaying the results

[0985] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[0986] The formatted test questions, explanations, and emotion evaluation results are displayed in the user interface, and are output to the screen so that the user can check them.

[0987] User

[0988] The user is the entity that inputs the conditions for creating test questions and uses the system to obtain original test questions. In addition, the emotion engine feeds back the user's emotions during learning to the system, improving the quality of the learning experience.

[0989] 1. Enter the conditions

[0990] The user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions) through the terminal.

[0991] 2. Emotional Feedback

[0992] The emotion engine recognizes the user's operating status and facial expressions, and transmits them to the server as emotion data.

[0993] 3. Check the results

[0994] The response from the server is displayed on the terminal. The test questions, explanations, and emotion evaluation results are displayed on the screen for the user to review.

[0995] To revise the test questions and explanations as needed, enter the criteria again and submit a new request.

[0996] Specific examples

[0997] For example, if a user wants to create five multiple-choice questions at an intermediate level in mathematics and wants advice on how to reduce stress levels while studying, the process is as follows:

[0998] 1. Users

[0999] Enter "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions, and send a request from your device. The emotion engine will also capture your learning progress.

[1000] 2. Terminal

[1001] The request data and emotion data are sent to the server.

[1002] 3. Server

[1003] The system analyzes request data, retrieves appropriate data from a database of past questions, and adjusts the difficulty of questions based on emotional data.

[1004] A generative AI model is invoked to generate new test questions and explanations.

[1005] The test questions, explanations, and emotional assessment results are formatted and sent back to the terminal.

[1006] 4. Terminal

[1007] The test questions, explanations, and emotion evaluation results returned from the server are received and displayed on the user interface.

[1008] 5. Users

[1009] Check the test questions, explanations, and emotional assessment results, and receive appropriate advice based on your learning progress.

[1010] If necessary, re-requests will be made to adjust the content and difficulty of the test questions.

[1011] The processing flow will be explained below.

[1012] server

[1013] Step 1:

[1014] The API endpoint receives requests from users. The request data includes test subject, question format, difficulty level, number of questions, and sentiment data.

[1015] Step 2:

[1016] The received request data and emotion data are parsed and each parameter (exam subject, question format, difficulty level, number of questions, emotion data) is extracted.

[1017] Step 3:

[1018] Access the appropriate past exam question database based on the exam subject. Use a database query to extract past exam question data that matches the input criteria.

[1019] Step 4:

[1020] The extracted data is preprocessed and converted into a format that can be input to the generative AI model, for example by tokenizing the text data and adding necessary metadata.

[1021] Step 5:

[1022] The difficulty of the generated test questions is adjusted based on the emotional data. For example, if the user is feeling stressed, the difficulty level is set slightly lower.

[1023] Step 6:

[1024] A generative AI model is then invoked to generate new test questions and explanations using the preprocessed data as input, with each question containing four options and a correct answer.

[1025] Step 7:

[1026] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[1027] Step 8:

[1028] Feedback information including evaluations by the emotion recognition engine is generated and sent back to the terminal along with the test questions.

[1029] Terminal

[1030] Step 1:

[1031] Through the user interface, the user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[1032] Step 2:

[1033] The input conditions are compiled into a single JSON request data. The request also includes the user's emotion data recognized by the emotion engine. For example, the format is { "Subject": "Mathematics", "Format": "Multiple Choice", "Difficulty": "Intermediate", "Number of Questions": 5, "Emotion": "Stress"}.

[1034] Step 3:

[1035] Send the request data to the server using an HTTP request (POST method).

[1036] Step 4:

[1037] Response data (test questions, explanations, and emotion evaluation results) returned from the server is received as an HTTP response (200 OK).

[1038] Step 5:

[1039] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[1040] Step 6:

[1041] The formatted test questions, explanations, and emotion evaluation results are displayed in the user interface, and are output to the screen so that the user can check them.

[1042] User

[1043] Step 1:

[1044] Operate the terminal's user interface and enter the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[1045] Step 2:

[1046] Once you have completed the input, click the send button on your device to send the request.

[1047] Step 3:

[1048] The emotion engine recognizes the user's operating status and facial expressions in real time and generates emotion data, which is included in the request.

[1049] Step 4:

[1050] The request sent from the terminal is processed by the server, and the generated test questions, explanations, and emotion evaluation results are displayed on the terminal.

[1051] Step 5:

[1052] Check the test questions and explanations displayed on the screen and receive appropriate study advice based on the emotional assessment results.

[1053] Step 6:

[1054] If necessary, re-requests will be made to adjust the content and difficulty of the test questions.

[1055] Example 2

[1056] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1057] There is a need to reduce the burden on educational institutions and test question creators and improve the quality of users' learning experience.However, there is no system that can take into account the user's emotional state while studying and adjust the difficulty of questions individually, making it difficult to maintain learning motivation and provide adaptive education.

[1058] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1059] In this invention, the server includes a generation means for automatically generating test questions for specific subjects, an acquisition means for acquiring data based on requested conditions, a call means for calling a generation AI model that generates new test questions from the acquired data, an output means for formatting and outputting the generated test questions and explanations, an emotion analysis means for recognizing the user's emotions and analyzing the data, a difficulty adjustment means for adjusting the difficulty of the questions based on the emotion data, and a feedback means for returning the generated test questions and explanations to the user and generating feedback including the emotion evaluation results. This reduces the burden on the question creator, improves the quality of the user's learning experience, and enables the difficulty of questions to be adjusted appropriately according to the emotional state of each user.

[1060] A "specific subject" refers to a specific field or academic area that serves as the subject matter for generating test questions.

[1061] A "test question" is material containing questions or tasks used to assess a user's knowledge and understanding.

[1062] "Generation means" refers to the processes and techniques used to automatically create test questions, primarily using generative AI models.

[1063] "Acquisition means" refers to the process or method of searching for and acquiring relevant data or information based on specific conditions.

[1064] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new test questions and explanations.

[1065] "Invocation means" refers to the process or method for running a generative AI model using acquired data to obtain the required output.

[1066] "Output means" refers to the process or method for outputting the generated test questions and explanations in a specified format.

[1067] "Emotion analysis means" refers to techniques and processes for detecting and analyzing a user's emotional state.

[1068] "Difficulty adjustment means" refers to a process or method for appropriately adjusting the difficulty of test questions based on emotional data.

[1069] "Feedback means" refers to a process or technology for returning the generated test questions, explanations, and emotional evaluation results to the user and providing feedback information.

[1070] "User interface" refers to the interface for exchanging information between the system and the user, and displays test questions, explanations, etc.

[1071] "Communication means" refers to the communication technology and protocol used to request and respond to test question generation between the server and the terminal.

[1072] "Input means" refers to a device or interface that allows a user to input conditions for generating test questions.

[1073] "Transmission means" refers to the process or method for sending a request to a server based on the conditions entered by the user.

[1074] MODE FOR CARRYING OUT THE INVENTION

[1075] This invention combines a system for automatically generating test questions for specific subjects with an emotion engine that recognizes user emotions. This system aims to reduce the burden on question creators in educational institutions and qualification exams and improve the quality of users' learning experience.

[1076] server

[1077] A server is a combination of hardware and software that performs multiple functions. In particular, it includes the following elements:

[1078] 1. Generation means

[1079] The server uses a generative AI model to automatically generate test questions for specific subjects. The generative AI model uses artificial intelligence algorithms to generate new questions and their explanations.

[1080] 2. Acquisition method

[1081] The server retrieves data from information sources such as a database of past exam questions based on the conditions requested by the user. To do this, it uses SQL queries to search for and retrieve data that matches the request conditions.

[1082] 3. Calling means

[1083] The server preprocesses the acquired data and converts it into a format that can be input to the generative AI model. It also calls the generative AI model to generate new test questions and explanations. Specifically, it passes the necessary data to the generative AI model as prompts.

[1084] 4. Output Method

[1085] The server formats the generated test questions and explanations in an appropriate format (e.g., JSON or HTML), outputs the data, and temporarily stores it before returning it to the user.

[1086] 5. Emotion analysis method

[1087] The server receives the user's emotional data and uses an emotion engine to analyze it, which detects various emotional states such as the user's stress level and concentration level.

[1088] 6. Difficulty adjustment method

[1089] The server adjusts the difficulty of the questions based on the emotional data, for example, setting the difficulty of the questions lower if the user indicates a high stress level.

[1090] 7. Feedback channels

[1091] The server returns the generated test questions and explanations, as well as the emotional evaluation results, to the user, providing feedback information, so that the user can receive appropriate feedback based on their learning progress and emotional state.

[1092] Terminal

[1093] The device is responsible for the interface with the user, accepting input, sending requests to the server, and displaying the results. It also recognizes the user's emotions and sends the data to the server.

[1094] 1. Submit a request

[1095] The terminal uses an HTTP request to transmit the test question creation conditions and emotion data received from the user to the server.

[1096] 2. Receiving a response

[1097] The terminal receives the response data returned from the server, which includes the test questions, explanations, and emotion evaluation results.

[1098] 3. Displaying the results

[1099] The terminal analyzes the received response data and displays it on the user interface. For example, it formats and displays test questions and explanations using HTML templates.

[1100] User

[1101] To use the system to obtain the original test questions, users must:

[1102] 1. Enter the conditions

[1103] The user inputs the conditions for creating test questions (for example, test subjects, question format, difficulty level, number of questions) through the terminal.

[1104] 2. Emotional Feedback

[1105] During learning, the user's emotional data is captured by the emotion engine through their operating status and facial expressions.

[1106] 3. Check the results

[1107] The user can check the test questions, explanations, and emotion evaluation results returned in response from the server by displaying them on the terminal. If necessary, the user can send a re-request to adjust the content and difficulty of the test questions.

[1108] Specific examples

[1109] For example, if a user wants to create five intermediate-level multiple-choice questions in math and get advice on how to reduce stress levels while studying, the process might go like this:

[1110] 1. Users

[1111] The user enters "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions into the device and sends a request. At this time, the emotion engine captures the user's emotions while studying.

[1112] 2. Terminal

[1113] The request data and emotion data are sent to the server.

[1114] 3. Server

[1115] The system analyzes the request data, retrieves appropriate data from a database of past exam questions, and adjusts the difficulty of the questions based on the emotion data. It then uses a generative AI model to generate new exam questions and explanations. The generated questions, explanations, and emotion evaluation results are formatted and sent back to the device.

[1116] 4. Terminal

[1117] Receives the data returned from the server and displays it in the user interface.

[1118] 5. Users

[1119] Review the test questions, explanations, and emotional evaluation results. If necessary, make a request again to adjust the content and difficulty of the test questions.

[1120] Examples of specific prompts to use

[1121] Create five multiple-choice questions for intermediate level math based on the following criteria, taking into account the user's stress level while studying:

[1122] Exam subject: Mathematics

[1123] Question format: 4 choices

[1124] Difficulty: Intermediate

[1125] Number of questions: 5

[1126] keyword

[1127] Generative AI model, prompt sentence

[1128] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1129] Step 1:

[1130] Accepting user input

[1131] server

[1132] The server receives requests from users through an HTTP API endpoint. Specifically, it receives JSON data (e.g., {"subject": "math", "difficulty": "intermediate", "question_count": 5}) in which the user enters conditions such as test subject, question format, difficulty level, and number of questions. It also receives the user's emotion data at the same time.

[1133] Input: Test question creation conditions (JSON format) and emotion data

[1134] Data processing / calculation: Parse JSON data and extract each parameter (exam subject, question format, difficulty level, number of questions)

[1135] Output: Extracted parameters and emotion data

[1136] Step 2:

[1137] Database Reference

[1138] server

[1139] The server references the past exam question database based on the criteria received from the user, and uses an SQL query to retrieve data that matches the request criteria (e.g. SELECT FROM problem_database WHERE subject = 'math' AND difficulty = 'intermediate').

[1140] Input: Extracted parameters (exam subject, difficulty level, etc.)

[1141] Data processing / calculation: Executing SQL queries and filtering retrieved data

[1142] Output: Filtered data

[1143] Step 3:

[1144] Invoking a generative AI model

[1145] server

[1146] The server preprocesses the acquired data and converts it into a format suitable for input to the generative AI model. It adjusts the difficulty of the questions based on the emotional data (e.g., lowering the difficulty level if the stress level is high). It then calls the generative AI model to generate new test questions and explanations. The generated questions include four options and a correct answer.

[1147] Input: filtered data, sentiment data

[1148] Data processing / calculation: Preprocessing data, converting it to an input format for the generative AI model, and invoking the generative AI model using prompts.

[1149] Output: Generated test questions and explanations (e.g., ["Q1: Question text...", "A: Choice 1", "B: Choice 2", "C: Choice 3", "D: Choice 4", "Answer: B"])

[1150] Step 4:

[1151] Configuring the Output

[1152] server

[1153] The server formats the generated test questions and explanations into an appropriate format (e.g., JSON format), and also generates feedback information including the emotion evaluation results, which is temporarily stored in a database for sending back to the user.

[1154] Input: Generated test questions and explanations, emotional evaluation results

[1155] Data processing / calculation: formatting test questions and explanations, generating feedback information

[1156] Output: Formatted exam questions, explanations and feedback information (JSON format)

[1157] Step 5:

[1158] Send request

[1159] Terminal

[1160] The terminal receives the user's input and sends the request data and emotion data to the server using an HTTP request.

[1161] Input: User-entered test question creation conditions and emotion data

[1162] Data processing / calculation: Converting input data to JSON format, creating and sending HTTP requests

[1163] Output: Request data sent to the server (JSON format)

[1164] Step 6:

[1165] Response reception

[1166] Terminal

[1167] The device receives the response data from the server. After receiving the HTTP status code 200 (OK), it acquires the test questions, explanations, and emotion evaluation results contained in the response body.

[1168] Input: Response data from the server

[1169] Data processing / calculation: Response data analysis (JSON parsing)

[1170] Output: Analyzed test questions, explanations, and emotion evaluation results

[1171] Step 7:

[1172] Displaying the results

[1173] Terminal

[1174] The device analyzes the received response data and displays it on the user interface. It uses HTML templates to format and display test questions, explanations, and emotion evaluation results.

[1175] Input: Analyzed test questions, explanations, and emotion evaluation results

[1176] Data processing / calculation: Rendering using HTML templates

[1177] Output: Formatted test questions, explanations, and emotional evaluation results (displayed on screen)

[1178] Step 8:

[1179] Entering conditions

[1180] User

[1181] The user inputs the conditions for creating test questions (test subjects, question format, difficulty level, number of questions) through the terminal. After the conditions are entered, the user clicks a button to send the request.

[1182] Input: Conditions for creating test questions

[1183] Data processing / calculation: None

[1184] Output: What you type appears on the terminal

[1185] Step 9:

[1186] Emotional feedback

[1187] User

[1188] The emotion engine captures the operation status and facial expressions during learning, and the emotion data is sent to the server.

[1189] Input: User operation status, facial expressions, etc.

[1190] Data processing / calculation: Analysis by emotion engine, data capture

[1191] Output: Captured emotion data

[1192] Step 10:

[1193] Checking the results

[1194] User

[1195] The test questions, explanations, and emotional evaluation results returned from the server are displayed on the device. If necessary, a request is made again to adjust the content and difficulty of the test questions.

[1196] Input: Test questions, explanations, and emotional evaluation results displayed on the terminal

[1197] Data processing / calculation: None

[1198] Output: User confirmation and re-request if necessary

[1199] (Application example 2)

[1200] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the smart glasses 214 will be referred to as a "terminal."

[1201] Conventional test question generation systems provide questions uniformly without considering the user's emotions, which reduces the user's learning efficiency. In addition, since feedback based on emotional data is not provided, users tend to feel stressed and may lose motivation to study.

[1202] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1203] In this invention, the server includes a generation means for automatically generating test questions for specific subjects, an acquisition means for acquiring data based on requested conditions, a call means for calling a generation AI model for generating new test questions from the acquired data, an output means for formatting and outputting the generated test questions and explanations, an emotion recognition means for recognizing a user's emotions, a difficulty adjustment means for adjusting the difficulty of the questions based on the emotion data, and a feedback generation means for generating feedback information based on the emotion data. This makes it possible to provide test questions of an appropriate difficulty level that takes the user's emotions into consideration, improving learning efficiency and maintaining the user's motivation to learn.

[1204] The "means for automatically generating test questions for a specific subject" is a function for automatically generating test questions based on the subjects specified by the user.

[1205] The "acquisition means for acquiring data based on requested conditions" is a function for acquiring necessary data from a database according to conditions input by a user.

[1206] The "calling means for calling the generative AI model" is a function that executes the AI ​​model to generate new test questions using the acquired data.

[1207] The "output means for formatting and outputting the generated test questions and explanations" is a function for formatting the newly generated test questions and their explanations in an appropriate format and outputting them.

[1208] The "emotion recognition means for recognizing the user's emotions" is a function for analyzing the user's facial expressions, actions, etc. to grasp the user's emotional state.

[1209] The "difficulty level adjusting means for adjusting the difficulty level of questions based on emotional data" is a function for providing test questions of a difficulty level appropriate for the user based on the recognized emotional data.

[1210] The "feedback generating means for generating feedback information based on emotional data" is a function for providing appropriate advice and feedback based on the emotional state of the user.

[1211] The "display means for displaying the generated test questions and explanations on the user interface" is a function for displaying the generated test questions and explanations on the screen so that the user can check them.

[1212] "Communication means for requesting and responding to test question generation between the server and the terminal" is a function for sending and receiving requests for test question generation and the generated results between the server and the terminal.

[1213] The "input means for the user to input conditions for generating test questions" is a function for the user to input conditions such as the subject, difficulty level, and format of the test questions that he or she wants to generate.

[1214] The "transmission means for transmitting a request to the server based on the input conditions" is a function for transmitting the conditions input by the user to the server and requesting the generation of test questions.

[1215] The "emotion data transmission means for acquiring emotion data in real time and transmitting it to the server" is a function for collecting the user's emotions in real time and transmitting that data to the server.

[1216] This invention combines a system that automatically generates test questions for specific subjects with an emotion engine that recognizes the user's emotions. This system includes a generation means, an acquisition means, a calling means, an output means, an emotion recognition means, a difficulty level adjustment means, and a feedback generation means. It is particularly characterized by using emotion data to adjust the difficulty of questions and provide appropriate feedback.

[1217] System Overview

[1218] server

[1219] The server has the following features:

[1220] Generation method: Automatically generate test questions for specified subjects.

[1221] Acquisition method: Retrieves the necessary data from the database based on a user request.

[1222] Calling method: Input the acquired data into the generative AI model to generate new test questions.

[1223] Output method: Format the generated test questions and explanations and output them in an appropriate format.

[1224] Emotion recognition means: Recognizes the user's emotions using cameras and sensors and acquires them as data.

[1225] Difficulty adjustment method: Adjust the difficulty of the questions based on emotional data.

[1226] Feedback generation means: Generates feedback information based on emotion data and provides it to the user.

[1227] The hardware used includes servers, databases, and computing resources for running the AI ​​model, while the software includes an emotion recognition engine (EmotionEngine) and a generative AI model (QuestionGenerator).

[1228] Terminal

[1229] The terminal is responsible for:

[1230] Input means: The user inputs the conditions for generating test questions.

[1231] Transmission method: The input conditions are sent to the server as a request, along with the emotion data.

[1232] Display: Displays the generated exam questions, explanations, and feedback.

[1233] For example, smartphones, smart glasses, head-mounted displays, etc. The emotion recognition engine works using the cameras and sensors built into these devices.

[1234] User

[1235] The user does the following:

[1236] Input of conditions: The user inputs the conditions for creating test questions (subject, difficulty level, question format, number) through the terminal.

[1237] Emotion recognition: The learning process is captured by the emotion recognition engine and the emotional data is sent to the server.

[1238] Review feedback: Review the generated exam questions, explanations, and sentiment-based advice.

[1239] Specific examples

[1240] For example, if a user wants to generate five multiple-choice questions for intermediate level math, they can input the following prompts into the generative AI model.

[1241] "Generate five intermediate level multiple choice questions in mathematics. The user's emotion is 'feeling somewhat stressed.' Adjust the difficulty of the questions taking this emotion into account."

[1242] This allows users to receive test questions of an appropriate difficulty level according to their own feelings, which is expected to improve learning efficiency and motivation.

[1243] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1244] Step 1:

[1245] The user inputs the conditions for generating test questions. The user inputs the test subjects, difficulty level, question format, and number of questions using a smartphone or other device. This input data becomes the basis for creating a request to generate test questions.

[1246] Step 2:

[1247] The emotion recognition engine recognizes the user's emotions in real time. It uses the device's built-in camera and sensors to capture the user's facial expressions and behaviors and extract emotional data, which is then used in subsequent processing steps.

[1248] Step 3:

[1249] The input conditions and emotion data are sent to the server. A test question generation request and emotion data are sent to the server using the terminal's transmission means. The input data in this process are the user's condition information acquired in (Step 1) and the emotion information acquired in (Step 2).

[1250] Step 4:

[1251] The server analyzes the request data and retrieves the necessary data. The server analyzes the request data and emotion data and retrieves related data such as past questions from the database based on that. Appropriate questions are extracted based on the request data as input.

[1252] Step 5:

[1253] The server calls the generative AI model to generate new test questions. The acquired data is preprocessed and input to the generative AI model along with the prompt. The model generates new test questions and explanations. The input data is the preprocessed data and the prompt, and the generated output is new test questions and their explanations.

[1254] Step 6:

[1255] Format the generated test questions and explanations and return them to the user. Format the generated data into an appropriate format and send it to the device. Input data is the output data from the generative AI model, which is converted into a format that can be sent to the device.

[1256] Step 7:

[1257] The device displays the test questions, explanations, and emotional feedback on the user interface. The received data is analyzed and displayed in a format that the user can understand. The input data is the response data from the server, and the displayed content is the output.

[1258] Through the above procedure, the user can obtain appropriate test questions and feedback according to his / her own emotions. These processing steps detail the processing flow of a specific system and program based on the present invention.

[1259] The specific processing unit 290 transmits the result of the specific processing to the smart glasses 214. In the smart glasses 214, the control unit 46A causes the speaker 240 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1260] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1261] In the above embodiment, an example in which the specific processing is performed by the data processing device 12 has been given, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the smart glasses 214.

[1262] [Third embodiment]

[1263] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1264] 5, the data processing system 310 includes the data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[1265] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1266] The headset type terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a display 343. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the display 343 are also connected to the bus 52.

[1267] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1268] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1269] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1270] Fig. 6 shows an example of the main functions of the data processing device 12 and the headset type terminal 314. As shown in Fig. 6, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1271] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1272] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1273] In the headset type terminal 314, a reception output process is performed by the processor 46. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1274] Next, a description will be given of the identification process performed by the identification processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as the "server" and the headset type terminal 314 will be referred to as the "terminal."

[1275] This invention relates to a system that uses a generative AI model to create and explain test questions. The purpose of this system is to reduce the burden on question creators in educational institutions and qualification exams, and to provide high-quality test questions with appropriate difficulty.

[1276] System Overview

[1277] server

[1278] The server receives test question creation requests from users, acquires the appropriate data, and generates new test questions and explanations using a generative AI model. Specifically, it performs the following processes:

[1279] 1. Accepting user input

[1280] The server receives requests from users through API endpoints.

[1281] The request data includes detailed conditions such as test subjects, question format, difficulty level, and number of questions.

[1282] Analyze the received request data and extract each parameter.

[1283] 2. Referencing the database

[1284] The server retrieves appropriate question data from a past exam question database based on the specified exam subject.

[1285] A database query is used to extract data that matches the input criteria and perform additional filtering.

[1286] 3. Invoking the generative AI model

[1287] The server preprocesses the acquired data and converts it into a format that can be input into the generative AI model.

[1288] New test questions and explanations are generated using a generative AI model.

[1289] 4. Configuring the Output

[1290] The generated test questions and explanations are formatted and returned to the user.

[1291] Terminal

[1292] The terminal is responsible for the interface with the user, accepting input, sending requests to the server, and displaying the results. Specifically, it performs the following processes:

[1293] 1. Submit a request

[1294] The terminal receives the user's input and sends the conditions for creating test questions to the server as a request.

[1295] 2. Receiving a response

[1296] The test questions and explanations data returned from the server are received and displayed on the user interface.

[1297] User

[1298] The user is the entity that inputs the conditions for creating test questions and uses the system to obtain original test questions. Specifically, the user performs the following operations:

[1299] 1. Enter the conditions

[1300] The user inputs the conditions for creating test questions through the terminal and sends a request.

[1301] 2. Check the results

[1302] Check the test questions and explanations sent from the server and make a request again if necessary.

[1303] Specific examples

[1304] For example, if a user wants to create five multiple-choice questions for intermediate level math, the process is as follows:

[1305] 1. Users

[1306] The user enters "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions, and then sends a request from the device.

[1307] 2. Terminal

[1308] The terminal transmits the conditions entered by the user to the server.

[1309] 3. Server

[1310] The server analyzes the received conditions and retrieves the appropriate data from the past exam question database.

[1311] A generative AI model is called and new test questions and their explanations are generated based on the acquired data.

[1312] The generated test questions and explanations are formatted and sent back to the terminal.

[1313] 4. Terminal

[1314] The terminal receives the test questions and explanations returned from the server and displays them to the user.

[1315] 5. Users

[1316] The user checks the test questions and explanations displayed on the terminal and makes a request again if necessary.

[1317] This system allows teachers and instructors to efficiently create high-quality test questions, contributing to improving students' academic abilities. In addition, users can easily customize test questions, making it extremely flexible and fast to use in educational settings.

[1318] The processing flow will be explained below.

[1319] Program processing flow

[1320] server

[1321] Step 1:

[1322] The API endpoint receives requests from users. The request data includes detailed conditions such as test subject, question format, difficulty level, and number of questions.

[1323] Step 2:

[1324] Parse the received request data and extract each parameter (exam subject, question format, difficulty level, number of questions).

[1325] Step 3:

[1326] Access the appropriate past exam question database based on the exam subject. Use a database query to extract past exam question data that matches the input criteria.

[1327] Step 4:

[1328] The extracted data is preprocessed and converted into a format that can be input to the generative AI model, for example by tokenizing the text data and adding necessary metadata.

[1329] Step 5:

[1330] A generative AI model is then invoked to generate new test questions and explanations using the preprocessed data as input, with each question containing four options and a correct answer.

[1331] Step 6:

[1332] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[1333] Step 7:

[1334] The formatted data is sent to the terminal as an API response.

[1335] Terminal

[1336] Step 1:

[1337] Through the user interface, the user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[1338] Step 2:

[1339] The entered conditions are combined into a single JSON request data, for example, { "Subject": "Mathematics", "Format": "Multiple Choice", "Difficulty": "Intermediate", "Number of Questions": 5}.

[1340] Step 3:

[1341] Send the request data to the server using an HTTP request (POST method).

[1342] Step 4:

[1343] Receive the response data (test questions and explanations) sent back from the server as an HTTP response (200 OK).

[1344] Step 5:

[1345] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[1346] Step 6:

[1347] The formatted test questions and explanations are displayed in the user interface, and are output to the screen so that the user can check them.

[1348] User

[1349] Step 1:

[1350] Operate the terminal's user interface and enter the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[1351] Step 2:

[1352] Once you have completed the input, click the send button on your device to send the request.

[1353] Step 3:

[1354] Check that the response from the server is displayed on the terminal. The test questions and explanations are displayed on the screen so that the user can check them.

[1355] Step 4:

[1356] To revise the test questions and explanations as needed, enter the criteria again and submit a new request.

[1357] Example 1

[1358] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1359] Conventional test question creation not only requires a significant amount of time and effort from the question creator, but also has the problem of difficulty in maintaining appropriate level of difficulty and quality. Furthermore, efficiently utilizing a database of past questions to generate new test questions is also a technical challenge. There is a need to solve these problems and improve the efficiency and quality of test question creation in educational institutions and qualification examinations.

[1360] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1361] In this invention, the server includes input means for accepting a test question creation request from a user, analysis means for analyzing the received request data and extracting conditions such as test subject, question format, difficulty level, and number of questions, acquisition means for acquiring appropriate data from a past question database based on the specified test subject, preprocessing means for preprocessing the acquired data and converting it into a format to be input to the generative AI model, generation means for generating new test questions and explanations using the generative AI model, output means for formatting and outputting the generated test questions and explanations, and communication means for returning the output data to the user. This makes it possible to automatically generate test questions that meet the conditions specified by the user and quickly provide high-quality, validated questions and explanations.

[1362] A "user" is an entity that uses the system to make a request to create test questions.

[1363] A "server" is a computer system that receives requests from users, generates test questions and explanations, and returns them to the users.

[1364] A "terminal" is a device that a user uses to input test question creation conditions and check the results.

[1365] "Input means" is a function that accepts a test question creation request from a user.

[1366] The "analysis means" is a function that analyzes the request data received from the user and extracts conditions such as test subjects, question format, difficulty level, and number of questions.

[1367] "Acquisition means" is a function that acquires data from a past exam question database based on a specified exam subject.

[1368] "Preprocessing means" is a function that converts acquired data into a format that can be input into a generative AI model.

[1369] "Generation means" is a function that generates new test questions and explanations using a generative AI model.

[1370] "Output means" is a function that formats the generated test questions and explanations and provides them to the user.

[1371] "Communication means" refers to the function that sends and receives requests for test question generation between the server and the terminal.

[1372] A "database" is a collection of information in which past exam questions are stored.

[1373] A "generative AI model" is an artificial intelligence model used to automatically generate test questions.

[1374] A "prompt sentence" is an instruction sentence that causes a generative AI model to generate test questions.

[1375] The "user interface" is a display screen for displaying the generated test questions and explanations.

[1376] This invention relates to a system that leaves the creation of test questions and explanations to a generative AI model. Specifically, it shows the process of automatically generating test questions through interactions between a server, a terminal, and a user.

[1377] Server Operation

[1378] The server is a high-performance computer system that has the following functions:

[1379] 1. Input method:

[1380] The server accepts test question creation requests from users via an API endpoint. This input is sent in JSON format and includes the test subject, question format, difficulty level, number of questions, etc.

[1381] 2. Analysis method:

[1382] The server analyzes the received JSON data and extracts each parameter (exam subject, question format, difficulty level, number of questions). This analysis process is implemented using programming languages ​​such as Python and JavaScript.

[1383] 3. Acquisition method:

[1384] The server connects to a database of past exam questions and retrieves the appropriate data based on the selected exam subject. This database stores past exam questions and related information.

[1385] 4. Pretreatment methods:

[1386] The acquired data is converted into a format that can be input to the generative AI model, specifically by pre-processing the data to convert it into text-based prompts.

[1387] 5. Generation means:

[1388] Generate new test questions and their explanations using a generative AI model (e.g., OpenAI's GPT-4), using the following prompt:

[1389] "Please create a test based on the following criteria:

[1390] Subject: Mathematics

[1391] Question format: 4 choices

[1392] Difficulty: Intermediate

[1393] Number of questions: 5"

[1394] 6. Output Method:

[1395] The generated test questions and explanations are formatted and returned to the user in JSON format, which includes the question text, options, correct answers, explanations, etc.

[1396] 7. Means of communication:

[1397] The server sends the generated test questions and explanations to the user's device using the HTTP protocol.

[1398] Device behavior

[1399] The terminal provides an interface for users to create test questions and has the following functions:

[1400] 1. Input method:

[1401] The terminal accepts the conditions for creating test questions as input from the user, which is usually provided in the form of a form.

[1402] 2. Means of transmission:

[1403] The conditions entered by the user are converted into JSON format and sent to the server using an HTTP POST request.

[1404] 3. Receiving means:

[1405] The JSON data returned from the server is received and the test questions and explanations are displayed in the user interface, which can be provided as a browser-based UI or a dedicated application.

[1406] User Actions

[1407] 1. Enter the conditions:

[1408] The user inputs the conditions for creating test questions through a terminal, including the subject, question format, difficulty level, and number of questions.

[1409] 2. Check the results:

[1410] The user can check the test questions and explanations returned by the server and make a request again if necessary. For example, if the difficulty or quality of the questions does not meet their expectations, they can change the conditions and make a request again.

[1411] Specific examples

[1412] For example, to create five multiple-choice questions for intermediate level math, the user enters the following conditions into the input form on the device:

[1413] Subject: Mathematics

[1414] Question format: 4 choices

[1415] Difficulty: Intermediate

[1416] Number of questions: 5

[1417] The device converts this data into JSON format and sends it to the server. The server analyzes the input conditions and retrieves the appropriate question data from the database. Based on this, it invokes the generative AI model to generate new test questions and explanations, which are then formatted and sent back to the device. The device then displays the received data on the user interface so that the user can view it.

[1418] This system will improve the efficiency and quality of test question creation, and can be used flexibly in educational institutions and qualification exams.

[1419] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1420] Step 1:

[1421] Accepting user input

[1422] Specific explanation

[1423] The user enters the conditions for creating test questions into the input form on the terminal. These include the test subject (e.g., "Mathematics"), question format (e.g., "Multiple Choice"), difficulty level (e.g., "Intermediate"), and number of questions (e.g., "5"). When the user clicks the "Submit" button, these conditions are converted into JSON format.

[1424] input

[1425] Exam subjects, question format, difficulty level, number of questions

[1426] output

[1427] Request data in JSON format

[1428] Step 2:

[1429] Submitting a Request

[1430] Specific explanation

[1431] The terminal converts the conditions entered by the user into JSON format and sends it to the server as an HTTP POST request.

[1432] input

[1433] Request data in JSON format

[1434] output

[1435] HTTP POST request

[1436] Step 3:

[1437] Accepting a request

[1438] Specific explanation

[1439] The server receives requests from users via the API endpoint and parses the request data, extracting conditions such as test subject, question format, difficulty level, and number of questions from the JSON-formatted request data.

[1440] input

[1441] HTTP POST request

[1442] output

[1443] Extracted data on exam subjects, question format, difficulty level, and number of questions

[1444] Step 4:

[1445] Database Reference

[1446] Specific explanation

[1447] The server retrieves appropriate question data from the past exam database based on the extracted conditions. For example, it uses an SQL query to search for "intermediate" level questions in the "Mathematics" subject.

[1448] sql

[1449] SELECT FROM question_bank WHERE subject = 'Mathematics' AND difficulty = 'Intermediate' LIMIT 10;

[1450] input

[1451] Exam subjects and difficulty level

[1452] output

[1453] Appropriate problem data

[1454] Step 5:

[1455] Data Preprocessing

[1456] Specific explanation

[1457] The server converts the acquired data into a format that can be input to the generative AI model. For example, it formats the acquired problem data into a list format.

[1458] text

[1459] {"question": "What is the solution to x^2 + 2x + 1 = 0?", "choices": ["1", "-1", "0", "2"], "answer": "1"}

[1460] input

[1461] Appropriate problem data

[1462] output

[1463] Data formats that can be input to generative AI models

[1464] Step 6:

[1465] Invoking a generative AI model

[1466] Specific explanation

[1467] The server uses the converted data to send prompts to a generative AI model (e.g., GPT-4) to generate new test questions and explanations.

[1468] text

[1469] "Please create a test based on the following criteria:

[1470] Subject: Mathematics

[1471] Question format: 4 choices

[1472] Difficulty: Intermediate

[1473] Number of questions: 5"

[1474] input

[1475] Prompt statement

[1476] output

[1477] New exam questions and explanations

[1478] Step 7:

[1479] Configuring the Output

[1480] Specific explanation

[1481] The generated test questions and explanations are formatted and converted into JSON format for sending back to the user. For example, formatting it as follows:

[1482] json

[1483] {

[1484] "questions": [

[1485] {"question": "What is the solution to x^2 + 2x + 1 = 0?", "choices": ["1", "-1", "0", "2"], "answer": "1", "explanation": "Using the solution formula..."}

[1486] ]

[1487] }

[1488] input

[1489] New exam questions and explanations

[1490] output

[1491] JSON format data

[1492] Step 8:

[1493] Sending a response

[1494] Specific explanation

[1495] The server sends the formatted JSON data to the device as an HTTP response.

[1496] input

[1497] JSON format data

[1498] output

[1499] HTTP response

[1500] Step 9:

[1501] Receiving a response

[1502] Specific explanation

[1503] The terminal receives the response data from the server, analyzes it, and displays the test questions and explanations on the user interface.

[1504] input

[1505] HTTP response

[1506] output

[1507] Test questions and explanations displayed on the user interface

[1508] Step 10:

[1509] Checking the results

[1510] Specific explanation

[1511] The user can check the test questions and explanations displayed on the terminal and make a re-request if necessary.

[1512] input

[1513] Test questions and explanations displayed on the user interface

[1514] output

[1515] User confirmation / request

[1516] (Application example 1)

[1517] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1518] Conventional test question creation systems require test creators to expend a great deal of time and effort, making it difficult to quickly generate high-quality test questions. Furthermore, it is difficult to automatically generate new test questions by referencing past questions, which often results in inconsistent question quality and difficulty. Furthermore, there was a lack of a mechanism for instantly providing users with the generated test questions and explanations, making it difficult to provide appropriate learning support. There was a need to provide a system that could solve these issues and quickly and efficiently generate and provide test questions to users, particularly in online education and qualification exams.

[1519] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[1520] In this invention, the server includes a generation means for automatically generating test questions for a specific domain, an acquisition means for acquiring information based on requested conditions, a call means for calling a generative AI model for generating new test questions from the acquired information, an output means for formatting and outputting the generated test questions and explanations, a prompt generation means for receiving input from a user via a communication interface and creating prompt sentences for generating test questions that match the conditions, and a display means for displaying the generated test questions and explanations on a user interface. This enables users to easily and quickly generate and check high-quality test questions and explanations.

[1521] A "generation means" is a device or system that has the functionality to automatically generate test questions for a particular domain.

[1522] An "acquisition means" is a device or system for acquiring the necessary information or data based on a request.

[1523] A "calling means" is a device or system that has the function of calling a generative AI model and generating new test questions based on the acquired information.

[1524] The "output means" is a device or system for formatting the generated test questions and explanations into an appropriate format and providing them to an external party.

[1525] A "communications interface" is a device or system that allows data to be sent and received between a server and a user's terminal.

[1526] A "prompt generation means" is a device or system that has the function of creating prompt sentences to pose appropriate questions to the generative AI model based on input conditions from the user.

[1527] The "display means" is a device or system for displaying the generated test questions and explanations on a user interface.

[1528] This invention relates to a system that uses a generative AI model to create and explain test questions. The purpose of this system is to reduce the burden on question creators in educational institutions, qualification exams, and other settings, and to provide high-quality test questions. Here, we will explain a specific embodiment of the system that realizes this application example.

[1529] server

[1530] The server includes the following means:

[1531] 1. Generation means

[1532] This function automatically generates test questions for specific areas. Based on the user's requested tasks, it retrieves information from a database of previous questions and generates new questions.

[1533] 2. Acquisition method

[1534] This function retrieves appropriate information from the past exam question database based on user requests, allowing for optimal data collection.

[1535] 3. Calling means

[1536] This function calls a generative AI model based on the data collected by the acquisition means and generates new test questions. This AI model is usually a generative AI model such as OpenAI.

[1537] 4. Output Method

[1538] This function formats the generated test questions and explanations appropriately and outputs them to the user. This ensures that the generated content is presented to the user in a format that is easy for them to understand.

[1539] 5. Prompt Generation Methods

[1540] This function creates prompt sentences that ask appropriate questions to the generative AI model based on input conditions from the user, thereby eliciting appropriate responses from the AI ​​model.

[1541] 6. Display means

[1542] This function displays the generated test questions and explanations on the user interface.

[1543] Terminal

[1544] The terminal includes the following means:

[1545] 1. Communication Interface

[1546] This function sends and receives data between the server and the user, allowing for smooth requests and responses for test questions.

[1547] User

[1548] The user does the following:

[1549] 1. Input Method

[1550] The conditions for generating test questions are entered into the terminal and a request is sent.

[1551] 2. Transmission Method

[1552] A request is sent to the server based on the entered conditions.

[1553] 3. Display means

[1554] Check the generated test questions and explanations on the device interface.

[1555] For example, if a user wants to create five multiple-choice questions for the intermediate level of "Mathematics," he or she inputs the conditions into the terminal and sends a request to the server. The terminal transmits this information through the communication interface, and the server uses the acquisition means to collect appropriate data and creates a prompt sentence using the generation means and prompt generation means.

[1556] Here is an example prompt:

[1557] Create five multiple-choice questions and explanations for intermediate level mathematics.

[1558] The generative AI model generates new test questions and explanations based on this prompt, formats them using an output method, and returns them to the device. The user can then check the generated test questions and explanations on the device and make further requests as necessary. This allows users to create high-quality test questions efficiently and quickly.

[1559] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1560] Step 1:

[1561] The user uses the input means of the terminal to input conditions for generating test questions. Specifically, detailed conditions such as test subjects, question format, difficulty level, number of questions, etc. This input data is sent to the terminal as request data.

[1562] Step 2:

[1563] The terminal receives the request data and sends it to the server via a communication interface. At this time, the request data includes the test subject, question format, difficulty level, number of questions, etc. The terminal converts the request data into the required format and sends it to the server in JSON format, etc.

[1564] Step 3:

[1565] The server receives and analyzes the request data. As a result of the analysis, the test conditions desired by the user (test subject, question format, difficulty level, number of questions) are extracted. Based on the extracted conditions, the server references the past question database and retrieves past question data related to the specified test subject.

[1566] Step 4:

[1567] The retrieval means retrieves appropriate data from the past exam question database, including extracting questions that match specified criteria using a search query, and stores the extracted data in a temporary data store.

[1568] Step 5:

[1569] The server uses the prompt generation means to create a prompt sentence based on the user's input conditions. For example, a sentence such as "Please create five multiple-choice questions and explanations for intermediate level mathematics" is generated. This prompt sentence is prepared to be passed to the generative AI model.

[1570] Step 6:

[1571] The calling means calls the generative AI model based on the prompt sentence. Specifically, it sends the prompt sentence to the generative AI model, and generates new test questions and explanations in response. The generative AI model (e.g., OpenAI) generates appropriate questions and explanations according to the prompt sentence and returns the results to the server.

[1572] Step 7:

[1573] The generated test questions and explanations are formatted by the output means. The formatted data is converted into a format that is easy for the user to understand (e.g., text, PDF, HTML, etc.). The formatted data is then prepared in a format for transmission to the terminal.

[1574] Step 8:

[1575] The server sends the formatted test questions and explanations through a communications interface to the terminal, which receives them and prepares them for display on the user interface.

[1576] Step 9:

[1577] The terminal uses the display means to display the generated test questions and explanations on the user interface, allowing the user to check them on the terminal and make a re-request if necessary.

[1578] If the user changes the conditions and makes a request again, the process is repeated from step 1. This realizes a system that can quickly and efficiently provide high-quality test questions and explanations.

[1579] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[1580] This invention combines a system that automatically generates test questions for specific subjects with an emotion engine that recognizes user emotions. The purpose of this system is to reduce the burden on question creators in educational institutions and qualification exams, and to improve the quality of users' learning experience.

[1581] System Overview

[1582] server

[1583] The server receives test question creation requests from users, retrieves appropriate data, and generates new test questions and explanations using a generative AI model. It also uses an emotion engine to recognize user emotions and adjust the difficulty of the test questions.

[1584] 1. Accepting user input

[1585] The server receives requests from users through API endpoints. The request data includes detailed conditions such as test subject, question format, difficulty level, number of questions, etc. It also includes emotion recognition data.

[1586] Parse the received request data and emotion data and extract each parameter.

[1587] 2. Referencing the database

[1588] The server retrieves the appropriate question data from the past exam database based on the specified exam subject. It uses a database query to extract data that matches the input criteria and performs additional filtering.

[1589] 3. Invoking the generative AI model

[1590] The server preprocesses the acquired data and converts it into a format that can be input to the generative AI model. It also adjusts the difficulty of the questions based on the emotional data.

[1591] It uses a generative AI model to generate new test questions and explanations, each with four options and a correct answer.

[1592] 4. Configuring the Output

[1593] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[1594] Feedback information including an evaluation by the emotion recognition engine is generated and sent back to the user along with the test questions.

[1595] Terminal

[1596] The device is responsible for interfacing with the user, accepting input, sending requests to the server, and displaying the results. It also contains an emotion engine that recognizes the user's emotions and sends the data to the server.

[1597] 1. Submit a request

[1598] The terminal receives user input, sends the conditions for creating test questions to the server as a request, and also recognizes the user's emotions using an emotion engine and sends the data to the server.

[1599] 2. Receiving a response

[1600] Response data (test questions, explanations, and emotion evaluation results) is received from the server as an HTTP response (200 OK).

[1601] 3. Displaying the results

[1602] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[1603] The formatted test questions, explanations, and emotion evaluation results are displayed in the user interface, and are output to the screen so that the user can check them.

[1604] User

[1605] The user is the entity that inputs the conditions for creating test questions and uses the system to obtain original test questions. In addition, the emotion engine feeds back the user's emotions during learning to the system, improving the quality of the learning experience.

[1606] 1. Enter the conditions

[1607] The user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions) through the terminal.

[1608] 2. Emotional Feedback

[1609] The emotion engine recognizes the user's operating status and facial expressions, and transmits them to the server as emotion data.

[1610] 3. Check the results

[1611] The response from the server is displayed on the terminal. The test questions, explanations, and emotion evaluation results are displayed on the screen for the user to review.

[1612] To revise the test questions and explanations as needed, enter the criteria again and submit a new request.

[1613] Specific examples

[1614] For example, if a user wants to create five multiple-choice questions at an intermediate level in mathematics and wants advice on how to reduce stress levels while studying, the process is as follows:

[1615] 1. Users

[1616] Enter "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions, and send a request from your device. The emotion engine will also capture your learning progress.

[1617] 2. Terminal

[1618] The request data and emotion data are sent to the server.

[1619] 3. Server

[1620] The system analyzes request data, retrieves appropriate data from a database of past questions, and adjusts the difficulty of questions based on emotional data.

[1621] A generative AI model is invoked to generate new test questions and explanations.

[1622] The test questions, explanations, and emotional assessment results are formatted and sent back to the terminal.

[1623] 4. Terminal

[1624] The test questions, explanations, and emotion evaluation results returned from the server are received and displayed on the user interface.

[1625] 5. Users

[1626] Check the test questions, explanations, and emotional assessment results, and receive appropriate advice based on your learning progress.

[1627] If necessary, re-requests will be made to adjust the content and difficulty of the test questions.

[1628] The processing flow will be explained below.

[1629] server

[1630] Step 1:

[1631] The API endpoint receives requests from users. The request data includes test subject, question format, difficulty level, number of questions, and sentiment data.

[1632] Step 2:

[1633] The received request data and emotion data are parsed and each parameter (exam subject, question format, difficulty level, number of questions, emotion data) is extracted.

[1634] Step 3:

[1635] Access the appropriate past exam question database based on the exam subject. Use a database query to extract past exam question data that matches the input criteria.

[1636] Step 4:

[1637] The extracted data is preprocessed and converted into a format that can be input to the generative AI model, for example by tokenizing the text data and adding necessary metadata.

[1638] Step 5:

[1639] The difficulty of the generated test questions is adjusted based on the emotional data. For example, if the user is feeling stressed, the difficulty level is set slightly lower.

[1640] Step 6:

[1641] A generative AI model is then invoked to generate new test questions and explanations using the preprocessed data as input, with each question containing four options and a correct answer.

[1642] Step 7:

[1643] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[1644] Step 8:

[1645] Feedback information including evaluations by the emotion recognition engine is generated and sent back to the terminal along with the test questions.

[1646] Terminal

[1647] Step 1:

[1648] Through the user interface, the user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[1649] Step 2:

[1650] The input conditions are compiled into a single JSON request data. The request also includes the user's emotion data recognized by the emotion engine. For example, the format is { "Subject": "Mathematics", "Format": "Multiple Choice", "Difficulty": "Intermediate", "Number of Questions": 5, "Emotion": "Stress"}.

[1651] Step 3:

[1652] Send the request data to the server using an HTTP request (POST method).

[1653] Step 4:

[1654] Response data (test questions, explanations, and emotion evaluation results) returned from the server is received as an HTTP response (200 OK).

[1655] Step 5:

[1656] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[1657] Step 6:

[1658] The formatted test questions, explanations, and emotion evaluation results are displayed in the user interface, and are output to the screen so that the user can check them.

[1659] User

[1660] Step 1:

[1661] Operate the terminal's user interface and enter the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[1662] Step 2:

[1663] Once you have completed the input, click the send button on your device to send the request.

[1664] Step 3:

[1665] The emotion engine recognizes the user's operating status and facial expressions in real time and generates emotion data, which is included in the request.

[1666] Step 4:

[1667] The request sent from the terminal is processed by the server, and the generated test questions, explanations, and emotion evaluation results are displayed on the terminal.

[1668] Step 5:

[1669] Check the test questions and explanations displayed on the screen and receive appropriate study advice based on the emotional assessment results.

[1670] Step 6:

[1671] If necessary, re-requests will be made to adjust the content and difficulty of the test questions.

[1672] Example 2

[1673] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1674] There is a need to reduce the burden on educational institutions and test question creators and improve the quality of users' learning experience.However, there is no system that can take into account the user's emotional state while studying and adjust the difficulty of questions individually, making it difficult to maintain learning motivation and provide adaptive education.

[1675] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[1676] In this invention, the server includes a generation means for automatically generating test questions for specific subjects, an acquisition means for acquiring data based on requested conditions, a call means for calling a generation AI model that generates new test questions from the acquired data, an output means for formatting and outputting the generated test questions and explanations, an emotion analysis means for recognizing the user's emotions and analyzing the data, a difficulty adjustment means for adjusting the difficulty of the questions based on the emotion data, and a feedback means for returning the generated test questions and explanations to the user and generating feedback including the emotion evaluation results. This reduces the burden on the question creator, improves the quality of the user's learning experience, and enables the difficulty of questions to be adjusted appropriately according to the emotional state of each user.

[1677] A "specific subject" refers to a specific field or academic area that serves as the subject matter for generating test questions.

[1678] A "test question" is material containing questions or tasks used to assess a user's knowledge and understanding.

[1679] "Generation means" refers to the processes and techniques used to automatically create test questions, primarily using generative AI models.

[1680] "Acquisition means" refers to the process or method of searching for and acquiring relevant data or information based on specific conditions.

[1681] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new test questions and explanations.

[1682] "Invocation means" refers to the process or method for running a generative AI model using acquired data to obtain the required output.

[1683] "Output means" refers to the process or method for outputting the generated test questions and explanations in a specified format.

[1684] "Emotion analysis means" refers to techniques and processes for detecting and analyzing a user's emotional state.

[1685] "Difficulty adjustment means" refers to a process or method for appropriately adjusting the difficulty of test questions based on emotional data.

[1686] "Feedback means" refers to a process or technology for returning the generated test questions, explanations, and emotional evaluation results to the user and providing feedback information.

[1687] "User interface" refers to the interface for exchanging information between the system and the user, and displays test questions, explanations, etc.

[1688] "Communication means" refers to the communication technology and protocol used to request and respond to test question generation between the server and the terminal.

[1689] "Input means" refers to a device or interface that allows a user to input conditions for generating test questions.

[1690] "Transmission means" refers to the process or method for sending a request to a server based on the conditions entered by the user.

[1691] MODE FOR CARRYING OUT THE INVENTION

[1692] This invention combines a system for automatically generating test questions for specific subjects with an emotion engine that recognizes user emotions. This system aims to reduce the burden on question creators in educational institutions and qualification exams and improve the quality of users' learning experience.

[1693] server

[1694] A server is a combination of hardware and software that performs multiple functions. In particular, it includes the following elements:

[1695] 1. Generation means

[1696] The server uses a generative AI model to automatically generate test questions for specific subjects. The generative AI model uses artificial intelligence algorithms to generate new questions and their explanations.

[1697] 2. Acquisition method

[1698] The server retrieves data from information sources such as a database of past exam questions based on the conditions requested by the user. To do this, it uses SQL queries to search for and retrieve data that matches the request conditions.

[1699] 3. Calling means

[1700] The server preprocesses the acquired data and converts it into a format that can be input to the generative AI model. It also calls the generative AI model to generate new test questions and explanations. Specifically, it passes the necessary data to the generative AI model as prompts.

[1701] 4. Output Method

[1702] The server formats the generated test questions and explanations in an appropriate format (e.g., JSON or HTML), outputs the data, and temporarily stores it before returning it to the user.

[1703] 5. Emotion analysis method

[1704] The server receives the user's emotional data and uses an emotion engine to analyze it, which detects various emotional states such as the user's stress level and concentration level.

[1705] 6. Difficulty adjustment method

[1706] The server adjusts the difficulty of the questions based on the emotional data, for example, setting the difficulty of the questions lower if the user indicates a high stress level.

[1707] 7. Feedback channels

[1708] The server returns the generated test questions and explanations, as well as the emotional evaluation results, to the user, providing feedback information, so that the user can receive appropriate feedback based on their learning progress and emotional state.

[1709] Terminal

[1710] The device is responsible for the interface with the user, accepting input, sending requests to the server, and displaying the results. It also recognizes the user's emotions and sends the data to the server.

[1711] 1. Submit a request

[1712] The terminal uses an HTTP request to transmit the test question creation conditions and emotion data received from the user to the server.

[1713] 2. Receiving a response

[1714] The terminal receives the response data returned from the server, which includes the test questions, explanations, and emotion evaluation results.

[1715] 3. Displaying the results

[1716] The terminal analyzes the received response data and displays it on the user interface. For example, it formats and displays test questions and explanations using HTML templates.

[1717] User

[1718] To use the system to obtain the original test questions, users must:

[1719] 1. Enter the conditions

[1720] The user inputs the conditions for creating test questions (for example, test subjects, question format, difficulty level, number of questions) through the terminal.

[1721] 2. Emotional Feedback

[1722] During learning, the user's emotional data is captured by the emotion engine through their operating status and facial expressions.

[1723] 3. Check the results

[1724] The user can check the test questions, explanations, and emotion evaluation results returned in response from the server by displaying them on the terminal. If necessary, the user can send a re-request to adjust the content and difficulty of the test questions.

[1725] Specific examples

[1726] For example, if a user wants to create five intermediate-level multiple-choice questions in math and get advice on how to reduce stress levels while studying, the process might go like this:

[1727] 1. Users

[1728] The user enters "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions into the device and sends a request. At this time, the emotion engine captures the user's emotions while studying.

[1729] 2. Terminal

[1730] The request data and emotion data are sent to the server.

[1731] 3. Server

[1732] The system analyzes the request data, retrieves appropriate data from a database of past exam questions, and adjusts the difficulty of the questions based on the emotion data. It then uses a generative AI model to generate new exam questions and explanations. The generated questions, explanations, and emotion evaluation results are formatted and sent back to the device.

[1733] 4. Terminal

[1734] Receives the data returned from the server and displays it in the user interface.

[1735] 5. Users

[1736] Review the test questions, explanations, and emotional evaluation results. If necessary, make a request again to adjust the content and difficulty of the test questions.

[1737] Examples of specific prompts to use

[1738] Create five multiple-choice questions for intermediate level math based on the following criteria, taking into account the user's stress level while studying:

[1739] Exam subject: Mathematics

[1740] Question format: 4 choices

[1741] Difficulty: Intermediate

[1742] Number of questions: 5

[1743] keyword

[1744] Generative AI model, prompt sentence

[1745] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1746] Step 1:

[1747] Accepting user input

[1748] server

[1749] The server receives requests from users through an HTTP API endpoint. Specifically, it receives JSON data (e.g., {"subject": "math", "difficulty": "intermediate", "question_count": 5}) in which the user enters conditions such as test subject, question format, difficulty level, and number of questions. It also receives the user's emotion data at the same time.

[1750] Input: Test question creation conditions (JSON format) and emotion data

[1751] Data processing / calculation: Parse JSON data and extract each parameter (exam subject, question format, difficulty level, number of questions)

[1752] Output: Extracted parameters and emotion data

[1753] Step 2:

[1754] Database Reference

[1755] server

[1756] The server references the past exam question database based on the criteria received from the user, and uses an SQL query to retrieve data that matches the request criteria (e.g. SELECT FROM problem_database WHERE subject = 'math' AND difficulty = 'intermediate').

[1757] Input: Extracted parameters (exam subject, difficulty level, etc.)

[1758] Data processing / calculation: Executing SQL queries and filtering retrieved data

[1759] Output: Filtered data

[1760] Step 3:

[1761] Invoking a generative AI model

[1762] server

[1763] The server preprocesses the acquired data and converts it into a format suitable for input to the generative AI model. It adjusts the difficulty of the questions based on the emotional data (e.g., lowering the difficulty level if the stress level is high). It then calls the generative AI model to generate new test questions and explanations. The generated questions include four options and a correct answer.

[1764] Input: filtered data, sentiment data

[1765] Data processing / calculation: Preprocessing data, converting it to an input format for the generative AI model, and invoking the generative AI model using prompts.

[1766] Output: Generated test questions and explanations (e.g., ["Q1: Question text...", "A: Choice 1", "B: Choice 2", "C: Choice 3", "D: Choice 4", "Answer: B"])

[1767] Step 4:

[1768] Configuring the Output

[1769] server

[1770] The server formats the generated test questions and explanations into an appropriate format (e.g., JSON format), and also generates feedback information including the emotion evaluation results, which is temporarily stored in a database for sending back to the user.

[1771] Input: Generated test questions and explanations, emotional evaluation results

[1772] Data processing / calculation: formatting test questions and explanations, generating feedback information

[1773] Output: Formatted exam questions, explanations and feedback information (JSON format)

[1774] Step 5:

[1775] Send request

[1776] Terminal

[1777] The terminal receives the user's input and sends the request data and emotion data to the server using an HTTP request.

[1778] Input: User-entered test question creation conditions and emotion data

[1779] Data processing / calculation: Converting input data to JSON format, creating and sending HTTP requests

[1780] Output: Request data sent to the server (JSON format)

[1781] Step 6:

[1782] Response reception

[1783] Terminal

[1784] The device receives the response data from the server. After receiving the HTTP status code 200 (OK), it acquires the test questions, explanations, and emotion evaluation results contained in the response body.

[1785] Input: Response data from the server

[1786] Data processing / calculation: Response data analysis (JSON parsing)

[1787] Output: Analyzed test questions, explanations, and emotion evaluation results

[1788] Step 7:

[1789] Displaying the results

[1790] Terminal

[1791] The device analyzes the received response data and displays it on the user interface. It uses HTML templates to format and display test questions, explanations, and emotion evaluation results.

[1792] Input: Analyzed test questions, explanations, and emotion evaluation results

[1793] Data processing / calculation: Rendering using HTML templates

[1794] Output: Formatted test questions, explanations, and emotional evaluation results (displayed on screen)

[1795] Step 8:

[1796] Entering conditions

[1797] User

[1798] The user inputs the conditions for creating test questions (test subjects, question format, difficulty level, number of questions) through the terminal. After the conditions are entered, the user clicks a button to send the request.

[1799] Input: Conditions for creating test questions

[1800] Data processing / calculation: None

[1801] Output: What you type appears on the terminal

[1802] Step 9:

[1803] Emotional feedback

[1804] User

[1805] The emotion engine captures the operation status and facial expressions during learning, and the emotion data is sent to the server.

[1806] Input: User operation status, facial expressions, etc.

[1807] Data processing / calculation: Analysis by emotion engine, data capture

[1808] Output: Captured emotion data

[1809] Step 10:

[1810] Checking the results

[1811] User

[1812] The test questions, explanations, and emotional evaluation results returned from the server are displayed on the device. If necessary, a request is made again to adjust the content and difficulty of the test questions.

[1813] Input: Test questions, explanations, and emotional evaluation results displayed on the terminal

[1814] Data processing / calculation: None

[1815] Output: User confirmation and re-request if necessary

[1816] (Application example 2)

[1817] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the headset type terminal 314 will be referred to as a "terminal."

[1818] Conventional test question generation systems provide questions uniformly without considering the user's emotions, which reduces the user's learning efficiency. In addition, since feedback based on emotional data is not provided, users tend to feel stressed and may lose motivation to study.

[1819] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[1820] In this invention, the server includes a generation means for automatically generating test questions for specific subjects, an acquisition means for acquiring data based on requested conditions, a call means for calling a generation AI model for generating new test questions from the acquired data, an output means for formatting and outputting the generated test questions and explanations, an emotion recognition means for recognizing a user's emotions, a difficulty adjustment means for adjusting the difficulty of the questions based on the emotion data, and a feedback generation means for generating feedback information based on the emotion data. This makes it possible to provide test questions of an appropriate difficulty level that takes the user's emotions into consideration, improving learning efficiency and maintaining the user's motivation to learn.

[1821] The "means for automatically generating test questions for a specific subject" is a function for automatically generating test questions based on the subjects specified by the user.

[1822] The "acquisition means for acquiring data based on requested conditions" is a function for acquiring necessary data from a database according to conditions input by a user.

[1823] The "calling means for calling the generative AI model" is a function that executes the AI ​​model to generate new test questions using the acquired data.

[1824] The "output means for formatting and outputting the generated test questions and explanations" is a function for formatting the newly generated test questions and their explanations in an appropriate format and outputting them.

[1825] The "emotion recognition means for recognizing the user's emotions" is a function for analyzing the user's facial expressions, actions, etc. to grasp the user's emotional state.

[1826] The "difficulty level adjusting means for adjusting the difficulty level of questions based on emotional data" is a function for providing test questions of a difficulty level appropriate for the user based on the recognized emotional data.

[1827] The "feedback generating means for generating feedback information based on emotional data" is a function for providing appropriate advice and feedback based on the emotional state of the user.

[1828] The "display means for displaying the generated test questions and explanations on the user interface" is a function for displaying the generated test questions and explanations on the screen so that the user can check them.

[1829] "Communication means for requesting and responding to test question generation between the server and the terminal" is a function for sending and receiving requests for test question generation and the generated results between the server and the terminal.

[1830] The "input means for the user to input conditions for generating test questions" is a function for the user to input conditions such as the subject, difficulty level, and format of the test questions that he or she wants to generate.

[1831] The "transmission means for transmitting a request to the server based on the input conditions" is a function for transmitting the conditions input by the user to the server and requesting the generation of test questions.

[1832] The "emotion data transmission means for acquiring emotion data in real time and transmitting it to the server" is a function for collecting the user's emotions in real time and transmitting that data to the server.

[1833] This invention combines a system that automatically generates test questions for specific subjects with an emotion engine that recognizes the user's emotions. This system includes a generation means, an acquisition means, a calling means, an output means, an emotion recognition means, a difficulty level adjustment means, and a feedback generation means. It is particularly characterized by using emotion data to adjust the difficulty of questions and provide appropriate feedback.

[1834] System Overview

[1835] server

[1836] The server has the following features:

[1837] Generation method: Automatically generate test questions for specified subjects.

[1838] Acquisition method: Retrieves the necessary data from the database based on a user request.

[1839] Calling method: Input the acquired data into the generative AI model to generate new test questions.

[1840] Output method: Format the generated test questions and explanations and output them in an appropriate format.

[1841] Emotion recognition means: Recognizes the user's emotions using cameras and sensors and acquires them as data.

[1842] Difficulty adjustment method: Adjust the difficulty of the questions based on emotional data.

[1843] Feedback generation means: Generates feedback information based on emotion data and provides it to the user.

[1844] The hardware used includes servers, databases, and computing resources for running the AI ​​model, while the software includes an emotion recognition engine (EmotionEngine) and a generative AI model (QuestionGenerator).

[1845] Terminal

[1846] The terminal is responsible for:

[1847] Input means: The user inputs the conditions for generating test questions.

[1848] Transmission method: The input conditions are sent to the server as a request, along with the emotion data.

[1849] Display: Displays the generated exam questions, explanations, and feedback.

[1850] For example, smartphones, smart glasses, head-mounted displays, etc. The emotion recognition engine works using the cameras and sensors built into these devices.

[1851] User

[1852] The user does the following:

[1853] Input of conditions: The user inputs the conditions for creating test questions (subject, difficulty level, question format, number) through the terminal.

[1854] Emotion recognition: The learning process is captured by the emotion recognition engine and the emotional data is sent to the server.

[1855] Review feedback: Review the generated exam questions, explanations, and sentiment-based advice.

[1856] Specific examples

[1857] For example, if a user wants to generate five multiple-choice questions for intermediate level math, they can input the following prompts into the generative AI model.

[1858] "Generate five intermediate level multiple choice questions in mathematics. The user's emotion is 'feeling somewhat stressed.' Adjust the difficulty of the questions taking this emotion into account."

[1859] This allows users to receive test questions of an appropriate difficulty level according to their own feelings, which is expected to improve learning efficiency and motivation.

[1860] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1861] Step 1:

[1862] The user inputs the conditions for generating test questions. The user inputs the test subjects, difficulty level, question format, and number of questions using a smartphone or other device. This input data becomes the basis for creating a request to generate test questions.

[1863] Step 2:

[1864] The emotion recognition engine recognizes the user's emotions in real time. It uses the device's built-in camera and sensors to capture the user's facial expressions and behaviors and extract emotional data, which is then used in subsequent processing steps.

[1865] Step 3:

[1866] The input conditions and emotion data are sent to the server. A test question generation request and emotion data are sent to the server using the terminal's transmission means. The input data in this process are the user's condition information acquired in (Step 1) and the emotion information acquired in (Step 2).

[1867] Step 4:

[1868] The server analyzes the request data and retrieves the necessary data. The server analyzes the request data and emotion data and retrieves related data such as past questions from the database based on that. Appropriate questions are extracted based on the request data as input.

[1869] Step 5:

[1870] The server calls the generative AI model to generate new test questions. The acquired data is preprocessed and input to the generative AI model along with the prompt. The model generates new test questions and explanations. The input data is the preprocessed data and the prompt, and the generated output is new test questions and their explanations.

[1871] Step 6:

[1872] Format the generated test questions and explanations and return them to the user. Format the generated data into an appropriate format and send it to the device. Input data is the output data from the generative AI model, which is converted into a format that can be sent to the device.

[1873] Step 7:

[1874] The device displays the test questions, explanations, and emotional feedback on the user interface. The received data is analyzed and displayed in a format that the user can understand. The input data is the response data from the server, and the displayed content is the output.

[1875] Through the above procedure, the user can obtain appropriate test questions and feedback according to his / her own emotions. These processing steps detail the processing flow of a specific system and program based on the present invention.

[1876] The specific processing unit 290 transmits the result of the specific processing to the headset type terminal 314. In the headset type terminal 314, the control unit 46A causes the speaker 240 and the display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating a user input regarding the result of the specific processing. The control unit 46A transmits audio data indicating the user input acquired by the microphone 238 to the data processing device 12. In the data processing device 12, the specific processing unit 290 acquires the audio data.

[1877] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (Internet Search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search <url: https: gemini.google.com ?hl="ja">) and other generation AIs. The data generation model 58 is obtained by performing deep learning on a neural network. A prompt including an instruction is input to the data generation model 58, and inference data such as voice data indicating voice, text data indicating text, and image data indicating an image is also input. The data generation model 58 performs inference on the input inference data in accordance with the instruction indicated by the prompt, and outputs the inference result in a data format such as voice data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1878] In the above embodiment, an example was given in which the specific processing is performed by the data processing device 12, but the technology of the present disclosure is not limited to this, and the specific processing may be performed by the headset type terminal 314.

[1879] [Fourth embodiment]

[1880] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1881] 7, a data processing system 410 includes a data processing device 12 and a robot 414. An example of the data processing device 12 is a server.

[1882] The data processing device 12 includes a computer 22, a database 24, and a communication I / F 26. The computer 22 is an example of a "computer" according to the technology of the present disclosure. The computer 22 includes a processor 28, a RAM 30, and a storage 32. The processor 28, the RAM 30, and the storage 32 are connected to a bus 34. The database 24 and the communication I / F 26 are also connected to the bus 34. The communication I / F 26 is connected to a network 54. Examples of the network 54 include a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[1883] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication I / F 44, and a control target 443. The computer 36 includes a processor 46, a RAM 48, and a storage 50. The processor 46, the RAM 48, and the storage 50 are connected to a bus 52. The microphone 238, the speaker 240, the camera 42, and the control target 443 are also connected to the bus 52.

[1884] The microphone 238 receives instructions and the like from the user 20 by receiving voice uttered by the user 20. The microphone 238 captures the voice uttered by the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio in accordance with instructions from the processor 46.

[1885] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an imaging element such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the surroundings of user 20 (for example, an imaging range defined by an angle of view equivalent to the field of vision of a typical healthy person).

[1886] The communication I / F 44 is connected to a network 54. The communication I / Fs 44 and 26 are responsible for the exchange of various information between the processor 46 and the processor 28 via the network 54. The exchange of various information between the processor 46 and the processor 28 using the communication I / Fs 44 and 26 is carried out in a secure state.

[1887] The control object 443 includes a display device, LEDs in the eyes, and motors for driving the arms, hands, and feet. The posture and gestures of the robot 414 are controlled by controlling the motors of the arms, hands, and feet. Some of the emotions of the robot 414 can be expressed by controlling these motors. In addition, the facial expressions of the robot 414 can also be expressed by controlling the light emission state of the LEDs in the eyes of the robot 414.

[1888] Fig. 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Fig. 8, in the data processing device 12, a specific process is performed by the processor 28. A specific process program 56 is stored in the storage 32.

[1889] The specific processing program 56 is an example of a "program" according to the technology of the present disclosure. The processor 28 reads the specific processing program 56 from the storage 32 and executes the read specific processing program 56 on the RAM 30. The specific processing is realized by the processor 28 operating as a specific processing unit 290 in accordance with the specific processing program 56 executed on the RAM 30.

[1890] The storage 32 stores a data generation model 58 and an emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[1891] In the robot 414, the processor 46 performs the reception output process. A reception output program 60 is stored in the storage 50. The processor 46 reads the reception output program 60 from the storage 50 and executes the read reception output program 60 on the RAM 48. The reception output process is realized by the processor 46 operating as the control unit 46A in accordance with the reception output program 60 executed on the RAM 48.

[1892] Next, a description will be given of the specific processing performed by the specific processing unit 290 of the data processing device 12. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1893] This invention relates to a system that uses a generative AI model to create and explain test questions. The purpose of this system is to reduce the burden on question creators in educational institutions and qualification exams, and to provide high-quality test questions with appropriate difficulty.

[1894] System Overview

[1895] server

[1896] The server receives test question creation requests from users, acquires the appropriate data, and generates new test questions and explanations using a generative AI model. Specifically, it performs the following processes:

[1897] 1. Accepting user input

[1898] The server receives requests from users through API endpoints.

[1899] The request data includes detailed conditions such as test subjects, question format, difficulty level, and number of questions.

[1900] Analyze the received request data and extract each parameter.

[1901] 2. Referencing the database

[1902] The server retrieves appropriate question data from a past exam question database based on the specified exam subject.

[1903] A database query is used to extract data that matches the input criteria and perform additional filtering.

[1904] 3. Invoking the generative AI model

[1905] The server preprocesses the acquired data and converts it into a format that can be input into the generative AI model.

[1906] New test questions and explanations are generated using a generative AI model.

[1907] 4. Configuring the Output

[1908] The generated test questions and explanations are formatted and returned to the user.

[1909] Terminal

[1910] The terminal is responsible for the interface with the user, accepting input, sending requests to the server, and displaying the results. Specifically, it performs the following processes:

[1911] 1. Submit a request

[1912] The terminal receives the user's input and sends the conditions for creating test questions to the server as a request.

[1913] 2. Receiving a response

[1914] The test questions and explanations data returned from the server are received and displayed on the user interface.

[1915] User

[1916] The user is the entity that inputs the conditions for creating test questions and uses the system to obtain original test questions. Specifically, the user performs the following operations:

[1917] 1. Enter the conditions

[1918] The user inputs the conditions for creating test questions through the terminal and sends a request.

[1919] 2. Check the results

[1920] Check the test questions and explanations sent from the server and make a request again if necessary.

[1921] Specific examples

[1922] For example, if a user wants to create five multiple-choice questions for intermediate level math, the process is as follows:

[1923] 1. Users

[1924] The user enters "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions, and then sends a request from the device.

[1925] 2. Terminal

[1926] The terminal transmits the conditions entered by the user to the server.

[1927] 3. Server

[1928] The server analyzes the received conditions and retrieves the appropriate data from the past exam question database.

[1929] A generative AI model is called and new test questions and their explanations are generated based on the acquired data.

[1930] The generated test questions and explanations are formatted and sent back to the terminal.

[1931] 4. Terminal

[1932] The terminal receives the test questions and explanations returned from the server and displays them to the user.

[1933] 5. Users

[1934] The user checks the test questions and explanations displayed on the terminal and makes a request again if necessary.

[1935] This system allows teachers and instructors to efficiently create high-quality test questions, contributing to improving students' academic abilities. In addition, users can easily customize test questions, making it extremely flexible and fast to use in educational settings.

[1936] The processing flow will be explained below.

[1937] Program processing flow

[1938] server

[1939] Step 1:

[1940] The API endpoint receives requests from users. The request data includes detailed conditions such as test subject, question format, difficulty level, and number of questions.

[1941] Step 2:

[1942] Parse the received request data and extract each parameter (exam subject, question format, difficulty level, number of questions).

[1943] Step 3:

[1944] Access the appropriate past exam question database based on the exam subject. Use a database query to extract past exam question data that matches the input criteria.

[1945] Step 4:

[1946] The extracted data is preprocessed and converted into a format that can be input to the generative AI model, for example by tokenizing the text data and adding necessary metadata.

[1947] Step 5:

[1948] A generative AI model is then invoked to generate new test questions and explanations using the preprocessed data as input, with each question containing four options and a correct answer.

[1949] Step 6:

[1950] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[1951] Step 7:

[1952] The formatted data is sent to the terminal as an API response.

[1953] Terminal

[1954] Step 1:

[1955] Through the user interface, the user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[1956] Step 2:

[1957] The entered conditions are combined into a single JSON request data, for example, { "Subject": "Mathematics", "Format": "Multiple Choice", "Difficulty": "Intermediate", "Number of Questions": 5}.

[1958] Step 3:

[1959] Send the request data to the server using an HTTP request (POST method).

[1960] Step 4:

[1961] Receive the response data (test questions and explanations) sent back from the server as an HTTP response (200 OK).

[1962] Step 5:

[1963] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[1964] Step 6:

[1965] The formatted test questions and explanations are displayed in the user interface, and are output to the screen so that the user can check them.

[1966] User

[1967] Step 1:

[1968] Operate the terminal's user interface and enter the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[1969] Step 2:

[1970] Once you have completed the input, click the send button on your device to send the request.

[1971] Step 3:

[1972] Check that the response from the server is displayed on the terminal. The test questions and explanations are displayed on the screen so that the user can check them.

[1973] Step 4:

[1974] To revise the test questions and explanations as needed, enter the criteria again and submit a new request.

[1975] Example 1

[1976] Next, a description will be given of Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[1977] Conventional test question creation not only requires a significant amount of time and effort from the question creator, but also has the problem of difficulty in maintaining appropriate level of difficulty and quality. Furthermore, efficiently utilizing a database of past questions to generate new test questions is also a technical challenge. There is a need to solve these problems and improve the efficiency and quality of test question creation in educational institutions and qualification examinations.

[1978] The specific processing by the specific processing unit 290 of the data processing device 12 in the first embodiment is realized by the following means.

[1979] In this invention, the server includes input means for accepting a test question creation request from a user, analysis means for analyzing the received request data and extracting conditions such as test subject, question format, difficulty level, and number of questions, acquisition means for acquiring appropriate data from a past question database based on the specified test subject, preprocessing means for preprocessing the acquired data and converting it into a format to be input to the generative AI model, generation means for generating new test questions and explanations using the generative AI model, output means for formatting and outputting the generated test questions and explanations, and communication means for returning the output data to the user. This makes it possible to automatically generate test questions that meet the conditions specified by the user and quickly provide high-quality, validated questions and explanations.

[1980] A "user" is an entity that uses the system to make a request to create test questions.

[1981] A "server" is a computer system that receives requests from users, generates test questions and explanations, and returns them to the users.

[1982] A "terminal" is a device that a user uses to input test question creation conditions and check the results.

[1983] "Input means" is a function that accepts a test question creation request from a user.

[1984] The "analysis means" is a function that analyzes the request data received from the user and extracts conditions such as test subjects, question format, difficulty level, and number of questions.

[1985] "Acquisition means" is a function that acquires data from a past exam question database based on a specified exam subject.

[1986] "Preprocessing means" is a function that converts acquired data into a format that can be input into a generative AI model.

[1987] "Generation means" is a function that generates new test questions and explanations using a generative AI model.

[1988] "Output means" is a function that formats the generated test questions and explanations and provides them to the user.

[1989] "Communication means" refers to the function that sends and receives requests for test question generation between the server and the terminal.

[1990] A "database" is a collection of information in which past exam questions are stored.

[1991] A "generative AI model" is an artificial intelligence model used to automatically generate test questions.

[1992] A "prompt sentence" is an instruction sentence that causes a generative AI model to generate test questions.

[1993] The "user interface" is a display screen for displaying the generated test questions and explanations.

[1994] This invention relates to a system that leaves the creation of test questions and explanations to a generative AI model. Specifically, it shows the process of automatically generating test questions through interactions between a server, a terminal, and a user.

[1995] Server Operation

[1996] The server is a high-performance computer system that has the following functions:

[1997] 1. Input method:

[1998] The server accepts test question creation requests from users via an API endpoint. This input is sent in JSON format and includes the test subject, question format, difficulty level, number of questions, etc.

[1999] 2. Analysis method:

[2000] The server analyzes the received JSON data and extracts each parameter (exam subject, question format, difficulty level, number of questions). This analysis process is implemented using programming languages ​​such as Python and JavaScript.

[2001] 3. Acquisition method:

[2002] The server connects to a database of past exam questions and retrieves the appropriate data based on the selected exam subject. This database stores past exam questions and related information.

[2003] 4. Pretreatment methods:

[2004] The acquired data is converted into a format that can be input to the generative AI model, specifically by pre-processing the data to convert it into text-based prompts.

[2005] 5. Generation means:

[2006] Generate new test questions and their explanations using a generative AI model (e.g., OpenAI's GPT-4), using the following prompt:

[2007] "Please create a test based on the following criteria:

[2008] Subject: Mathematics

[2009] Question format: 4 choices

[2010] Difficulty: Intermediate

[2011] Number of questions: 5"

[2012] 6. Output Method:

[2013] The generated test questions and explanations are formatted and returned to the user in JSON format, which includes the question text, options, correct answers, explanations, etc.

[2014] 7. Means of communication:

[2015] The server sends the generated test questions and explanations to the user's device using the HTTP protocol.

[2016] Device behavior

[2017] The terminal provides an interface for users to create test questions and has the following functions:

[2018] 1. Input method:

[2019] The terminal accepts the conditions for creating test questions as input from the user, which is usually provided in the form of a form.

[2020] 2. Means of transmission:

[2021] The conditions entered by the user are converted into JSON format and sent to the server using an HTTP POST request.

[2022] 3. Receiving means:

[2023] The JSON data returned from the server is received and the test questions and explanations are displayed in the user interface, which can be provided as a browser-based UI or a dedicated application.

[2024] User Actions

[2025] 1. Enter the conditions:

[2026] The user inputs the conditions for creating test questions through a terminal, including the subject, question format, difficulty level, and number of questions.

[2027] 2. Check the results:

[2028] The user can check the test questions and explanations returned by the server and make a request again if necessary. For example, if the difficulty or quality of the questions does not meet their expectations, they can change the conditions and make a request again.

[2029] Specific examples

[2030] For example, to create five multiple-choice questions for intermediate level math, the user enters the following conditions into the input form on the device:

[2031] Subject: Mathematics

[2032] Question format: 4 choices

[2033] Difficulty: Intermediate

[2034] Number of questions: 5

[2035] The device converts this data into JSON format and sends it to the server. The server analyzes the input conditions and retrieves the appropriate question data from the database. Based on this, it invokes the generative AI model to generate new test questions and explanations, which are then formatted and sent back to the device. The device then displays the received data on the user interface so that the user can view it.

[2036] This system will improve the efficiency and quality of test question creation, and can be used flexibly in educational institutions and qualification exams.

[2037] The flow of the identification process in the first embodiment will be described with reference to FIG.

[2038] Step 1:

[2039] Accepting user input

[2040] Specific explanation

[2041] The user enters the conditions for creating test questions into the input form on the terminal. These include the test subject (e.g., "Mathematics"), question format (e.g., "Multiple Choice"), difficulty level (e.g., "Intermediate"), and number of questions (e.g., "5"). When the user clicks the "Submit" button, these conditions are converted into JSON format.

[2042] input

[2043] Exam subjects, question format, difficulty level, number of questions

[2044] output

[2045] Request data in JSON format

[2046] Step 2:

[2047] Submitting a Request

[2048] Specific explanation

[2049] The terminal converts the conditions entered by the user into JSON format and sends it to the server as an HTTP POST request.

[2050] input

[2051] Request data in JSON format

[2052] output

[2053] HTTP POST request

[2054] Step 3:

[2055] Accepting a request

[2056] Specific explanation

[2057] The server receives requests from users via the API endpoint and parses the request data, extracting conditions such as test subject, question format, difficulty level, and number of questions from the JSON-formatted request data.

[2058] input

[2059] HTTP POST request

[2060] output

[2061] Extracted data on exam subjects, question format, difficulty level, and number of questions

[2062] Step 4:

[2063] Database Reference

[2064] Specific explanation

[2065] The server retrieves appropriate question data from the past exam database based on the extracted conditions. For example, it uses an SQL query to search for "intermediate" level questions in the "Mathematics" subject.

[2066] sql

[2067] SELECT FROM question_bank WHERE subject = 'Mathematics' AND difficulty = 'Intermediate' LIMIT 10;

[2068] input

[2069] Exam subjects and difficulty level

[2070] output

[2071] Appropriate problem data

[2072] Step 5:

[2073] Data Preprocessing

[2074] Specific explanation

[2075] The server converts the acquired data into a format that can be input to the generative AI model. For example, it formats the acquired problem data into a list format.

[2076] text

[2077] {"question": "What is the solution to x^2 + 2x + 1 = 0?", "choices": ["1", "-1", "0", "2"], "answer": "1"}

[2078] input

[2079] Appropriate problem data

[2080] output

[2081] Data formats that can be input to generative AI models

[2082] Step 6:

[2083] Invoking a generative AI model

[2084] Specific explanation

[2085] The server uses the converted data to send prompts to a generative AI model (e.g., GPT-4) to generate new test questions and explanations.

[2086] text

[2087] "Please create a test based on the following criteria:

[2088] Subject: Mathematics

[2089] Question format: 4 choices

[2090] Difficulty: Intermediate

[2091] Number of questions: 5"

[2092] input

[2093] Prompt statement

[2094] output

[2095] New exam questions and explanations

[2096] Step 7:

[2097] Configuring the Output

[2098] Specific explanation

[2099] The generated test questions and explanations are formatted and converted into JSON format for sending back to the user. For example, formatting it as follows:

[2100] json

[2101] {

[2102] "questions": [

[2103] {"question": "What is the solution to x^2 + 2x + 1 = 0?", "choices": ["1", "-1", "0", "2"], "answer": "1", "explanation": "Using the solution formula..."}

[2104] ]

[2105] }

[2106] input

[2107] New exam questions and explanations

[2108] output

[2109] JSON format data

[2110] Step 8:

[2111] Sending a response

[2112] Specific explanation

[2113] The server sends the formatted JSON data to the device as an HTTP response.

[2114] input

[2115] JSON format data

[2116] output

[2117] HTTP response

[2118] Step 9:

[2119] Receiving a response

[2120] Specific explanation

[2121] The terminal receives the response data from the server, analyzes it, and displays the test questions and explanations on the user interface.

[2122] input

[2123] HTTP response

[2124] output

[2125] Test questions and explanations displayed on the user interface

[2126] Step 10:

[2127] Checking the results

[2128] Specific explanation

[2129] The user can check the test questions and explanations displayed on the terminal and make a re-request if necessary.

[2130] input

[2131] Test questions and explanations displayed on the user interface

[2132] output

[2133] User confirmation / request

[2134] (Application example 1)

[2135] Next, a description will be given of Application Example 1. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2136] Conventional test question creation systems require test creators to expend a great deal of time and effort, making it difficult to quickly generate high-quality test questions. Furthermore, it is difficult to automatically generate new test questions by referencing past questions, which often results in inconsistent question quality and difficulty. Furthermore, there was a lack of a mechanism for instantly providing users with the generated test questions and explanations, making it difficult to provide appropriate learning support. There was a need to provide a system that could solve these issues and quickly and efficiently generate and provide test questions to users, particularly in online education and qualification exams.

[2137] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 1 is realized by the following means.

[2138] In this invention, the server includes a generation means for automatically generating test questions for a specific domain, an acquisition means for acquiring information based on requested conditions, a call means for calling a generative AI model for generating new test questions from the acquired information, an output means for formatting and outputting the generated test questions and explanations, a prompt generation means for receiving input from a user via a communication interface and creating prompt sentences for generating test questions that match the conditions, and a display means for displaying the generated test questions and explanations on a user interface. This enables users to easily and quickly generate and check high-quality test questions and explanations.

[2139] A "generation means" is a device or system that has the functionality to automatically generate test questions for a particular domain.

[2140] An "acquisition means" is a device or system for acquiring the necessary information or data based on a request.

[2141] A "calling means" is a device or system that has the function of calling a generative AI model and generating new test questions based on the acquired information.

[2142] The "output means" is a device or system for formatting the generated test questions and explanations into an appropriate format and providing them to an external party.

[2143] A "communications interface" is a device or system that allows data to be sent and received between a server and a user's terminal.

[2144] A "prompt generation means" is a device or system that has the function of creating prompt sentences to pose appropriate questions to the generative AI model based on input conditions from the user.

[2145] The "display means" is a device or system for displaying the generated test questions and explanations on a user interface.

[2146] This invention relates to a system that uses a generative AI model to create and explain test questions. The purpose of this system is to reduce the burden on question creators in educational institutions, qualification exams, and other settings, and to provide high-quality test questions. Here, we will explain a specific embodiment of the system that realizes this application example.

[2147] server

[2148] The server includes the following means:

[2149] 1. Generation means

[2150] This function automatically generates test questions for specific areas. Based on the user's requested tasks, it retrieves information from a database of previous questions and generates new questions.

[2151] 2. Acquisition method

[2152] This function retrieves appropriate information from the past exam question database based on user requests, allowing for optimal data collection.

[2153] 3. Calling means

[2154] This function calls a generative AI model based on the data collected by the acquisition means and generates new test questions. This AI model is usually a generative AI model such as OpenAI.

[2155] 4. Output Method

[2156] This function formats the generated test questions and explanations appropriately and outputs them to the user. This ensures that the generated content is presented to the user in a format that is easy for them to understand.

[2157] 5. Prompt Generation Methods

[2158] This function creates prompt sentences that ask appropriate questions to the generative AI model based on input conditions from the user, thereby eliciting appropriate responses from the AI ​​model.

[2159] 6. Display means

[2160] This function displays the generated test questions and explanations on the user interface.

[2161] Terminal

[2162] The terminal includes the following means:

[2163] 1. Communication Interface

[2164] This function sends and receives data between the server and the user, allowing for smooth requests and responses for test questions.

[2165] User

[2166] The user does the following:

[2167] 1. Input Method

[2168] The conditions for generating test questions are entered into the terminal and a request is sent.

[2169] 2. Transmission Method

[2170] A request is sent to the server based on the entered conditions.

[2171] 3. Display means

[2172] Check the generated test questions and explanations on the device interface.

[2173] For example, if a user wants to create five multiple-choice questions for the intermediate level of "Mathematics," he or she inputs the conditions into the terminal and sends a request to the server. The terminal transmits this information through the communication interface, and the server uses the acquisition means to collect appropriate data and creates a prompt sentence using the generation means and prompt generation means.

[2174] Here is an example prompt:

[2175] Create five multiple-choice questions and explanations for intermediate level mathematics.

[2176] The generative AI model generates new test questions and explanations based on this prompt, formats them using an output method, and returns them to the device. The user can then check the generated test questions and explanations on the device and make further requests as necessary. This allows users to create high-quality test questions efficiently and quickly.

[2177] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[2178] Step 1:

[2179] The user uses the input means of the terminal to input conditions for generating test questions. Specifically, detailed conditions such as test subjects, question format, difficulty level, number of questions, etc. This input data is sent to the terminal as request data.

[2180] Step 2:

[2181] The terminal receives the request data and sends it to the server via a communication interface. At this time, the request data includes the test subject, question format, difficulty level, number of questions, etc. The terminal converts the request data into the required format and sends it to the server in JSON format, etc.

[2182] Step 3:

[2183] The server receives and analyzes the request data. As a result of the analysis, the test conditions desired by the user (test subject, question format, difficulty level, number of questions) are extracted. Based on the extracted conditions, the server references the past question database and retrieves past question data related to the specified test subject.

[2184] Step 4:

[2185] The retrieval means retrieves appropriate data from the past exam question database, including extracting questions that match specified criteria using a search query, and stores the extracted data in a temporary data store.

[2186] Step 5:

[2187] The server uses the prompt generation means to create a prompt sentence based on the user's input conditions. For example, a sentence such as "Please create five multiple-choice questions and explanations for intermediate level mathematics" is generated. This prompt sentence is prepared to be passed to the generative AI model.

[2188] Step 6:

[2189] The calling means calls the generative AI model based on the prompt sentence. Specifically, it sends the prompt sentence to the generative AI model, and generates new test questions and explanations in response. The generative AI model (e.g., OpenAI) generates appropriate questions and explanations according to the prompt sentence and returns the results to the server.

[2190] Step 7:

[2191] The generated test questions and explanations are formatted by the output means. The formatted data is converted into a format that is easy for the user to understand (e.g., text, PDF, HTML, etc.). The formatted data is then prepared in a format for transmission to the terminal.

[2192] Step 8:

[2193] The server sends the formatted test questions and explanations through a communications interface to the terminal, which receives them and prepares them for display on the user interface.

[2194] Step 9:

[2195] The terminal uses the display means to display the generated test questions and explanations on the user interface, allowing the user to check them on the terminal and make a re-request if necessary.

[2196] If the user changes the conditions and makes a request again, the process is repeated from step 1. This realizes a system that can quickly and efficiently provide high-quality test questions and explanations.

[2197] Furthermore, an emotion engine that estimates the user's emotion may be further combined. That is, the identification processing unit 290 may estimate the user's emotion using the emotion identification model 59, and perform identification processing using the user's emotion.

[2198] This invention combines a system that automatically generates test questions for specific subjects with an emotion engine that recognizes user emotions. The purpose of this system is to reduce the burden on question creators in educational institutions and qualification exams, and to improve the quality of users' learning experience.

[2199] System Overview

[2200] server

[2201] The server receives test question creation requests from users, retrieves appropriate data, and generates new test questions and explanations using a generative AI model. It also uses an emotion engine to recognize user emotions and adjust the difficulty of the test questions.

[2202] 1. Accepting user input

[2203] The server receives requests from users through API endpoints. The request data includes detailed conditions such as test subject, question format, difficulty level, number of questions, etc. It also includes emotion recognition data.

[2204] Parse the received request data and emotion data and extract each parameter.

[2205] 2. Referencing the database

[2206] The server retrieves the appropriate question data from the past exam database based on the specified exam subject. It uses a database query to extract data that matches the input criteria and performs additional filtering.

[2207] 3. Invoking the generative AI model

[2208] The server preprocesses the acquired data and converts it into a format that can be input to the generative AI model. It also adjusts the difficulty of the questions based on the emotional data.

[2209] It uses a generative AI model to generate new test questions and explanations, each with four options and a correct answer.

[2210] 4. Configuring the Output

[2211] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[2212] Feedback information including an evaluation by the emotion recognition engine is generated and sent back to the user along with the test questions.

[2213] Terminal

[2214] The device is responsible for interfacing with the user, accepting input, sending requests to the server, and displaying the results. It also contains an emotion engine that recognizes the user's emotions and sends the data to the server.

[2215] 1. Submit a request

[2216] The terminal receives user input, sends the conditions for creating test questions to the server as a request, and also recognizes the user's emotions using an emotion engine and sends the data to the server.

[2217] 2. Receiving a response

[2218] Response data (test questions, explanations, and emotion evaluation results) is received from the server as an HTTP response (200 OK).

[2219] 3. Displaying the results

[2220] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[2221] The formatted test questions, explanations, and emotion evaluation results are displayed in the user interface, and are output to the screen so that the user can check them.

[2222] User

[2223] The user is the entity that inputs the conditions for creating test questions and uses the system to obtain original test questions. In addition, the emotion engine feeds back the user's emotions during learning to the system, improving the quality of the learning experience.

[2224] 1. Enter the conditions

[2225] The user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions) through the terminal.

[2226] 2. Emotional Feedback

[2227] The emotion engine recognizes the user's operating status and facial expressions, and transmits them to the server as emotion data.

[2228] 3. Check the results

[2229] The response from the server is displayed on the terminal. The test questions, explanations, and emotion evaluation results are displayed on the screen for the user to review.

[2230] To revise the test questions and explanations as needed, enter the criteria again and submit a new request.

[2231] Specific examples

[2232] For example, if a user wants to create five multiple-choice questions at an intermediate level in mathematics and wants advice on how to reduce stress levels while studying, the process is as follows:

[2233] 1. Users

[2234] Enter "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions, and send a request from your device. The emotion engine will also capture your learning progress.

[2235] 2. Terminal

[2236] The request data and emotion data are sent to the server.

[2237] 3. Server

[2238] The system analyzes request data, retrieves appropriate data from a database of past questions, and adjusts the difficulty of questions based on emotional data.

[2239] A generative AI model is invoked to generate new test questions and explanations.

[2240] The test questions, explanations, and emotional assessment results are formatted and sent back to the terminal.

[2241] 4. Terminal

[2242] The test questions, explanations, and emotion evaluation results returned from the server are received and displayed on the user interface.

[2243] 5. Users

[2244] Check the test questions, explanations, and emotional assessment results, and receive appropriate advice based on your learning progress.

[2245] If necessary, re-requests will be made to adjust the content and difficulty of the test questions.

[2246] The processing flow will be explained below.

[2247] server

[2248] Step 1:

[2249] The API endpoint receives requests from users. The request data includes test subject, question format, difficulty level, number of questions, and sentiment data.

[2250] Step 2:

[2251] The received request data and emotion data are parsed and each parameter (exam subject, question format, difficulty level, number of questions, emotion data) is extracted.

[2252] Step 3:

[2253] Access the appropriate past exam question database based on the exam subject. Use a database query to extract past exam question data that matches the input criteria.

[2254] Step 4:

[2255] The extracted data is preprocessed and converted into a format that can be input to the generative AI model, for example by tokenizing the text data and adding necessary metadata.

[2256] Step 5:

[2257] The difficulty of the generated test questions is adjusted based on the emotional data. For example, if the user is feeling stressed, the difficulty level is set slightly lower.

[2258] Step 6:

[2259] A generative AI model is then invoked to generate new test questions and explanations using the preprocessed data as input, with each question containing four options and a correct answer.

[2260] Step 7:

[2261] Format the generated questions and explanations into an appropriate format (e.g., JSON format). Temporarily save the formatted data.

[2262] Step 8:

[2263] Feedback information including evaluations by the emotion recognition engine is generated and sent back to the terminal along with the test questions.

[2264] Terminal

[2265] Step 1:

[2266] Through the user interface, the user inputs the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[2267] Step 2:

[2268] The input conditions are compiled into a single JSON request data. The request also includes the user's emotion data recognized by the emotion engine. For example, the format is { "Subject": "Mathematics", "Format": "Multiple Choice", "Difficulty": "Intermediate", "Number of Questions": 5, "Emotion": "Stress"}.

[2269] Step 3:

[2270] Send the request data to the server using an HTTP request (POST method).

[2271] Step 4:

[2272] Response data (test questions, explanations, and emotion evaluation results) returned from the server is received as an HTTP response (200 OK).

[2273] Step 5:

[2274] Parses the received response data and formats it for display in the user interface, for example by rendering the display content using HTML templates.

[2275] Step 6:

[2276] The formatted test questions, explanations, and emotion evaluation results are displayed in the user interface, and are output to the screen so that the user can check them.

[2277] User

[2278] Step 1:

[2279] Operate the terminal's user interface and enter the conditions for creating test questions (test subject, question format, difficulty level, number of questions).

[2280] Step 2:

[2281] Once you have completed the input, click the send button on your device to send the request.

[2282] Step 3:

[2283] The emotion engine recognizes the user's operating status and facial expressions in real time and generates emotion data, which is included in the request.

[2284] Step 4:

[2285] The request sent from the terminal is processed by the server, and the generated test questions, explanations, and emotion evaluation results are displayed on the terminal.

[2286] Step 5:

[2287] Check the test questions and explanations displayed on the screen and receive appropriate study advice based on the emotional assessment results.

[2288] Step 6:

[2289] If necessary, re-requests will be made to adjust the content and difficulty of the test questions.

[2290] Example 2

[2291] Next, a description will be given of Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2292] There is a need to reduce the burden on educational institutions and test question creators and improve the quality of users' learning experience.However, there is no system that can take into account the user's emotional state while studying and adjust the difficulty of questions individually, making it difficult to maintain learning motivation and provide adaptive education.

[2293] The specific processing by the specific processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means.

[2294] In this invention, the server includes a generation means for automatically generating test questions for specific subjects, an acquisition means for acquiring data based on requested conditions, a call means for calling a generation AI model that generates new test questions from the acquired data, an output means for formatting and outputting the generated test questions and explanations, an emotion analysis means for recognizing the user's emotions and analyzing the data, a difficulty adjustment means for adjusting the difficulty of the questions based on the emotion data, and a feedback means for returning the generated test questions and explanations to the user and generating feedback including the emotion evaluation results. This reduces the burden on the question creator, improves the quality of the user's learning experience, and enables the difficulty of questions to be adjusted appropriately according to the emotional state of each user.

[2295] A "specific subject" refers to a specific field or academic area that serves as the subject matter for generating test questions.

[2296] A "test question" is material containing questions or tasks used to assess a user's knowledge and understanding.

[2297] "Generation means" refers to the processes and techniques used to automatically create test questions, primarily using generative AI models.

[2298] "Acquisition means" refers to the process or method of searching for and acquiring relevant data or information based on specific conditions.

[2299] A "generative AI model" refers to an algorithm or system that uses artificial intelligence to generate new test questions and explanations.

[2300] "Invocation means" refers to the process or method for running a generative AI model using acquired data to obtain the required output.

[2301] "Output means" refers to the process or method for outputting the generated test questions and explanations in a specified format.

[2302] "Emotion analysis means" refers to techniques and processes for detecting and analyzing a user's emotional state.

[2303] "Difficulty adjustment means" refers to a process or method for appropriately adjusting the difficulty of test questions based on emotional data.

[2304] "Feedback means" refers to a process or technology for returning the generated test questions, explanations, and emotional evaluation results to the user and providing feedback information.

[2305] "User interface" refers to the interface for exchanging information between the system and the user, and displays test questions, explanations, etc.

[2306] "Communication means" refers to the communication technology and protocol used to request and respond to test question generation between the server and the terminal.

[2307] "Input means" refers to a device or interface that allows a user to input conditions for generating test questions.

[2308] "Transmission means" refers to the process or method for sending a request to a server based on the conditions entered by the user.

[2309] MODE FOR CARRYING OUT THE INVENTION

[2310] This invention combines a system for automatically generating test questions for specific subjects with an emotion engine that recognizes user emotions. This system aims to reduce the burden on question creators in educational institutions and qualification exams and improve the quality of users' learning experience.

[2311] server

[2312] A server is a combination of hardware and software that performs multiple functions. In particular, it includes the following elements:

[2313] 1. Generation means

[2314] The server uses a generative AI model to automatically generate test questions for specific subjects. The generative AI model uses artificial intelligence algorithms to generate new questions and their explanations.

[2315] 2. Acquisition method

[2316] The server retrieves data from information sources such as a database of past exam questions based on the conditions requested by the user. To do this, it uses SQL queries to search for and retrieve data that matches the request conditions.

[2317] 3. Calling means

[2318] The server preprocesses the acquired data and converts it into a format that can be input to the generative AI model. It also calls the generative AI model to generate new test questions and explanations. Specifically, it passes the necessary data to the generative AI model as prompts.

[2319] 4. Output Method

[2320] The server formats the generated test questions and explanations in an appropriate format (e.g., JSON or HTML), outputs the data, and temporarily stores it before returning it to the user.

[2321] 5. Emotion analysis method

[2322] The server receives the user's emotional data and uses an emotion engine to analyze it, which detects various emotional states such as the user's stress level and concentration level.

[2323] 6. Difficulty adjustment method

[2324] The server adjusts the difficulty of the questions based on the emotional data, for example, setting the difficulty of the questions lower if the user indicates a high stress level.

[2325] 7. Feedback channels

[2326] The server returns the generated test questions and explanations, as well as the emotional evaluation results, to the user, providing feedback information, so that the user can receive appropriate feedback based on their learning progress and emotional state.

[2327] Terminal

[2328] The device is responsible for the interface with the user, accepting input, sending requests to the server, and displaying the results. It also recognizes the user's emotions and sends the data to the server.

[2329] 1. Submit a request

[2330] The terminal uses an HTTP request to transmit the test question creation conditions and emotion data received from the user to the server.

[2331] 2. Receiving a response

[2332] The terminal receives the response data returned from the server, which includes the test questions, explanations, and emotion evaluation results.

[2333] 3. Displaying the results

[2334] The terminal analyzes the received response data and displays it on the user interface. For example, it formats and displays test questions and explanations using HTML templates.

[2335] User

[2336] To use the system to obtain the original test questions, users must:

[2337] 1. Enter the conditions

[2338] The user inputs the conditions for creating test questions (for example, test subjects, question format, difficulty level, number of questions) through the terminal.

[2339] 2. Emotional Feedback

[2340] During learning, the user's emotional data is captured by the emotion engine through their operating status and facial expressions.

[2341] 3. Check the results

[2342] The user can check the test questions, explanations, and emotion evaluation results returned in response from the server by displaying them on the terminal. If necessary, the user can send a re-request to adjust the content and difficulty of the test questions.

[2343] Specific examples

[2344] For example, if a user wants to create five intermediate-level multiple-choice questions in math and get advice on how to reduce stress levels while studying, the process might go like this:

[2345] 1. Users

[2346] The user enters "Mathematics" as the subject, "Multiple Choice" as the question format, "Intermediate" as the difficulty level, and "5" as the number of questions into the device and sends a request. At this time, the emotion engine captures the user's emotions while studying.

[2347] 2. Terminal

[2348] The request data and emotion data are sent to the server.

[2349] 3. Server

[2350] The system analyzes the request data, retrieves appropriate data from a database of past exam questions, and adjusts the difficulty of the questions based on the emotion data. It then uses a generative AI model to generate new exam questions and explanations. The generated questions, explanations, and emotion evaluation results are formatted and sent back to the device.

[2351] 4. Terminal

[2352] Receives the data returned from the server and displays it in the user interface.

[2353] 5. Users

[2354] Review the test questions, explanations, and emotional evaluation results. If necessary, make a request again to adjust the content and difficulty of the test questions.

[2355] Examples of specific prompts to use

[2356] Create five multiple-choice questions for intermediate level math based on the following criteria, taking into account the user's stress level while studying:

[2357] Exam subject: Mathematics

[2358] Question format: 4 choices

[2359] Difficulty: Intermediate

[2360] Number of questions: 5

[2361] keyword

[2362] Generative AI model, prompt sentence

[2363] The flow of the identification process in the second embodiment will be described with reference to FIG.

[2364] Step 1:

[2365] Accepting user input

[2366] server

[2367] The server receives requests from users through an HTTP API endpoint. Specifically, it receives JSON data (e.g., {"subject": "math", "difficulty": "intermediate", "question_count": 5}) in which the user enters conditions such as test subject, question format, difficulty level, and number of questions. It also receives the user's emotion data at the same time.

[2368] Input: Test question creation conditions (JSON format) and emotion data

[2369] Data processing / calculation: Parse JSON data and extract each parameter (exam subject, question format, difficulty level, number of questions)

[2370] Output: Extracted parameters and emotion data

[2371] Step 2:

[2372] Database Reference

[2373] server

[2374] The server references the past exam question database based on the criteria received from the user, and uses an SQL query to retrieve data that matches the request criteria (e.g. SELECT FROM problem_database WHERE subject = 'math' AND difficulty = 'intermediate').

[2375] Input: Extracted parameters (exam subject, difficulty level, etc.)

[2376] Data processing / calculation: Executing SQL queries and filtering retrieved data

[2377] Output: Filtered data

[2378] Step 3:

[2379] Invoking a generative AI model

[2380] server

[2381] The server preprocesses the acquired data and converts it into a format suitable for input to the generative AI model. It adjusts the difficulty of the questions based on the emotional data (e.g., lowering the difficulty level if the stress level is high). It then calls the generative AI model to generate new test questions and explanations. The generated questions include four options and a correct answer.

[2382] Input: filtered data, sentiment data

[2383] Data processing / calculation: Preprocessing data, converting it to an input format for the generative AI model, and invoking the generative AI model using prompts.

[2384] Output: Generated test questions and explanations (e.g., ["Q1: Question text...", "A: Choice 1", "B: Choice 2", "C: Choice 3", "D: Choice 4", "Answer: B"])

[2385] Step 4:

[2386] Configuring the Output

[2387] server

[2388] The server formats the generated test questions and explanations into an appropriate format (e.g., JSON format), and also generates feedback information including the emotion evaluation results, which is temporarily stored in a database for sending back to the user.

[2389] Input: Generated test questions and explanations, emotional evaluation results

[2390] Data processing / calculation: formatting test questions and explanations, generating feedback information

[2391] Output: Formatted exam questions, explanations and feedback information (JSON format)

[2392] Step 5:

[2393] Send request

[2394] Terminal

[2395] The terminal receives the user's input and sends the request data and emotion data to the server using an HTTP request.

[2396] Input: User-entered test question creation conditions and emotion data

[2397] Data processing / calculation: Converting input data to JSON format, creating and sending HTTP requests

[2398] Output: Request data sent to the server (JSON format)

[2399] Step 6:

[2400] Response reception

[2401] Terminal

[2402] The device receives the response data from the server. After receiving the HTTP status code 200 (OK), it acquires the test questions, explanations, and emotion evaluation results contained in the response body.

[2403] Input: Response data from the server

[2404] Data processing / calculation: Response data analysis (JSON parsing)

[2405] Output: Analyzed test questions, explanations, and emotion evaluation results

[2406] Step 7:

[2407] Displaying the results

[2408] Terminal

[2409] The device analyzes the received response data and displays it on the user interface. It uses HTML templates to format and display test questions, explanations, and emotion evaluation results.

[2410] Input: Analyzed test questions, explanations, and emotion evaluation results

[2411] Data processing / calculation: Rendering using HTML templates

[2412] Output: Formatted test questions, explanations, and emotional evaluation results (displayed on screen)

[2413] Step 8:

[2414] Entering conditions

[2415] User

[2416] The user inputs the conditions for creating test questions (test subjects, question format, difficulty level, number of questions) through the terminal. After the conditions are entered, the user clicks a button to send the request.

[2417] Input: Conditions for creating test questions

[2418] Data processing / calculation: None

[2419] Output: What you type appears on the terminal

[2420] Step 9:

[2421] Emotional feedback

[2422] User

[2423] The emotion engine captures the operation status and facial expressions during learning, and the emotion data is sent to the server.

[2424] Input: User operation status, facial expressions, etc.

[2425] Data processing / calculation: Analysis by emotion engine, data capture

[2426] Output: Captured emotion data

[2427] Step 10:

[2428] Checking the results

[2429] User

[2430] The test questions, explanations, and emotional evaluation results returned from the server are displayed on the device. If necessary, a request is made again to adjust the content and difficulty of the test questions.

[2431] Input: Test questions, explanations, and emotional evaluation results displayed on the terminal

[2432] Data processing / calculation: None

[2433] Output: User confirmation and re-request if necessary

[2434] (Application example 2)

[2435] Next, a description will be given of Application Example 2. In the following description, the data processing device 12 will be referred to as a "server" and the robot 414 will be referred to as a "terminal."

[2436] Conventional test question generation systems provide questions uniformly without considering the user's emotions, which reduces the user's learning efficiency. In addition, since feedback based on emotional data is not provided, users tend to feel stressed and may lose motivation to study.

[2437] The specific processing by the specific processing unit 290 of the data processing device 12 in the application example 2 is realized by the following means.

[2438] In this invention, the server includes a generation means for automatically generating test questions for specific subjects, an acquisition means for acquiring data based on requested conditions, a call means for calling a generation AI model for generating new test questions from the acquired data, an output means for formatting and outputting the generated test questions and explanations, an emotion recognition means for recognizing a user's emotions, a difficulty adjustment means for adjusting the difficulty of the questions based on the emotion data, and a feedback generation means for generating feedback information based on the emotion data. This makes it possible to provide test questions of an appropriate difficulty level that takes the user's emotions into consideration, improving learning efficiency and maintaining the user's motivation to learn.

[2439] The "means for automatically generating test questions for a specific subject" is a function for automatically generating test questions based on the subjects specified by the user.

[2440] The "acquisition means for acquiring data based on requested conditions" is a function for acquiring necessary data from a database according to conditions input by a user.

[2441] The "calling means for calling the generative AI model" is a function that executes the AI ​​model to generate new test questions using the acquired data.

[2442] The "output means for formatting and outputting the generated test questions and explanations" is a function for formatting the newly generated test questions and their explanations in an appropriate format and outputting them.

[2443] The "emotion recognition means for recognizing the user's emotions" is a function for analyzing the user's facial expressions, actions, etc. to grasp the user's emotional state.

[2444] The "difficulty level adjusting means for adjusting the difficulty level of questions based on emotional data" is a function for providing test questions of a difficulty level appropriate for the user based on the recognized emotional data.

[2445] The "feedback generating means for generating feedback information based on emotional data" is a function for providing appropriate advice and feedback based on the emotional state of the user.

[2446] The "display means for displaying the generated test questions and explanations on the user interface" is a function for displaying the generated test questions and explanations on the screen so that the user can check them.

[2447] "Communication means for requesting and responding to test question generation between the server and the terminal" is a function for sending and receiving requests for test question generation and the generated results between the server and the terminal.

[2448] The "input means for the user to input conditions for generating test questions" is a function for the user to input conditions such as the subject, difficulty level, and format of the test questions that he or she wants to generate.

[2449] The "transmission means for transmitting a request to the server based on the input conditions" is a function for transmitting the conditions input by the user to the server and requesting the generation of test questions.

[2450] The "emotion data transmission means for acquiring emotion data in real time and transmitting it to the server" is a function for collecting the user's emotions in real time and transmitting that data to the server.

[2451] This invention combines a system that automatically generates test questions for specific subjects with an emotion engine that recognizes the user's emotions. This system includes a generation means, an acquisition means, a calling means, an output means, an emotion recognition means, a difficulty level adjustment means, and a feedback generation means. It is particularly characterized by using emotion data to adjust the difficulty of questions and provide appropriate feedback.

[2452] System Overview

[2453] server

[2454] The server has the following features:

[2455] Generation method: Automatically generate test questions for specified subjects.

[2456] Acquisition method: Retrieves the necessary data from the database based on a user request.

[2457] Calling method: Input the acquired data into the generative AI model to generate new test questions.

[2458] Output method: Format the generated test questions and explanations and output them in an appropriate format.

[2459] Emotion recognition means: Recognizes the user's emotions using cameras and sensors and acquires them as data.

[2460] Difficulty adjustment method: Adjust the difficulty of the questions based on emotional data.

[2461] Feedback generation means: Generates feedback information based on emotion data and provides it to the user.

[2462] The hardware used includes servers, databases, and computing resources for running the AI ​​model, while the software includes an emotion recognition engine (EmotionEngine) and a generative AI model (QuestionGenerator).

[2463] Terminal

[2464] The terminal is responsible for:

[2465] Input means: The user inputs the conditions for generating test questions.

[2466] Transmission method: The input conditions are sent to the server as a request, along with the emotion data.

[2467] Display: Displays the generated exam questions, explanations, and feedback.

[2468] For example, smartphones, smart glasses, head-mounted displays, etc. The emotion recognition engine works using the cameras and sensors built into these devices.

[2469] User

[2470] The user does the following:

[2471] Input of conditions: The user inputs the conditions for creating test questions (subject, difficulty level, question format, number) through the terminal.

[2472] Emotion recognition: The learning process is captured by the emotion recognition engine and the emotional data is sent to the server.

[2473] Review feedback: Review the generated exam questions, explanations, and sentiment-based advice.

[2474] Specific examples

[2475] For example, if a user wants to generate five multiple-choice questions for intermediate level math, they can input the following prompts into the generative AI model.

[2476] "Generate five intermediate level multiple choice questions in mathematics. The user's emotion is 'feeling somewhat stressed.' Adjust the difficulty of the questions taking this emotion into account."

[2477] This allows users to receive test questions of an appropriate difficulty level according to their own feelings, which is expected to improve learning efficiency and motivation.

[2478] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[2479] Step 1:

[2480] The user inputs the conditions for generating test questions. The user inputs the test subjects, difficulty level, question format, and number of questions using a smartphone or other device. This input data becomes the basis for creating a request to generate test questions.

[2481] Step 2:

[2482] The emotion recognition engine recognizes the user's emotions in real time. It uses the device's built-in camera and sensors to capture the user's facial expressions and behaviors and extract emotional data, which is then used in subsequent processing steps.

[2483] Step 3:

[2484] The input conditions and emotion data are sent to the server. A test question generation request and emotion data are sent to the server using the terminal's transmission means. The input data in this process are the user's condition information acquired in (Step 1) and the emotion information acquired in (Step 2).

[2485] Step 4:

[2486] The server analyzes the request data and retrieves the necessary data. The server analyzes the request data and emotion data and retrieves related data such as past questions from the database based on that. Appropriate questions are extracted based on the request data as input.

[2487] Step 5:

[2488] The server calls the generative AI model to generate new test questions. The acquired data is preprocessed and input to the generative AI model along with the prompt. The model generates new test questions and explanations. The input data is the preprocessed data and the prompt, and the generated output is new test questions and their explanations.

[2489] Step 6:

[2490] Format the generated test questions and explanations and return them to the user. Format the generated data into an appropriate format and send it to the device. Input data is the output data from the generative AI model, which is converted into a format that can be sent to the device.

[2491] Step 7:

[2492] The device displays the test questions, explanations, and emotional feedback on the user interface. The received data is analyzed and displayed in a format that the user can understand. The input data is the response data from the server, and the displayed content is the output.

[2493] Through the above procedure, the user can obtain appropriate test questions and feedback according to his / her own emotions. These processing steps detail the processing flow of a specific system and program based on the present invention.

[2494] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the control target 443 to output the result of the specific processing. The micropho...

Claims

1. A generation means for automatically generating test questions for a specific subject; an acquisition means for acquiring data based on requested conditions; A calling means for calling a generation AI model that generates new test questions from the acquired data; The system includes an output means for formatting and outputting the generated test questions and explanations.

2. 10. The system of claim 1, The system further includes a display means for displaying the generated test questions and explanations on a user interface.

3. A communication means for requesting and responding to test question generation between the server and the terminal; an input means for a user to input conditions for generating test questions; 2. The system according to claim 1, further comprising a transmitting means for transmitting a request to the server based on the input conditions.

Images