System

Abstract

A system is provided.SOLUTION: A system comprising: means for receiving and storing educational material provided by educational institutions in a database; means for pre-processing the educational material using natural language processing techniques to extract important keywords and concepts; means for organizing the educational material based on a range of questions specified by a teacher; means for generating test questions from the range of questions using a generative AI model; means for formatting the test questions in both paper and web formats; and means for automatically scoring student answers.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] Currently, teachers at schools and cram schools are overwhelmed with many tasks, including lesson preparation, student guidance, test creation and grading, and club activities. Creating and grading test questions in particular requires a significant amount of time and effort, placing a heavy burden on teachers. While efficient analysis and feedback of test results is crucial for improving the quality of education, there are limitations to doing this manually. A method is needed to improve this situation, reduce teachers' workload, and improve the quality of education. [Means for solving the problem]

[0005] The present invention solves the above-mentioned problems by providing a system that includes: means for receiving teaching material data provided by educational institutions and storing it in a database; means for preprocessing the teaching material data using natural language processing technology to extract important keywords and concepts; means for organizing the teaching material data based on the test scope specified by the teacher; means for generating test questions from the test scope using a generative AI model; means for formatting the test questions in both paper and web formats; means for automatically scoring student answers; and means for aggregating and analyzing performance data and providing feedback to teachers. The system of the present invention automates the entire process, from automatically generating test questions to scoring, and analyzing and providing feedback on the results, thereby significantly reducing the burden on teachers and improving the quality of education.

[0006] "Teaching material data" refers to digital data of textbooks and reference books provided by educational institutions.

[0007] A "database" is a system for efficiently managing, searching, and retrieving stored digital data.

[0008] "Natural language processing technology" is a technology that uses computers to understand, interpret, and generate human language.

[0009] A "generative AI model" is an algorithm or program that uses artificial intelligence (AI) to generate new information.

[0010] The "test area" refers to a specific chapter or page of a textbook or reference book that a teacher specifies when generating test questions.

[0011] "Test questions" are test questions in the form of multiple choice or essay that are generated based on the specified question scope.

[0012] "Paper media" refers to documents and exam papers printed on paper.

[0013] A "web format" is a digital format that is accessible through a browser or application on the Internet.

[0014] "Automated scoring" is the process of using a computer program to evaluate test answers and calculate a score.

[0015] "Performance Data" refers to information about the scores obtained as a result of a student's answers to a test.

[0016] "Feedback" refers to providing information that will be useful for the next lesson or test based on analyzed grade data. [Brief explanation of the drawings]

[0017] [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

[0018] 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.

[0019] First, the terms used in the following description will be explained.

[0020] 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).

[0021] 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.

[0022] 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.

[0023] 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.

[0024] 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."

[0025] [First embodiment]

[0026] FIG. 1 shows an example of the configuration of a data processing system 10 according to the first embodiment.

[0027] 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.

[0028] 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).

[0029] 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.

[0030] 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.

[0031] 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.

[0032] 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.

[0033] FIG. 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0034] 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.

[0035] 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.

[0036] 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.

[0037] 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."

[0038] The following describes an embodiment of the present invention. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs scoring and result analysis. The detailed operational process required for this will be explained step by step.

[0039] Importing teaching material data

[0040] server

[0041] The system receives and stores the educational material data (PDF, Docx, etc.) provided by educational institutions. It then converts the data into text format using natural language processing technology. This allows it to extract important keywords and concepts and store them in a database.

[0042] Specifying the scope of questions

[0043] User (Teacher)

[0044] The teacher uses a terminal to input the specific chapters and pages to be tested, and this information is sent from the terminal to the server.

[0045] Terminal

[0046] The specified range is sent to the server as data.

[0047] server

[0048] The received range information is analyzed, and the teaching material data within the specified range is extracted from the database.

[0049] Generate test questions

[0050] server

[0051] Based on the specified range of teaching material data, the system selects an appropriate question format, such as fill-in-the-blank, essay, or calculation questions. The generative AI model analyzes trends from past entrance exams and mock exams to create test questions, which are then formatted in both paper and web formats.

[0052] Conducting the test

[0053] User (student)

[0054] Students access the test via a web browser and enter their answers. If the test is on paper, they can scan the answers and upload them to the server.

[0055] Scoring

[0056] server

[0057] For web-based tests, automatic scoring is performed in real time as answers are submitted. For essay-style questions, natural language processing technology is used for scoring. For paper-based tests, scanned data is converted to text using OCR technology and then automatically scored.

[0058] Analysis of results and feedback

[0059] server

[0060] The system compiles each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. It identifies the difficulty of each question and common error patterns and generates an analysis report. Teachers can use this feedback to identify points that should be emphasized in the next lesson.

[0061] Terminal

[0062] Feedback reports are sent to teachers' devices to help them plan their next lessons.

[0063] In this way, this system automates everything from creating test questions to grading, analyzing the results, and providing feedback, greatly improving the efficiency of educational activities. Furthermore, the questions generated reflect past exam trends, making it possible to improve the quality of education.

[0064] The processing flow will be explained below.

[0065] Step 1:

[0066] server

[0067] Receives educational material data provided by educational institutions, including formats such as PDF and Docx. The received data is stored in a database.

[0068] Step 2:

[0069] server

[0070] Natural language processing technology is applied to the saved teaching material data to convert it into text format. After conversion, important keywords and concepts are extracted and this information is stored in a database as structured data.

[0071] Step 3:

[0072] User (Teacher)

[0073] The teacher uses a terminal to input specific chapters and page numbers to specify the range of questions to be asked, and the information is sent from the terminal to the server.

[0074] Step 4:

[0075] server

[0076] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[0077] Step 5:

[0078] server

[0079] The generative AI model generates test questions based on the scope of the questions, referencing data from past entrance exams and mock exams to create questions that reflect important points.

[0080] Step 6:

[0081] server

[0082] The generated test questions are formatted to be compatible with both paper and web formats. Print data is created for paper media, and online test data is created for web formats.

[0083] Step 7:

[0084] User (student)

[0085] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[0086] Step 8:

[0087] server

[0088] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[0089] Step 9:

[0090] server

[0091] Compile the performance data of each student, calculate the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, etc., and create an analysis report.

[0092] Step 10:

[0093] server

[0094] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test.

[0095] Step 11:

[0096] Terminal

[0097] Feedback reports are sent to teachers' devices and made available for viewing, allowing teachers to use them to plan their lessons.

[0098] This allows the system to efficiently carry out a series of steps, from automatically generating test questions, to grading, analyzing the results, and providing feedback, thereby improving the quality of education while significantly reducing the burden on teachers.

[0099] Example 1

[0100] 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."

[0101] Conventional educational systems require a great deal of manual work, from importing teaching material data to creating test questions, grading, and analyzing the results, placing a heavy burden on teachers. It was also difficult to generate test questions that reflected past exam trends, or to provide real-time grading and feedback. This raised concerns that the quality and efficiency of education would decline.

[0102] 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.

[0103] In this invention, the server includes means for receiving teaching material data provided by educational institutions and storing it in a database, means for preprocessing the teaching material data using natural language processing technology to extract important keywords and concepts, means for organizing the teaching material data based on the test scope specified by the teacher, means for generating test questions from the test scope using a generative AI model, means for formatting the test questions in formats such as LaTeX or HTML depending on the question format, means for providing the generated test questions in both paper and web formats, means for automatically scoring the results of students' answers, and means for aggregating and analyzing performance data and providing feedback to teachers. This automates the entire process from importing teaching material data to generating test questions, scoring, and analyzing and providing feedback on the results, reducing the burden on teachers and improving the quality and efficiency of education.

[0104] "Educational institutions" refers to organizations that provide education and training, such as schools, universities, and cram schools.

[0105] "Teaching material data" refers to materials for educational purposes provided by educational institutions, including file formats such as PDF and Docx.

[0106] "Database" refers to a system designed to efficiently manage and retrieve data within an organization.

[0107] "Natural language processing technology" refers to the field of technology in which computers understand, interpret, and generate human language.

[0108] "Keywords" refer to words and phrases that are considered particularly important within a document and are used for information retrieval and data classification.

[0109] "Test area" refers to the area set for test questions, such as specific chapters or page numbers designated by the teacher.

[0110] A "generative AI model" refers to an algorithm or model that uses artificial intelligence technology to generate data or analyze information.

[0111] "Test questions" refer to questions or tasks designed to assess a student's understanding or knowledge.

[0112] "Formatting" refers to the process of converting documents or data into an appropriate format to make them easier to read.

[0113] "Web format" refers to a format that is provided over the Internet and can be viewed and manipulated in a browser.

[0114] "Paper media" refers to information or data printed on paper.

[0115] "Automatic scoring" refers to the process of mechanically evaluating and scoring students' answers using a special program.

[0116] "Academic performance data" refers to data that records student test results and grades in numerical or written form.

[0117] "Feedback" refers to assessments and information provided by systems and teachers to help improve students and lessons.

[0118] The following describes an embodiment of the present invention. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. The detailed operational process required for this purpose is specifically shown.

[0119] Importing teaching material data

[0120] server

[0121] The server receives learning material data (PDF, Docx, etc.) provided by educational institutions via HTTP requests and stores it in local storage. It then converts the learning material data into text format using Google® OCR API or Amazon Textract. From the converted text, natural language processing technology (e.g., spaCy or NLTK) is used to extract important keywords and concepts, which are then stored in a MySQL® database. As a specific example, if a school provides a PDF file as an English learning material, the server receives the PDF, converts it into text format using Google's OCR API, and stores the extracted words and phrases in a database.

[0122] Example prompt sentence:

[0123] The system should receive English learning materials in PDF format provided by educational institutions, convert them into text using Google OCR API, extract important words and phrases from the text, and store them in a database.

[0124] Specifying the scope of questions

[0125] User (Teacher)

[0126] Teachers log in to the system using their device's browser and enter the question scope by specifying a specific chapter or page number from the interface.

[0127] Terminal

[0128] The data for the question area specified by the teacher is obtained from the input form and sent to the server as an API request.

[0129] server

[0130] Analyze the received API request, query the database to retrieve the learning material data within the specified range, and store the retrieved learning material data in a temporary storage area.

[0131] As a specific example, when a teacher specifies Chapter 5 of English as the subject of questions, the terminal sends that information to the server, and the server retrieves data corresponding to Chapter 5 from the database.

[0132] Example prompt sentence:

[0133] If a teacher specifies Chapter 5 of English as the subject of the test, that information is sent to the server, and the server retrieves the relevant data from the database.

[0134] Generate test questions

[0135] server

[0136] Test questions are generated using a generative AI model (e.g., GPT-3 (registered trademark)) based on the specified range of teaching material data. Questions can be fill-in-the-blank, essay-style, calculation-based, or other formats, and are generated in the appropriate format (LaTeX or HTML). The generated questions are output in both web and paper formats.

[0137] As a concrete example, the server uses GPT-3 to generate fill-in-the-blank questions on the theme of Chapter 5 of English and formats them in HTML format.

[0138] Example prompt sentence:

[0139] Use GPT-3 to generate fill-in-the-blank questions on the theme of English Chapter 5. Format the questions in HTML format so that they can be output in both web and paper formats.

[0140] Conducting the test

[0141] User (student)

[0142] Students access the test via a web browser on their device and enter their answers. For paper-based tests, students fill out the answer sheets, scan them, and upload them to the server from their device.

[0143] For example, students can take an English test online using a browser on their device and enter their answers, or if the answers are paper, they can scan them and upload them to the server.

[0144] Example prompt sentence:

[0145] Students can answer the test online using their device's browser. If the test is paper, they can write their answers, scan them, and upload them to the server from their device.

[0146] Scoring

[0147] server

[0148] Web-based answers are automatically scored as soon as they are submitted. Multiple-choice questions are scored by comparing them with preset correct answer data. Essay questions are evaluated using natural language processing technology (e.g., BERT). Paper-based tests are automatically scored after uploaded scanned data is converted to text using OCR technology (e.g., Tesseract).

[0149] For example, the server analyzes the answer data received and scores multiple-choice questions by comparing them with the correct answer data. For essay questions, natural language analysis is performed using the BERT model to calculate a score. Scanned data is converted into text using an OCR engine, and the results are scored in the same way.

[0150] Example prompt sentence:

[0151] Please use BERT to grade the essay questions that students answer online. For paper-based documents, please scan the data and convert it to text using OCR, and grade it in the same way.

[0152] Analysis of results and feedback

[0153] server

[0154] After all answers have been graded, the data is aggregated and analyzed for correct answer rates and incorrect answer trends. The difficulty level of each question and common error patterns are identified, and an analysis report is generated. This report is generated in JSON or PDF format and sent to the teacher's device.

[0155] As a specific example, the server analyzes the grade data, identifies common trends in incorrect answers, generates a PDF report, and sends the report to the teacher's device.

[0156] Example prompt sentence:

[0157] Please compile the grade data and analyze the percentage of correct answers and trends in incorrect answers. After the analysis, the server will identify common incorrect answer patterns, generate a report, and send it to the teacher's device.

[0158] This system automates the entire process, from importing teaching material data to generating test questions, grading answers, and analyzing and providing feedback on the results, thereby reducing the burden on teachers and improving the quality and efficiency of education.

[0159] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0160] Step 1:

[0161] Importing teaching material data

[0162] The user (educational institution) uploads learning material data (PDF, Docx, etc.) to the server. The input data is in PDF or Docx file format. The server receives the learning material data provided by the educational institution as an HTTP request and stores it in local storage. The server then converts the learning material data into text format using Google's OCR API or Amazon Textract. The converted text is analyzed using natural language processing technology (e.g., spaCy or NLTK) to extract important keywords and concepts. This extracted data is then stored in a MySQL database.

[0163] Specific behavior:

[0164] Teaching material data provided by educational institutions is uploaded.

[0165] The teaching material data received by the server is saved in the "Received" folder in the local storage.

[0166] Calls Google's OCR API to convert PDF and Docx to text.

[0167] The converted text is analyzed using a natural language processing library to extract keywords and concepts.

[0168] The extracted data is stored in the "Keywords" table in the MySQL database.

[0169] Input: Teaching material data (PDF, Docx, etc.)

[0170] Output: Keywords and concepts stored in the database

[0171] Step 2:

[0172] Specifying the scope of questions

[0173] The user (teacher) logs into the system using the browser on their device and selects a specific chapter or page number from the interface. The input data is the subject matter of the test specified by the teacher. The device sends the specified subject matter information to the server as an API request. The server analyzes the received API request and queries the database to retrieve the teaching material data that falls within the specified range. The retrieved teaching material data is stored in a temporary storage area.

[0174] Specific behavior:

[0175] The teacher accesses the "Specify Question Area" page in a browser and specifies the chapter and page number.

[0176] The device converts the specified question scope information into a JSON format API request and sends it to the server.

[0177] The server runs an SQL query against the database to retrieve the data corresponding to the specified chapter or page.

[0178] The acquired data is stored in a temporary storage area.

[0179] Input: Question range information specified by the teacher

[0180] Output: Teaching material data stored in the temporary storage area

[0181] Step 3:

[0182] Generate test questions

[0183] The server generates test questions using a generative AI model (e.g., GPT-3) based on the specified range of teaching material data. The input data is the specified range of teaching material data. The generative AI model receives this and generates test questions in an appropriate format, such as fill-in-the-blank questions, essay questions, or calculation questions. The generated test questions are formatted in formats such as LaTeX or HTML and output in each test format (web format and paper format).

[0184] Specific behavior:

[0185] The server acquires the range-specified teaching material data from the temporary storage area.

[0186] The GPT-3 model is called and test questions are generated using the teaching material data as input.

[0187] The generated questions are formatted into LaTeX or HTML depending on the question format.

[0188] Save formatted test questions in web and paper formats.

[0189] Input: Range-specified teaching material data

[0190] Output: Formatted test questions (LaTeX, HTML)

[0191] Step 4:

[0192] Conducting the test

[0193] Users (students) access the web-based test using a browser on their device and enter their answers. The input data is the student's answers. In the case of paper-based tests, students fill out the answer sheets, scan them, and upload them to the server from their device.

[0194] Specific behavior:

[0195] Students access the test page via the HTTPS protocol and enter their answers in their browser.

[0196] The entered answer data is sent to the server.

[0197] If the answer sheet is paper, scan it and convert it to JPEG or PDF format.

[0198] The converted file is sent to the server via an upload form.

[0199] Input: Student answers (web or paper)

[0200] Output: Answer data sent to the server

[0201] Step 5:

[0202] Scoring

[0203] The server automatically grades the web-based answers as they are sent. The input data is the student's answer data. Multiple-choice questions are compared with preset correct answer data and automatically graded. Essay questions are evaluated using natural language processing technology such as BERT. For paper-based tests, uploaded scanned data is converted to text using OCR technology (e.g., Tesseract) and then graded automatically.

[0204] Specific behavior:

[0205] The server analyzes the received answer data and scores the multiple-choice questions by comparing them with the correct answer data.

[0206] For essay questions, the BERT model is used to perform natural language analysis and calculate a score.

[0207] The scanned data is run through an OCR engine, converted into text, and scored.

[0208] Input: Student answer data (web format, paper scanned data)

[0209] Output:Scoring results

[0210] Step 6:

[0211] Analysis of results and feedback

[0212] After all answers have been graded, the server aggregates the grade data and analyzes the percentage of correct answers and trends in incorrect answers. The input data is the students' grade data. The server identifies the difficulty level and incorrect answer patterns for each question, generates an analysis report, and saves it in JSON or PDF format. The generated feedback report is sent to the teacher's device.

[0213] Specific behavior:

[0214] The server retrieves the performance data and performs data analysis using Python's pandas library.

[0215] Analyze the difficulty of the questions and common error patterns and generate a report.

[0216] The report is saved in JSON or PDF format and sent to the teacher's device using the SMTP protocol.

[0217] Input: Student performance data

[0218] Output: Analysis report (JSON, PDF format)

[0219] (Application example 1)

[0220] 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."

[0221] In today's educational environment, teachers' workloads are increasing, often making it difficult to provide effective teaching. Tasks such as determining the scope of study, creating test questions, grading answers, and providing feedback on results require time and effort. Therefore, there is an urgent need to automate these tasks and reduce the burden on teachers. Furthermore, with the spread of online education, there is also a need to create an environment where students can use smartphones to study efficiently. Therefore, a system is needed to significantly improve the efficiency of all educational activities and improve the quality of education.

[0222] 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.

[0223] In this invention, the server includes means for receiving teaching material data provided by educational institutions and storing it in a database, means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, means for organizing the teaching material data based on the test scope specified by the teacher, means for generating test questions from the test scope using a generative AI model, means for formatting the test questions in both paper and web format, means for automatically scoring the results of students' answers, means for aggregating and analyzing performance data and providing feedback to teachers, means for specifying the test scope using a smartphone app and sending information to the server, and means for students to access web-based tests and submit answers using their smartphones. This automates all educational activities, significantly reducing the burden on teachers and enabling efficient and effective education.

[0224] An "educational institution" is an organization such as a school or university that provides a place for learning and research.

[0225] "Teaching material data" refers to data on learning materials provided by educational institutions, including electronic files in PDF and Docx formats.

[0226] A "database" is a system for efficiently storing and managing teaching material data, grade data, etc.

[0227] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and is used for preprocessing text data and extracting important keywords.

[0228] "Important keywords and concepts" are words and ideas in the teaching material data that are considered to be particularly important for learning.

[0229] A "teacher" is a professional who educates students, and is responsible for specifying the scope of test questions and guiding educational activities.

[0230] "Test Coverage" refers to the specific chapters or areas that are covered when creating test questions.

[0231] A "generative AI model" is a program that automatically generates test questions using artificial intelligence technology.

[0232] "Test questions" are questions used to evaluate students' learning outcomes, and include fill-in-the-blank questions and essay-style questions.

[0233] "Paper media" refers to printed paper forms, including forms on which handwritten answers are written.

[0234] "Web format" refers to an online format delivered over the Internet and accessible through a browser or app.

[0235] "Automatic marking" is the process by which a computer automatically evaluates and scores student answers.

[0236] "Performance data" refers to numerical and evaluation data that represent test results.

[0237] "Feedback" refers to evaluations and guidance comments provided based on performance data.

[0238] A "smartphone" is a type of mobile phone that can connect to the Internet and run applications.

[0239] A "server" is a computer system that provides information over a network.

[0240] "OCR technology" is a technology that extracts text information from an image and is used to convert handwritten answers written on paper into text data.

[0241] The following describes in detail an embodiment of the present invention. This system uses a smartphone app and a server based on educational material data provided by an educational institution to efficiently generate test questions, grade answers, and provide feedback on performance data.

[0242] Importing teaching material data

[0243] server

[0244] Teaching material data (PDF, Docx, etc.) provided by educational institutions is received and stored in a database. During this process, an API is set up using Python and Django to receive the teaching material data. After the teaching material data is saved, it is converted into text format using natural language processing technology, and important keywords and concepts are extracted and stored in the database. Specifically, PDF data is converted into text using PDFMiner, and Docx data is converted into text using python-docx.

[0245] Specifying the scope of questions

[0246] User (Teacher)

[0247] Teachers use a smartphone app to input specific chapters and pages to specify the test area. This information is sent from the app to a server. The smartphone app is developed using a cross-platform framework such as React Native or Flutter (registered trademark).

[0248] server

[0249] The received range information is analyzed, and the teaching material data within the specified range is extracted from the database.

[0250] Generate test questions

[0251] server

[0252] Based on the specified range, the generative AI model (using TENSORFLOW®) analyzes past test data and question trends to automatically generate test questions. During this process, the generative AI model is input with the following prompt sentence:

[0253] Example prompt sentence:

[0254] "Scope: Questions will be based on Chapter 2, Overview of Ancient Greek Philosophy. Multiple choice, fill-in-the-blank, and essay questions will be created. The exam format has been optimized taking into account past exam questions."

[0255] The generated test questions are formatted and saved in both paper and web formats.

[0256] Conducting the test

[0257] User (student)

[0258] Students access the web-based test via a smartphone app and enter their answers, which are then sent to a server in real time.

[0259] Scoring

[0260] server

[0261] For web-based tests, automatic scoring is performed in real time as the answers are submitted. For essay-style questions, scoring is performed using natural language processing technology. For paper-based tests, the scanned data is converted to text using OCR technology (using OpenCV) and then automatically scored.

[0262] Analysis of results and feedback

[0263] server

[0264] The system compiles each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. It identifies the difficulty of each question and common error patterns and generates an analysis report. The feedback report is sent to the teacher's device and used as a reference for planning the next lesson.

[0265] This system automates all teaching activities, significantly reducing the burden on teachers, and generates questions that reflect past exam trends, improving the quality of teaching.

[0266] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0267] Step 1:

[0268] Importing teaching material data

[0269] The server receives learning material data (PDF, Docx, etc.) provided by educational institutions and stores it in a database. Specifically, an API endpoint is set up using the Django framework to receive the learning material data. The received data is converted to text format using PDFMiner for PDF and python-docx for Docx. The input here is a PDF or Docx file, and the output is text data.

[0270] Step 2:

[0271] Preprocessing of teaching material data

[0272] The server preprocesses the converted text data using natural language processing techniques to extract important keywords and concepts. This is done using natural language processing libraries such as NLTK. Specific operations include tokenizing the text data, morphological analysis, removing stop words, and extracting feature words. The input is the text data obtained in step 1, and the output is a list of keywords and concepts.

[0273] Step 3:

[0274] Specifying the scope of questions

[0275] The user (teacher) uses a smartphone app to input the test range, such as a specific chapter or page. The specified range information is sent from the device to the server. The input is the test range information specified by the teacher, and the output is the range information stored on the server. The smartphone app is developed using React Native and Flutter.

[0276] Step 4:

[0277] Analysis of the test scope

[0278] The server analyzes the received range information and retrieves the corresponding teaching material data from the database. Based on the specific chapters and pages included in the range information, it searches for and extracts related text data. The input is the range information obtained in step 3, and the output is text data belonging to the specific range.

[0279] Step 5:

[0280] Generate test questions

[0281] The server uses a generative AI model (using TensorFlow) to automatically generate appropriate test questions (multiple choice, fill-in-the-blank, essay, etc.) from the test range. Specifically, the questions are automatically created using a specific algorithm, taking into account past test data and question trends. During this process, the following prompt sentence is input into the generative AI model.

[0282] Example prompt: "Scope: Questions will be drawn from Chapter 2, Overview of Ancient Greek Philosophy. Multiple choice, fill-in-the-blank, and essay questions will be created. The exam format will be optimized based on past exam questions."

[0283] The input is a specific range of text data and a prompt, and the output is automatically generated test questions.

[0284] Step 6:

[0285] Test question formatting

[0286] The server formats the generated test questions in both paper and web formats. In the case of web format, it converts them into a format that can be displayed in a browser using HTML and CSS, and in the case of paper format, it converts them into a format for printing. The input is the test questions created in step 5, and the output is the formatted test question data.

[0287] Step 7:

[0288] Conducting the test

[0289] Users (students) access web-based tests through a smartphone app, enter their answers, and submit them. The smartphone app provides a browser-based display and input interface. During this process, answers are sent to the server in real time. The input is the student's answer, which is the data sent to the server in real time. The output is the submitted answer data.

[0290] Step 8:

[0291] Automatic scoring

[0292] The server automatically grades the answer data submitted by students in real time. Multiple choice and fill-in-the-blank questions are automatically evaluated by comparing them with preset answers, and essay questions are graded using natural language processing technology. The input is the submitted answer data, and the output is the graded results.

[0293] Step 9:

[0294] Result collection and feedback

[0295] The server aggregates each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. Data analysis tools (such as Pandas and NumPy) are used for the analysis. A feedback report is generated based on the performance data and sent to the teacher's device. The input is the automatically graded result data, and the output is the feedback report.

[0296] 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.

[0297] An embodiment of the present invention will be described below. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. The detailed operational process required for this will be explained step by step. In addition, by combining it with an emotion engine that recognizes the user's emotions, a more advanced feedback function can be achieved.

[0298] Importing teaching material data

[0299] server

[0300] The system receives and stores the educational material data (PDF, Docx, etc.) provided by educational institutions. It then converts the data into text format using natural language processing technology. This allows it to extract important keywords and concepts and store them in a database.

[0301] Specifying the scope of questions

[0302] User (Teacher)

[0303] The teacher uses a terminal to input specific chapter and page numbers to specify the test area, and this information is sent from the terminal to the server.

[0304] Terminal

[0305] The specified range is sent to the server as data.

[0306] server

[0307] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[0308] Generate test questions

[0309] server

[0310] Based on the exam scope, a generative AI model generates test questions. It references data from past entrance exams and mock exams to create questions that reflect key points. It then formats the questions in both paper and web formats.

[0311] Conducting the test

[0312] User (student)

[0313] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[0314] Scoring

[0315] server

[0316] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[0317] Emotion recognition

[0318] server

[0319] The emotion engine analyzes language patterns and facial expression data (if required) in students' responses sent from their devices to recognize their emotional state. This information is stored in a database along with their grades.

[0320] Analysis of results and feedback

[0321] server

[0322] The system compiles each student's academic and emotional data, calculating the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, and emotional trends, and creates an analysis report.

[0323] server

[0324] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test administration. In addition, emotional data is taken into account, and feedback is reflected in which questions students are stressed or excited about.

[0325] Providing feedback

[0326] Terminal

[0327] Feedback reports are sent to teachers' devices and made available for viewing. Teachers can use this information to plan lessons and identify which issues specific students responded to emotionally, allowing them to provide individualized support.

[0328] Specific examples

[0329] For example, when administering a test on "Calculus" in Mathematics II, the system works as follows: When a teacher specifies the "Calculus chapter" as the test area, the generative AI model creates multiple-choice and essay questions from that area. Students answer the online test, and the emotion engine identifies areas of stress from their facial expressions and language patterns. Based on the obtained performance and emotion data, the server generates a detailed analysis report, which the teacher can use to understand points to emphasize in the next lesson and individual student needs.

[0330] In this way, the system of the present invention comprehensively covers everything from automatic generation of test questions to scoring, result analysis, and feedback incorporating emotion recognition, thereby improving the quality of education.

[0331] The processing flow will be explained below.

[0332] Step 1:

[0333] server

[0334] Receives teaching material data (PDF, Docx, etc.) provided by educational institutions and stores it in a database.

[0335] Step 2:

[0336] server

[0337] Natural language processing technology is applied to the received teaching material data to convert it into text format. After conversion, important keywords and concepts are extracted and this information is stored in a database as structured data.

[0338] Step 3:

[0339] User (Teacher)

[0340] The teacher uses a terminal to input specific chapter and page numbers to specify the subject of the questions, and this information is sent from the terminal to the server.

[0341] Step 4:

[0342] Terminal

[0343] The specified range is sent to the server as data.

[0344] Step 5:

[0345] server

[0346] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[0347] Step 6:

[0348] server

[0349] The generative AI model generates test questions based on the scope of the questions, referencing data from past entrance exams and mock exams to create questions that reflect important points.

[0350] Step 7:

[0351] server

[0352] The generated test questions are formatted to be compatible with both paper and web formats. Print data is created for paper media, and online test data is created for web formats.

[0353] Step 8:

[0354] User (student)

[0355] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[0356] Step 9:

[0357] server

[0358] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[0359] Step 10:

[0360] server

[0361] An emotion engine is used to recognize a student's emotional state by analyzing language patterns and facial expression data (if necessary) in the student's answers.

[0362] Step 11:

[0363] server

[0364] The system compiles each student's academic and emotional data, calculating the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, and emotional trends, and creates an analysis report.

[0365] Step 12:

[0366] server

[0367] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test administration. In addition, emotional data is taken into account, and feedback is reflected in which questions students are stressed or excited about.

[0368] Step 13:

[0369] Terminal

[0370] Feedback reports are sent to teachers' devices and made available for viewing. Teachers can use this information to plan lessons and identify which issues specific students responded to emotionally, allowing them to provide individualized support.

[0371] Specific examples

[0372] For example, when conducting a test on "Calculus" in Mathematics II, the system operates as follows.

[0373] In steps 1 to 5, the teacher specifies the "calculus chapter" as the test area, and the generative AI model creates multiple-choice and essay questions from that area.

[0374] In steps 6 to 8, students complete a web-based test that is automatically scored in real time. Essay questions are also scored using natural language processing technology.

[0375] In steps 9 to 12, the emotion engine analyzes the students' emotional state and creates an analysis report along with their grade data. This report is sent to the teacher's device and can be used as a reference for the next lesson or test.

[0376] In this way, the system of the present invention comprehensively performs everything from automatic generation of test questions to grading, result analysis, and feedback that incorporates emotion recognition, thereby improving the quality of education.

[0377] Example 2

[0378] 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."

[0379] Conventional educational systems have had problems with the time and effort required to import teaching material data, create test questions, automate grading, and provide feedback. Furthermore, feedback that ignores the student's emotional state can hinder effective education. The present invention aims to solve these problems and enable efficient and effective education.

[0380] The specification process by the specification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for receiving teaching material data provided by an educational institution and storing it in a database, a means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, and a means for recognizing the emotional state of a student using an emotion engine and storing it together with grade data. This makes it possible to efficiently use teaching materials provided by educational institutions and provide feedback that takes into consideration the student's emotions.

[0381] "Teaching material data" refers to data containing educational information provided by educational institutions, and is provided in formats such as PDF and Docx.

[0382] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language, and is used for preprocessing text data and extracting important keywords and concepts.

[0383] A "generative AI model" is a model that uses machine learning algorithms to generate test questions, and is a technology that creates new questions based on past data and teacher specifications.

[0384] The "test scope" refers to the scope of specific educational content designated by the teacher when creating test questions, and is designated by a specific chapter, page number, or the like.

[0385] "Test questions" are questions or assignments given to assess students' understanding, and are available in multiple choice or essay formats and in paper and web formats.

[0386] "Automatic scoring" is a technology in which a computer automatically evaluates and scores students' answers, and it can be applied to multiple-choice and essay questions.

[0387] The "emotion engine" is a technology that analyzes language patterns and facial expression data collected from devices to recognize students' emotional states.

[0388] "Academic performance data" refers to information recorded as numerical or statistical data on student test results, and is used to evaluate education.

[0389] "Feedback" is information provided to teachers based on test results and emotional data, and is used to improve the quality of education.

[0390] "OCR technology" is a technology that extracts character information from scanned image data and converts it into text data.

[0391] The present invention is a system that receives learning material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. Furthermore, by combining it with an emotion engine that recognizes students' emotions, a more advanced feedback function is realized. Detailed embodiments of this system are described below.

[0392] Importing teaching material data

[0393] server

[0394] The system receives and stores educational material data (PDF, Docx, etc.) provided by educational institutions. It then uses Optical Character Recognition (OCR) technology and the Natural Language Toolkit (NLTK) library to convert this data into text format using natural language processing techniques. It then extracts important keywords and concepts from the converted text and stores them in a database.

[0395] Specific operation example

[0396] The server receives PDF files containing teaching material data for "high school mathematics," converts them into text using OCR technology, and then uses the NLTK library to extract important keywords such as "calculus" and "function" from the text and store them in a database.

[0397] Specifying the scope of questions

[0398] User (Teacher)

[0399] The teacher uses a device to input chapter and page numbers to identify the test topics, and this information is sent from the device to the server.

[0400] Terminal

[0401] Send the question range data to the server.

[0402] server

[0403] The system analyzes the received information on the scope of the exam, extracts the teaching material data within the specified scope from the database, and identifies important points and topics.

[0404] Specific operation example

[0405] The teacher types "Mathematics II Calculus chapter" into the interface, and the device sends the data to the server, which extracts the relevant content from the database.

[0406] Generate test questions

[0407] server

[0408] Generative AI models (e.g., natural language generation models) are used to generate test questions based on the specified subject area. Data from past entrance exams and mock exams is also referenced to ensure high quality questions. The generated questions are formatted in multiple-choice and essay formats for both paper and web versions.

[0409] Specific operation example

[0410] The server generates problems related to the "Chapter on Calculus" by providing a prompt to the AI ​​model. For example, the prompt might be, "Please create three problems to find differential coefficients and two written questions to find definite integrals." The generated problems are provided in both web page and PDF formats.

[0411] Conducting the test

[0412] User (student)

[0413] Students access the test via a web browser and enter their answers. If the test is paper, they scan the answers and upload them to the server.

[0414] Specific operation example

[0415] Students access the test using a web browser on their device and answer questions, or, for paper tests, scan their answers and upload them to the server.

[0416] Scoring

[0417] server

[0418] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper answers are scanned and converted to text using OCR technology, and then graded automatically.

[0419] Specific operation example

[0420] The server instantly grades answers entered online. Written questions are also graded using natural language processing algorithms. Paper answers are also converted to text using OCR technology and automatically graded.

[0421] Emotion recognition

[0422] server

[0423] An emotion engine is used to recognize students' emotional states by analyzing language patterns and facial expression data in their answers, and this information is stored in a database along with their performance data.

[0424] Specific operation example

[0425] The server analyzes the student's answers and uses an emotion engine to analyze language patterns to determine "difficulty" and "stress." If necessary, it also uses the device's camera to analyze facial expression data and grasp the student's overall emotional state.

[0426] Analysis of results and feedback

[0427] server

[0428] Each student's academic performance and emotional data are compiled, and an analysis report is generated by calculating the correct answer rate, incorrect answer rate, difficulty level of each question, emotional trends, etc.

[0429] Specific operation example

[0430] The server analyzes each student's answer data and emotional data and creates a report containing specific information such as "Many students felt stressed on question 3."

[0431] Providing feedback

[0432] Terminal

[0433] Feedback reports are sent to teachers' devices for viewing, including which questions elicited emotional reactions from students.

[0434] Specific operation example

[0435] The teacher checks the feedback report provided on the device and, based on the information that "many students are having difficulty with Problem 3," explains Problem 3 again in the next lesson.

[0436] In this way, the system of the present invention comprehensively covers everything from importing teaching material data to generating test questions, marking, recognizing emotions, analyzing results, and providing feedback, thereby realizing advanced educational support.

[0437] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0438] Step 1: Importing teaching material data

[0439] server

[0440] The server receives and stores the learning material data provided by educational institutions. This data is usually provided in PDF or Docx format. OCR technology and the NLTK library are used to convert the received learning material data into text format. Important keywords and concepts are extracted from the converted text and stored in a database.

[0441] Input PDF or Docx teaching materials data provided by the educational institution

[0442] Output: Text-formatted teaching material data and extracted keyword and concept data

[0443] Specific operation example

[0444] The server receives PDF files containing teaching materials for high school mathematics, converts them into text using OCR technology, and then analyzes the text using the NLTK library to extract keywords such as "calculus" and "function," which are then stored in a database.

[0445] Step 2: Specify the question scope

[0446] User (Teacher)

[0447] The teacher uses a terminal to input chapter and page numbers to specify the test area, and this data is sent to the server via the terminal.

[0448] Input Question information entered by the teacher on the device

[0449] Output Sending question scope information to the server

[0450] Terminal

[0451] The device receives data on the subject matter specified by the teacher and sends it to the server.

[0452] server

[0453] The server analyzes the received information on the test scope, extracts the relevant study material data from the database, and then identifies important points and topics.

[0454] Input question range information

[0455] Output: Extracted teaching material data and key points

[0456] Specific operation example

[0457] When a teacher types "calculus chapter in Mathematics II" into the device interface, the device sends this information to the server, which retrieves the corresponding teaching material data from the database and extracts key points.

[0458] Step 3: Generate test questions

[0459] server

[0460] The server uses a generative AI model to generate test questions based on the specified test scope. Questions are usually created by referencing data from past entrance exams and mock exams. Test questions include both multiple-choice and essay questions. The questions are formatted in both paper and web formats.

[0461] Input question scope information and related teaching material data

[0462] Output Generated test questions

[0463] Specific operation example

[0464] Based on the "Chapter on Calculus," the server instructs the generative AI model to "create three problems to find differential coefficients and two written problems to find definite integrals." The generative AI model generates problems from the specified range and outputs them in PDF and web page format.

[0465] Step 4: Testing

[0466] User (student)

[0467] Students access the web-based test from their devices and enter their answers. If the test is paper-based, they write their answers, scan them, and upload them to the server from their devices.

[0468] Input Test questions provided by the server

[0469] Output Student Answers

[0470] Specific operation example

[0471] Students access the test on a web browser using their device and enter their answers to each question, or if the test is paper, scan the answers and send them to the server.

[0472] Step 5: Scoring

[0473] server

[0474] The server automatically grades web-based answers in real time. For essay questions, natural language processing technology is used to grade. For paper-based answers, scanned data is converted to text using OCR technology and then graded automatically.

[0475] Input student answer data

[0476] Output Scoring results

[0477] Specific operation example

[0478] The server automatically grades multiple-choice questions answered online. Essay questions are also evaluated using natural language processing algorithms. Paper-based answers are also converted to text using OCR technology and graded automatically.

[0479] Step 6: Recognize emotions

[0480] server

[0481] An emotion engine is used to recognize students' emotional states by analyzing language patterns and facial expression data in their answers, and this information is stored in a database along with their performance data.

[0482] Input: Student answers and facial expression data

[0483] Output: Student emotional state data

[0484] Specific operation example

[0485] The server analyzes the answer data and identifies "difficulty" and "stress" from language patterns. It also analyzes facial expression data collected by the device's camera to identify the student's emotional state.

[0486] Step 7: Analyze results and provide feedback

[0487] server

[0488] Each student's academic performance and emotional data are compiled, and an analysis report is generated by calculating the correct answer rate, incorrect answer rate, difficulty level of each question, emotional trends, etc.

[0489] Input: Achievement data and emotion data

[0490] Output Analysis Report

[0491] Specific operation example

[0492] Based on the grade data and emotional data, the server creates a specific analysis report, such as "Many students feel stressed about question 3."

[0493] Step 8: Provide feedback

[0494] Terminal

[0495] Feedback reports are sent to teachers' devices for viewing, including which questions elicited emotional reactions from students.

[0496] Input Analysis Report

[0497] Output Teacher feedback report

[0498] Specific operation example

[0499] The teacher reviews the feedback report provided on the device and takes measures such as "paying particular attention to question 3" in the next lesson.

[0500] (Application example 2)

[0501] 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."

[0502] Traditional test question generation systems based on teaching material data provided by educational institutions are effective in improving student learning outcomes in the field of education. However, particularly in manufacturing, different applications are required, such as product quality inspection and identification of defective products. Existing systems are specialized for education and do not address the real-time quality control and operator emotion recognition feedback required in manufacturing. Therefore, there is a need for a means to ensure quality and effectively manage operators in manufacturing.

[0503] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving teaching material data provided by an educational institution and storing it in a database, means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, means for generating test questions from a test range using a generative AI model, means for collecting production data and inspecting product quality in real time, means for identifying defective products and analyzing and reporting the results, and means for recognizing operator emotions and reflecting the data in feedback. This enables automation of quality control processes and operator stress management in manufacturing sites, as well as learning support in the education field.

[0504] "Teaching material data" refers to learning materials provided by educational institutions, specifically documents provided in PDF or Docx format.

[0505] "Natural language processing technology" refers to technology that enables computers to understand, interpret, and generate human language.

[0506] "Important keywords and concepts" refer to words and ideas that deserve particular attention in the teaching material data.

[0507] A "generative AI model" refers to an algorithm or system that learns from large amounts of data and then makes inferences and generates information based on new data.

[0508] "Test questions" refers to questions or question sets used to assess the level of understanding of the learning content.

[0509] "Automated scoring" refers to the process of using machines or software to automatically evaluate and score answers.

[0510] "Aggregation and analysis" refers to the process of collecting data, processing it using statistical methods, etc., and deriving meaning from it.

[0511] "Real-time product quality inspection" refers to the process of instantly evaluating the quality of products during the manufacturing process.

[0512] "Identification of defective products" refers to identifying products that do not meet quality standards based on inspection criteria.

[0513] "Emotion recognition" refers to the process of determining the emotional state of an operator or student from their facial expressions and behavior.

[0514] "Feedback" refers to providing information that is useful for next actions based on the results and data of a system or process.

[0515] This system automates quality control and operator emotion recognition in manufacturing sites. Specifically, the system is composed of the following hardware and software:

[0516] First, the server must have high-performance data processing capabilities, and it is desirable to use an NVIDIA GPU server. This server has the ability to receive and store data from the production line in real time. The data types used are diverse, including image data, temperature data, and pressure data.

[0517] Data analysis is performed using Python with either TensorFlow or PyTorch from Plads, and Hugging Face's Transformers library is used to train and infer the generative AI model, which automatically generates quality control metrics based on the various data received.

[0518] In addition, the server uses Amazon Rekognition's facial recognition and emotion analysis functions to analyze the operator's emotions in real time, allowing it to understand the stress and other emotional states the operator experiences while working and manage the operator's state as well as product quality.

[0519] As a specific example, consider the case of inspecting the quality of products in a factory. Images, temperature data, and pressure data of products produced on the production line are sent to a server in real time. The server analyzes this data, and a generative AI model automatically sets quality inspection standards. Based on the inspection standards, the products are inspected and passed or failed products are identified. If a failed product is found, the results are analyzed in detail and reported.

[0520] Amazon Rekognition also analyzes the operator's facial expressions based on the inspection results and data from the work, and estimates their emotional state. For example, if an operator is feeling high stress while processing a defective product, that data will also be reflected in the feedback.

[0521] Example prompt sentence:

[0522] "What are the main inspection items for product quality?"

[0523] "Estimate the stress level that the operator experiences from this task."

[0524] In this way, the system of the present invention enhances the quality control process in manufacturing sites and also manages the emotions of operators, thereby improving not only product quality but also worker efficiency and satisfaction.

[0525] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[0526] Step 1:

[0527] The server receives and stores production data, such as product images, temperature data, and pressure data, in real time from the production line. First, it stores the input data from sensors and cameras in a database, which then forms a basic data set for subsequent data analysis.

[0528] Step 2:

[0529] The server performs preprocessing to analyze the received data. Specific operations include noise removal from the image data and smoothing of temperature and pressure data. The preprocessed data is then input into a generative AI model to generate inspection standards.

[0530] Step 3:

[0531] Using the generative AI model, the server sets the product quality inspection criteria. It uses preprocessed product images, temperature, and pressure data as input data, and generates pass / fail criteria based on the criteria learned by the model. The output is specific inspection items and their pass / fail criteria.

[0532] Step 4:

[0533] The server inspects the product based on the generated inspection standards. Specifically, it analyzes product images and takes into account temperature and pressure data to determine whether or not there are any defects. The output is the inspection results, and if there are any defects, detailed data is provided.

[0534] Step 5:

[0535] The server uses Amazon Rekognition for facial recognition and emotion analysis to recognize the operator's emotions in real time. The operator's video data is input, and the emotion engine analyzes their stress level and other emotional states. The output is the operator's emotional state, which is then saved as data.

[0536] Step 6:

[0537] The server integrates product inspection results and operator sentiment data and generates a detailed analysis report based on them. Using the inspection results and sentiment data as input, the analysis tool aggregates and analyzes the data. As output, a detailed analysis report is created for management to use in the next round of inspections and to improve operator work.

[0538] Step 7:

[0539] The terminal provides the generated analysis report to the manager. Specifically, the report is visualized and displayed in a format that is easy for the manager to understand. Based on this, the manager can plan measures to improve quality control and the operator's working environment.

[0540] Through the above processing steps, the system highly automates quality control and operator emotional management on the production line.

[0541] 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.

[0542] The data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of the data generation model 58 is ChatGPT (registered trademark) (Internet search engine).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (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.

[0543] 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.

[0544] [Second embodiment]

[0545] FIG. 3 shows an example of the configuration of a data processing system 210 according to the second embodiment.

[0546] 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.

[0547] 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).

[0548] 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.

[0549] 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.

[0550] 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).

[0551] 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.

[0552] 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.

[0553] 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.

[0554] 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.

[0555] In the smart glasses 214, 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.

[0556] 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."

[0557] The following describes an embodiment of the present invention. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs scoring and result analysis. The detailed operational process required for this will be explained step by step.

[0558] Importing teaching material data

[0559] server

[0560] The system receives and stores the educational material data (PDF, Docx, etc.) provided by educational institutions. It then converts the data into text format using natural language processing technology. This allows it to extract important keywords and concepts and store them in a database.

[0561] Specifying the scope of questions

[0562] User (Teacher)

[0563] The teacher uses a terminal to input the specific chapters and pages to be tested, and this information is sent from the terminal to the server.

[0564] Terminal

[0565] The specified range is sent to the server as data.

[0566] server

[0567] The received range information is analyzed, and the teaching material data within the specified range is extracted from the database.

[0568] Generate test questions

[0569] server

[0570] Based on the specified range of teaching material data, the system selects an appropriate question format, such as fill-in-the-blank, essay, or calculation questions. The generative AI model analyzes trends from past entrance exams and mock exams to create test questions, which are then formatted in both paper and web formats.

[0571] Conducting the test

[0572] User (student)

[0573] Students access the test via a web browser and enter their answers. If the test is on paper, they can scan the answers and upload them to the server.

[0574] Scoring

[0575] server

[0576] For web-based tests, automatic scoring is performed in real time as answers are submitted. For essay-style questions, natural language processing technology is used for scoring. For paper-based tests, scanned data is converted to text using OCR technology and then automatically scored.

[0577] Analysis of results and feedback

[0578] server

[0579] The system compiles each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. It identifies the difficulty of each question and common error patterns and generates an analysis report. Teachers can use this feedback to identify points that should be emphasized in the next lesson.

[0580] Terminal

[0581] Feedback reports are sent to teachers' devices to help them plan their next lessons.

[0582] In this way, this system automates everything from creating test questions to grading, analyzing the results, and providing feedback, greatly improving the efficiency of educational activities. Furthermore, the questions generated reflect past exam trends, making it possible to improve the quality of education.

[0583] The processing flow will be explained below.

[0584] Step 1:

[0585] server

[0586] Receives educational material data provided by educational institutions, including formats such as PDF and Docx. The received data is stored in a database.

[0587] Step 2:

[0588] server

[0589] Natural language processing technology is applied to the saved teaching material data to convert it into text format. After conversion, important keywords and concepts are extracted and this information is stored in a database as structured data.

[0590] Step 3:

[0591] User (Teacher)

[0592] The teacher uses a terminal to input specific chapters and page numbers to specify the range of questions to be asked, and the information is sent from the terminal to the server.

[0593] Step 4:

[0594] server

[0595] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[0596] Step 5:

[0597] server

[0598] The generative AI model generates test questions based on the scope of the questions, referencing data from past entrance exams and mock exams to create questions that reflect important points.

[0599] Step 6:

[0600] server

[0601] The generated test questions are formatted to be compatible with both paper and web formats. Print data is created for paper media, and online test data is created for web formats.

[0602] Step 7:

[0603] User (student)

[0604] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[0605] Step 8:

[0606] server

[0607] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[0608] Step 9:

[0609] server

[0610] Compile the performance data of each student, calculate the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, etc., and create an analysis report.

[0611] Step 10:

[0612] server

[0613] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test.

[0614] Step 11:

[0615] Terminal

[0616] Feedback reports are sent to teachers' devices and made available for viewing, allowing teachers to use them to plan their lessons.

[0617] This allows the system to efficiently carry out a series of steps, from automatically generating test questions, to grading, analyzing the results, and providing feedback, thereby improving the quality of education while significantly reducing the burden on teachers.

[0618] Example 1

[0619] 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."

[0620] Conventional educational systems require a great deal of manual work, from importing teaching material data to creating test questions, grading, and analyzing the results, placing a heavy burden on teachers. It was also difficult to generate test questions that reflected past exam trends, or to provide real-time grading and feedback. This raised concerns that the quality and efficiency of education would decline.

[0621] 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.

[0622] In this invention, the server includes means for receiving teaching material data provided by educational institutions and storing it in a database, means for preprocessing the teaching material data using natural language processing technology to extract important keywords and concepts, means for organizing the teaching material data based on the test scope specified by the teacher, means for generating test questions from the test scope using a generative AI model, means for formatting the test questions in formats such as LaTeX or HTML depending on the question format, means for providing the generated test questions in both paper and web formats, means for automatically scoring the results of students' answers, and means for aggregating and analyzing performance data and providing feedback to teachers. This automates the entire process from importing teaching material data to generating test questions, scoring, and analyzing and providing feedback on the results, reducing the burden on teachers and improving the quality and efficiency of education.

[0623] "Educational institutions" refers to organizations that provide education and training, such as schools, universities, and cram schools.

[0624] "Teaching material data" refers to materials for educational purposes provided by educational institutions, including file formats such as PDF and Docx.

[0625] "Database" refers to a system designed to efficiently manage and retrieve data within an organization.

[0626] "Natural language processing technology" refers to the field of technology in which computers understand, interpret, and generate human language.

[0627] "Keywords" refer to words and phrases that are considered particularly important within a document and are used for information retrieval and data classification.

[0628] "Test area" refers to the area set for test questions, such as specific chapters or page numbers designated by the teacher.

[0629] A "generative AI model" refers to an algorithm or model that uses artificial intelligence technology to generate data or analyze information.

[0630] "Test questions" refer to questions or tasks designed to assess a student's understanding or knowledge.

[0631] "Formatting" refers to the process of converting documents or data into an appropriate format to make them easier to read.

[0632] "Web format" refers to a format that is provided over the Internet and can be viewed and manipulated in a browser.

[0633] "Paper media" refers to information or data printed on paper.

[0634] "Automatic scoring" refers to the process of mechanically evaluating and scoring students' answers using a special program.

[0635] "Academic performance data" refers to data that records student test results and grades in numerical or written form.

[0636] "Feedback" refers to assessments and information provided by systems and teachers to help improve students and lessons.

[0637] The following describes an embodiment of the present invention. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. The detailed operational process required for this purpose is specifically shown.

[0638] Importing teaching material data

[0639] server

[0640] The server receives learning material data (PDF, Docx, etc.) provided by educational institutions via HTTP requests and stores it in local storage. It then converts the learning material data into text format using Google's OCR API or Amazon Textract. From the converted text, natural language processing technology (e.g., spaCy or NLTK) is used to extract important keywords and concepts, which are then stored in a MySQL database. As a concrete example, if a school provides a PDF file as an English learning material, the server receives the PDF, converts it into text format using Google's OCR API, and stores the extracted words and phrases in the database.

[0641] Example prompt sentence:

[0642] The system should receive English learning materials in PDF format provided by educational institutions, convert them into text using Google OCR API, extract important words and phrases from the text, and store them in a database.

[0643] Specifying the scope of questions

[0644] User (Teacher)

[0645] Teachers log in to the system using their device's browser and enter the question scope by specifying a specific chapter or page number from the interface.

[0646] Terminal

[0647] The data for the question area specified by the teacher is obtained from the input form and sent to the server as an API request.

[0648] server

[0649] Analyze the received API request, query the database to retrieve the learning material data within the specified range, and store the retrieved learning material data in a temporary storage area.

[0650] As a specific example, when a teacher specifies Chapter 5 of English as the subject of questions, the terminal sends that information to the server, and the server retrieves data corresponding to Chapter 5 from the database.

[0651] Example prompt sentence:

[0652] If a teacher specifies Chapter 5 of English as the subject of the test, that information is sent to the server, and the server retrieves the relevant data from the database.

[0653] Generate test questions

[0654] server

[0655] Test questions are generated using a generative AI model (e.g., GPT-3) based on a specified range of teaching material data. Questions can be fill-in-the-blank, essay-style, calculation-based, or other formats, and are generated in the appropriate format (LaTeX or HTML). The generated questions are output in both web and paper formats.

[0656] As a concrete example, the server uses GPT-3 to generate fill-in-the-blank questions on the theme of Chapter 5 of English and formats them in HTML format.

[0657] Example prompt sentence:

[0658] Use GPT-3 to generate fill-in-the-blank questions on the theme of English Chapter 5. Format the questions in HTML format so that they can be output in both web and paper formats.

[0659] Conducting the test

[0660] User (student)

[0661] Students access the test via a web browser on their device and enter their answers. For paper-based tests, students fill out the answer sheets, scan them, and upload them to the server from their device.

[0662] For example, students can take an English test online using a browser on their device and enter their answers, or if the answers are paper, they can scan them and upload them to the server.

[0663] Example prompt sentence:

[0664] Students can answer the test online using their device's browser. If the test is paper, they can write their answers, scan them, and upload them to the server from their device.

[0665] Scoring

[0666] server

[0667] Web-based answers are automatically scored as soon as they are submitted. Multiple-choice questions are scored by comparing them with preset correct answer data. Essay questions are evaluated using natural language processing technology (e.g., BERT). Paper-based tests are automatically scored after uploaded scanned data is converted to text using OCR technology (e.g., Tesseract).

[0668] For example, the server analyzes the answer data received and scores multiple-choice questions by comparing them with the correct answer data. For essay questions, natural language analysis is performed using the BERT model to calculate a score. Scanned data is converted into text using an OCR engine, and the results are scored in the same way.

[0669] Example prompt sentence:

[0670] Please use BERT to grade the essay questions that students answer online. For paper-based documents, please scan the data and convert it to text using OCR, and grade it in the same way.

[0671] Analysis of results and feedback

[0672] server

[0673] After all answers have been graded, the data is aggregated and analyzed for correct answer rates and incorrect answer trends. The difficulty level of each question and common error patterns are identified, and an analysis report is generated. This report is generated in JSON or PDF format and sent to the teacher's device.

[0674] As a specific example, the server analyzes the grade data, identifies common trends in incorrect answers, generates a PDF report, and sends the report to the teacher's device.

[0675] Example prompt sentence:

[0676] Please compile the grade data and analyze the percentage of correct answers and trends in incorrect answers. After the analysis, the server will identify common incorrect answer patterns, generate a report, and send it to the teacher's device.

[0677] This system automates the entire process, from importing teaching material data to generating test questions, grading answers, and analyzing and providing feedback on the results, thereby reducing the burden on teachers and improving the quality and efficiency of education.

[0678] The flow of the identification process in the first embodiment will be described with reference to FIG.

[0679] Step 1:

[0680] Importing teaching material data

[0681] The user (educational institution) uploads learning material data (PDF, Docx, etc.) to the server. The input data is in PDF or Docx file format. The server receives the learning material data provided by the educational institution as an HTTP request and stores it in local storage. The server then converts the learning material data into text format using Google's OCR API or Amazon Textract. The converted text is analyzed using natural language processing technology (e.g., spaCy or NLTK) to extract important keywords and concepts. This extracted data is then stored in a MySQL database.

[0682] Specific behavior:

[0683] Teaching material data provided by educational institutions is uploaded.

[0684] The teaching material data received by the server is saved in the "Received" folder in the local storage.

[0685] Calls Google's OCR API to convert PDF and Docx to text.

[0686] The converted text is analyzed using a natural language processing library to extract keywords and concepts.

[0687] The extracted data is stored in the "Keywords" table in the MySQL database.

[0688] Input: Teaching material data (PDF, Docx, etc.)

[0689] Output: Keywords and concepts stored in the database

[0690] Step 2:

[0691] Specifying the scope of questions

[0692] The user (teacher) logs into the system using the browser on their device and selects a specific chapter or page number from the interface. The input data is the subject matter of the test specified by the teacher. The device sends the specified subject matter information to the server as an API request. The server analyzes the received API request and queries the database to retrieve the teaching material data that falls within the specified range. The retrieved teaching material data is stored in a temporary storage area.

[0693] Specific behavior:

[0694] The teacher accesses the "Specify Question Area" page in a browser and specifies the chapter and page number.

[0695] The device converts the specified question scope information into a JSON format API request and sends it to the server.

[0696] The server runs an SQL query against the database to retrieve the data corresponding to the specified chapter or page.

[0697] The acquired data is stored in a temporary storage area.

[0698] Input: Question range information specified by the teacher

[0699] Output: Teaching material data stored in the temporary storage area

[0700] Step 3:

[0701] Generate test questions

[0702] The server generates test questions using a generative AI model (e.g., GPT-3) based on the specified range of teaching material data. The input data is the specified range of teaching material data. The generative AI model receives this and generates test questions in an appropriate format, such as fill-in-the-blank questions, essay questions, or calculation questions. The generated test questions are formatted in formats such as LaTeX or HTML and output in each test format (web format and paper format).

[0703] Specific behavior:

[0704] The server acquires the range-specified teaching material data from the temporary storage area.

[0705] The GPT-3 model is called and test questions are generated using the teaching material data as input.

[0706] The generated questions are formatted into LaTeX or HTML depending on the question format.

[0707] Save formatted test questions in web and paper formats.

[0708] Input: Range-specified teaching material data

[0709] Output: Formatted test questions (LaTeX, HTML)

[0710] Step 4:

[0711] Conducting the test

[0712] Users (students) access the web-based test using a browser on their device and enter their answers. The input data is the student's answers. In the case of paper-based tests, students fill out the answer sheets, scan them, and upload them to the server from their device.

[0713] Specific behavior:

[0714] Students access the test page via the HTTPS protocol and enter their answers in their browser.

[0715] The entered answer data is sent to the server.

[0716] If the answer sheet is paper, scan it and convert it to JPEG or PDF format.

[0717] The converted file is sent to the server via an upload form.

[0718] Input: Student answers (web or paper)

[0719] Output: Answer data sent to the server

[0720] Step 5:

[0721] Scoring

[0722] The server automatically grades the web-based answers as they are sent. The input data is the student's answer data. Multiple-choice questions are compared with preset correct answer data and automatically graded. Essay questions are evaluated using natural language processing technology such as BERT. For paper-based tests, uploaded scanned data is converted to text using OCR technology (e.g., Tesseract) and then graded automatically.

[0723] Specific behavior:

[0724] The server analyzes the received answer data and scores the multiple-choice questions by comparing them with the correct answer data.

[0725] For essay questions, the BERT model is used to perform natural language analysis and calculate a score.

[0726] The scanned data is run through an OCR engine, converted into text, and scored.

[0727] Input: Student answer data (web format, paper scanned data)

[0728] Output:Scoring results

[0729] Step 6:

[0730] Analysis of results and feedback

[0731] After all answers have been graded, the server aggregates the grade data and analyzes the percentage of correct answers and trends in incorrect answers. The input data is the students' grade data. The server identifies the difficulty level and incorrect answer patterns for each question, generates an analysis report, and saves it in JSON or PDF format. The generated feedback report is sent to the teacher's device.

[0732] Specific behavior:

[0733] The server retrieves the performance data and performs data analysis using Python's pandas library.

[0734] Analyze the difficulty of the questions and common error patterns and generate a report.

[0735] The report is saved in JSON or PDF format and sent to the teacher's device using the SMTP protocol.

[0736] Input: Student performance data

[0737] Output: Analysis report (JSON, PDF format)

[0738] (Application example 1)

[0739] 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."

[0740] In today's educational environment, teachers' workloads are increasing, often making it difficult to provide effective teaching. Tasks such as determining the scope of study, creating test questions, grading answers, and providing feedback on results require time and effort. Therefore, there is an urgent need to automate these tasks and reduce the burden on teachers. Furthermore, with the spread of online education, there is also a need to create an environment where students can use smartphones to study efficiently. Therefore, a system is needed to significantly improve the efficiency of all educational activities and improve the quality of education.

[0741] 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.

[0742] In this invention, the server includes means for receiving teaching material data provided by educational institutions and storing it in a database, means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, means for organizing the teaching material data based on the test scope specified by the teacher, means for generating test questions from the test scope using a generative AI model, means for formatting the test questions in both paper and web format, means for automatically scoring the results of students' answers, means for aggregating and analyzing performance data and providing feedback to teachers, means for specifying the test scope using a smartphone app and sending information to the server, and means for students to access web-based tests and submit answers using their smartphones. This automates all educational activities, significantly reducing the burden on teachers and enabling efficient and effective education.

[0743] An "educational institution" is an organization such as a school or university that provides a place for learning and research.

[0744] "Teaching material data" refers to data on learning materials provided by educational institutions, including electronic files in PDF and Docx formats.

[0745] A "database" is a system for efficiently storing and managing teaching material data, grade data, etc.

[0746] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and is used for preprocessing text data and extracting important keywords.

[0747] "Important keywords and concepts" are words and ideas in the teaching material data that are considered to be particularly important for learning.

[0748] A "teacher" is a professional who educates students, and is responsible for specifying the scope of test questions and guiding educational activities.

[0749] "Test Coverage" refers to the specific chapters or areas that are covered when creating test questions.

[0750] A "generative AI model" is a program that automatically generates test questions using artificial intelligence technology.

[0751] "Test questions" are questions used to evaluate students' learning outcomes, and include fill-in-the-blank questions and essay-style questions.

[0752] "Paper media" refers to printed paper forms, including forms on which handwritten answers are written.

[0753] "Web format" refers to an online format delivered over the Internet and accessible through a browser or app.

[0754] "Automatic marking" is the process by which a computer automatically evaluates and scores student answers.

[0755] "Performance data" refers to numerical and evaluation data that represent test results.

[0756] "Feedback" refers to evaluations and guidance comments provided based on performance data.

[0757] A "smartphone" is a type of mobile phone that can connect to the Internet and run applications.

[0758] A "server" is a computer system that provides information over a network.

[0759] "OCR technology" is a technology that extracts text information from an image and is used to convert handwritten answers written on paper into text data.

[0760] The following describes in detail an embodiment of the present invention. This system uses a smartphone app and a server based on educational material data provided by an educational institution to efficiently generate test questions, grade answers, and provide feedback on performance data.

[0761] Importing teaching material data

[0762] server

[0763] Teaching material data (PDF, Docx, etc.) provided by educational institutions is received and stored in a database. During this process, an API is set up using Python and Django to receive the teaching material data. After the teaching material data is saved, it is converted into text format using natural language processing technology, and important keywords and concepts are extracted and stored in the database. Specifically, PDF data is converted into text using PDFMiner, and Docx data is converted into text using python-docx.

[0764] Specifying the scope of questions

[0765] User (Teacher)

[0766] Teachers use a smartphone app to input specific chapters and pages to specify the test area, and the app sends this information to a server. The smartphone app is developed using a cross-platform framework such as React Native or Flutter.

[0767] server

[0768] The received range information is analyzed, and the teaching material data within the specified range is extracted from the database.

[0769] Generate test questions

[0770] server

[0771] Based on the specified range, a generative AI model (using TensorFlow) analyzes past test data and question trends to automatically generate test questions. During this process, the generative AI model is given prompts such as the following:

[0772] Example prompt sentence:

[0773] "Scope: Questions will be based on Chapter 2, Overview of Ancient Greek Philosophy. Multiple choice, fill-in-the-blank, and essay questions will be created. The exam format has been optimized taking into account past exam questions."

[0774] The generated test questions are formatted and saved in both paper and web formats.

[0775] Conducting the test

[0776] User (student)

[0777] Students access the web-based test via a smartphone app and enter their answers, which are then sent to a server in real time.

[0778] Scoring

[0779] server

[0780] For web-based tests, automatic scoring is performed in real time as the answers are submitted. For essay-style questions, scoring is performed using natural language processing technology. For paper-based tests, the scanned data is converted to text using OCR technology (using OpenCV) and then automatically scored.

[0781] Analysis of results and feedback

[0782] server

[0783] The system compiles each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. It identifies the difficulty of each question and common error patterns and generates an analysis report. The feedback report is sent to the teacher's device and used as a reference for planning the next lesson.

[0784] This system automates all teaching activities, significantly reducing the burden on teachers, and generates questions that reflect past exam trends, improving the quality of teaching.

[0785] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[0786] Step 1:

[0787] Importing teaching material data

[0788] The server receives learning material data (PDF, Docx, etc.) provided by educational institutions and stores it in a database. Specifically, an API endpoint is set up using the Django framework to receive the learning material data. The received data is converted to text format using PDFMiner for PDF and python-docx for Docx. The input here is a PDF or Docx file, and the output is text data.

[0789] Step 2:

[0790] Preprocessing of teaching material data

[0791] The server preprocesses the converted text data using natural language processing techniques to extract important keywords and concepts. This is done using natural language processing libraries such as NLTK. Specific operations include tokenizing the text data, morphological analysis, removing stop words, and extracting feature words. The input is the text data obtained in step 1, and the output is a list of keywords and concepts.

[0792] Step 3:

[0793] Specifying the scope of questions

[0794] The user (teacher) uses a smartphone app to input the test range, such as a specific chapter or page. The specified range information is sent from the device to the server. The input is the test range information specified by the teacher, and the output is the range information stored on the server. The smartphone app is developed using React Native and Flutter.

[0795] Step 4:

[0796] Analysis of the test scope

[0797] The server analyzes the received range information and retrieves the corresponding teaching material data from the database. Based on the specific chapters and pages included in the range information, it searches for and extracts related text data. The input is the range information obtained in step 3, and the output is text data belonging to the specific range.

[0798] Step 5:

[0799] Generate test questions

[0800] The server uses a generative AI model (using TensorFlow) to automatically generate appropriate test questions (multiple choice, fill-in-the-blank, essay, etc.) from the test range. Specifically, the questions are automatically created using a specific algorithm, taking into account past test data and question trends. During this process, the following prompt sentence is input into the generative AI model.

[0801] Example prompt: "Scope: Questions will be drawn from Chapter 2, Overview of Ancient Greek Philosophy. Multiple choice, fill-in-the-blank, and essay questions will be created. The exam format will be optimized based on past exam questions."

[0802] The input is a specific range of text data and a prompt, and the output is automatically generated test questions.

[0803] Step 6:

[0804] Test question formatting

[0805] The server formats the generated test questions in both paper and web formats. In the case of web format, it converts them into a format that can be displayed in a browser using HTML and CSS, and in the case of paper format, it converts them into a format for printing. The input is the test questions created in step 5, and the output is the formatted test question data.

[0806] Step 7:

[0807] Conducting the test

[0808] Users (students) access web-based tests through a smartphone app, enter their answers, and submit them. The smartphone app provides a browser-based display and input interface. During this process, answers are sent to the server in real time. The input is the student's answer, which is the data sent to the server in real time. The output is the submitted answer data.

[0809] Step 8:

[0810] Automatic scoring

[0811] The server automatically grades the answer data submitted by students in real time. Multiple choice and fill-in-the-blank questions are automatically evaluated by comparing them with preset answers, and essay questions are graded using natural language processing technology. The input is the submitted answer data, and the output is the graded results.

[0812] Step 9:

[0813] Result collection and feedback

[0814] The server aggregates each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. Data analysis tools (such as Pandas and NumPy) are used for the analysis. A feedback report is generated based on the performance data and sent to the teacher's device. The input is the automatically graded result data, and the output is the feedback report.

[0815] 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.

[0816] An embodiment of the present invention will be described below. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. The detailed operational process required for this will be explained step by step. In addition, by combining it with an emotion engine that recognizes the user's emotions, a more advanced feedback function can be achieved.

[0817] Importing teaching material data

[0818] server

[0819] The system receives and stores the educational material data (PDF, Docx, etc.) provided by educational institutions. It then converts the data into text format using natural language processing technology. This allows it to extract important keywords and concepts and store them in a database.

[0820] Specifying the scope of questions

[0821] User (Teacher)

[0822] The teacher uses a terminal to input specific chapter and page numbers to specify the test area, and this information is sent from the terminal to the server.

[0823] Terminal

[0824] The specified range is sent to the server as data.

[0825] server

[0826] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[0827] Generate test questions

[0828] server

[0829] Based on the exam scope, a generative AI model generates test questions. It references data from past entrance exams and mock exams to create questions that reflect key points. It then formats the questions in both paper and web formats.

[0830] Conducting the test

[0831] User (student)

[0832] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[0833] Scoring

[0834] server

[0835] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[0836] Emotion recognition

[0837] server

[0838] The emotion engine analyzes language patterns and facial expression data (if required) in students' responses sent from their devices to recognize their emotional state. This information is stored in a database along with their grades.

[0839] Analysis of results and feedback

[0840] server

[0841] The system compiles each student's academic and emotional data, calculating the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, and emotional trends, and creates an analysis report.

[0842] server

[0843] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test administration. In addition, emotional data is taken into account, and feedback is reflected in which questions students are stressed or excited about.

[0844] Providing feedback

[0845] Terminal

[0846] Feedback reports are sent to teachers' devices and made available for viewing. Teachers can use this information to plan lessons and identify which issues specific students responded to emotionally, allowing them to provide individualized support.

[0847] Specific examples

[0848] For example, when administering a test on "Calculus" in Mathematics II, the system works as follows: When a teacher specifies the "Calculus chapter" as the test area, the generative AI model creates multiple-choice and essay questions from that area. Students answer the online test, and the emotion engine identifies areas of stress from their facial expressions and language patterns. Based on the obtained performance and emotion data, the server generates a detailed analysis report, which the teacher can use to understand points to emphasize in the next lesson and individual student needs.

[0849] In this way, the system of the present invention comprehensively covers everything from automatic generation of test questions to scoring, result analysis, and feedback incorporating emotion recognition, thereby improving the quality of education.

[0850] The processing flow will be explained below.

[0851] Step 1:

[0852] server

[0853] Receives teaching material data (PDF, Docx, etc.) provided by educational institutions and stores it in a database.

[0854] Step 2:

[0855] server

[0856] Natural language processing technology is applied to the received teaching material data to convert it into text format. After conversion, important keywords and concepts are extracted and this information is stored in a database as structured data.

[0857] Step 3:

[0858] User (Teacher)

[0859] The teacher uses a terminal to input specific chapter and page numbers to specify the subject of the questions, and this information is sent from the terminal to the server.

[0860] Step 4:

[0861] Terminal

[0862] The specified range is sent to the server as data.

[0863] Step 5:

[0864] server

[0865] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[0866] Step 6:

[0867] server

[0868] The generative AI model generates test questions based on the scope of the questions, referencing data from past entrance exams and mock exams to create questions that reflect important points.

[0869] Step 7:

[0870] server

[0871] The generated test questions are formatted to be compatible with both paper and web formats. Print data is created for paper media, and online test data is created for web formats.

[0872] Step 8:

[0873] User (student)

[0874] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[0875] Step 9:

[0876] server

[0877] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[0878] Step 10:

[0879] server

[0880] An emotion engine is used to recognize a student's emotional state by analyzing language patterns and facial expression data (if necessary) in the student's answers.

[0881] Step 11:

[0882] server

[0883] The system compiles each student's academic and emotional data, calculating the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, and emotional trends, and creates an analysis report.

[0884] Step 12:

[0885] server

[0886] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test administration. In addition, emotional data is taken into account, and feedback is reflected in which questions students are stressed or excited about.

[0887] Step 13:

[0888] Terminal

[0889] Feedback reports are sent to teachers' devices and made available for viewing. Teachers can use this information to plan lessons and identify which issues specific students responded to emotionally, allowing them to provide individualized support.

[0890] Specific examples

[0891] For example, when conducting a test on "Calculus" in Mathematics II, the system operates as follows.

[0892] In steps 1 to 5, the teacher specifies the "calculus chapter" as the test area, and the generative AI model creates multiple-choice and essay questions from that area.

[0893] In steps 6 to 8, students complete a web-based test that is automatically scored in real time. Essay questions are also scored using natural language processing technology.

[0894] In steps 9 to 12, the emotion engine analyzes the students' emotional state and creates an analysis report along with their grade data. This report is sent to the teacher's device and can be used as a reference for the next lesson or test.

[0895] In this way, the system of the present invention comprehensively performs everything from automatic generation of test questions to grading, result analysis, and feedback that incorporates emotion recognition, thereby improving the quality of education.

[0896] Example 2

[0897] 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."

[0898] Conventional educational systems have had problems with the time and effort required to import teaching material data, create test questions, automate grading, and provide feedback. Furthermore, feedback that ignores the student's emotional state can hinder effective education. The present invention aims to solve these problems and enable efficient and effective education.

[0899] The specification process by the specification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for receiving teaching material data provided by an educational institution and storing it in a database, a means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, and a means for recognizing the emotional state of a student using an emotion engine and storing it together with grade data. This makes it possible to efficiently use teaching materials provided by educational institutions and provide feedback that takes into consideration the student's emotions.

[0900] "Teaching material data" refers to data containing educational information provided by educational institutions, and is provided in formats such as PDF and Docx.

[0901] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language, and is used for preprocessing text data and extracting important keywords and concepts.

[0902] A "generative AI model" is a model that uses machine learning algorithms to generate test questions, and is a technology that creates new questions based on past data and teacher specifications.

[0903] The "test scope" refers to the scope of specific educational content designated by the teacher when creating test questions, and is designated by a specific chapter, page number, or the like.

[0904] "Test questions" are questions or assignments given to assess students' understanding, and are available in multiple choice or essay formats and in paper and web formats.

[0905] "Automatic scoring" is a technology in which a computer automatically evaluates and scores students' answers, and it can be applied to multiple-choice and essay questions.

[0906] The "emotion engine" is a technology that analyzes language patterns and facial expression data collected from devices to recognize students' emotional states.

[0907] "Academic performance data" refers to information recorded as numerical or statistical data on student test results, and is used to evaluate education.

[0908] "Feedback" is information provided to teachers based on test results and emotional data, and is used to improve the quality of education.

[0909] "OCR technology" is a technology that extracts character information from scanned image data and converts it into text data.

[0910] The present invention is a system that receives learning material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. Furthermore, by combining it with an emotion engine that recognizes students' emotions, a more advanced feedback function is realized. Detailed embodiments of this system are described below.

[0911] Importing teaching material data

[0912] server

[0913] The system receives and stores educational material data (PDF, Docx, etc.) provided by educational institutions. It then uses Optical Character Recognition (OCR) technology and the Natural Language Toolkit (NLTK) library to convert this data into text format using natural language processing techniques. It then extracts important keywords and concepts from the converted text and stores them in a database.

[0914] Specific operation example

[0915] The server receives PDF files containing teaching material data for "high school mathematics," converts them into text using OCR technology, and then uses the NLTK library to extract important keywords such as "calculus" and "function" from the text and store them in a database.

[0916] Specifying the scope of questions

[0917] User (Teacher)

[0918] The teacher uses a device to input chapter and page numbers to identify the test topics, and this information is sent from the device to the server.

[0919] Terminal

[0920] Send the question range data to the server.

[0921] server

[0922] The system analyzes the received information on the scope of the exam, extracts the teaching material data within the specified scope from the database, and identifies important points and topics.

[0923] Specific operation example

[0924] The teacher types "Mathematics II Calculus chapter" into the interface, and the device sends the data to the server, which extracts the relevant content from the database.

[0925] Generate test questions

[0926] server

[0927] Generative AI models (e.g., natural language generation models) are used to generate test questions based on the specified subject area. Data from past entrance exams and mock exams is also referenced to ensure high quality questions. The generated questions are formatted in multiple-choice and essay formats for both paper and web versions.

[0928] Specific operation example

[0929] The server generates problems related to the "Chapter on Calculus" by providing a prompt to the AI ​​model. For example, the prompt might be, "Please create three problems to find differential coefficients and two written questions to find definite integrals." The generated problems are provided in both web page and PDF formats.

[0930] Conducting the test

[0931] User (student)

[0932] Students access the test via a web browser and enter their answers. If the test is paper, they scan the answers and upload them to the server.

[0933] Specific operation example

[0934] Students access the test using a web browser on their device and answer questions, or, for paper tests, scan their answers and upload them to the server.

[0935] Scoring

[0936] server

[0937] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper answers are scanned and converted to text using OCR technology, and then graded automatically.

[0938] Specific operation example

[0939] The server instantly grades answers entered online. Written questions are also graded using natural language processing algorithms. Paper answers are also converted to text using OCR technology and automatically graded.

[0940] Emotion recognition

[0941] server

[0942] An emotion engine is used to recognize students' emotional states by analyzing language patterns and facial expression data in their answers, and this information is stored in a database along with their performance data.

[0943] Specific operation example

[0944] The server analyzes the student's answers and uses an emotion engine to analyze language patterns to determine "difficulty" and "stress." If necessary, it also uses the device's camera to analyze facial expression data and grasp the student's overall emotional state.

[0945] Analysis of results and feedback

[0946] server

[0947] Each student's academic performance and emotional data are compiled, and an analysis report is generated by calculating the correct answer rate, incorrect answer rate, difficulty level of each question, emotional trends, etc.

[0948] Specific operation example

[0949] The server analyzes each student's answer data and emotional data and creates a report containing specific information such as "Many students felt stressed on question 3."

[0950] Providing feedback

[0951] Terminal

[0952] Feedback reports are sent to teachers' devices for viewing, including which questions elicited emotional reactions from students.

[0953] Specific operation example

[0954] The teacher checks the feedback report provided on the device and, based on the information that "many students are having difficulty with Problem 3," explains Problem 3 again in the next lesson.

[0955] In this way, the system of the present invention comprehensively covers everything from importing teaching material data to generating test questions, marking, recognizing emotions, analyzing results, and providing feedback, thereby realizing advanced educational support.

[0956] The flow of the identification process in the second embodiment will be described with reference to FIG.

[0957] Step 1: Importing teaching material data

[0958] server

[0959] The server receives and stores the learning material data provided by educational institutions. This data is usually provided in PDF or Docx format. OCR technology and the NLTK library are used to convert the received learning material data into text format. Important keywords and concepts are extracted from the converted text and stored in a database.

[0960] Input PDF or Docx teaching materials data provided by the educational institution

[0961] Output: Text-formatted teaching material data and extracted keyword and concept data

[0962] Specific operation example

[0963] The server receives PDF files containing teaching materials for high school mathematics, converts them into text using OCR technology, and then analyzes the text using the NLTK library to extract keywords such as "calculus" and "function," which are then stored in a database.

[0964] Step 2: Specify the question scope

[0965] User (Teacher)

[0966] The teacher uses a terminal to input chapter and page numbers to specify the test area, and this data is sent to the server via the terminal.

[0967] Input Question information entered by the teacher on the device

[0968] Output Sending question scope information to the server

[0969] Terminal

[0970] The device receives data on the subject matter specified by the teacher and sends it to the server.

[0971] server

[0972] The server analyzes the received information on the test scope, extracts the relevant study material data from the database, and then identifies important points and topics.

[0973] Input question range information

[0974] Output: Extracted teaching material data and key points

[0975] Specific operation example

[0976] When a teacher types "calculus chapter in Mathematics II" into the device interface, the device sends this information to the server, which retrieves the corresponding teaching material data from the database and extracts key points.

[0977] Step 3: Generate test questions

[0978] server

[0979] The server uses a generative AI model to generate test questions based on the specified test scope. Questions are usually created by referencing data from past entrance exams and mock exams. Test questions include both multiple-choice and essay questions. The questions are formatted in both paper and web formats.

[0980] Input question scope information and related teaching material data

[0981] Output Generated test questions

[0982] Specific operation example

[0983] Based on the "Chapter on Calculus," the server instructs the generative AI model to "create three problems to find differential coefficients and two written problems to find definite integrals." The generative AI model generates problems from the specified range and outputs them in PDF and web page format.

[0984] Step 4: Testing

[0985] User (student)

[0986] Students access the web-based test from their devices and enter their answers. If the test is paper-based, they write their answers, scan them, and upload them to the server from their devices.

[0987] Input Test questions provided by the server

[0988] Output Student Answers

[0989] Specific operation example

[0990] Students access the test on a web browser using their device and enter their answers to each question, or if the test is paper, scan the answers and send them to the server.

[0991] Step 5: Scoring

[0992] server

[0993] The server automatically grades web-based answers in real time. For essay questions, natural language processing technology is used to grade. For paper-based answers, scanned data is converted to text using OCR technology and then graded automatically.

[0994] Input student answer data

[0995] Output Scoring results

[0996] Specific operation example

[0997] The server automatically grades multiple-choice questions answered online. Essay questions are also evaluated using natural language processing algorithms. Paper-based answers are also converted to text using OCR technology and graded automatically.

[0998] Step 6: Recognize emotions

[0999] server

[1000] An emotion engine is used to recognize students' emotional states by analyzing language patterns and facial expression data in their answers, and this information is stored in a database along with their performance data.

[1001] Input: Student answers and facial expression data

[1002] Output: Student emotional state data

[1003] Specific operation example

[1004] The server analyzes the answer data and identifies "difficulty" and "stress" from language patterns. It also analyzes facial expression data collected by the device's camera to identify the student's emotional state.

[1005] Step 7: Analyze results and provide feedback

[1006] server

[1007] Each student's academic performance and emotional data are compiled, and an analysis report is generated by calculating the correct answer rate, incorrect answer rate, difficulty level of each question, emotional trends, etc.

[1008] Input: Achievement data and emotion data

[1009] Output Analysis Report

[1010] Specific operation example

[1011] Based on the grade data and emotional data, the server creates a specific analysis report, such as "Many students feel stressed about question 3."

[1012] Step 8: Provide feedback

[1013] Terminal

[1014] Feedback reports are sent to teachers' devices for viewing, including which questions elicited emotional reactions from students.

[1015] Input Analysis Report

[1016] Output Teacher feedback report

[1017] Specific operation example

[1018] The teacher reviews the feedback report provided on the device and takes measures such as "paying particular attention to question 3" in the next lesson.

[1019] (Application example 2)

[1020] 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."

[1021] Traditional test question generation systems based on teaching material data provided by educational institutions are effective in improving student learning outcomes in the field of education. However, particularly in manufacturing, different applications are required, such as product quality inspection and identification of defective products. Existing systems are specialized for education and do not address the real-time quality control and operator emotion recognition feedback required in manufacturing. Therefore, there is a need for a means to ensure quality and effectively manage operators in manufacturing.

[1022] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving teaching material data provided by an educational institution and storing it in a database, means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, means for generating test questions from a test range using a generative AI model, means for collecting production data and inspecting product quality in real time, means for identifying defective products and analyzing and reporting the results, and means for recognizing operator emotions and reflecting the data in feedback. This enables automation of quality control processes and operator stress management in manufacturing sites, as well as learning support in the education field.

[1023] "Teaching material data" refers to learning materials provided by educational institutions, specifically documents provided in PDF or Docx format.

[1024] "Natural language processing technology" refers to technology that enables computers to understand, interpret, and generate human language.

[1025] "Important keywords and concepts" refer to words and ideas that deserve particular attention in the teaching material data.

[1026] A "generative AI model" refers to an algorithm or system that learns from large amounts of data and then makes inferences and generates information based on new data.

[1027] "Test questions" refers to questions or question sets used to assess the level of understanding of the learning content.

[1028] "Automated scoring" refers to the process of using machines or software to automatically evaluate and score answers.

[1029] "Aggregation and analysis" refers to the process of collecting data, processing it using statistical methods, etc., and deriving meaning from it.

[1030] "Real-time product quality inspection" refers to the process of instantly evaluating the quality of products during the manufacturing process.

[1031] "Identification of defective products" refers to identifying products that do not meet quality standards based on inspection criteria.

[1032] "Emotion recognition" refers to the process of determining the emotional state of an operator or student from their facial expressions and behavior.

[1033] "Feedback" refers to providing information that is useful for next actions based on the results and data of a system or process.

[1034] This system automates quality control and operator emotion recognition in manufacturing sites. Specifically, the system is composed of the following hardware and software:

[1035] First, the server must have high-performance data processing capabilities, and it is desirable to use an NVIDIA GPU server. This server has the ability to receive and store data from the production line in real time. The data types used are diverse, including image data, temperature data, and pressure data.

[1036] Data analysis is performed using Python with either TensorFlow or PyTorch from Plads, and Hugging Face's Transformers library is used to train and infer the generative AI model, which automatically generates quality control metrics based on the various data received.

[1037] In addition, the server uses Amazon Rekognition's facial recognition and emotion analysis functions to analyze the operator's emotions in real time, allowing it to understand the stress and other emotional states the operator experiences while working and manage the operator's state as well as product quality.

[1038] As a specific example, consider the case of inspecting the quality of products in a factory. Images, temperature data, and pressure data of products produced on the production line are sent to a server in real time. The server analyzes this data, and a generative AI model automatically sets quality inspection standards. Based on the inspection standards, the products are inspected and passed or failed products are identified. If a failed product is found, the results are analyzed in detail and reported.

[1039] Amazon Rekognition also analyzes the operator's facial expressions based on the inspection results and data from the work, and estimates their emotional state. For example, if an operator is feeling high stress while processing a defective product, that data will also be reflected in the feedback.

[1040] Example prompt sentence:

[1041] "What are the main inspection items for product quality?"

[1042] "Estimate the stress level that the operator experiences from this task."

[1043] In this way, the system of the present invention enhances the quality control process in manufacturing sites and also manages the emotions of operators, thereby improving not only product quality but also worker efficiency and satisfaction.

[1044] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1045] Step 1:

[1046] The server receives and stores production data, such as product images, temperature data, and pressure data, in real time from the production line. First, it stores the input data from sensors and cameras in a database, which then forms a basic data set for subsequent data analysis.

[1047] Step 2:

[1048] The server performs preprocessing to analyze the received data. Specific operations include noise removal from the image data and smoothing of temperature and pressure data. The preprocessed data is then input into a generative AI model to generate inspection standards.

[1049] Step 3:

[1050] Using the generative AI model, the server sets the product quality inspection criteria. It uses preprocessed product images, temperature, and pressure data as input data, and generates pass / fail criteria based on the criteria learned by the model. The output is specific inspection items and their pass / fail criteria.

[1051] Step 4:

[1052] The server inspects the product based on the generated inspection standards. Specifically, it analyzes product images and takes into account temperature and pressure data to determine whether or not there are any defects. The output is the inspection results, and if there are any defects, detailed data is provided.

[1053] Step 5:

[1054] The server uses Amazon Rekognition for facial recognition and emotion analysis to recognize the operator's emotions in real time. The operator's video data is input, and the emotion engine analyzes their stress level and other emotional states. The output is the operator's emotional state, which is then saved as data.

[1055] Step 6:

[1056] The server integrates product inspection results and operator sentiment data and generates a detailed analysis report based on them. Using the inspection results and sentiment data as input, the analysis tool aggregates and analyzes the data. As output, a detailed analysis report is created for management to use in the next round of inspections and to improve operator work.

[1057] Step 7:

[1058] The terminal provides the generated analysis report to the manager. Specifically, the report is visualized and displayed in a format that is easy for the manager to understand. Based on this, the manager can plan measures to improve quality control and the operator's working environment.

[1059] Through the above processing steps, the system highly automates quality control and operator emotional management on the production line.

[1060] 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.

[1061] 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.

[1062] 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.

[1063] [Third embodiment]

[1064] FIG. 5 shows an example of the configuration of a data processing system 310 according to the third embodiment.

[1065] 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.

[1066] 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).

[1067] 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.

[1068] 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.

[1069] 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).

[1070] 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.

[1071] 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.

[1072] 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.

[1073] 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.

[1074] 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.

[1075] 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."

[1076] The following describes an embodiment of the present invention. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs scoring and result analysis. The detailed operational process required for this will be explained step by step.

[1077] Importing teaching material data

[1078] server

[1079] The system receives and stores the educational material data (PDF, Docx, etc.) provided by educational institutions. It then converts the data into text format using natural language processing technology. This allows it to extract important keywords and concepts and store them in a database.

[1080] Specifying the scope of questions

[1081] User (Teacher)

[1082] The teacher uses a terminal to input the specific chapters and pages to be tested, and this information is sent from the terminal to the server.

[1083] Terminal

[1084] The specified range is sent to the server as data.

[1085] server

[1086] The received range information is analyzed, and the teaching material data within the specified range is extracted from the database.

[1087] Generate test questions

[1088] server

[1089] Based on the specified range of teaching material data, the system selects an appropriate question format, such as fill-in-the-blank, essay, or calculation questions. The generative AI model analyzes trends from past entrance exams and mock exams to create test questions, which are then formatted in both paper and web formats.

[1090] Conducting the test

[1091] User (student)

[1092] Students access the test via a web browser and enter their answers. If the test is on paper, they can scan the answers and upload them to the server.

[1093] Scoring

[1094] server

[1095] For web-based tests, automatic scoring is performed in real time as answers are submitted. For essay-style questions, natural language processing technology is used for scoring. For paper-based tests, scanned data is converted to text using OCR technology and then automatically scored.

[1096] Analysis of results and feedback

[1097] server

[1098] The system compiles each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. It identifies the difficulty of each question and common error patterns and generates an analysis report. Teachers can use this feedback to identify points that should be emphasized in the next lesson.

[1099] Terminal

[1100] Feedback reports are sent to teachers' devices to help them plan their next lessons.

[1101] In this way, this system automates everything from creating test questions to grading, analyzing the results, and providing feedback, greatly improving the efficiency of educational activities. Furthermore, the questions generated reflect past exam trends, making it possible to improve the quality of education.

[1102] The processing flow will be explained below.

[1103] Step 1:

[1104] server

[1105] Receives educational material data provided by educational institutions, including formats such as PDF and Docx. The received data is stored in a database.

[1106] Step 2:

[1107] server

[1108] Natural language processing technology is applied to the saved teaching material data to convert it into text format. After conversion, important keywords and concepts are extracted and this information is stored in a database as structured data.

[1109] Step 3:

[1110] User (Teacher)

[1111] The teacher uses a terminal to input specific chapters and page numbers to specify the range of questions to be asked, and the information is sent from the terminal to the server.

[1112] Step 4:

[1113] server

[1114] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[1115] Step 5:

[1116] server

[1117] The generative AI model generates test questions based on the scope of the questions, referencing data from past entrance exams and mock exams to create questions that reflect important points.

[1118] Step 6:

[1119] server

[1120] The generated test questions are formatted to be compatible with both paper and web formats. Print data is created for paper media, and online test data is created for web formats.

[1121] Step 7:

[1122] User (student)

[1123] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[1124] Step 8:

[1125] server

[1126] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[1127] Step 9:

[1128] server

[1129] Compile the performance data of each student, calculate the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, etc., and create an analysis report.

[1130] Step 10:

[1131] server

[1132] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test.

[1133] Step 11:

[1134] Terminal

[1135] Feedback reports are sent to teachers' devices and made available for viewing, allowing teachers to use them to plan their lessons.

[1136] This allows the system to efficiently carry out a series of steps, from automatically generating test questions, to grading, analyzing the results, and providing feedback, thereby improving the quality of education while significantly reducing the burden on teachers.

[1137] Example 1

[1138] 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."

[1139] Conventional educational systems require a great deal of manual work, from importing teaching material data to creating test questions, grading, and analyzing the results, placing a heavy burden on teachers. It was also difficult to generate test questions that reflected past exam trends, or to provide real-time grading and feedback. This raised concerns that the quality and efficiency of education would decline.

[1140] 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.

[1141] In this invention, the server includes means for receiving teaching material data provided by educational institutions and storing it in a database, means for preprocessing the teaching material data using natural language processing technology to extract important keywords and concepts, means for organizing the teaching material data based on the test scope specified by the teacher, means for generating test questions from the test scope using a generative AI model, means for formatting the test questions in formats such as LaTeX or HTML depending on the question format, means for providing the generated test questions in both paper and web formats, means for automatically scoring the results of students' answers, and means for aggregating and analyzing performance data and providing feedback to teachers. This automates the entire process from importing teaching material data to generating test questions, scoring, and analyzing and providing feedback on the results, reducing the burden on teachers and improving the quality and efficiency of education.

[1142] "Educational institutions" refers to organizations that provide education and training, such as schools, universities, and cram schools.

[1143] "Teaching material data" refers to materials for educational purposes provided by educational institutions, including file formats such as PDF and Docx.

[1144] "Database" refers to a system designed to efficiently manage and retrieve data within an organization.

[1145] "Natural language processing technology" refers to the field of technology in which computers understand, interpret, and generate human language.

[1146] "Keywords" refer to words and phrases that are considered particularly important within a document and are used for information retrieval and data classification.

[1147] "Test area" refers to the area set for test questions, such as specific chapters or page numbers designated by the teacher.

[1148] A "generative AI model" refers to an algorithm or model that uses artificial intelligence technology to generate data or analyze information.

[1149] "Test questions" refer to questions or tasks designed to assess a student's understanding or knowledge.

[1150] "Formatting" refers to the process of converting documents or data into an appropriate format to make them easier to read.

[1151] "Web format" refers to a format that is provided over the Internet and can be viewed and manipulated in a browser.

[1152] "Paper media" refers to information or data printed on paper.

[1153] "Automatic scoring" refers to the process of mechanically evaluating and scoring students' answers using a special program.

[1154] "Academic performance data" refers to data that records student test results and grades in numerical or written form.

[1155] "Feedback" refers to assessments and information provided by systems and teachers to help improve students and lessons.

[1156] The following describes an embodiment of the present invention. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. The detailed operational process required for this purpose is specifically shown.

[1157] Importing teaching material data

[1158] server

[1159] The server receives learning material data (PDF, Docx, etc.) provided by educational institutions via HTTP requests and stores it in local storage. It then converts the learning material data into text format using Google's OCR API or Amazon Textract. From the converted text, natural language processing technology (e.g., spaCy or NLTK) is used to extract important keywords and concepts, which are then stored in a MySQL database. As a concrete example, if a school provides a PDF file as an English learning material, the server receives the PDF, converts it into text format using Google's OCR API, and stores the extracted words and phrases in the database.

[1160] Example prompt sentence:

[1161] The system should receive English learning materials in PDF format provided by educational institutions, convert them into text using Google OCR API, extract important words and phrases from the text, and store them in a database.

[1162] Specifying the scope of questions

[1163] User (Teacher)

[1164] Teachers log in to the system using their device's browser and enter the question scope by specifying a specific chapter or page number from the interface.

[1165] Terminal

[1166] The data for the question area specified by the teacher is obtained from the input form and sent to the server as an API request.

[1167] server

[1168] Analyze the received API request, query the database to retrieve the learning material data within the specified range, and store the retrieved learning material data in a temporary storage area.

[1169] As a specific example, when a teacher specifies Chapter 5 of English as the subject of questions, the terminal sends that information to the server, and the server retrieves data corresponding to Chapter 5 from the database.

[1170] Example prompt sentence:

[1171] If a teacher specifies Chapter 5 of English as the subject of the test, that information is sent to the server, and the server retrieves the relevant data from the database.

[1172] Generate test questions

[1173] server

[1174] Test questions are generated using a generative AI model (e.g., GPT-3) based on a specified range of teaching material data. Questions can be fill-in-the-blank, essay-style, calculation-based, or other formats, and are generated in the appropriate format (LaTeX or HTML). The generated questions are output in both web and paper formats.

[1175] As a concrete example, the server uses GPT-3 to generate fill-in-the-blank questions on the theme of Chapter 5 of English and formats them in HTML format.

[1176] Example prompt sentence:

[1177] Use GPT-3 to generate fill-in-the-blank questions on the theme of English Chapter 5. Format the questions in HTML format so that they can be output in both web and paper formats.

[1178] Conducting the test

[1179] User (student)

[1180] Students access the test via a web browser on their device and enter their answers. For paper-based tests, students fill out the answer sheets, scan them, and upload them to the server from their device.

[1181] For example, students can take an English test online using a browser on their device and enter their answers, or if the answers are paper, they can scan them and upload them to the server.

[1182] Example prompt sentence:

[1183] Students can answer the test online using their device's browser. If the test is paper, they can write their answers, scan them, and upload them to the server from their device.

[1184] Scoring

[1185] server

[1186] Web-based answers are automatically scored as soon as they are submitted. Multiple-choice questions are scored by comparing them with preset correct answer data. Essay questions are evaluated using natural language processing technology (e.g., BERT). Paper-based tests are automatically scored after uploaded scanned data is converted to text using OCR technology (e.g., Tesseract).

[1187] For example, the server analyzes the answer data received and scores multiple-choice questions by comparing them with the correct answer data. For essay questions, natural language analysis is performed using the BERT model to calculate a score. Scanned data is converted into text using an OCR engine, and the results are scored in the same way.

[1188] Example prompt sentence:

[1189] Please use BERT to grade the essay questions that students answer online. For paper-based documents, please scan the data and convert it to text using OCR, and grade it in the same way.

[1190] Analysis of results and feedback

[1191] server

[1192] After all answers have been graded, the data is aggregated and analyzed for correct answer rates and incorrect answer trends. The difficulty level of each question and common error patterns are identified, and an analysis report is generated. This report is generated in JSON or PDF format and sent to the teacher's device.

[1193] As a specific example, the server analyzes the grade data, identifies common trends in incorrect answers, generates a PDF report, and sends the report to the teacher's device.

[1194] Example prompt sentence:

[1195] Please compile the grade data and analyze the percentage of correct answers and trends in incorrect answers. After the analysis, the server will identify common incorrect answer patterns, generate a report, and send it to the teacher's device.

[1196] This system automates the entire process, from importing teaching material data to generating test questions, grading answers, and analyzing and providing feedback on the results, thereby reducing the burden on teachers and improving the quality and efficiency of education.

[1197] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1198] Step 1:

[1199] Importing teaching material data

[1200] The user (educational institution) uploads learning material data (PDF, Docx, etc.) to the server. The input data is in PDF or Docx file format. The server receives the learning material data provided by the educational institution as an HTTP request and stores it in local storage. The server then converts the learning material data into text format using Google's OCR API or Amazon Textract. The converted text is analyzed using natural language processing technology (e.g., spaCy or NLTK) to extract important keywords and concepts. This extracted data is then stored in a MySQL database.

[1201] Specific behavior:

[1202] Teaching material data provided by educational institutions is uploaded.

[1203] The teaching material data received by the server is saved in the "Received" folder in the local storage.

[1204] Calls Google's OCR API to convert PDF and Docx to text.

[1205] The converted text is analyzed using a natural language processing library to extract keywords and concepts.

[1206] The extracted data is stored in the "Keywords" table in the MySQL database.

[1207] Input: Teaching material data (PDF, Docx, etc.)

[1208] Output: Keywords and concepts stored in the database

[1209] Step 2:

[1210] Specifying the scope of questions

[1211] The user (teacher) logs into the system using the browser on their device and selects a specific chapter or page number from the interface. The input data is the subject matter of the test specified by the teacher. The device sends the specified subject matter information to the server as an API request. The server analyzes the received API request and queries the database to retrieve the teaching material data that falls within the specified range. The retrieved teaching material data is stored in a temporary storage area.

[1212] Specific behavior:

[1213] The teacher accesses the "Specify Question Area" page in a browser and specifies the chapter and page number.

[1214] The device converts the specified question scope information into a JSON format API request and sends it to the server.

[1215] The server runs an SQL query against the database to retrieve the data corresponding to the specified chapter or page.

[1216] The acquired data is stored in a temporary storage area.

[1217] Input: Question range information specified by the teacher

[1218] Output: Teaching material data stored in the temporary storage area

[1219] Step 3:

[1220] Generate test questions

[1221] The server generates test questions using a generative AI model (e.g., GPT-3) based on the specified range of teaching material data. The input data is the specified range of teaching material data. The generative AI model receives this and generates test questions in an appropriate format, such as fill-in-the-blank questions, essay questions, or calculation questions. The generated test questions are formatted in formats such as LaTeX or HTML and output in each test format (web format and paper format).

[1222] Specific behavior:

[1223] The server acquires the range-specified teaching material data from the temporary storage area.

[1224] The GPT-3 model is called and test questions are generated using the teaching material data as input.

[1225] The generated questions are formatted into LaTeX or HTML depending on the question format.

[1226] Save formatted test questions in web and paper formats.

[1227] Input: Range-specified teaching material data

[1228] Output: Formatted test questions (LaTeX, HTML)

[1229] Step 4:

[1230] Conducting the test

[1231] Users (students) access the web-based test using a browser on their device and enter their answers. The input data is the student's answers. In the case of paper-based tests, students fill out the answer sheets, scan them, and upload them to the server from their device.

[1232] Specific behavior:

[1233] Students access the test page via the HTTPS protocol and enter their answers in their browser.

[1234] The entered answer data is sent to the server.

[1235] If the answer sheet is paper, scan it and convert it to JPEG or PDF format.

[1236] The converted file is sent to the server via an upload form.

[1237] Input: Student answers (web or paper)

[1238] Output: Answer data sent to the server

[1239] Step 5:

[1240] Scoring

[1241] The server automatically grades the web-based answers as they are sent. The input data is the student's answer data. Multiple-choice questions are compared with preset correct answer data and automatically graded. Essay questions are evaluated using natural language processing technology such as BERT. For paper-based tests, uploaded scanned data is converted to text using OCR technology (e.g., Tesseract) and then graded automatically.

[1242] Specific behavior:

[1243] The server analyzes the received answer data and scores the multiple-choice questions by comparing them with the correct answer data.

[1244] For essay questions, the BERT model is used to perform natural language analysis and calculate a score.

[1245] The scanned data is run through an OCR engine, converted into text, and scored.

[1246] Input: Student answer data (web format, paper scanned data)

[1247] Output:Scoring results

[1248] Step 6:

[1249] Analysis of results and feedback

[1250] After all answers have been graded, the server aggregates the grade data and analyzes the percentage of correct answers and trends in incorrect answers. The input data is the students' grade data. The server identifies the difficulty level and incorrect answer patterns for each question, generates an analysis report, and saves it in JSON or PDF format. The generated feedback report is sent to the teacher's device.

[1251] Specific behavior:

[1252] The server retrieves the performance data and performs data analysis using Python's pandas library.

[1253] Analyze the difficulty of the questions and common error patterns and generate a report.

[1254] The report is saved in JSON or PDF format and sent to the teacher's device using the SMTP protocol.

[1255] Input: Student performance data

[1256] Output: Analysis report (JSON, PDF format)

[1257] (Application example 1)

[1258] 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."

[1259] In today's educational environment, teachers' workloads are increasing, often making it difficult to provide effective teaching. Tasks such as determining the scope of study, creating test questions, grading answers, and providing feedback on results require time and effort. Therefore, there is an urgent need to automate these tasks and reduce the burden on teachers. Furthermore, with the spread of online education, there is also a need to create an environment where students can use smartphones to study efficiently. Therefore, a system is needed to significantly improve the efficiency of all educational activities and improve the quality of education.

[1260] 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.

[1261] In this invention, the server includes means for receiving teaching material data provided by educational institutions and storing it in a database, means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, means for organizing the teaching material data based on the test scope specified by the teacher, means for generating test questions from the test scope using a generative AI model, means for formatting the test questions in both paper and web format, means for automatically scoring the results of students' answers, means for aggregating and analyzing performance data and providing feedback to teachers, means for specifying the test scope using a smartphone app and sending information to the server, and means for students to access web-based tests and submit answers using their smartphones. This automates all educational activities, significantly reducing the burden on teachers and enabling efficient and effective education.

[1262] An "educational institution" is an organization such as a school or university that provides a place for learning and research.

[1263] "Teaching material data" refers to data on learning materials provided by educational institutions, including electronic files in PDF and Docx formats.

[1264] A "database" is a system for efficiently storing and managing teaching material data, grade data, etc.

[1265] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and is used for preprocessing text data and extracting important keywords.

[1266] "Important keywords and concepts" are words and ideas in the teaching material data that are considered to be particularly important for learning.

[1267] A "teacher" is a professional who educates students, and is responsible for specifying the scope of test questions and guiding educational activities.

[1268] "Test Coverage" refers to the specific chapters or areas that are covered when creating test questions.

[1269] A "generative AI model" is a program that automatically generates test questions using artificial intelligence technology.

[1270] "Test questions" are questions used to evaluate students' learning outcomes, and include fill-in-the-blank questions and essay-style questions.

[1271] "Paper media" refers to printed paper forms, including forms on which handwritten answers are written.

[1272] "Web format" refers to an online format delivered over the Internet and accessible through a browser or app.

[1273] "Automatic marking" is the process by which a computer automatically evaluates and scores student answers.

[1274] "Performance data" refers to numerical and evaluation data that represent test results.

[1275] "Feedback" refers to evaluations and guidance comments provided based on performance data.

[1276] A "smartphone" is a type of mobile phone that can connect to the Internet and run applications.

[1277] A "server" is a computer system that provides information over a network.

[1278] "OCR technology" is a technology that extracts text information from an image and is used to convert handwritten answers written on paper into text data.

[1279] The following describes in detail an embodiment of the present invention. This system uses a smartphone app and a server based on educational material data provided by an educational institution to efficiently generate test questions, grade answers, and provide feedback on performance data.

[1280] Importing teaching material data

[1281] server

[1282] Teaching material data (PDF, Docx, etc.) provided by educational institutions is received and stored in a database. During this process, an API is set up using Python and Django to receive the teaching material data. After the teaching material data is saved, it is converted into text format using natural language processing technology, and important keywords and concepts are extracted and stored in the database. Specifically, PDF data is converted into text using PDFMiner, and Docx data is converted into text using python-docx.

[1283] Specifying the scope of questions

[1284] User (Teacher)

[1285] Teachers use a smartphone app to input specific chapters and pages to specify the test area, and the app sends this information to a server. The smartphone app is developed using a cross-platform framework such as React Native or Flutter.

[1286] server

[1287] The received range information is analyzed, and the teaching material data within the specified range is extracted from the database.

[1288] Generate test questions

[1289] server

[1290] Based on the specified range, a generative AI model (using TensorFlow) analyzes past test data and question trends to automatically generate test questions. During this process, the generative AI model is given prompts such as the following:

[1291] Example prompt sentence:

[1292] "Scope: Questions will be based on Chapter 2, Overview of Ancient Greek Philosophy. Multiple choice, fill-in-the-blank, and essay questions will be created. The exam format has been optimized taking into account past exam questions."

[1293] The generated test questions are formatted and saved in both paper and web formats.

[1294] Conducting the test

[1295] User (student)

[1296] Students access the web-based test via a smartphone app and enter their answers, which are then sent to a server in real time.

[1297] Scoring

[1298] server

[1299] For web-based tests, automatic scoring is performed in real time as the answers are submitted. For essay-style questions, scoring is performed using natural language processing technology. For paper-based tests, the scanned data is converted to text using OCR technology (using OpenCV) and then automatically scored.

[1300] Analysis of results and feedback

[1301] server

[1302] The system compiles each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. It identifies the difficulty of each question and common error patterns and generates an analysis report. The feedback report is sent to the teacher's device and used as a reference for planning the next lesson.

[1303] This system automates all teaching activities, significantly reducing the burden on teachers, and generates questions that reflect past exam trends, improving the quality of teaching.

[1304] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1305] Step 1:

[1306] Importing teaching material data

[1307] The server receives learning material data (PDF, Docx, etc.) provided by educational institutions and stores it in a database. Specifically, an API endpoint is set up using the Django framework to receive the learning material data. The received data is converted to text format using PDFMiner for PDF and python-docx for Docx. The input here is a PDF or Docx file, and the output is text data.

[1308] Step 2:

[1309] Preprocessing of teaching material data

[1310] The server preprocesses the converted text data using natural language processing techniques to extract important keywords and concepts. This is done using natural language processing libraries such as NLTK. Specific operations include tokenizing the text data, morphological analysis, removing stop words, and extracting feature words. The input is the text data obtained in step 1, and the output is a list of keywords and concepts.

[1311] Step 3:

[1312] Specifying the scope of questions

[1313] The user (teacher) uses a smartphone app to input the test range, such as a specific chapter or page. The specified range information is sent from the device to the server. The input is the test range information specified by the teacher, and the output is the range information stored on the server. The smartphone app is developed using React Native and Flutter.

[1314] Step 4:

[1315] Analysis of the test scope

[1316] The server analyzes the received range information and retrieves the corresponding teaching material data from the database. Based on the specific chapters and pages included in the range information, it searches for and extracts related text data. The input is the range information obtained in step 3, and the output is text data belonging to the specific range.

[1317] Step 5:

[1318] Generate test questions

[1319] The server uses a generative AI model (using TensorFlow) to automatically generate appropriate test questions (multiple choice, fill-in-the-blank, essay, etc.) from the test range. Specifically, the questions are automatically created using a specific algorithm, taking into account past test data and question trends. During this process, the following prompt sentence is input into the generative AI model.

[1320] Example prompt: "Scope: Questions will be drawn from Chapter 2, Overview of Ancient Greek Philosophy. Multiple choice, fill-in-the-blank, and essay questions will be created. The exam format will be optimized based on past exam questions."

[1321] The input is a specific range of text data and a prompt, and the output is automatically generated test questions.

[1322] Step 6:

[1323] Test question formatting

[1324] The server formats the generated test questions in both paper and web formats. In the case of web format, it converts them into a format that can be displayed in a browser using HTML and CSS, and in the case of paper format, it converts them into a format for printing. The input is the test questions created in step 5, and the output is the formatted test question data.

[1325] Step 7:

[1326] Conducting the test

[1327] Users (students) access web-based tests through a smartphone app, enter their answers, and submit them. The smartphone app provides a browser-based display and input interface. During this process, answers are sent to the server in real time. The input is the student's answer, which is the data sent to the server in real time. The output is the submitted answer data.

[1328] Step 8:

[1329] Automatic scoring

[1330] The server automatically grades the answer data submitted by students in real time. Multiple choice and fill-in-the-blank questions are automatically evaluated by comparing them with preset answers, and essay questions are graded using natural language processing technology. The input is the submitted answer data, and the output is the graded results.

[1331] Step 9:

[1332] Result collection and feedback

[1333] The server aggregates each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. Data analysis tools (such as Pandas and NumPy) are used for the analysis. A feedback report is generated based on the performance data and sent to the teacher's device. The input is the automatically graded result data, and the output is the feedback report.

[1334] 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.

[1335] An embodiment of the present invention will be described below. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. The detailed operational process required for this will be explained step by step. In addition, by combining it with an emotion engine that recognizes the user's emotions, a more advanced feedback function can be achieved.

[1336] Importing teaching material data

[1337] server

[1338] The system receives and stores the educational material data (PDF, Docx, etc.) provided by educational institutions. It then converts the data into text format using natural language processing technology. This allows it to extract important keywords and concepts and store them in a database.

[1339] Specifying the scope of questions

[1340] User (Teacher)

[1341] The teacher uses a terminal to input specific chapter and page numbers to specify the test area, and this information is sent from the terminal to the server.

[1342] Terminal

[1343] The specified range is sent to the server as data.

[1344] server

[1345] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[1346] Generate test questions

[1347] server

[1348] Based on the exam scope, a generative AI model generates test questions. It references data from past entrance exams and mock exams to create questions that reflect key points. It then formats the questions in both paper and web formats.

[1349] Conducting the test

[1350] User (student)

[1351] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[1352] Scoring

[1353] server

[1354] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[1355] Emotion recognition

[1356] server

[1357] The emotion engine analyzes language patterns and facial expression data (if required) in students' responses sent from their devices to recognize their emotional state. This information is stored in a database along with their grades.

[1358] Analysis of results and feedback

[1359] server

[1360] The system compiles each student's academic and emotional data, calculating the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, and emotional trends, and creates an analysis report.

[1361] server

[1362] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test administration. In addition, emotional data is taken into account, and feedback is reflected in which questions students are stressed or excited about.

[1363] Providing feedback

[1364] Terminal

[1365] Feedback reports are sent to teachers' devices and made available for viewing. Teachers can use this information to plan lessons and identify which issues specific students responded to emotionally, allowing them to provide individualized support.

[1366] Specific examples

[1367] For example, when administering a test on "Calculus" in Mathematics II, the system works as follows: When a teacher specifies the "Calculus chapter" as the test area, the generative AI model creates multiple-choice and essay questions from that area. Students answer the online test, and the emotion engine identifies areas of stress from their facial expressions and language patterns. Based on the obtained performance and emotion data, the server generates a detailed analysis report, which the teacher can use to understand points to emphasize in the next lesson and individual student needs.

[1368] In this way, the system of the present invention comprehensively covers everything from automatic generation of test questions to scoring, result analysis, and feedback incorporating emotion recognition, thereby improving the quality of education.

[1369] The processing flow will be explained below.

[1370] Step 1:

[1371] server

[1372] Receives teaching material data (PDF, Docx, etc.) provided by educational institutions and stores it in a database.

[1373] Step 2:

[1374] server

[1375] Natural language processing technology is applied to the received teaching material data to convert it into text format. After conversion, important keywords and concepts are extracted and this information is stored in a database as structured data.

[1376] Step 3:

[1377] User (Teacher)

[1378] The teacher uses a terminal to input specific chapter and page numbers to specify the subject of the questions, and this information is sent from the terminal to the server.

[1379] Step 4:

[1380] Terminal

[1381] The specified range is sent to the server as data.

[1382] Step 5:

[1383] server

[1384] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[1385] Step 6:

[1386] server

[1387] The generative AI model generates test questions based on the scope of the questions, referencing data from past entrance exams and mock exams to create questions that reflect important points.

[1388] Step 7:

[1389] server

[1390] The generated test questions are formatted to be compatible with both paper and web formats. Print data is created for paper media, and online test data is created for web formats.

[1391] Step 8:

[1392] User (student)

[1393] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[1394] Step 9:

[1395] server

[1396] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[1397] Step 10:

[1398] server

[1399] An emotion engine is used to recognize a student's emotional state by analyzing language patterns and facial expression data (if necessary) in the student's answers.

[1400] Step 11:

[1401] server

[1402] The system compiles each student's academic and emotional data, calculating the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, and emotional trends, and creates an analysis report.

[1403] Step 12:

[1404] server

[1405] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test administration. In addition, emotional data is taken into account, and feedback is reflected in which questions students are stressed or excited about.

[1406] Step 13:

[1407] Terminal

[1408] Feedback reports are sent to teachers' devices and made available for viewing. Teachers can use this information to plan lessons and identify which issues specific students responded to emotionally, allowing them to provide individualized support.

[1409] Specific examples

[1410] For example, when conducting a test on "Calculus" in Mathematics II, the system operates as follows.

[1411] In steps 1 to 5, the teacher specifies the "calculus chapter" as the test area, and the generative AI model creates multiple-choice and essay questions from that area.

[1412] In steps 6 to 8, students complete a web-based test that is automatically scored in real time. Essay questions are also scored using natural language processing technology.

[1413] In steps 9 to 12, the emotion engine analyzes the students' emotional state and creates an analysis report along with their grade data. This report is sent to the teacher's device and can be used as a reference for the next lesson or test.

[1414] In this way, the system of the present invention comprehensively performs everything from automatic generation of test questions to grading, result analysis, and feedback that incorporates emotion recognition, thereby improving the quality of education.

[1415] Example 2

[1416] 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."

[1417] Conventional educational systems have had problems with the time and effort required to import teaching material data, create test questions, automate grading, and provide feedback. Furthermore, feedback that ignores the student's emotional state can hinder effective education. The present invention aims to solve these problems and enable efficient and effective education.

[1418] The specification process by the specification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for receiving teaching material data provided by an educational institution and storing it in a database, a means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, and a means for recognizing the emotional state of a student using an emotion engine and storing it together with grade data. This makes it possible to efficiently use teaching materials provided by educational institutions and provide feedback that takes into consideration the student's emotions.

[1419] "Teaching material data" refers to data containing educational information provided by educational institutions, and is provided in formats such as PDF and Docx.

[1420] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language, and is used for preprocessing text data and extracting important keywords and concepts.

[1421] A "generative AI model" is a model that uses machine learning algorithms to generate test questions, and is a technology that creates new questions based on past data and teacher specifications.

[1422] The "test scope" refers to the scope of specific educational content designated by the teacher when creating test questions, and is designated by a specific chapter, page number, or the like.

[1423] "Test questions" are questions or assignments given to assess students' understanding, and are available in multiple choice or essay formats and in paper and web formats.

[1424] "Automatic scoring" is a technology in which a computer automatically evaluates and scores students' answers, and it can be applied to multiple-choice and essay questions.

[1425] The "emotion engine" is a technology that analyzes language patterns and facial expression data collected from devices to recognize students' emotional states.

[1426] "Academic performance data" refers to information recorded as numerical or statistical data on student test results, and is used to evaluate education.

[1427] "Feedback" is information provided to teachers based on test results and emotional data, and is used to improve the quality of education.

[1428] "OCR technology" is a technology that extracts character information from scanned image data and converts it into text data.

[1429] The present invention is a system that receives learning material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. Furthermore, by combining it with an emotion engine that recognizes students' emotions, a more advanced feedback function is realized. Detailed embodiments of this system are described below.

[1430] Importing teaching material data

[1431] server

[1432] The system receives and stores educational material data (PDF, Docx, etc.) provided by educational institutions. It then uses Optical Character Recognition (OCR) technology and the Natural Language Toolkit (NLTK) library to convert this data into text format using natural language processing techniques. It then extracts important keywords and concepts from the converted text and stores them in a database.

[1433] Specific operation example

[1434] The server receives PDF files containing teaching material data for "high school mathematics," converts them into text using OCR technology, and then uses the NLTK library to extract important keywords such as "calculus" and "function" from the text and store them in a database.

[1435] Specifying the scope of questions

[1436] User (Teacher)

[1437] The teacher uses a device to input chapter and page numbers to identify the test topics, and this information is sent from the device to the server.

[1438] Terminal

[1439] Send the question range data to the server.

[1440] server

[1441] The system analyzes the received information on the scope of the exam, extracts the teaching material data within the specified scope from the database, and identifies important points and topics.

[1442] Specific operation example

[1443] The teacher types "Mathematics II Calculus chapter" into the interface, and the device sends the data to the server, which extracts the relevant content from the database.

[1444] Generate test questions

[1445] server

[1446] Generative AI models (e.g., natural language generation models) are used to generate test questions based on the specified subject area. Data from past entrance exams and mock exams is also referenced to ensure high quality questions. The generated questions are formatted in multiple-choice and essay formats for both paper and web versions.

[1447] Specific operation example

[1448] The server generates problems related to the "Chapter on Calculus" by providing a prompt to the AI ​​model. For example, the prompt might be, "Please create three problems to find differential coefficients and two written questions to find definite integrals." The generated problems are provided in both web page and PDF formats.

[1449] Conducting the test

[1450] User (student)

[1451] Students access the test via a web browser and enter their answers. If the test is paper, they scan the answers and upload them to the server.

[1452] Specific operation example

[1453] Students access the test using a web browser on their device and answer questions, or, for paper tests, scan their answers and upload them to the server.

[1454] Scoring

[1455] server

[1456] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper answers are scanned and converted to text using OCR technology, and then graded automatically.

[1457] Specific operation example

[1458] The server instantly grades answers entered online. Written questions are also graded using natural language processing algorithms. Paper answers are also converted to text using OCR technology and automatically graded.

[1459] Emotion recognition

[1460] server

[1461] An emotion engine is used to recognize students' emotional states by analyzing language patterns and facial expression data in their answers, and this information is stored in a database along with their performance data.

[1462] Specific operation example

[1463] The server analyzes the student's answers and uses an emotion engine to analyze language patterns to determine "difficulty" and "stress." If necessary, it also uses the device's camera to analyze facial expression data and grasp the student's overall emotional state.

[1464] Analysis of results and feedback

[1465] server

[1466] Each student's academic performance and emotional data are compiled, and an analysis report is generated by calculating the correct answer rate, incorrect answer rate, difficulty level of each question, emotional trends, etc.

[1467] Specific operation example

[1468] The server analyzes each student's answer data and emotional data and creates a report containing specific information such as "Many students felt stressed on question 3."

[1469] Providing feedback

[1470] Terminal

[1471] Feedback reports are sent to teachers' devices for viewing, including which questions elicited emotional reactions from students.

[1472] Specific operation example

[1473] The teacher checks the feedback report provided on the device and, based on the information that "many students are having difficulty with Problem 3," explains Problem 3 again in the next lesson.

[1474] In this way, the system of the present invention comprehensively covers everything from importing teaching material data to generating test questions, marking, recognizing emotions, analyzing results, and providing feedback, thereby realizing advanced educational support.

[1475] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1476] Step 1: Importing teaching material data

[1477] server

[1478] The server receives and stores the learning material data provided by educational institutions. This data is usually provided in PDF or Docx format. OCR technology and the NLTK library are used to convert the received learning material data into text format. Important keywords and concepts are extracted from the converted text and stored in a database.

[1479] Input PDF or Docx teaching materials data provided by the educational institution

[1480] Output: Text-formatted teaching material data and extracted keyword and concept data

[1481] Specific operation example

[1482] The server receives PDF files containing teaching materials for high school mathematics, converts them into text using OCR technology, and then analyzes the text using the NLTK library to extract keywords such as "calculus" and "function," which are then stored in a database.

[1483] Step 2: Specify the question scope

[1484] User (Teacher)

[1485] The teacher uses a terminal to input chapter and page numbers to specify the test area, and this data is sent to the server via the terminal.

[1486] Input Question information entered by the teacher on the device

[1487] Output Sending question scope information to the server

[1488] Terminal

[1489] The device receives data on the subject matter specified by the teacher and sends it to the server.

[1490] server

[1491] The server analyzes the received information on the test scope, extracts the relevant study material data from the database, and then identifies important points and topics.

[1492] Input question range information

[1493] Output: Extracted teaching material data and key points

[1494] Specific operation example

[1495] When a teacher types "calculus chapter in Mathematics II" into the device interface, the device sends this information to the server, which retrieves the corresponding teaching material data from the database and extracts key points.

[1496] Step 3: Generate test questions

[1497] server

[1498] The server uses a generative AI model to generate test questions based on the specified test scope. Questions are usually created by referencing data from past entrance exams and mock exams. Test questions include both multiple-choice and essay questions. The questions are formatted in both paper and web formats.

[1499] Input question scope information and related teaching material data

[1500] Output Generated test questions

[1501] Specific operation example

[1502] Based on the "Chapter on Calculus," the server instructs the generative AI model to "create three problems to find differential coefficients and two written problems to find definite integrals." The generative AI model generates problems from the specified range and outputs them in PDF and web page format.

[1503] Step 4: Testing

[1504] User (student)

[1505] Students access the web-based test from their devices and enter their answers. If the test is paper-based, they write their answers, scan them, and upload them to the server from their devices.

[1506] Input Test questions provided by the server

[1507] Output Student Answers

[1508] Specific operation example

[1509] Students access the test on a web browser using their device and enter their answers to each question, or if the test is paper, scan the answers and send them to the server.

[1510] Step 5: Scoring

[1511] server

[1512] The server automatically grades web-based answers in real time. For essay questions, natural language processing technology is used to grade. For paper-based answers, scanned data is converted to text using OCR technology and then graded automatically.

[1513] Input student answer data

[1514] Output Scoring results

[1515] Specific operation example

[1516] The server automatically grades multiple-choice questions answered online. Essay questions are also evaluated using natural language processing algorithms. Paper-based answers are also converted to text using OCR technology and graded automatically.

[1517] Step 6: Recognize emotions

[1518] server

[1519] An emotion engine is used to recognize students' emotional states by analyzing language patterns and facial expression data in their answers, and this information is stored in a database along with their performance data.

[1520] Input: Student answers and facial expression data

[1521] Output: Student emotional state data

[1522] Specific operation example

[1523] The server analyzes the answer data and identifies "difficulty" and "stress" from language patterns. It also analyzes facial expression data collected by the device's camera to identify the student's emotional state.

[1524] Step 7: Analyze results and provide feedback

[1525] server

[1526] Each student's academic performance and emotional data are compiled, and an analysis report is generated by calculating the correct answer rate, incorrect answer rate, difficulty level of each question, emotional trends, etc.

[1527] Input: Achievement data and emotion data

[1528] Output Analysis Report

[1529] Specific operation example

[1530] Based on the grade data and emotional data, the server creates a specific analysis report, such as "Many students feel stressed about question 3."

[1531] Step 8: Provide feedback

[1532] Terminal

[1533] Feedback reports are sent to teachers' devices for viewing, including which questions elicited emotional reactions from students.

[1534] Input Analysis Report

[1535] Output Teacher feedback report

[1536] Specific operation example

[1537] The teacher reviews the feedback report provided on the device and takes measures such as "paying particular attention to question 3" in the next lesson.

[1538] (Application example 2)

[1539] 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."

[1540] Traditional test question generation systems based on teaching material data provided by educational institutions are effective in improving student learning outcomes in the field of education. However, particularly in manufacturing, different applications are required, such as product quality inspection and identification of defective products. Existing systems are specialized for education and do not address the real-time quality control and operator emotion recognition feedback required in manufacturing. Therefore, there is a need for a means to ensure quality and effectively manage operators in manufacturing.

[1541] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving teaching material data provided by an educational institution and storing it in a database, means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, means for generating test questions from a test range using a generative AI model, means for collecting production data and inspecting product quality in real time, means for identifying defective products and analyzing and reporting the results, and means for recognizing operator emotions and reflecting the data in feedback. This enables automation of quality control processes and operator stress management in manufacturing sites, as well as learning support in the education field.

[1542] "Teaching material data" refers to learning materials provided by educational institutions, specifically documents provided in PDF or Docx format.

[1543] "Natural language processing technology" refers to technology that enables computers to understand, interpret, and generate human language.

[1544] "Important keywords and concepts" refer to words and ideas that deserve particular attention in the teaching material data.

[1545] A "generative AI model" refers to an algorithm or system that learns from large amounts of data and then makes inferences and generates information based on new data.

[1546] "Test questions" refers to questions or question sets used to assess the level of understanding of the learning content.

[1547] "Automated scoring" refers to the process of using machines or software to automatically evaluate and score answers.

[1548] "Aggregation and analysis" refers to the process of collecting data, processing it using statistical methods, etc., and deriving meaning from it.

[1549] "Real-time product quality inspection" refers to the process of instantly evaluating the quality of products during the manufacturing process.

[1550] "Identification of defective products" refers to identifying products that do not meet quality standards based on inspection criteria.

[1551] "Emotion recognition" refers to the process of determining the emotional state of an operator or student from their facial expressions and behavior.

[1552] "Feedback" refers to providing information that is useful for next actions based on the results and data of a system or process.

[1553] This system automates quality control and operator emotion recognition in manufacturing sites. Specifically, the system is composed of the following hardware and software:

[1554] First, the server must have high-performance data processing capabilities, and it is desirable to use an NVIDIA GPU server. This server has the ability to receive and store data from the production line in real time. The data types used are diverse, including image data, temperature data, and pressure data.

[1555] Data analysis is performed using Python with either TensorFlow or PyTorch from Plads, and Hugging Face's Transformers library is used to train and infer the generative AI model, which automatically generates quality control metrics based on the various data received.

[1556] In addition, the server uses Amazon Rekognition's facial recognition and emotion analysis functions to analyze the operator's emotions in real time, allowing it to understand the stress and other emotional states the operator experiences while working and manage the operator's state as well as product quality.

[1557] As a specific example, consider the case of inspecting the quality of products in a factory. Images, temperature data, and pressure data of products produced on the production line are sent to a server in real time. The server analyzes this data, and a generative AI model automatically sets quality inspection standards. Based on the inspection standards, the products are inspected and passed or failed products are identified. If a failed product is found, the results are analyzed in detail and reported.

[1558] Amazon Rekognition also analyzes the operator's facial expressions based on the inspection results and data from the work, and estimates their emotional state. For example, if an operator is feeling high stress while processing a defective product, that data will also be reflected in the feedback.

[1559] Example prompt sentence:

[1560] "What are the main inspection items for product quality?"

[1561] "Estimate the stress level that the operator experiences from this task."

[1562] In this way, the system of the present invention enhances the quality control process in manufacturing sites and also manages the emotions of operators, thereby improving not only product quality but also worker efficiency and satisfaction.

[1563] The flow of the specific processing in the application example 2 will be described with reference to FIG.

[1564] Step 1:

[1565] The server receives and stores production data, such as product images, temperature data, and pressure data, in real time from the production line. First, it stores the input data from sensors and cameras in a database, which then forms a basic data set for subsequent data analysis.

[1566] Step 2:

[1567] The server performs preprocessing to analyze the received data. Specific operations include noise removal from the image data and smoothing of temperature and pressure data. The preprocessed data is then input into a generative AI model to generate inspection standards.

[1568] Step 3:

[1569] Using the generative AI model, the server sets the product quality inspection criteria. It uses preprocessed product images, temperature, and pressure data as input data, and generates pass / fail criteria based on the criteria learned by the model. The output is specific inspection items and their pass / fail criteria.

[1570] Step 4:

[1571] The server inspects the product based on the generated inspection standards. Specifically, it analyzes product images and takes into account temperature and pressure data to determine whether or not there are any defects. The output is the inspection results, and if there are any defects, detailed data is provided.

[1572] Step 5:

[1573] The server uses Amazon Rekognition for facial recognition and emotion analysis to recognize the operator's emotions in real time. The operator's video data is input, and the emotion engine analyzes their stress level and other emotional states. The output is the operator's emotional state, which is then saved as data.

[1574] Step 6:

[1575] The server integrates product inspection results and operator sentiment data and generates a detailed analysis report based on them. Using the inspection results and sentiment data as input, the analysis tool aggregates and analyzes the data. As output, a detailed analysis report is created for management to use in the next round of inspections and to improve operator work.

[1576] Step 7:

[1577] The terminal provides the generated analysis report to the manager. Specifically, the report is visualized and displayed in a format that is easy for the manager to understand. Based on this, the manager can plan measures to improve quality control and the operator's working environment.

[1578] Through the above processing steps, the system highly automates quality control and operator emotional management on the production line.

[1579] 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.

[1580] 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.

[1581] 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.

[1582] [Fourth embodiment]

[1583] FIG. 7 shows an example of the configuration of a data processing system 410 according to the fourth embodiment.

[1584] 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.

[1585] 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).

[1586] 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.

[1587] 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.

[1588] 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).

[1589] 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.

[1590] 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.

[1591] 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.

[1592] 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.

[1593] 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.

[1594] 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.

[1595] 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."

[1596] The following describes an embodiment of the present invention. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs scoring and result analysis. The detailed operational process required for this will be explained step by step.

[1597] Importing teaching material data

[1598] server

[1599] The system receives and stores the educational material data (PDF, Docx, etc.) provided by educational institutions. It then converts the data into text format using natural language processing technology. This allows it to extract important keywords and concepts and store them in a database.

[1600] Specifying the scope of questions

[1601] User (Teacher)

[1602] The teacher uses a terminal to input the specific chapters and pages to be tested, and this information is sent from the terminal to the server.

[1603] Terminal

[1604] The specified range is sent to the server as data.

[1605] server

[1606] The received range information is analyzed, and the teaching material data within the specified range is extracted from the database.

[1607] Generate test questions

[1608] server

[1609] Based on the specified range of teaching material data, the system selects an appropriate question format, such as fill-in-the-blank, essay, or calculation questions. The generative AI model analyzes trends from past entrance exams and mock exams to create test questions, which are then formatted in both paper and web formats.

[1610] Conducting the test

[1611] User (student)

[1612] Students access the test via a web browser and enter their answers. If the test is on paper, they can scan the answers and upload them to the server.

[1613] Scoring

[1614] server

[1615] For web-based tests, automatic scoring is performed in real time as answers are submitted. For essay-style questions, natural language processing technology is used for scoring. For paper-based tests, scanned data is converted to text using OCR technology and then automatically scored.

[1616] Analysis of results and feedback

[1617] server

[1618] The system compiles each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. It identifies the difficulty of each question and common error patterns and generates an analysis report. Teachers can use this feedback to identify points that should be emphasized in the next lesson.

[1619] Terminal

[1620] Feedback reports are sent to teachers' devices to help them plan their next lessons.

[1621] In this way, this system automates everything from creating test questions to grading, analyzing the results, and providing feedback, greatly improving the efficiency of educational activities. Furthermore, the questions generated reflect past exam trends, making it possible to improve the quality of education.

[1622] The processing flow will be explained below.

[1623] Step 1:

[1624] server

[1625] Receives educational material data provided by educational institutions, including formats such as PDF and Docx. The received data is stored in a database.

[1626] Step 2:

[1627] server

[1628] Natural language processing technology is applied to the saved teaching material data to convert it into text format. After conversion, important keywords and concepts are extracted and this information is stored in a database as structured data.

[1629] Step 3:

[1630] User (Teacher)

[1631] The teacher uses a terminal to input specific chapters and page numbers to specify the range of questions to be asked, and the information is sent from the terminal to the server.

[1632] Step 4:

[1633] server

[1634] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[1635] Step 5:

[1636] server

[1637] The generative AI model generates test questions based on the scope of the questions, referencing data from past entrance exams and mock exams to create questions that reflect important points.

[1638] Step 6:

[1639] server

[1640] The generated test questions are formatted to be compatible with both paper and web formats. Print data is created for paper media, and online test data is created for web formats.

[1641] Step 7:

[1642] User (student)

[1643] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[1644] Step 8:

[1645] server

[1646] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[1647] Step 9:

[1648] server

[1649] Compile the performance data of each student, calculate the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, etc., and create an analysis report.

[1650] Step 10:

[1651] server

[1652] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test.

[1653] Step 11:

[1654] Terminal

[1655] Feedback reports are sent to teachers' devices and made available for viewing, allowing teachers to use them to plan their lessons.

[1656] This allows the system to efficiently carry out a series of steps, from automatically generating test questions, to grading, analyzing the results, and providing feedback, thereby improving the quality of education while significantly reducing the burden on teachers.

[1657] Example 1

[1658] 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."

[1659] Conventional educational systems require a great deal of manual work, from importing teaching material data to creating test questions, grading, and analyzing the results, placing a heavy burden on teachers. It was also difficult to generate test questions that reflected past exam trends, or to provide real-time grading and feedback. This raised concerns that the quality and efficiency of education would decline.

[1660] 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.

[1661] In this invention, the server includes means for receiving teaching material data provided by educational institutions and storing it in a database, means for preprocessing the teaching material data using natural language processing technology to extract important keywords and concepts, means for organizing the teaching material data based on the test scope specified by the teacher, means for generating test questions from the test scope using a generative AI model, means for formatting the test questions in formats such as LaTeX or HTML depending on the question format, means for providing the generated test questions in both paper and web formats, means for automatically scoring the results of students' answers, and means for aggregating and analyzing performance data and providing feedback to teachers. This automates the entire process from importing teaching material data to generating test questions, scoring, and analyzing and providing feedback on the results, reducing the burden on teachers and improving the quality and efficiency of education.

[1662] "Educational institutions" refers to organizations that provide education and training, such as schools, universities, and cram schools.

[1663] "Teaching material data" refers to materials for educational purposes provided by educational institutions, including file formats such as PDF and Docx.

[1664] "Database" refers to a system designed to efficiently manage and retrieve data within an organization.

[1665] "Natural language processing technology" refers to the field of technology in which computers understand, interpret, and generate human language.

[1666] "Keywords" refer to words and phrases that are considered particularly important within a document and are used for information retrieval and data classification.

[1667] "Test area" refers to the area set for test questions, such as specific chapters or page numbers designated by the teacher.

[1668] A "generative AI model" refers to an algorithm or model that uses artificial intelligence technology to generate data or analyze information.

[1669] "Test questions" refer to questions or tasks designed to assess a student's understanding or knowledge.

[1670] "Formatting" refers to the process of converting documents or data into an appropriate format to make them easier to read.

[1671] "Web format" refers to a format that is provided over the Internet and can be viewed and manipulated in a browser.

[1672] "Paper media" refers to information or data printed on paper.

[1673] "Automatic scoring" refers to the process of mechanically evaluating and scoring students' answers using a special program.

[1674] "Academic performance data" refers to data that records student test results and grades in numerical or written form.

[1675] "Feedback" refers to assessments and information provided by systems and teachers to help improve students and lessons.

[1676] The following describes an embodiment of the present invention. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. The detailed operational process required for this purpose is specifically shown.

[1677] Importing teaching material data

[1678] server

[1679] The server receives learning material data (PDF, Docx, etc.) provided by educational institutions via HTTP requests and stores it in local storage. It then converts the learning material data into text format using Google's OCR API or Amazon Textract. From the converted text, natural language processing technology (e.g., spaCy or NLTK) is used to extract important keywords and concepts, which are then stored in a MySQL database. As a concrete example, if a school provides a PDF file as an English learning material, the server receives the PDF, converts it into text format using Google's OCR API, and stores the extracted words and phrases in the database.

[1680] Example prompt sentence:

[1681] The system should receive English learning materials in PDF format provided by educational institutions, convert them into text using Google OCR API, extract important words and phrases from the text, and store them in a database.

[1682] Specifying the scope of questions

[1683] User (Teacher)

[1684] Teachers log in to the system using their device's browser and enter the question scope by specifying a specific chapter or page number from the interface.

[1685] Terminal

[1686] The data for the question area specified by the teacher is obtained from the input form and sent to the server as an API request.

[1687] server

[1688] Analyze the received API request, query the database to retrieve the learning material data within the specified range, and store the retrieved learning material data in a temporary storage area.

[1689] As a specific example, when a teacher specifies Chapter 5 of English as the subject of questions, the terminal sends that information to the server, and the server retrieves data corresponding to Chapter 5 from the database.

[1690] Example prompt sentence:

[1691] If a teacher specifies Chapter 5 of English as the subject of the test, that information is sent to the server, and the server retrieves the relevant data from the database.

[1692] Generate test questions

[1693] server

[1694] Test questions are generated using a generative AI model (e.g., GPT-3) based on a specified range of teaching material data. Questions can be fill-in-the-blank, essay-style, calculation-based, or other formats, and are generated in the appropriate format (LaTeX or HTML). The generated questions are output in both web and paper formats.

[1695] As a concrete example, the server uses GPT-3 to generate fill-in-the-blank questions on the theme of Chapter 5 of English and formats them in HTML format.

[1696] Example prompt sentence:

[1697] Use GPT-3 to generate fill-in-the-blank questions on the theme of English Chapter 5. Format the questions in HTML format so that they can be output in both web and paper formats.

[1698] Conducting the test

[1699] User (student)

[1700] Students access the test via a web browser on their device and enter their answers. For paper-based tests, students fill out the answer sheets, scan them, and upload them to the server from their device.

[1701] For example, students can take an English test online using a browser on their device and enter their answers, or if the answers are paper, they can scan them and upload them to the server.

[1702] Example prompt sentence:

[1703] Students can answer the test online using their device's browser. If the test is paper, they can write their answers, scan them, and upload them to the server from their device.

[1704] Scoring

[1705] server

[1706] Web-based answers are automatically scored as soon as they are submitted. Multiple-choice questions are scored by comparing them with preset correct answer data. Essay questions are evaluated using natural language processing technology (e.g., BERT). Paper-based tests are automatically scored after uploaded scanned data is converted to text using OCR technology (e.g., Tesseract).

[1707] For example, the server analyzes the answer data received and scores multiple-choice questions by comparing them with the correct answer data. For essay questions, natural language analysis is performed using the BERT model to calculate a score. Scanned data is converted into text using an OCR engine, and the results are scored in the same way.

[1708] Example prompt sentence:

[1709] Please use BERT to grade the essay questions that students answer online. For paper-based documents, please scan the data and convert it to text using OCR, and grade it in the same way.

[1710] Analysis of results and feedback

[1711] server

[1712] After all answers have been graded, the data is aggregated and analyzed for correct answer rates and incorrect answer trends. The difficulty level of each question and common error patterns are identified, and an analysis report is generated. This report is generated in JSON or PDF format and sent to the teacher's device.

[1713] As a specific example, the server analyzes the grade data, identifies common trends in incorrect answers, generates a PDF report, and sends the report to the teacher's device.

[1714] Example prompt sentence:

[1715] Please compile the grade data and analyze the percentage of correct answers and trends in incorrect answers. After the analysis, the server will identify common incorrect answer patterns, generate a report, and send it to the teacher's device.

[1716] This system automates the entire process, from importing teaching material data to generating test questions, grading answers, and analyzing and providing feedback on the results, thereby reducing the burden on teachers and improving the quality and efficiency of education.

[1717] The flow of the identification process in the first embodiment will be described with reference to FIG.

[1718] Step 1:

[1719] Importing teaching material data

[1720] The user (educational institution) uploads learning material data (PDF, Docx, etc.) to the server. The input data is in PDF or Docx file format. The server receives the learning material data provided by the educational institution as an HTTP request and stores it in local storage. The server then converts the learning material data into text format using Google's OCR API or Amazon Textract. The converted text is analyzed using natural language processing technology (e.g., spaCy or NLTK) to extract important keywords and concepts. This extracted data is then stored in a MySQL database.

[1721] Specific behavior:

[1722] Teaching material data provided by educational institutions is uploaded.

[1723] The teaching material data received by the server is saved in the "Received" folder in the local storage.

[1724] Calls Google's OCR API to convert PDF and Docx to text.

[1725] The converted text is analyzed using a natural language processing library to extract keywords and concepts.

[1726] The extracted data is stored in the "Keywords" table in the MySQL database.

[1727] Input: Teaching material data (PDF, Docx, etc.)

[1728] Output: Keywords and concepts stored in the database

[1729] Step 2:

[1730] Specifying the scope of questions

[1731] The user (teacher) logs into the system using the browser on their device and selects a specific chapter or page number from the interface. The input data is the subject matter of the test specified by the teacher. The device sends the specified subject matter information to the server as an API request. The server analyzes the received API request and queries the database to retrieve the teaching material data that falls within the specified range. The retrieved teaching material data is stored in a temporary storage area.

[1732] Specific behavior:

[1733] The teacher accesses the "Specify Question Area" page in a browser and specifies the chapter and page number.

[1734] The device converts the specified question scope information into a JSON format API request and sends it to the server.

[1735] The server runs an SQL query against the database to retrieve the data corresponding to the specified chapter or page.

[1736] The acquired data is stored in a temporary storage area.

[1737] Input: Question range information specified by the teacher

[1738] Output: Teaching material data stored in the temporary storage area

[1739] Step 3:

[1740] Generate test questions

[1741] The server generates test questions using a generative AI model (e.g., GPT-3) based on the specified range of teaching material data. The input data is the specified range of teaching material data. The generative AI model receives this and generates test questions in an appropriate format, such as fill-in-the-blank questions, essay questions, or calculation questions. The generated test questions are formatted in formats such as LaTeX or HTML and output in each test format (web format and paper format).

[1742] Specific behavior:

[1743] The server acquires the range-specified teaching material data from the temporary storage area.

[1744] The GPT-3 model is called and test questions are generated using the teaching material data as input.

[1745] The generated questions are formatted into LaTeX or HTML depending on the question format.

[1746] Save formatted test questions in web and paper formats.

[1747] Input: Range-specified teaching material data

[1748] Output: Formatted test questions (LaTeX, HTML)

[1749] Step 4:

[1750] Conducting the test

[1751] Users (students) access the web-based test using a browser on their device and enter their answers. The input data is the student's answers. In the case of paper-based tests, students fill out the answer sheets, scan them, and upload them to the server from their device.

[1752] Specific behavior:

[1753] Students access the test page via the HTTPS protocol and enter their answers in their browser.

[1754] The entered answer data is sent to the server.

[1755] If the answer sheet is paper, scan it and convert it to JPEG or PDF format.

[1756] The converted file is sent to the server via an upload form.

[1757] Input: Student answers (web or paper)

[1758] Output: Answer data sent to the server

[1759] Step 5:

[1760] Scoring

[1761] The server automatically grades the web-based answers as they are sent. The input data is the student's answer data. Multiple-choice questions are compared with preset correct answer data and automatically graded. Essay questions are evaluated using natural language processing technology such as BERT. For paper-based tests, uploaded scanned data is converted to text using OCR technology (e.g., Tesseract) and then graded automatically.

[1762] Specific behavior:

[1763] The server analyzes the received answer data and scores the multiple-choice questions by comparing them with the correct answer data.

[1764] For essay questions, the BERT model is used to perform natural language analysis and calculate a score.

[1765] The scanned data is run through an OCR engine, converted into text, and scored.

[1766] Input: Student answer data (web format, paper scanned data)

[1767] Output:Scoring results

[1768] Step 6:

[1769] Analysis of results and feedback

[1770] After all answers have been graded, the server aggregates the grade data and analyzes the percentage of correct answers and trends in incorrect answers. The input data is the students' grade data. The server identifies the difficulty level and incorrect answer patterns for each question, generates an analysis report, and saves it in JSON or PDF format. The generated feedback report is sent to the teacher's device.

[1771] Specific behavior:

[1772] The server retrieves the performance data and performs data analysis using Python's pandas library.

[1773] Analyze the difficulty of the questions and common error patterns and generate a report.

[1774] The report is saved in JSON or PDF format and sent to the teacher's device using the SMTP protocol.

[1775] Input: Student performance data

[1776] Output: Analysis report (JSON, PDF format)

[1777] (Application example 1)

[1778] 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."

[1779] In today's educational environment, teachers' workloads are increasing, often making it difficult to provide effective teaching. Tasks such as determining the scope of study, creating test questions, grading answers, and providing feedback on results require time and effort. Therefore, there is an urgent need to automate these tasks and reduce the burden on teachers. Furthermore, with the spread of online education, there is also a need to create an environment where students can use smartphones to study efficiently. Therefore, a system is needed to significantly improve the efficiency of all educational activities and improve the quality of education.

[1780] 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.

[1781] In this invention, the server includes means for receiving teaching material data provided by educational institutions and storing it in a database, means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, means for organizing the teaching material data based on the test scope specified by the teacher, means for generating test questions from the test scope using a generative AI model, means for formatting the test questions in both paper and web format, means for automatically scoring the results of students' answers, means for aggregating and analyzing performance data and providing feedback to teachers, means for specifying the test scope using a smartphone app and sending information to the server, and means for students to access web-based tests and submit answers using their smartphones. This automates all educational activities, significantly reducing the burden on teachers and enabling efficient and effective education.

[1782] An "educational institution" is an organization such as a school or university that provides a place for learning and research.

[1783] "Teaching material data" refers to data on learning materials provided by educational institutions, including electronic files in PDF and Docx formats.

[1784] A "database" is a system for efficiently storing and managing teaching material data, grade data, etc.

[1785] "Natural language processing technology" is a technology that enables computers to understand and analyze human language, and is used for preprocessing text data and extracting important keywords.

[1786] "Important keywords and concepts" are words and ideas in the teaching material data that are considered to be particularly important for learning.

[1787] A "teacher" is a professional who educates students, and is responsible for specifying the scope of test questions and guiding educational activities.

[1788] "Test Coverage" refers to the specific chapters or areas that are covered when creating test questions.

[1789] A "generative AI model" is a program that automatically generates test questions using artificial intelligence technology.

[1790] "Test questions" are questions used to evaluate students' learning outcomes, and include fill-in-the-blank questions and essay-style questions.

[1791] "Paper media" refers to printed paper forms, including forms on which handwritten answers are written.

[1792] "Web format" refers to an online format delivered over the Internet and accessible through a browser or app.

[1793] "Automatic marking" is the process by which a computer automatically evaluates and scores student answers.

[1794] "Performance data" refers to numerical and evaluation data that represent test results.

[1795] "Feedback" refers to evaluations and guidance comments provided based on performance data.

[1796] A "smartphone" is a type of mobile phone that can connect to the Internet and run applications.

[1797] A "server" is a computer system that provides information over a network.

[1798] "OCR technology" is a technology that extracts text information from an image and is used to convert handwritten answers written on paper into text data.

[1799] The following describes in detail an embodiment of the present invention. This system uses a smartphone app and a server based on educational material data provided by an educational institution to efficiently generate test questions, grade answers, and provide feedback on performance data.

[1800] Importing teaching material data

[1801] server

[1802] Teaching material data (PDF, Docx, etc.) provided by educational institutions is received and stored in a database. During this process, an API is set up using Python and Django to receive the teaching material data. After the teaching material data is saved, it is converted into text format using natural language processing technology, and important keywords and concepts are extracted and stored in the database. Specifically, PDF data is converted into text using PDFMiner, and Docx data is converted into text using python-docx.

[1803] Specifying the scope of questions

[1804] User (Teacher)

[1805] Teachers use a smartphone app to input specific chapters and pages to specify the test area, and the app sends this information to a server. The smartphone app is developed using a cross-platform framework such as React Native or Flutter.

[1806] server

[1807] The received range information is analyzed, and the teaching material data within the specified range is extracted from the database.

[1808] Generate test questions

[1809] server

[1810] Based on the specified range, a generative AI model (using TensorFlow) analyzes past test data and question trends to automatically generate test questions. During this process, the generative AI model is given prompts such as the following:

[1811] Example prompt sentence:

[1812] "Scope: Questions will be based on Chapter 2, Overview of Ancient Greek Philosophy. Multiple choice, fill-in-the-blank, and essay questions will be created. The exam format has been optimized taking into account past exam questions."

[1813] The generated test questions are formatted and saved in both paper and web formats.

[1814] Conducting the test

[1815] User (student)

[1816] Students access the web-based test via a smartphone app and enter their answers, which are then sent to a server in real time.

[1817] Scoring

[1818] server

[1819] For web-based tests, automatic scoring is performed in real time as the answers are submitted. For essay-style questions, scoring is performed using natural language processing technology. For paper-based tests, the scanned data is converted to text using OCR technology (using OpenCV) and then automatically scored.

[1820] Analysis of results and feedback

[1821] server

[1822] The system compiles each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. It identifies the difficulty of each question and common error patterns and generates an analysis report. The feedback report is sent to the teacher's device and used as a reference for planning the next lesson.

[1823] This system automates all teaching activities, significantly reducing the burden on teachers, and generates questions that reflect past exam trends, improving the quality of teaching.

[1824] The flow of the specific processing in the application example 1 will be described with reference to FIG.

[1825] Step 1:

[1826] Importing teaching material data

[1827] The server receives learning material data (PDF, Docx, etc.) provided by educational institutions and stores it in a database. Specifically, an API endpoint is set up using the Django framework to receive the learning material data. The received data is converted to text format using PDFMiner for PDF and python-docx for Docx. The input here is a PDF or Docx file, and the output is text data.

[1828] Step 2:

[1829] Preprocessing of teaching material data

[1830] The server preprocesses the converted text data using natural language processing techniques to extract important keywords and concepts. This is done using natural language processing libraries such as NLTK. Specific operations include tokenizing the text data, morphological analysis, removing stop words, and extracting feature words. The input is the text data obtained in step 1, and the output is a list of keywords and concepts.

[1831] Step 3:

[1832] Specifying the scope of questions

[1833] The user (teacher) uses a smartphone app to input the test range, such as a specific chapter or page. The specified range information is sent from the device to the server. The input is the test range information specified by the teacher, and the output is the range information stored on the server. The smartphone app is developed using React Native and Flutter.

[1834] Step 4:

[1835] Analysis of the test scope

[1836] The server analyzes the received range information and retrieves the corresponding teaching material data from the database. Based on the specific chapters and pages included in the range information, it searches for and extracts related text data. The input is the range information obtained in step 3, and the output is text data belonging to the specific range.

[1837] Step 5:

[1838] Generate test questions

[1839] The server uses a generative AI model (using TensorFlow) to automatically generate appropriate test questions (multiple choice, fill-in-the-blank, essay, etc.) from the test range. Specifically, the questions are automatically created using a specific algorithm, taking into account past test data and question trends. During this process, the following prompt sentence is input into the generative AI model.

[1840] Example prompt: "Scope: Questions will be drawn from Chapter 2, Overview of Ancient Greek Philosophy. Multiple choice, fill-in-the-blank, and essay questions will be created. The exam format will be optimized based on past exam questions."

[1841] The input is a specific range of text data and a prompt, and the output is automatically generated test questions.

[1842] Step 6:

[1843] Test question formatting

[1844] The server formats the generated test questions in both paper and web formats. In the case of web format, it converts them into a format that can be displayed in a browser using HTML and CSS, and in the case of paper format, it converts them into a format for printing. The input is the test questions created in step 5, and the output is the formatted test question data.

[1845] Step 7:

[1846] Conducting the test

[1847] Users (students) access web-based tests through a smartphone app, enter their answers, and submit them. The smartphone app provides a browser-based display and input interface. During this process, answers are sent to the server in real time. The input is the student's answer, which is the data sent to the server in real time. The output is the submitted answer data.

[1848] Step 8:

[1849] Automatic scoring

[1850] The server automatically grades the answer data submitted by students in real time. Multiple choice and fill-in-the-blank questions are automatically evaluated by comparing them with preset answers, and essay questions are graded using natural language processing technology. The input is the submitted answer data, and the output is the graded results.

[1851] Step 9:

[1852] Result collection and feedback

[1853] The server aggregates each student's performance data and analyzes the percentage of correct answers and trends in incorrect answers. Data analysis tools (such as Pandas and NumPy) are used for the analysis. A feedback report is generated based on the performance data and sent to the teacher's device. The input is the automatically graded result data, and the output is the feedback report.

[1854] 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.

[1855] An embodiment of the present invention will be described below. This system receives educational material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. The detailed operational process required for this will be explained step by step. In addition, by combining it with an emotion engine that recognizes the user's emotions, a more advanced feedback function can be achieved.

[1856] Importing teaching material data

[1857] server

[1858] The system receives and stores the educational material data (PDF, Docx, etc.) provided by educational institutions. It then converts the data into text format using natural language processing technology. This allows it to extract important keywords and concepts and store them in a database.

[1859] Specifying the scope of questions

[1860] User (Teacher)

[1861] The teacher uses a terminal to input specific chapter and page numbers to specify the test area, and this information is sent from the terminal to the server.

[1862] Terminal

[1863] The specified range is sent to the server as data.

[1864] server

[1865] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[1866] Generate test questions

[1867] server

[1868] Based on the exam scope, a generative AI model generates test questions. It references data from past entrance exams and mock exams to create questions that reflect key points. It then formats the questions in both paper and web formats.

[1869] Conducting the test

[1870] User (student)

[1871] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[1872] Scoring

[1873] server

[1874] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[1875] Emotion recognition

[1876] server

[1877] The emotion engine analyzes language patterns and facial expression data (if required) in students' responses sent from their devices to recognize their emotional state. This information is stored in a database along with their grades.

[1878] Analysis of results and feedback

[1879] server

[1880] The system compiles each student's academic and emotional data, calculating the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, and emotional trends, and creates an analysis report.

[1881] server

[1882] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test administration. In addition, emotional data is taken into account, and feedback is reflected in which questions students are stressed or excited about.

[1883] Providing feedback

[1884] Terminal

[1885] Feedback reports are sent to teachers' devices and made available for viewing. Teachers can use this information to plan lessons and identify which issues specific students responded to emotionally, allowing them to provide individualized support.

[1886] Specific examples

[1887] For example, when administering a test on "Calculus" in Mathematics II, the system works as follows: When a teacher specifies the "Calculus chapter" as the test area, the generative AI model creates multiple-choice and essay questions from that area. Students answer the online test, and the emotion engine identifies areas of stress from their facial expressions and language patterns. Based on the obtained performance and emotion data, the server generates a detailed analysis report, which the teacher can use to understand points to emphasize in the next lesson and individual student needs.

[1888] In this way, the system of the present invention comprehensively covers everything from automatic generation of test questions to scoring, result analysis, and feedback incorporating emotion recognition, thereby improving the quality of education.

[1889] The processing flow will be explained below.

[1890] Step 1:

[1891] server

[1892] Receives teaching material data (PDF, Docx, etc.) provided by educational institutions and stores it in a database.

[1893] Step 2:

[1894] server

[1895] Natural language processing technology is applied to the received teaching material data to convert it into text format. After conversion, important keywords and concepts are extracted and this information is stored in a database as structured data.

[1896] Step 3:

[1897] User (Teacher)

[1898] The teacher uses a terminal to input specific chapter and page numbers to specify the subject of the questions, and this information is sent from the terminal to the server.

[1899] Step 4:

[1900] Terminal

[1901] The specified range is sent to the server as data.

[1902] Step 5:

[1903] server

[1904] The received information on the test scope is analyzed, and the learning material data within the specified scope is extracted from the database. The extracted data is then organized and the important points are identified.

[1905] Step 6:

[1906] server

[1907] The generative AI model generates test questions based on the scope of the questions, referencing data from past entrance exams and mock exams to create questions that reflect important points.

[1908] Step 7:

[1909] server

[1910] The generated test questions are formatted to be compatible with both paper and web formats. Print data is created for paper media, and online test data is created for web formats.

[1911] Step 8:

[1912] User (student)

[1913] Students access the test via a web browser and enter their answers. For paper tests, students scan the answers and upload them to the server.

[1914] Step 9:

[1915] server

[1916] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper-based answers are scanned and converted to text using OCR technology, which is then graded automatically.

[1917] Step 10:

[1918] server

[1919] An emotion engine is used to recognize a student's emotional state by analyzing language patterns and facial expression data (if necessary) in the student's answers.

[1920] Step 11:

[1921] server

[1922] The system compiles each student's academic and emotional data, calculating the percentage of correct answers, the percentage of incorrect answers, the difficulty of each question, and emotional trends, and creates an analysis report.

[1923] Step 12:

[1924] server

[1925] Based on the generated analysis report, feedback is created for teachers, providing information that will be useful for the next lesson or test administration. In addition, emotional data is taken into account, and feedback is reflected in which questions students are stressed or excited about.

[1926] Step 13:

[1927] Terminal

[1928] Feedback reports are sent to teachers' devices and made available for viewing. Teachers can use this information to plan lessons and identify which issues specific students responded to emotionally, allowing them to provide individualized support.

[1929] Specific examples

[1930] For example, when conducting a test on "Calculus" in Mathematics II, the system operates as follows.

[1931] In steps 1 to 5, the teacher specifies the "calculus chapter" as the test area, and the generative AI model creates multiple-choice and essay questions from that area.

[1932] In steps 6 to 8, students complete a web-based test that is automatically scored in real time. Essay questions are also scored using natural language processing technology.

[1933] In steps 9 to 12, the emotion engine analyzes the students' emotional state and creates an analysis report along with their grade data. This report is sent to the teacher's device and can be used as a reference for the next lesson or test.

[1934] In this way, the system of the present invention comprehensively performs everything from automatic generation of test questions to grading, result analysis, and feedback that incorporates emotion recognition, thereby improving the quality of education.

[1935] Example 2

[1936] 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."

[1937] Conventional educational systems have had problems with the time and effort required to import teaching material data, create test questions, automate grading, and provide feedback. Furthermore, feedback that ignores the student's emotional state can hinder effective education. The present invention aims to solve these problems and enable efficient and effective education.

[1938] The specification process by the specification processing unit 290 of the data processing device 12 in the second embodiment is realized by the following means. In this invention, the server includes a means for receiving teaching material data provided by an educational institution and storing it in a database, a means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, and a means for recognizing the emotional state of a student using an emotion engine and storing it together with grade data. This makes it possible to efficiently use teaching materials provided by educational institutions and provide feedback that takes into consideration the student's emotions.

[1939] "Teaching material data" refers to data containing educational information provided by educational institutions, and is provided in formats such as PDF and Docx.

[1940] "Natural language processing technology" is a technology that enables computers to understand, analyze, and generate human language, and is used for preprocessing text data and extracting important keywords and concepts.

[1941] A "generative AI model" is a model that uses machine learning algorithms to generate test questions, and is a technology that creates new questions based on past data and teacher specifications.

[1942] The "test scope" refers to the scope of specific educational content designated by the teacher when creating test questions, and is designated by a specific chapter, page number, or the like.

[1943] "Test questions" are questions or assignments given to assess students' understanding, and are available in multiple choice or essay formats and in paper and web formats.

[1944] "Automatic scoring" is a technology in which a computer automatically evaluates and scores students' answers, and it can be applied to multiple-choice and essay questions.

[1945] The "emotion engine" is a technology that analyzes language patterns and facial expression data collected from devices to recognize students' emotional states.

[1946] "Academic performance data" refers to information recorded as numerical or statistical data on student test results, and is used to evaluate education.

[1947] "Feedback" is information provided to teachers based on test results and emotional data, and is used to improve the quality of education.

[1948] "OCR technology" is a technology that extracts character information from scanned image data and converts it into text data.

[1949] The present invention is a system that receives learning material data provided by educational institutions, automatically generates test questions, and performs grading and result analysis. Furthermore, by combining it with an emotion engine that recognizes students' emotions, a more advanced feedback function is realized. Detailed embodiments of this system are described below.

[1950] Importing teaching material data

[1951] server

[1952] The system receives and stores educational material data (PDF, Docx, etc.) provided by educational institutions. It then uses Optical Character Recognition (OCR) technology and the Natural Language Toolkit (NLTK) library to convert this data into text format using natural language processing techniques. It then extracts important keywords and concepts from the converted text and stores them in a database.

[1953] Specific operation example

[1954] The server receives PDF files containing teaching material data for "high school mathematics," converts them into text using OCR technology, and then uses the NLTK library to extract important keywords such as "calculus" and "function" from the text and store them in a database.

[1955] Specifying the scope of questions

[1956] User (Teacher)

[1957] The teacher uses a device to input chapter and page numbers to identify the test topics, and this information is sent from the device to the server.

[1958] Terminal

[1959] Send the question range data to the server.

[1960] server

[1961] The system analyzes the received information on the scope of the exam, extracts the teaching material data within the specified scope from the database, and identifies important points and topics.

[1962] Specific operation example

[1963] The teacher types "Mathematics II Calculus chapter" into the interface, and the device sends the data to the server, which extracts the relevant content from the database.

[1964] Generate test questions

[1965] server

[1966] Generative AI models (e.g., natural language generation models) are used to generate test questions based on the specified subject area. Data from past entrance exams and mock exams is also referenced to ensure high quality questions. The generated questions are formatted in multiple-choice and essay formats for both paper and web versions.

[1967] Specific operation example

[1968] The server generates problems related to the "Chapter on Calculus" by providing a prompt to the AI ​​model. For example, the prompt might be, "Please create three problems to find differential coefficients and two written questions to find definite integrals." The generated problems are provided in both web page and PDF formats.

[1969] Conducting the test

[1970] User (student)

[1971] Students access the test via a web browser and enter their answers. If the test is paper, they scan the answers and upload them to the server.

[1972] Specific operation example

[1973] Students access the test using a web browser on their device and answer questions, or, for paper tests, scan their answers and upload them to the server.

[1974] Scoring

[1975] server

[1976] Web-based answers are automatically graded in real time. Essay questions are graded using natural language processing technology. Paper answers are scanned and converted to text using OCR technology, and then graded automatically.

[1977] Specific operation example

[1978] The server instantly grades answers entered online. Written questions are also graded using natural language processing algorithms. Paper answers are also converted to text using OCR technology and automatically graded.

[1979] Emotion recognition

[1980] server

[1981] An emotion engine is used to recognize students' emotional states by analyzing language patterns and facial expression data in their answers, and this information is stored in a database along with their performance data.

[1982] Specific operation example

[1983] The server analyzes the student's answers and uses an emotion engine to analyze language patterns to determine "difficulty" and "stress." If necessary, it also uses the device's camera to analyze facial expression data and grasp the student's overall emotional state.

[1984] Analysis of results and feedback

[1985] server

[1986] Each student's academic performance and emotional data are compiled, and an analysis report is generated by calculating the correct answer rate, incorrect answer rate, difficulty level of each question, emotional trends, etc.

[1987] Specific operation example

[1988] The server analyzes each student's answer data and emotional data and creates a report containing specific information such as "Many students felt stressed on question 3."

[1989] Providing feedback

[1990] Terminal

[1991] Feedback reports are sent to teachers' devices for viewing, including which questions elicited emotional reactions from students.

[1992] Specific operation example

[1993] The teacher checks the feedback report provided on the device and, based on the information that "many students are having difficulty with Problem 3," explains Problem 3 again in the next lesson.

[1994] In this way, the system of the present invention comprehensively covers everything from importing teaching material data to generating test questions, marking, recognizing emotions, analyzing results, and providing feedback, thereby realizing advanced educational support.

[1995] The flow of the identification process in the second embodiment will be described with reference to FIG.

[1996] Step 1: Importing teaching material data

[1997] server

[1998] The server receives and stores the learning material data provided by educational institutions. This data is usually provided in PDF or Docx format. OCR technology and the NLTK library are used to convert the received learning material data into text format. Important keywords and concepts are extracted from the converted text and stored in a database.

[1999] Input PDF or Docx teaching materials data provided by the educational institution

[2000] Output: Text-formatted teaching material data and extracted keyword and concept data

[2001] Specific operation example

[2002] The server receives PDF files containing teaching materials for high school mathematics, converts them into text using OCR technology, and then analyzes the text using the NLTK library to extract keywords such as "calculus" and "function," which are then stored in a database.

[2003] Step 2: Specify the question scope

[2004] User (Teacher)

[2005] The teacher uses a terminal to input chapter and page numbers to specify the test area, and this data is sent to the server via the terminal.

[2006] Input Question information entered by the teacher on the device

[2007] Output Sending question scope information to the server

[2008] Terminal

[2009] The device receives data on the subject matter specified by the teacher and sends it to the server.

[2010] server

[2011] The server analyzes the received information on the test scope, extracts the relevant study material data from the database, and then identifies important points and topics.

[2012] Input question range information

[2013] Output: Extracted teaching material data and key points

[2014] Specific operation example

[2015] When a teacher types "calculus chapter in Mathematics II" into the device interface, the device sends this information to the server, which retrieves the corresponding teaching material data from the database and extracts key points.

[2016] Step 3: Generate test questions

[2017] server

[2018] The server uses a generative AI model to generate test questions based on the specified test scope. Questions are usually created by referencing data from past entrance exams and mock exams. Test questions include both multiple-choice and essay questions. The questions are formatted in both paper and web formats.

[2019] Input question scope information and related teaching material data

[2020] Output Generated test questions

[2021] Specific operation example

[2022] Based on the "Chapter on Calculus," the server instructs the generative AI model to "create three problems to find differential coefficients and two written problems to find definite integrals." The generative AI model generates problems from the specified range and outputs them in PDF and web page format.

[2023] Step 4: Testing

[2024] User (student)

[2025] Students access the web-based test from their devices and enter their answers. If the test is paper-based, they write their answers, scan them, and upload them to the server from their devices.

[2026] Input Test questions provided by the server

[2027] Output Student Answers

[2028] Specific operation example

[2029] Students access the test on a web browser using their device and enter their answers to each question, or if the test is paper, scan the answers and send them to the server.

[2030] Step 5: Scoring

[2031] server

[2032] The server automatically grades web-based answers in real time. For essay questions, natural language processing technology is used to grade. For paper-based answers, scanned data is converted to text using OCR technology and then graded automatically.

[2033] Input student answer data

[2034] Output Scoring results

[2035] Specific operation example

[2036] The server automatically grades multiple-choice questions answered online. Essay questions are also evaluated using natural language processing algorithms. Paper-based answers are also converted to text using OCR technology and graded automatically.

[2037] Step 6: Recognize emotions

[2038] server

[2039] An emotion engine is used to recognize students' emotional states by analyzing language patterns and facial expression data in their answers, and this information is stored in a database along with their performance data.

[2040] Input: Student answers and facial expression data

[2041] Output: Student emotional state data

[2042] Specific operation example

[2043] The server analyzes the answer data and identifies "difficulty" and "stress" from language patterns. It also analyzes facial expression data collected by the device's camera to identify the student's emotional state.

[2044] Step 7: Analyze results and provide feedback

[2045] server

[2046] Each student's academic performance and emotional data are compiled, and an analysis report is generated by calculating the correct answer rate, incorrect answer rate, difficulty level of each question, emotional trends, etc.

[2047] Input: Achievement data and emotion data

[2048] Output Analysis Report

[2049] Specific operation example

[2050] Based on the grade data and emotional data, the server creates a specific analysis report, such as "Many students feel stressed about question 3."

[2051] Step 8: Provide feedback

[2052] Terminal

[2053] Feedback reports are sent to teachers' devices for viewing, including which questions elicited emotional reactions from students.

[2054] Input Analysis Report

[2055] Output Teacher feedback report

[2056] Specific operation example

[2057] The teacher reviews the feedback report provided on the device and takes measures such as "paying particular attention to question 3" in the next lesson.

[2058] (Application example 2)

[2059] 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."

[2060] Traditional test question generation systems based on teaching material data provided by educational institutions are effective in improving student learning outcomes in the field of education. However, particularly in manufacturing, different applications are required, such as product quality inspection and identification of defective products. Existing systems are specialized for education and do not address the real-time quality control and operator emotion recognition feedback required in manufacturing. Therefore, there is a need for a means to ensure quality and effectively manage operators in manufacturing.

[2061] The specific processing by the specific processing unit 290 of the data processing device 12 in Application Example 2 is realized by the following means. In this invention, the server includes means for receiving teaching material data provided by an educational institution and storing it in a database, means for preprocessing the teaching material data using natural language processing technology and extracting important keywords and concepts, means for generating test questions from a test range using a generative AI model, means for collecting production data and inspecting product quality in real time, means for identifying defective products and analyzing and reporting the results, and means for recognizing operator emotions and reflecting the data in feedback. This enables automation of quality control processes and operator stress management in manufacturing sites, as well as learning support in the education field.

[2062] "Teaching material data" refers to learning materials provided by educational institutions, specifically documents provided in PDF or Docx format.

[2063] "Natural language processing technology" refers to technology that enables computers to understand, interpret, and generate human language.

[2064] "Important keywords and concepts" refer to words and ideas that deserve particular attention in the teaching material data.

[2065] A "generative AI model" refers to an algorithm or system that learns from large amounts of data and then makes inferences and generates information based on new data.

[2066] "Test questions" refers to questions or question sets used to assess the level of understanding of the learning content.

[2067] "Automated scoring" refers to the process of using machines or software to automatically evaluate and score answers.

[2068] "Aggregation and analysis" refers to the process of collecting data, processing it using statistical methods, etc., and deriving meaning from it.

[2069] "Real-time product quality inspection" refers to the process of instantly evaluating the quality of products during the manufacturing process.

[2070] "Identification of defective products" refers to identifying products that do not meet quality standards based on inspection criteria.

[2071] "Emotion recognition" refers to the process of determining the emotional state of an operator or student from their facial expressions and behavior.

[2072] "Feedback" refers to providing information that is useful for next actions based on the results and data of a system or process.

[2073] This system automates quality control and operator emotion recognition in manufacturing sites. Specifically, the system is composed of the following hardware and software:

[2074] First, the server must have high-performance data processing capabilities, and it is desirable to use an NVIDIA GPU server. This server has the ability to receive and store data from the production line in real time. The data types used are diverse, including image data, temperature data, and pressure data.

[2075] Data analysis is performed using Python with either TensorFlow or PyTorch from Plads, and Hugging Face's Transformers library is used to train and infer the generative AI model, which automatically generates quality control metrics based on the various data received.

[2076] In addition, the server uses Amazon Rekognition's facial recognition and emotion analysis functions to analyze the operator's emotions in real time, allowing it to understand the stress and other emotional states the operator experiences while working and manage the operator's state as well as product quality.

[2077] A...

Claims

1. a means for receiving teaching material data provided by an educational institution and storing the data in a database; A method for preprocessing teaching material data using natural language processing technology and extracting important keywords and concepts. A means for organizing teaching material data based on the range of questions specified by the teacher; A means for generating test questions from a range of questions using a generative AI model; A means of formatting test questions in both paper and web formats; A means for automatically scoring the student's answers; a means of compiling and analyzing performance data and providing feedback to teachers; A system including:

2. 10. The system of claim 1, further comprising means for grading the student's written responses using natural language processing techniques.

3. The system according to claim 1, further comprising means for converting scanned data into text using OCR technology and automatically scoring answers to tests administered on paper.

Images