system

Abstract

We provide the system. [Solution] A means of collecting qualification exam data from an internal database, A method for preprocessing collected data and training a generative AI, A means for automatically generating examination questions for the first and second examinations, A means of distributing the exam and evaluating the examinees' answers in real time, A means of automatically grading test-takers' answers and generating feedback, Analyzing test-taker performance data to improve future exam questions and feedback, A system that includes this.

Description

Technical Field

[0004] , , ,

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

Background Art

[0002] Patent Document 1 discloses a method for controlling a persona chatbot, which is performed by at least one processor, and includes steps of receiving a user utterance, adding the user utterance to a prompt including an instruction sentence related to an explanation of a chatbot character, encoding the prompt, and inputting the encoded prompt into a language model to generate a chatbot utterance as a response to the user utterance.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] In a conventional qualification examination system, a large number of human resources were required from the creation of examination questions to the implementation, scoring, and provision of feedback, so there was a problem that a huge amount of time and cost were required for each examination. In addition, there was also a problem that the evaluation criteria were not consistent and there were variations in the performance evaluation of examinees. Furthermore, since the quality of feedback also depends on the skills of the person in charge, there were cases where appropriate improvement points could not be presented to some examinees.

Means for Solving the Problems

[0005] To solve the above problems, the present invention provides the following means. First, it provides means for collecting qualification exam data from an internal database. This data is preprocessed, and missing or abnormal data is supplemented or corrected as necessary. Next, it provides means for training a generative AI to automatically generate exam questions for the first and second exams. It also provides means for distributing the exams and evaluating examinees' answers in real time, and provides means for automatic scoring and feedback generation. Furthermore, it provides means for collecting and analyzing examinee performance data to improve exam questions and feedback content for future exams. In this way, the present invention automates the entire examination process, thereby reducing human resources and improving the consistency of evaluations, while also enabling the rapid provision of high-quality feedback.

[0006] An "internal database" is a system used to centrally manage and store data held within a company or organization.

[0007] "Qualification examination data" refers to all information related to qualification examinations, including exam questions, answers, exam results, examinee information, and evaluation criteria.

[0008] "Preprocessing" is the process of shaping and transforming data into a format suitable for analysis and learning, and includes imputing missing values ​​and correcting outliers.

[0009] "Generative AI" is an artificial intelligence technology that automatically generates new information and content based on collected and learned data.

[0010] The "first-stage examination" is an initial test designed to assess the applicant's basic knowledge, and typically includes multiple-choice and short-answer questions.

[0011] The "second examination" is a more advanced test designed to assess the applicant's applied skills and practical abilities, and includes scenario-based problems and case studies.

[0012] "Automatic test question generation" is a process that uses artificial intelligence technology to automatically create test questions.

[0013] "Real-time evaluation" is a process that immediately analyzes and evaluates the test-takers' answers.

[0014] "Automated scoring" is a process that uses artificial intelligence or specialized software to automatically grade exam answers.

[0015] "Feedback" refers to response information that includes evaluation results, areas for improvement, and advice regarding the test-taker's answers.

[0016] "Performance data" refers to data that includes evaluation metrics such as the test results and the quality of answers of test takers.

[0017] "Analysis" is the process of thoroughly examining and evaluating collected data to derive patterns and trends.

[0018] "Improvement" refers to making changes or modifications to existing methods or processes to make them more effective and efficient. [Brief explanation of the drawing]

[0019] [Figure 1] This is a conceptual diagram showing an example of the configuration of a data processing system according to the first embodiment. [Figure 2] This is a conceptual diagram showing an example of the essential functions of a data processing device and a smart device according to the first embodiment. [Figure 3] This is a conceptual diagram showing an example of the configuration of a data processing system according to the second embodiment. [Figure 4] This is a conceptual diagram showing an example of the main functions of a data processing device and smart glasses according to the second embodiment. [Figure 5] This is a conceptual diagram showing an example of the configuration of a data processing system according to the third embodiment. [Figure 6] This is a conceptual diagram showing an example of the main functions of a data processing device and a headset-type terminal according to the third embodiment. [Figure 7]It is a conceptual diagram showing an example of the configuration of a data processing system according to the fourth embodiment. [Figure 8] It is a conceptual diagram showing an example of the main functions of a data processing device and a robot according to the fourth embodiment. [Figure 9] It shows an emotion map to which a plurality of emotions are mapped. [Figure 10] It shows an emotion map to which a plurality of emotions are mapped. [Figure 11] It is a sequence diagram showing the processing flow of the data processing system in Example 1. [Figure 12] It is a sequence diagram showing the processing flow of the data processing system in Application Example 1. [Figure 13] It is a sequence diagram showing the processing flow of the data processing system in Example 2 when an emotion engine is combined. [Figure 14] It is a sequence diagram showing the processing flow of the data processing system in Application Example 2 when an emotion engine is combined.

Modes for Carrying Out the Invention

[0020] Hereinafter, an example of an embodiment of a system according to the technology of the present disclosure will be described with reference to the accompanying drawings.

[0021] First, the terms used in the following description will be explained.

[0022] In the following embodiments, the signed processor (hereinafter simply referred to as "processor") may be a single arithmetic unit or a combination of multiple arithmetic units. Furthermore, the processor may be a single type of arithmetic unit or a combination of multiple types of arithmetic units. Examples of arithmetic units include CPU (Central Processing Unit), GPU (Graphics Processing Unit), GPGPU (General-Purpose computing on Graphics Processing Units), and APU (Accelerated Processing Unit).

[0023] In the following embodiments, signed RAM (Random Access Memory) is a memory that temporarily stores information and is used as work memory by the processor.

[0024] In the following embodiments, the signed storage is one or more non-volatile storage devices that store various programs and various parameters. Examples of non-volatile storage devices include flash memory (SSD (Solid State Drive)), magnetic disks (e.g., hard disks), or magnetic tapes.

[0025] In the following embodiments, the signed communication interface (I / F) is an interface that includes a communication processor and an antenna, etc. The communication interface manages communication between multiple computers. Examples of communication standards applicable to the communication interface include wireless communication standards such as 5G (5th Generation Mobile Communication System), Wi-Fi (registered trademark), or Bluetooth (registered trademark).

[0026] In the following embodiments, "A and / or B" is synonymous with "at least one of A and B." That is, "A and / or B" means that it may be A alone, or B alone, or a combination of A and B. Furthermore, in this specification, the same concept as "A and / or B" applies when expressing three or more things linked by "and / or."

[0027] [First Embodiment]

[0028] Figure 1 shows an example of the configuration of the data processing system 10 according to the first embodiment.

[0029] As shown in Figure 1, the 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.

[0030] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0031] The smart device 14 comprises a computer 36, a reception device 38, an output device 40, a camera 42, and a communication interface 44. The computer 36 comprises a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The reception device 38, output device 40, and camera 42 are also connected to the bus 52.

[0032] The reception device 38 is equipped with a touch panel 38A and a microphone 38B, etc., and receives user input. The touch panel 38A receives user input by detecting contact with an object (e.g., a pen or finger). The microphone 38B receives user input by detecting the user's voice. The control unit 46A transmits data indicating the user input received by the touch panel 38A and 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.

[0033] 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 perceptible to the user 20 (e.g., audio and / or text). The display 40A displays visible information such as text and images according to instructions from the processor 46. The speaker 40B outputs audio according to instructions from the processor 46. The camera 42 is a small digital camera equipped with an optical system such as a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor.

[0034] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various types of information between processor 46 and processor 28 via network 54.

[0035] Figure 2 shows an example of the main functions of the data processing device 12 and the smart device 14.

[0036] As shown in Figure 2, in the data processing device 12, a specific processing 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" related to the technology of this 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 according to the specific processing program 56 executed on the RAM 30.

[0037] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0038] In the smart device 14, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. The reception output program 60 is used in conjunction with a 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0039] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0040] The system of the present invention aims to automate qualification examinations, with a server at the center executing multiple functions. The specific procedure for implementing the present invention is shown below.

[0041] Collection and preprocessing of qualification exam data

[0042] The server collects data related to certification exams from internal databases within the company. This includes past exam questions, examinee results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. Here, it checks for any missing or abnormal data and supplements or corrects it as needed. This preprocessed data is then used directly as training data for generative AI.

[0043] Learning of generative AI

[0044] The server trains a generative AI based on pre-processed data. During this process, the AI ​​learns the evaluation criteria for certification exams, patterns in past exam questions, and the quality of answers. This creates a foundation for the AI ​​to evaluate test-takers' responses and generate appropriate exam questions.

[0045] Automatic generation of exam questions

[0046] Once the training is complete, the generative AI automatically generates questions for the first and second exams according to the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge. Specific examples include questions such as, "What are the basic steps of the sales process?" The second exam includes scenario-based questions and case studies to assess practical skills. For example, it generates practical questions such as, "What is the best way to handle a customer complaint?"

[0047] Exam distribution and administration

[0048] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers through their device. During the exam, the server receives the examinee's answers in real time and performs automatic evaluation. Additional questions may be automatically generated based on the accuracy rate and quality of the answers. This process continues into the second exam after the first exam is completed.

[0049] Grading and feedback

[0050] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. The automated scoring system evaluates each answer and calculates a total score. Based on the evaluation results, the server generates detailed feedback. This feedback includes an evaluation of each question, the test-taker's strengths, and areas for improvement. The server compiles this feedback for each test-taker and sends it to the user (test-taker) via email.

[0051] Performance data analysis and feedback improvement

[0052] The server collects and analyzes test-taker performance data. This analysis is then used to improve future exam questions and feedback. In addition, the server collects reviews from test-takers and incorporates them as training data for generative AI, thereby improving the overall system quality.

[0053] As described above, the server-centered automated system streamlines the execution of certification exams and enables the rapid provision of high-quality feedback. This system not only saves human resources but also contributes to improved consistency in evaluations and enhanced skills among test-takers.

[0054] The following describes the processing flow.

[0055] Step 1:

[0056] The server collects data related to the certification exam from its internal database. Specifically, it retrieves past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, it checks for any missing or abnormal data and supplements or corrects it as needed. This pre-processed data is then used as training data for generative AI.

[0057] Step 2:

[0058] The server trains the generative AI based on pre-processed data. Specifically, it trains the AI ​​on evaluation criteria for certification exams, patterns of past exam questions, and the quality of answers. Through this learning process, the generative AI builds a foundation for creating appropriate exam questions and evaluating answers.

[0059] Step 3:

[0060] The server uses generative AI to automatically generate questions for the first and second exams. For the first exam, it generates multiple-choice and short-answer questions to assess basic knowledge. For the second exam, it generates scenario-based questions and case studies to assess practical skills. For example, it generates questions such as, "What are the basic steps of the sales process?" and "What is the best way to handle a customer complaint?"

[0061] Step 4:

[0062] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers using their device. During the exam, the server receives the examinee's answers in real time and performs automatic evaluation. Based on this evaluation, additional questions may be automatically generated as needed.

[0063] Step 5:

[0064] Once a user completes the first exam, the server immediately delivers the questions for the second exam. The user then takes the second exam within the same Zoom session. The answers to the second exam are also evaluated in real time, and the evaluation results are continuously fed back to the user.

[0065] Step 6:

[0066] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. Using an automated scoring system, it evaluates the accuracy and quality of each answer and calculates an overall score. Based on the evaluation results, detailed feedback is generated. The feedback includes an evaluation of each question, the test-taker's strengths, and areas for improvement.

[0067] Step 7:

[0068] The server compiles this feedback for each test taker and sends it to the user (test taker) via email. The feedback is also recorded in the company's internal database and used as reference material for future tests and evaluations.

[0069] Step 8:

[0070] The server collects and analyzes test-takers' performance data. This analysis is used to improve future test questions and feedback. Furthermore, the server collects reviews from test-takers and incorporates them as AI training data, thereby improving the overall quality of the system.

[0071] Through the steps outlined above, this system automates the entire certification examination process, reducing human resources and improving the consistency of evaluations. It also enables the rapid provision of high-quality feedback to test-takers.

[0072] (Example 1)

[0073] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0074] Traditional certification examination systems require significant human resources and are inefficient because the creation of exam questions, evaluation of answers, and provision of feedback are all done manually. Furthermore, consistency in evaluation and quality of feedback are prone to inconsistencies. Additionally, the analysis of test-taker performance data and the improvement of feedback based on that analysis are not adequately performed. This hinders the improvement of test-takers' abilities and fair evaluation.

[0075] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0076] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data to supplement and correct any deficiencies or anomalies, means for training a generative AI based on the preprocessed data, means for automatically generating exam questions for the first and second exams in a specified format, means for distributing the exams and evaluating examinees' answers in real time, means for automatically scoring examinees' answers, generating detailed feedback, and sending it via email, and means for analyzing examinee performance data to improve the next exam questions and feedback. This enables automation and efficiency of the exam, improved consistency of evaluation, and rapid provision of high-quality feedback.

[0077] An "internal database" refers to a collection of information and data related to certification exams that is maintained within a specific organization or system.

[0078] "Qualification exam data" refers to a variety of data related to the exam, including past exam questions, examinee results, evaluation criteria, and examinee performance data.

[0079] "Preprocessing" refers to the process of performing tasks such as imputing missing values, correcting outliers, and formatting the collected data.

[0080] "Generative AI" refers to artificial intelligence systems that learn from collected and pre-processed data and automatically generate new test questions and answers.

[0081] The term "first-stage examination" refers to the initial stage of an exam that includes multiple-choice and short-answer questions designed to assess basic knowledge.

[0082] The term "second examination" refers to the second stage of the examination, which includes scenario-based problems and case studies designed to assess practical skills.

[0083] "Exam delivery" refers to the process of providing exam questions to test takers and having them take the exam.

[0084] "Real-time evaluation" refers to a process where the answers entered by test-takers during the exam are immediately analyzed and evaluated.

[0085] "Automated scoring" refers to a process that automatically evaluates and scores test-takers' answers to exam questions.

[0086] "Feedback" refers to information generated based on a test-taker's exam results, including evaluation results and suggestions for improvement.

[0087] "Performance data" refers to data related to test-takers' behavior during the exam, the quality of their answers, and the accuracy of their responses.

[0088] "Analysis" refers to the process of examining collected data and extracting meaningful patterns and trends.

[0089] The system of this invention aims to automate qualification examinations and performs various processes centered around a server. This system is implemented in the following specific steps.

[0090] Collection and preprocessing of qualification exam data

[0091] The server first connects to an internal database to collect data related to the certification exam. This includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. Specifically, it checks for missing or abnormal data and then fills in or corrects them. For example, missing values ​​are filled in with the mean, and abnormal values ​​are corrected appropriately. Python's Pandas library and SQL queries are used for this preprocessing.

[0092] Learning of generative AI

[0093] The pre-processed data is used as training data for generative AI. The server uses this data to train the AI ​​model. The AI ​​models used include advanced generative AIs such as GPT-4®. Deep learning frameworks such as TENSORFLOW® and PyTorch are utilized in this process, and the model learns patterns and evaluation criteria from past exam questions. As a result, the AI ​​acquires the ability to automatically generate the next exam questions.

[0094] Automatic generation of exam questions

[0095] Once the generative AI has completed its training, it automatically generates questions for the first and second exams according to the specified exam format. For example, the first exam generates multiple-choice and short-answer questions to assess basic knowledge. On the other hand, the second exam includes scenario-based questions and case studies to assess practical skills.

[0096] Example of a prompt

[0097] "What are the basic steps of the sales process?"

[0098] "What is the best way to handle a customer complaint that arises during customer service?"

[0099] Exam distribution and administration

[0100] Next, the server handles the delivery and administration of the exam. Users (examinees) log in to ZOOM from their devices at the designated exam date and time. The server delivers the first exam question set through the ZOOM session, and users answer it. Answers are sent to the server in real time, and the server evaluates them immediately. In some cases, additional questions may be automatically generated based on the accuracy rate and quality of the answers.

[0101] Grading and feedback

[0102] After the exam is completed, the server compiles all the answers and evaluates them using an automated scoring system. Based on the evaluation results, detailed feedback is generated and sent to the user (test taker) via email. This feedback includes an evaluation of each question, the test taker's strengths, and areas for improvement.

[0103] Performance data analysis and feedback improvement

[0104] The server collects and analyzes test-takers' performance data. This analysis is used to improve the creation of future test questions and feedback. Furthermore, reviews from test-takers are collected and incorporated as training data for generative AI, thereby improving the overall quality of the system.

[0105] As described above, the present invention's automated qualification examination system automates a series of processes centered on a server, including data collection, preprocessing, training of a generative AI, automatic generation of examination questions, distribution and administration of the examination, scoring and generation of feedback, and analysis and improvement of performance data, thereby realizing efficient and high-quality examination management.

[0106] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0107] Step 1: Start data collection

[0108] The server establishes a connection to the internal database and executes an SQL query to extract "certification exam data" (past exam questions, examinee results, evaluation criteria, etc.). For example, it might execute a query like "SELECT FROM ExamData WHERE Year BETWEEN(registered trademark) 2018 AND 2023".

[0109] Input: Connection information for the internal database

[0110] Output: Extraction results of qualification exam data

[0111] Step 2: Data checking and preprocessing

[0112] The server reads the collected data using Python's Pandas library and checks for missing values ​​and outliers. For example, it checks for missing values ​​using df.isnull().sum() and imputes them with df.fillna(df.mean()). If outliers exist, it corrects them using an appropriate method.

[0113] Input: Certification exam data extracted in Step 1

[0114] Output: Preprocessed data

[0115] Step 3: Preparing training data for generative AI

[0116] The server organizes the preprocessed data as training data for generative AI and converts it into an applicable format. For example, it selects the necessary features and converts the data into numerical vectors.

[0117] Input: Preprocessed data obtained in Step 2

[0118] Output: Training dataset

[0119] Step 4: Training the generative AI model

[0120] The server uses deep learning frameworks such as TensorFlow or PyTorch to train generative AI models. For example, it trains the model using the form model.fit(X_train, y_train, epochs=100).

[0121] Input: Training dataset prepared in Step 3

[0122] Output: Trained generative AI model

[0123] Step 5: Automatic generation of primary exam questions

[0124] The server uses a pre-trained generative AI model to automatically generate questions for the first stage of the exam. For example, it might create a multiple-choice question such as, "What are the basic steps of the sales process?"

[0125] Input: Pre-trained generative AI model

[0126] Output: First Exam Question Set

[0127] Step 6: Automatic generation of secondary examination questions

[0128] The server uses a pre-trained generative AI model to automatically generate questions for the second stage of the exam. For example, it creates scenario questions such as, "What is the best way to handle a customer complaint that arises during customer service?"

[0129] Input: Pre-trained generative AI model

[0130] Output: Secondary Exam Question Set

[0131] Step 7: Notification of exam date and time

[0132] The server will send a notification email to the test taker containing the date, time, and link to join the ZOOM session.

[0133] Input: Candidate information, exam schedule

[0134] Output: Notification email

[0135] Step 8: Exam delivery and real-time collection of test-takers' responses

[0136] The user (examinee) logs into ZOOM from their device at the designated date and time. The server delivers the first-stage exam question set via the ZOOM session, and the user enters their answers. The server collects and evaluates the answers sent from the device in real time.

[0137] Input: First-stage exam question set, examinee's answers

[0138] Output: Real-time evaluation results

[0139] Step 9: Generating additional questions based on accuracy and quality.

[0140] The server evaluates the accuracy rate and quality of answers based on the responses received in real time, and generates additional questions as needed.

[0141] Input: Test-takers' responses, real-time evaluation results

[0142] Output: Additional questions

[0143] Step 10: Aggregation of test results and generation of feedback

[0144] After the exam is completed, the server compiles all the answers and calculates an overall score using an automated scoring system. Next, it generates detailed feedback and sends it via email.

[0145] Input: All responses from test takers

[0146] Output: Overall score, feedback email

[0147] Step 11: Accumulate and analyze performance data

[0148] The server collects and analyzes test-takers' performance data. This analysis is used to create future exam questions and improve feedback.

[0149] Input: Test-taker's performance data

[0150] Output: analysis results, improvement suggestions

[0151] Through the processing steps described above, the system of the present invention can automate and streamline qualification examinations and provide high-quality feedback quickly.

[0152] (Application Example 1)

[0153] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart device 14 will be referred to as the "terminal."

[0154] Traditional certification exams and training programs require significant human resources for tasks such as creating exam questions, administering exams, grading, and providing feedback. This often leads to inefficient exam administration and the risk of biased human evaluation. Furthermore, in on-site employee training, such as in factories, individual evaluation and feedback are difficult to provide, hindering effective training. To address these challenges, an efficient and consistent automated evaluation system is needed.

[0155] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0156] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and training a generative AI, means for automatically generating exam questions for the first and second exams, means for distributing the exams and evaluating examinees' answers in real time, means for automatically scoring examinees' answers and generating feedback, means for analyzing examinee performance data and improving the next exam questions and feedback, and means for a robot to generate training questions similar to qualification exams and provide real-time feedback during factory employee training. This streamlines the employee training and evaluation process and enables the provision of consistent, high-quality feedback.

[0157] An "internal database" is a database that stores data related to certification exams and allows users to retrieve that data as needed.

[0158] "Qualification exam data" refers to all data related to the administration of qualification exams, including past exam questions, examinee results, evaluation criteria, and examinee performance data.

[0159] "Preprocessing" is the process of supplementing and correcting any deficiencies or anomalies in the collected data and converting it into a format suitable for learning by generative AI.

[0160] "Generative AI" is a type of artificial intelligence that uses large amounts of data to train models and generate test questions.

[0161] The "first-stage examination" is an exam that includes multiple-choice and short-answer questions designed to assess basic knowledge.

[0162] The "second examination" is a test that includes scenario-based problems and case studies, designed to evaluate practical skills.

[0163] "Automatic scoring" is a process that automatically evaluates the test-takers' answers and calculates their scores.

[0164] "Feedback" refers to providing detailed comments on the test-taker's answers, along with an evaluation of their responses, including their strengths and areas for improvement.

[0165] "Performance data" refers to the overall results a test-taker demonstrated during the exam, along with various analytical data based on those results.

[0166] "Factory employee training" refers to the process of educating and providing factory employees with the necessary skills and knowledge to perform their duties.

[0167] "Real-time evaluation" is a process in which, each time a test-taker submits their answer, it is immediately evaluated and the results are provided.

[0168] "Training problems" are questions or tasks presented to assess an employee's abilities.

[0169] To implement this invention, it is first necessary to collect data related to qualification exams from an internal database. The server retrieves data such as past exam questions, examinee results, and evaluation criteria from the database and preprocesses this data. In this preprocessing, any missing or abnormal data is supplemented and corrected, and the data is standardized.

[0170] Next, the server uses the pre-processed data to train a generative AI. A natural language processing (NLP) model, such as GPT-3®, is suitable as a generative AI model. Based on the training data, the AI ​​model learns the question patterns and evaluation criteria of certification exams, enabling the automatic generation of exam questions.

[0171] Once the training is complete, the generative AI generates questions for the first and second exams based on the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge, including specific examples such as "What are basic safety procedures?" The second exam generates scenario questions and case studies to assess practical skills, such as "Explain how to respond if an emergency occurs during work."

[0172] The generated test questions are delivered to factory employees' terminals and robots via a server. Employees provide answers to the terminals or robots, and the server evaluates the answers in real time. An AI-powered automated scoring system is used to evaluate the answers, and feedback is generated based on this evaluation. The feedback includes not only whether the answer is correct or incorrect, but also the employee's strengths and areas for improvement.

[0173] Furthermore, the server collects employee performance data and analyzes it to improve the quality of future exam questions and feedback. This analysis utilizes statistical analysis tools (such as Python's Pandas and NumPy) to identify data trends and develop more effective training programs.

[0174] As a concrete example, a factory robot presents an employee with a training question such as, "Explain how to use the emergency stop button," and the employee answers using voice or controls. The AI ​​evaluates the answer in real time and provides feedback such as, "The procedure itself is correct, but safety checks are missing." In this way, employees can learn practical skills on the spot while being efficiently evaluated.

[0175] Examples of prompt messages include the following:

[0176] Please generate the following training questions regarding factory machine operation:

[0177] 1. Explain how to operate the machine according to the safety procedures.

[0178] 2. Please explain how to use the emergency stop button.

[0179] 3. Please describe the procedure for routine maintenance checks.

[0180] The above describes the specific procedures and system configuration necessary for implementing this invention.

[0181] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0182] Step 1:

[0183] The server collects data related to certification exams from an internal database. The data retrieved from the database includes past exam questions, examinee results, evaluation criteria, and examinee performance data. All data related to the certification exams is retrieved from the database as input. The collected dataset is passed to a preprocessing module as output.

[0184] Step 2:

[0185] The server preprocesses the collected data. Preprocessing includes imputing missing values, correcting anomalous data, and normalization. The input is the data collected in step 1, and the output is the preprocessed dataset. At this stage, data cleansing and normalization are performed using data preprocessing libraries (e.g., Pandas, NumPy).

[0186] Step 3:

[0187] The server trains a generative AI using preprocessed data. Preprocessed data is used as input, and a generative AI model (e.g., GPT-3) is trained. The output is the trained AI model. Here, the training process is executed using an AI model training framework (e.g., TensorFlow, PyTorch).

[0188] Step 4:

[0189] The server automatically generates questions for the first and second exams using a pre-trained generative AI. The input consists of a pre-trained AI model and prompts, and the output is the automatically generated exam questions. Specifically, the following prompts are input to the AI ​​to generate the questions.

[0190] Please generate the following training questions regarding factory machine operation:

[0191] 1. Explain how to operate the machine according to the safety procedures.

[0192] 2. Please explain how to use the emergency stop button.

[0193] 3. Please describe the procedure for routine maintenance checks.

[0194] Step 5:

[0195] The server distributes the generated test questions to terminals and robots. The input is automatically generated test questions, and the output is the questions distributed to the terminals and robots. The terminals and robots then present these questions to factory workers.

[0196] Step 6:

[0197] The user (factory employee) provides answers to the presented problems using voice or controls. The input is the factory employee's answer, and the output is the answer data sent to the server. The answers are input through a voice recognition system and a sensor network.

[0198] Step 7:

[0199] The server evaluates test-takers' answers in real time. The input is user-submitted answer data, and the output is the evaluation result and feedback. Here, an AI-powered automated scoring system evaluates the answers and provides real-time feedback in text and audio formats.

[0200] Step 8:

[0201] The server collects employee performance data and analyzes it to improve the quality of future training programs and feedback. Real-time evaluation results are used as input, and the output is an improved training program and feedback. Statistical analysis tools (e.g., Python's Pandas, NumPy) are used to perform data analysis and incorporate the findings into future improvements.

[0202] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0203] The system of the present invention combines automation of qualification examinations with an emotion engine to recognize user emotions and reflect them in the feedback. Specific embodiments for carrying out the present invention are shown below.

[0204] Collection and preprocessing of qualification exam data

[0205] The server collects data related to certification exams from the company's internal database. This includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. During this process, it checks for missing or abnormal data and supplements or corrects it as needed. The preprocessed data is then used as training data for generative AI.

[0206] Learning of generative AI

[0207] The server trains a generative AI based on pre-processed data. During the training process, the AI ​​learns the evaluation criteria for the certification exam, patterns of past exam questions, and the quality of answers. This allows the AI ​​to build a foundation for accurately creating exam questions and appropriately evaluating answers.

[0208] Automatic generation of exam questions

[0209] Once the training is complete, the generative AI automatically generates questions for the first and second exams according to the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge. For example, one question might be, "What are the basic steps of the sales process?" The second exam generates scenario questions and case studies to assess practical skills. For example, one might be, "What is the best way to handle a customer complaint that arises during customer service?"

[0210] Exam distribution and administration

[0211] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers using their device. During the exam, the server receives the examinee's answers in real time and evaluates them using an emotion engine. The emotion engine recognizes emotions from the examinee's facial expressions and voice and uses this to perform real-time evaluations.

[0212] Scoring and feedback generation

[0213] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. Through an automated scoring system, it evaluates the accuracy and quality of each answer and calculates a total score. Based on the evaluation results, the server generates detailed feedback. This feedback includes evaluations of each question, the test-taker's strengths and areas for improvement, as well as comments on emotional states detected by the emotion engine. This feedback more accurately reflects the test-taker's performance and provides information to help improve next time.

[0214] Performance data analysis and improvement

[0215] The server stores and analyzes test-takers' performance data. This analysis is used to create future test questions and improve feedback. Furthermore, by incorporating emotion engine data as AI learning data, the system can provide guidance and support based on the test-takers' emotions.

[0216] Applications of the Emotion Engine

[0217] The emotion engine also detects the test-taker's stress level and concentration level. For example, if the stress level is high, the server will extend the answer time or provide additional support information. By providing questions and feedback that take emotional states into account, it is possible to help test-takers perform at their best.

[0218] This combination of systems allows for a highly automated qualification examination process, reflecting test-takers' emotions and stress levels in real time, thereby providing more appropriate and effective feedback. This not only reduces human resources and improves the consistency of evaluations, but also alleviates the mental burden on test-takers and maximizes learning effectiveness.

[0219] The following describes the processing flow.

[0220] Step 1:

[0221] The server collects data related to the certification exam from its internal database. This includes past exam questions, test-takers' results, and performance data for top and bottom-ranked candidates. Next, it checks for missing or abnormal data and supplements or corrects it as needed.

[0222] Step 2:

[0223] The server trains a generative AI based on pre-processed data. Specifically, it learns the evaluation criteria for certification exams, past exam question patterns, and the quality of answers. Through this process, the generative AI builds the foundation for creating exam questions and evaluating answers.

[0224] Step 3:

[0225] The server uses generative AI to automatically generate questions for the first and second exams. The first exam generates multiple-choice and short-answer questions to assess basic knowledge, while the second exam generates scenario-based questions and case studies to assess practical skills. Examples of such questions include, "What are the basic steps of the sales process?" and "What is the best way to handle a customer complaint?"

[0226] Step 4:

[0227] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers on their device.

[0228] Step 5:

[0229] During the exam, the server receives test-takers' answers in real time and evaluates them using an emotion engine. For example, it recognizes emotions from the test-taker's facial expressions and voice via camera and microphone, and analyzes their stress and tension levels.

[0230] Step 6:

[0231] Based on the analysis results from the emotion engine, the server generates additional questions and instructions in real time that take into account the test-taker's emotional state. For example, if a test-taker is feeling stressed, the difficulty level of the test questions may be adjusted or the time allotted for answering may be extended.

[0232] Step 7:

[0233] After the first exam is completed, the server immediately delivers the questions for the second exam. Users continue to answer the second exam questions within the same ZOOM session. The answers to the second exam are also evaluated in real time, and the evaluation results are continuously fed back through the sentiment engine.

[0234] Step 8:

[0235] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. An automated scoring system is used to evaluate the accuracy and quality of each answer and calculate the overall score.

[0236] Step 9:

[0237] The server generates detailed feedback based on the evaluation results. This feedback includes an evaluation of each problem, the test-taker's strengths and areas for improvement, as well as comments on the emotional state detected by the emotion engine. The server compiles this feedback for each test-taker and sends it to the user (test-taker) via email.

[0238] Step 10:

[0239] The server stores and analyzes test-takers' performance data. This analysis is used to create future test questions and improve feedback. Furthermore, incorporating emotion engine data as AI training data enables more accurate emotion recognition and feedback.

[0240] Through these steps, an automated certification exam system incorporating an emotion engine can reflect the test-taker's emotions in real time, providing high-quality feedback while reducing mental stress.

[0241] (Example 2)

[0242] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the smart device 14 as the "terminal".

[0243] In the process of certification examinations, there is a demand for automation and efficiency, as well as feedback that reflects the emotions and stress levels of test-takers in real time. However, in conventional systems, the generation and evaluation of exam questions are often done manually, which is inefficient. Furthermore, there has been a lack of technology to provide feedback that reflects the emotional state of test-takers. Therefore, providing accurate, real-time evaluation and feedback, and efficiently managing certification examinations are challenges.

[0244] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0245] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and supplementing / correcting missing or abnormal data, means for training a generative AI model based on the preprocessed data, means for automatically generating exam questions for the first and second exams using prompt sentences, means for receiving the exams and examinees' answers in real time and evaluating them using an emotion engine, means for automatically scoring examinees' answers and generating feedback, means for analyzing examinees' performance data and improving the next exam questions and feedback, and means for detecting the examinee's emotional state and evaluating stress levels and concentration levels. This makes it possible to highly automate the qualification exam process and provide accurate and effective feedback that reflects the examinee's emotions in real time.

[0246] An "internal database" is a centrally managed collection of information within a system, and its role is to store and provide data related to certification exams.

[0247] "Qualification exam data" refers to data containing information necessary for qualification exams, specifically including past exam questions, examinee results, evaluation criteria, and performance data.

[0248] "Data preprocessing" is the process of converting collected raw data into a format that can be used for analysis and machine learning, and includes supplementing and correcting missing or abnormal data.

[0249] A "generative AI model" is a type of artificial intelligence technology that learns from data and automatically generates new test questions based on the results of that learning.

[0250] A "prompt sentence" is an input sentence used by a generative AI model, and it is an instruction sentence that the AI ​​uses to generate answers or test questions.

[0251] The "emotion engine" is a system that recognizes and evaluates the emotional state of test takers in real time, detecting emotions based on data such as facial expressions and voice.

[0252] "Real-time evaluation" is a process that involves immediately analyzing and evaluating the test-taker's answers and emotional state during the exam, allowing for feedback and responses without delay.

[0253] "Automated scoring" is a process that uses algorithms or AI to mechanically evaluate test-takers' answers and calculate the accuracy rate and quality.

[0254] "Feedback" refers to evaluation information and suggestions for improvement generated based on the test taker's exam results and performance, and is intended to support the test taker in achieving better results in the next exam.

[0255] "Performance data" refers to data that includes all performance indicators related to the exam, such as the test-taker's response status during the exam, response speed, accuracy rate, and emotional state.

[0256] "Stress level" is an indicator that shows the degree of mental burden on test takers during the exam, and it quantitatively evaluates emotional states such as tension and anxiety.

[0257] "Concentration level" is an indicator used to evaluate the level of attention and focus of test-takers during an exam, and a high level of concentration is considered to lead to high performance.

[0258] This system is designed to highly automate the qualification examination process and provide real-time feedback that reflects the examinee's emotions and stress levels. The specific implementation procedure is as follows.

[0259] Collection and preprocessing of qualification exam data

[0260] The server collects data related to certification exams from the company's internal database. This data includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. The collected data is checked for missing or abnormal data and supplemented or corrected as needed. For example, the Python Pandas library is used to impute missing values ​​in the data.

[0261] Learning of generative AI

[0262] The server trains a generative AI model based on pre-processed data. Specifically, it uses past exam questions and test-takers' answers as input data and trains the AI ​​model using machine learning libraries such as TensorFlow and PyTorch. This allows the AI ​​to learn the evaluation criteria for the certification exam, patterns in past exam questions, and the quality of the answers.

[0263] Automatic generation of exam questions

[0264] Once the training is complete, the generative AI will automatically generate questions for the first and second exams according to the specified exam format. The first exam will generate multiple-choice and short-answer questions to assess basic knowledge. For example, it will use prompts like the following:

[0265] "What are the basic steps of the sales process?"

[0266] In the second stage of the exam, scenario-based problems and case studies are generated to assess practical skills. For example,

[0267] "What is the best way to handle a customer complaint that arises during customer service?"

[0268] Exam distribution and administration

[0269] The user (examinee) logs into the online conferencing system from their device at the designated exam date and time. The server delivers the first-stage exam question set through the online conferencing system. The user enters their answers using their device and sends them to the server. During the exam, the server uses Web-RTC technology to capture the examinee's facial expressions and collect data in real time.

[0270] Scoring and feedback generation

[0271] After the exam ends, the server compiles all the answers and begins evaluation through an automated scoring system. It assesses the accuracy and quality of each answer and calculates an overall score. Based on the evaluation results, the server generates detailed feedback. This feedback includes evaluations of each question, the test-taker's strengths and areas for improvement, as well as comments on emotional states detected by the emotion engine.

[0272] Performance data analysis and improvement

[0273] The server stores and analyzes test-takers' performance data. This analysis is used to improve future test questions and feedback. Furthermore, data acquired by the emotion engine is incorporated as AI learning data, enabling guidance and support based on the test-takers' emotions.

[0274] Applications of the Emotion Engine

[0275] The emotion engine detects the test-taker's stress level and concentration level in real time. For example, if the stress level is high, the server will provide an extension of the answer time or additional support information. Specific feedback includes:

[0276] "Your stress level appears high. We have extended your response time by 10 minutes."

[0277] This system, designed in this way, can efficiently automate the qualification examination process, reduce the mental burden on test-takers, and provide accurate and effective feedback.

[0278] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0279] Step 1:

[0280] The server collects past exam questions, test-taker results, evaluation criteria, and performance data for top and bottom-ranked test-takers from an internal database. The data to be collected is easily extracted using SQL queries. It receives connection information to the internal database and data retrieval conditions as input, and outputs the necessary exam data based on that. Specifically, it executes SQL queries to extract the required data and stores it on the server.

[0281] Step 2:

[0282] The server preprocesses the collected data. This includes checking for missing or abnormal data and supplementing or correcting them. It receives the raw data collected in step 1 as input and outputs preprocessed, clean data. Specifically, it uses the Python Pandas library to impute missing values ​​and correct invalid data.

[0283] Step 3:

[0284] The server trains a generative AI model based on preprocessed data. It receives the preprocessed data as input and uses it as training data. The output is a trained generative AI model. As specific operations, it defines and trains the model using machine learning libraries such as TensorFlow and PyTorch.

[0285] Step 4:

[0286] The trained generative AI model automatically generates questions for the primary and secondary tests. It receives the test format and the prompt text of basic knowledge questions as input and provides the generated test questions as output. An example of the prompt text is "What are the basic steps of the sales process?" As specific operations, it inputs the prompt text into the generative AI and generates test questions based on it.

[0287] Step 5:

[0288] The user (test taker) logs in to the online meeting system from their terminal at the specified test date and time. The server distributes the test questions, and the user inputs the answers. The input data is the test taker's answers and the facial expression data captured in real time. The output is the test taker's answer data and emotional state data. As specific operations, it sends the answers input from the terminal and the facial expression data collected using the WEB-RTC technology to the server.

[0289] Step 6:

[0290] After the test, the server aggregates all the answers and performs automatic scoring and evaluation. It receives the test taker's answer data as input, evaluates the correct rate and quality of each answer as output, and generates a comprehensive score. As specific operations, it uses Scikit-learn in Python to calculate the correct rate of the answers and performs evaluation using a dedicated algorithm.

[0291] Step 7:

[0292] The server generates feedback based on the aggregated results. It receives evaluation results and sentiment data from the sentiment engine as input, and generates a feedback document as output. Specifically, it creates detailed feedback including evaluations of each problem, the test-taker's strengths, areas for improvement, and comments on the sentiment state detected by the sentiment engine.

[0293] Step 8:

[0294] The server stores and analyzes test-taker performance data. It receives individual test-taker performance data as input and outputs the next exam questions and feedback for improvement. Specifically, it uses the Python Seaborn library to visualize the data and analyze trends.

[0295] Step 9:

[0296] The server uses an emotion engine to detect the test-taker's stress level and concentration level. It receives real-time facial expression and voice data as input and provides stress level and concentration level as output. Specifically, it uses facial expression and voice analysis algorithms to evaluate the emotional state and adjusts the test environment based on that.

[0297] (Application Example 2)

[0298] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as a "server" and the smart device 14 as a "terminal".

[0299] The current certification examination system suffers from a lack of automation and difficulty in providing feedback that takes into account the emotions and stress levels of test-takers. This increases the mental burden on test-takers, making it difficult for them to perform at their best. Furthermore, the inconsistency in exam evaluation and feedback makes it difficult to accurately identify areas for improvement.

[0300] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and training a generative AI, means for automatically generating exam questions for the first and second exams, means for distributing the exam and evaluating the examinee's answers in real time, means for automatically scoring the examinee's answers and generating feedback, means for analyzing the examinee's performance data and improving the next exam questions and feedback, and means for detecting the examinee's emotions and stress level using an emotion recognition engine and providing feedback that takes this into account. This makes it possible to automate the exam and provide appropriate feedback according to the examinee's emotions and stress level.

[0301] The "internal database" is a digital repository for collecting and storing qualification exam data, and for retrieving and updating it as needed.

[0302] "Preprocessing" is the process of detecting missing or anomaly data in the collected data, performing necessary completion or correction, and preparing it in a format suitable for training data for generative AI.

[0303] "Generative AI" refers to artificial intelligence that uses machine learning algorithms to learn and automatically generate questions for certification exams based on specific patterns and evaluation criteria.

[0304] The "first-stage examination" refers to a type of examination that includes multiple-choice and short-answer questions used to assess basic knowledge.

[0305] The "second examination" is a type of test that includes scenario-based questions and case studies used to assess practical skills.

[0306] "Real-time evaluation" is a process that instantly evaluates the answers entered by test-takers during the exam and provides necessary feedback on the spot.

[0307] "Automatic scoring" is a mechanism that mechanically evaluates the answers of test takers using generative AI or emotion recognition engines and calculates a comprehensive score.

[0308] "Feedback" is information for reference in the next test or learning by providing information such as evaluations, comments, areas for improvement, and strengths regarding the test taker's answers.

[0309] "Performance data" is data indicating the test performance ability of test takers, including the correct answer rate, quality, and emotional state of the answers shown by test takers during the test.

[0310] "Emotion recognition engine" is a technology for analyzing the expressions and voices of test takers to detect emotions and stress levels, and is used for real-time feedback.

[0311] The system of the present invention recognizes the emotions and stress levels of workers in factory work in real time, provides appropriate feedback to improve efficiency, and improves the working environment. Specific embodiments are shown below.

[0312] Data collection and preprocessing

[0313] The server first collects past work data and environmental data from the internal database. This data includes past work achievements, performance data of workers, operation data of machines used, etc. The collected data is preprocessed using libraries such as OpenCV and Dlib. In preprocessing, missing data is supplemented and abnormal data is corrected, and it is prepared as a learning dataset for generative AI.

[0314] Learning of generative AI

[0315] The server trains a generative AI based on pre-processed data. In this process, the AI ​​learns evaluation criteria for tasks and past performance patterns, building the foundation necessary for improving task efficiency and generating optimal work processes. The OpenAI® GPT model is used as the generative AI to generate specific feedback content based on prompt text.

[0316] Use of emotion recognition engine

[0317] The emotion recognition engine is used by the server to analyze the worker's facial expressions and voice data in real time. It utilizes OpenCV and Dlib for facial recognition, Google Cloud Speech-to-Text for speech recognition, and Affectiva and Emotion API (Microsoft®) for emotion analysis to detect the worker's emotions and stress levels.

[0318] Provide feedback

[0319] Based on data obtained from the emotion recognition engine, the server uses generative AI to provide appropriate feedback to workers in real time. This feedback covers a wide range of topics, including adjusting work pace, suggesting breaks, and offering advice to reduce stress.

[0320] Specific example:

[0321] For example, if "anger" is detected from the worker's facial expression, the OpenAI GPT model generates feedback using the following prompt message.

[0322] Based on the factory worker's sentiment data: 'Angry', generate appropriate feedback.

[0323] In response, an example of the generated feedback is a comment such as, "You appear to be feeling stressed right now. We recommend taking a 5-minute break."

[0324] Performance data analysis and improvement

[0325] The server accumulates worker performance data daily and performs periodic analysis. This data is used to optimize future work processes and improve feedback. Furthermore, by incorporating data from the emotion recognition engine as AI training data, it becomes possible to provide more precise guidance and support based on the worker's emotions.

[0326] The system described above can improve the efficiency of factory operations, provide appropriate feedback tailored to workers' emotions and stress levels, and significantly improve the work environment.

[0327] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0328] Step 1:

[0329] The server collects historical work data and environmental data from its internal database and performs preprocessing. In this step, missing and anomalous data are supplemented and corrected, and the data is prepared in a format suitable for use as a training dataset for generative AI. Specifically, OpenCV and Dlib are used to analyze and cleanse the data. The input is data obtained from the internal database, and the output is a preprocessed dataset.

[0330] Step 2:

[0331] The server trains a generative AI using preprocessed data. During the training process, the AI ​​learns evaluation criteria for tasks and past performance patterns. The OpenAI GPT model is used in this step. The input is a preprocessed dataset, and the output is a trained generative AI model.

[0332] Step 3:

[0333] The server uses an emotion recognition engine to analyze the worker's facial expressions and voice data in real time. It uses OpenCV and Dlib for facial recognition, Google Cloud Speech-to-Text for speech recognition, and Affectiva and Emotion API (Microsoft) for emotion analysis. Inputs include camera video and audio data, and output data indicating emotions and stress levels is generated.

[0334] Step 4:

[0335] The server uses generative AI to provide appropriate feedback to workers in real time, based on data obtained from the emotion recognition engine. It creates prompt messages and generates feedback content using OpenAI's GPT model. The input consists of emotion analysis data and prompt messages, and the output is a feedback message. For example, a prompt message such as "Generate appropriate feedback based on the factory worker's emotion data: 'Angry'" might be used.

[0336] Step 5:

[0337] The server provides workers with real-time generated feedback, including suggestions for adjusting work pace, taking breaks, and stress reduction advice. The input is generated feedback messages, and the output is specific actions or messages provided to the worker.

[0338] Step 6:

[0339] The server stores and analyzes worker performance data. This data is used to optimize future work processes and improve feedback. Real-time and historical performance data are used as input, and analysis results and improved feedback are generated as output. By feeding this analysis result back into the generative AI as new training data, more precise guidance and support become possible.

[0340] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating 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.

[0341] Data generation model 58 is a so-called generative AI (Artificial Intelligence). An example of data generation model 58 is ChatGPT (registered trademark) (Internet search).<URL: https: / / openai.com / blog / chatgpt> ), Gemini (registered trademark) (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0342] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart device 14.

[0343] [Second Embodiment]

[0344] Figure 3 shows an example of the configuration of the data processing system 210 according to the second embodiment.

[0345] As shown in Figure 3, the data processing system 210 includes a data processing device 12 and smart glasses 214. An example of the data processing device 12 is a server.

[0346] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0347] The smart glasses 214 include a computer 36, a microphone 238, a speaker 240, a camera 42, and a communication interface 44. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, and camera 42 are also connected to the bus 52.

[0348] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0349] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0350] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0351] Figure 4 shows an example of the main functions of the data processing device 12 and the smart glasses 214. As shown in Figure 4, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0352] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0353] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0354] In the smart glasses 214, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0355] Next, the identification processing performed by the identification processing unit 290 of the data processing device 12 will be described. 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".

[0356] The system of the present invention aims to automate qualification examinations, with a server at the center executing multiple functions. The specific procedure for implementing the present invention is shown below.

[0357] Collection and preprocessing of qualification exam data

[0358] The server collects data related to certification exams from internal databases within the company. This includes past exam questions, examinee results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. Here, it checks for any missing or abnormal data and supplements or corrects it as needed. This preprocessed data is then used directly as training data for generative AI.

[0359] Learning of generative AI

[0360] The server trains a generative AI based on pre-processed data. During this process, the AI ​​learns the evaluation criteria for certification exams, patterns in past exam questions, and the quality of answers. This creates a foundation for the AI ​​to evaluate test-takers' responses and generate appropriate exam questions.

[0361] Automatic generation of exam questions

[0362] Once the training is complete, the generative AI automatically generates questions for the first and second exams according to the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge. Specific examples include questions such as, "What are the basic steps of the sales process?" The second exam includes scenario-based questions and case studies to assess practical skills. For example, it generates practical questions such as, "What is the best way to handle a customer complaint?"

[0363] Exam distribution and administration

[0364] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers through their device. During the exam, the server receives the examinee's answers in real time and performs automatic evaluation. Additional questions may be automatically generated based on the accuracy rate and quality of the answers. This process continues into the second exam after the first exam is completed.

[0365] Grading and feedback

[0366] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. The automated scoring system evaluates each answer and calculates a total score. Based on the evaluation results, the server generates detailed feedback. This feedback includes an evaluation of each question, the test-taker's strengths, and areas for improvement. The server compiles this feedback for each test-taker and sends it to the user (test-taker) via email.

[0367] Performance data analysis and feedback improvement

[0368] The server collects and analyzes test-taker performance data. This analysis is then used to improve future exam questions and feedback. In addition, the server collects reviews from test-takers and incorporates them as training data for generative AI, thereby improving the overall system quality.

[0369] As described above, the server-centered automated system streamlines the execution of certification exams and enables the rapid provision of high-quality feedback. This system not only saves human resources but also contributes to improved consistency in evaluations and enhanced skills among test-takers.

[0370] The following describes the processing flow.

[0371] Step 1:

[0372] The server collects data related to the certification exam from its internal database. Specifically, it retrieves past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, it checks for any missing or abnormal data and supplements or corrects it as needed. This pre-processed data is then used as training data for generative AI.

[0373] Step 2:

[0374] The server trains the generative AI based on pre-processed data. Specifically, it trains the AI ​​on evaluation criteria for certification exams, patterns of past exam questions, and the quality of answers. Through this learning process, the generative AI builds a foundation for creating appropriate exam questions and evaluating answers.

[0375] Step 3:

[0376] The server uses generative AI to automatically generate questions for the first and second exams. For the first exam, it generates multiple-choice and short-answer questions to assess basic knowledge. For the second exam, it generates scenario-based questions and case studies to assess practical skills. For example, it generates questions such as, "What are the basic steps of the sales process?" and "What is the best way to handle a customer complaint?"

[0377] Step 4:

[0378] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers using their device. During the exam, the server receives the examinee's answers in real time and performs automatic evaluation. Based on this evaluation, additional questions may be automatically generated as needed.

[0379] Step 5:

[0380] Once a user completes the first exam, the server immediately delivers the questions for the second exam. The user then takes the second exam within the same Zoom session. The answers to the second exam are also evaluated in real time, and the evaluation results are continuously fed back to the user.

[0381] Step 6:

[0382] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. Using an automated scoring system, it evaluates the accuracy and quality of each answer and calculates an overall score. Based on the evaluation results, detailed feedback is generated. The feedback includes an evaluation of each question, the test-taker's strengths, and areas for improvement.

[0383] Step 7:

[0384] The server compiles this feedback for each test taker and sends it to the user (test taker) via email. The feedback is also recorded in the company's internal database and used as reference material for future tests and evaluations.

[0385] Step 8:

[0386] The server collects and analyzes test-takers' performance data. This analysis is used to improve future test questions and feedback. Furthermore, the server collects reviews from test-takers and incorporates them as AI training data, thereby improving the overall quality of the system.

[0387] Through the steps outlined above, this system automates the entire certification examination process, reducing human resources and improving the consistency of evaluations. It also enables the rapid provision of high-quality feedback to test-takers.

[0388] (Example 1)

[0389] Next, we will describe Example 1. 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."

[0390] Traditional certification examination systems require significant human resources and are inefficient because the creation of exam questions, evaluation of answers, and provision of feedback are all done manually. Furthermore, consistency in evaluation and quality of feedback are prone to inconsistencies. Additionally, the analysis of test-taker performance data and the improvement of feedback based on that analysis are not adequately performed. This hinders the improvement of test-takers' abilities and fair evaluation.

[0391] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0392] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data to supplement and correct any deficiencies or anomalies, means for training a generative AI based on the preprocessed data, means for automatically generating exam questions for the first and second exams in a specified format, means for distributing the exams and evaluating examinees' answers in real time, means for automatically scoring examinees' answers, generating detailed feedback, and sending it via email, and means for analyzing examinee performance data to improve the next exam questions and feedback. This enables automation and efficiency of the exam, improved consistency of evaluation, and rapid provision of high-quality feedback.

[0393] An "internal database" refers to a collection of information and data related to certification exams that is maintained within a specific organization or system.

[0394] "Qualification exam data" refers to a variety of data related to the exam, including past exam questions, examinee results, evaluation criteria, and examinee performance data.

[0395] "Preprocessing" refers to the process of performing tasks such as imputing missing values, correcting outliers, and formatting the collected data.

[0396] "Generative AI" refers to artificial intelligence systems that learn from collected and pre-processed data and automatically generate new test questions and answers.

[0397] The term "first-stage examination" refers to the initial stage of an exam that includes multiple-choice and short-answer questions designed to assess basic knowledge.

[0398] The term "second examination" refers to the second stage of the examination, which includes scenario-based problems and case studies designed to assess practical skills.

[0399] "Exam delivery" refers to the process of providing exam questions to test takers and having them take the exam.

[0400] "Real-time evaluation" refers to a process where the answers entered by test-takers during the exam are immediately analyzed and evaluated.

[0401] "Automated scoring" refers to a process that automatically evaluates and scores test-takers' answers to exam questions.

[0402] "Feedback" refers to information generated based on a test-taker's exam results, including evaluation results and suggestions for improvement.

[0403] "Performance data" refers to data related to test-takers' behavior during the exam, the quality of their answers, and the accuracy of their responses.

[0404] "Analysis" refers to the process of examining collected data and extracting meaningful patterns and trends.

[0405] The system of this invention aims to automate qualification examinations and performs various processes centered around a server. This system is implemented in the following specific steps.

[0406] Collection and preprocessing of qualification exam data

[0407] The server first connects to an internal database to collect data related to the certification exam. This includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. Specifically, it checks for missing or abnormal data and then fills in or corrects them. For example, missing values ​​are filled in with the mean, and abnormal values ​​are corrected appropriately. Python's Pandas library and SQL queries are used for this preprocessing.

[0408] Learning of generative AI

[0409] The pre-processed data is used as training data for generative AI. The server uses this data to train the AI ​​model. The AI ​​models used include advanced generative AIs such as GPT-4. In this process, deep learning frameworks such as TensorFlow and PyTorch are utilized, and the model learns patterns and evaluation criteria from past exam questions. As a result, the AI ​​acquires the ability to automatically generate the next exam questions.

[0410] Automatic generation of exam questions

[0411] Once the generative AI has completed its training, it automatically generates questions for the first and second exams according to the specified exam format. For example, the first exam generates multiple-choice and short-answer questions to assess basic knowledge. On the other hand, the second exam includes scenario-based questions and case studies to assess practical skills.

[0412] Example of a prompt

[0413] "What are the basic steps of the sales process?"

[0414] "What is the best way to handle a customer complaint that arises during customer service?"

[0415] Exam distribution and administration

[0416] Next, the server handles the delivery and administration of the exam. Users (examinees) log in to ZOOM from their devices at the designated exam date and time. The server delivers the first exam question set through the ZOOM session, and users answer it. Answers are sent to the server in real time, and the server evaluates them immediately. In some cases, additional questions may be automatically generated based on the accuracy rate and quality of the answers.

[0417] Grading and feedback

[0418] After the exam is completed, the server compiles all the answers and evaluates them using an automated scoring system. Based on the evaluation results, detailed feedback is generated and sent to the user (test taker) via email. This feedback includes an evaluation of each question, the test taker's strengths, and areas for improvement.

[0419] Performance data analysis and feedback improvement

[0420] The server collects and analyzes test-takers' performance data. This analysis is used to improve the creation of future test questions and feedback. Furthermore, reviews from test-takers are collected and incorporated as training data for generative AI, thereby improving the overall quality of the system.

[0421] As described above, the present invention's automated qualification examination system automates a series of processes centered on a server, including data collection, preprocessing, training of a generative AI, automatic generation of examination questions, distribution and administration of the examination, scoring and generation of feedback, and analysis and improvement of performance data, thereby realizing efficient and high-quality examination management.

[0422] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0423] Step 1: Start data collection

[0424] The server establishes a connection to the internal database and executes an SQL query to extract "certification exam data" (past exam questions, examinee results, evaluation criteria, etc.). For example, it might execute a query like "SELECT FROM ExamData WHERE Year BETWEEN 2018 AND 2023".

[0425] Input: Connection information for the internal database

[0426] Output: Extraction results of qualification exam data

[0427] Step 2: Data checking and preprocessing

[0428] The server reads the collected data using Python's Pandas library and checks for missing values ​​and outliers. For example, it checks for missing values ​​using df.isnull().sum() and imputes them with df.fillna(df.mean()). If outliers exist, it corrects them using an appropriate method.

[0429] Input: Certification exam data extracted in Step 1

[0430] Output: Preprocessed data

[0431] Step 3: Preparing training data for generative AI

[0432] The server organizes the preprocessed data as training data for generative AI and converts it into an applicable format. For example, it selects the necessary features and converts the data into numerical vectors.

[0433] Input: Preprocessed data obtained in Step 2

[0434] Output: Training dataset

[0435] Step 4: Training the generative AI model

[0436] The server uses deep learning frameworks such as TensorFlow or PyTorch to train generative AI models. For example, it trains the model using the form model.fit(X_train, y_train, epochs=100).

[0437] Input: Training dataset prepared in Step 3

[0438] Output: Trained generative AI model

[0439] Step 5: Automatic generation of primary exam questions

[0440] The server uses a pre-trained generative AI model to automatically generate questions for the first stage of the exam. For example, it might create a multiple-choice question such as, "What are the basic steps of the sales process?"

[0441] Input: Pre-trained generative AI model

[0442] Output: First Exam Question Set

[0443] Step 6: Automatic generation of secondary examination questions

[0444] The server uses a pre-trained generative AI model to automatically generate questions for the second stage of the exam. For example, it creates scenario questions such as, "What is the best way to handle a customer complaint that arises during customer service?"

[0445] Input: Pre-trained generative AI model

[0446] Output: Secondary Exam Question Set

[0447] Step 7: Notification of exam date and time

[0448] The server will send a notification email to the test taker containing the date, time, and link to join the ZOOM session.

[0449] Input: Candidate information, exam schedule

[0450] Output: Notification email

[0451] Step 8: Exam delivery and real-time collection of test-takers' responses

[0452] The user (examinee) logs into ZOOM from their device at the designated date and time. The server delivers the first-stage exam question set via the ZOOM session, and the user enters their answers. The server collects and evaluates the answers sent from the device in real time.

[0453] Input: First-stage exam question set, examinee's answers

[0454] Output: Real-time evaluation results

[0455] Step 9: Generating additional questions based on accuracy and quality.

[0456] The server evaluates the accuracy rate and quality of answers based on the responses received in real time, and generates additional questions as needed.

[0457] Input: Test-takers' responses, real-time evaluation results

[0458] Output: Additional questions

[0459] Step 10: Aggregation of test results and generation of feedback

[0460] After the exam is completed, the server compiles all the answers and calculates an overall score using an automated scoring system. Next, it generates detailed feedback and sends it via email.

[0461] Input: All responses from test takers

[0462] Output: Overall score, feedback email

[0463] Step 11: Accumulate and analyze performance data

[0464] The server collects and analyzes test-takers' performance data. This analysis is used to create future exam questions and improve feedback.

[0465] Input: Test-taker's performance data

[0466] Output: analysis results, improvement suggestions

[0467] Through the processing steps described above, the system of the present invention can automate and streamline qualification examinations and provide high-quality feedback quickly.

[0468] (Application Example 1)

[0469] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0470] Traditional certification exams and training programs require significant human resources for tasks such as creating exam questions, administering exams, grading, and providing feedback. This often leads to inefficient exam administration and the risk of biased human evaluation. Furthermore, in on-site employee training, such as in factories, individual evaluation and feedback are difficult to provide, hindering effective training. To address these challenges, an efficient and consistent automated evaluation system is needed.

[0471] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0472] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and training a generative AI, means for automatically generating exam questions for the first and second exams, means for distributing the exams and evaluating examinees' answers in real time, means for automatically scoring examinees' answers and generating feedback, means for analyzing examinee performance data and improving the next exam questions and feedback, and means for a robot to generate training questions similar to qualification exams and provide real-time feedback during factory employee training. This streamlines the employee training and evaluation process and enables the provision of consistent, high-quality feedback.

[0473] An "internal database" is a database that stores data related to certification exams and allows users to retrieve that data as needed.

[0474] "Qualification exam data" refers to all data related to the administration of qualification exams, including past exam questions, examinee results, evaluation criteria, and examinee performance data.

[0475] "Preprocessing" is the process of supplementing and correcting any deficiencies or anomalies in the collected data and converting it into a format suitable for learning by generative AI.

[0476] "Generative AI" is a type of artificial intelligence that uses large amounts of data to train models and generate test questions.

[0477] The "first-stage examination" is an exam that includes multiple-choice and short-answer questions designed to assess basic knowledge.

[0478] The "second examination" is a test that includes scenario-based problems and case studies, designed to evaluate practical skills.

[0479] "Automatic scoring" is a process that automatically evaluates the test-takers' answers and calculates their scores.

[0480] "Feedback" refers to providing detailed comments on the test-taker's answers, along with an evaluation of their responses, including their strengths and areas for improvement.

[0481] "Performance data" refers to the overall results a test-taker demonstrated during the exam, along with various analytical data based on those results.

[0482] "Factory employee training" refers to the process of educating and providing factory employees with the necessary skills and knowledge to perform their duties.

[0483] "Real-time evaluation" is a process in which, each time a test-taker submits their answer, it is immediately evaluated and the results are provided.

[0484] "Training problems" are questions or tasks presented to assess an employee's abilities.

[0485] To implement this invention, it is first necessary to collect data related to qualification exams from an internal database. The server retrieves data such as past exam questions, examinee results, and evaluation criteria from the database and preprocesses this data. In this preprocessing, any missing or abnormal data is supplemented and corrected, and the data is standardized.

[0486] Next, the server uses the pre-processed data to train a generative AI. A natural language processing (NLP) model, such as GPT-3, is suitable as a generative AI model. Based on the training data, the AI ​​model learns the question patterns and evaluation criteria of certification exams, enabling the automatic generation of exam questions.

[0487] Once the training is complete, the generative AI generates questions for the first and second exams based on the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge, including specific examples such as "What are basic safety procedures?" The second exam generates scenario questions and case studies to assess practical skills, such as "Explain how to respond if an emergency occurs during work."

[0488] The generated test questions are delivered to factory employees' terminals and robots via a server. Employees provide answers to the terminals or robots, and the server evaluates the answers in real time. An AI-powered automated scoring system is used to evaluate the answers, and feedback is generated based on this evaluation. The feedback includes not only whether the answer is correct or incorrect, but also the employee's strengths and areas for improvement.

[0489] Furthermore, the server collects employee performance data and analyzes it to improve the quality of future exam questions and feedback. This analysis utilizes statistical analysis tools (such as Python's Pandas and NumPy) to identify data trends and develop more effective training programs.

[0490] As a concrete example, a factory robot presents an employee with a training question such as, "Explain how to use the emergency stop button," and the employee answers using voice or controls. The AI ​​evaluates the answer in real time and provides feedback such as, "The procedure itself is correct, but safety checks are missing." In this way, employees can learn practical skills on the spot while being efficiently evaluated.

[0491] Examples of prompt messages include the following:

[0492] Please generate the following training questions regarding factory machine operation:

[0493] 1. Explain how to operate the machine according to the safety procedures.

[0494] 2. Please explain how to use the emergency stop button.

[0495] 3. Please describe the procedure for routine maintenance checks.

[0496] The above describes the specific procedures and system configuration necessary for implementing this invention.

[0497] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0498] Step 1:

[0499] The server collects data related to certification exams from an internal database. The data retrieved from the database includes past exam questions, examinee results, evaluation criteria, and examinee performance data. All data related to the certification exams is retrieved from the database as input. The collected dataset is passed to a preprocessing module as output.

[0500] Step 2:

[0501] The server preprocesses the collected data. Preprocessing includes imputing missing values, correcting anomalous data, and normalization. The input is the data collected in step 1, and the output is the preprocessed dataset. At this stage, data cleansing and normalization are performed using data preprocessing libraries (e.g., Pandas, NumPy).

[0502] Step 3:

[0503] The server trains a generative AI using preprocessed data. Preprocessed data is used as input, and a generative AI model (e.g., GPT-3) is trained. The output is the trained AI model. Here, the training process is executed using an AI model training framework (e.g., TensorFlow, PyTorch).

[0504] Step 4:

[0505] The server automatically generates questions for the first and second exams using a pre-trained generative AI. The input consists of a pre-trained AI model and prompts, and the output is the automatically generated exam questions. Specifically, the following prompts are input to the AI ​​to generate the questions.

[0506] Please generate the following training questions regarding factory machine operation:

[0507] 1. Explain how to operate the machine according to the safety procedures.

[0508] 2. Please explain how to use the emergency stop button.

[0509] 3. Please describe the procedure for routine maintenance checks.

[0510] Step 5:

[0511] The server distributes the generated test questions to terminals and robots. The input is automatically generated test questions, and the output is the questions distributed to the terminals and robots. The terminals and robots then present these questions to factory workers.

[0512] Step 6:

[0513] The user (factory employee) provides answers to the presented problems using voice or controls. The input is the factory employee's answer, and the output is the answer data sent to the server. The answers are input through a voice recognition system and a sensor network.

[0514] Step 7:

[0515] The server evaluates test-takers' answers in real time. The input is user-submitted answer data, and the output is the evaluation result and feedback. Here, an AI-powered automated scoring system evaluates the answers and provides real-time feedback in text and audio formats.

[0516] Step 8:

[0517] The server collects employee performance data and analyzes it to improve the quality of future training programs and feedback. Real-time evaluation results are used as input, and the output is an improved training program and feedback. Statistical analysis tools (e.g., Python's Pandas, NumPy) are used to perform data analysis and incorporate the findings into future improvements.

[0518] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0519] The system of the present invention combines automation of qualification examinations with an emotion engine to recognize user emotions and reflect them in the feedback. Specific embodiments for carrying out the present invention are shown below.

[0520] Collection and preprocessing of qualification exam data

[0521] The server collects data related to certification exams from the company's internal database. This includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. During this process, it checks for missing or abnormal data and supplements or corrects it as needed. The preprocessed data is then used as training data for generative AI.

[0522] Learning of generative AI

[0523] The server trains a generative AI based on pre-processed data. During the training process, the AI ​​learns the evaluation criteria for the certification exam, patterns of past exam questions, and the quality of answers. This allows the AI ​​to build a foundation for accurately creating exam questions and appropriately evaluating answers.

[0524] Automatic generation of exam questions

[0525] Once the training is complete, the generative AI automatically generates questions for the first and second exams according to the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge. For example, one question might be, "What are the basic steps of the sales process?" The second exam generates scenario questions and case studies to assess practical skills. For example, one might be, "What is the best way to handle a customer complaint that arises during customer service?"

[0526] Exam distribution and administration

[0527] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers using their device. During the exam, the server receives the examinee's answers in real time and evaluates them using an emotion engine. The emotion engine recognizes emotions from the examinee's facial expressions and voice and uses this to perform real-time evaluations.

[0528] Scoring and feedback generation

[0529] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. Through an automated scoring system, it evaluates the accuracy and quality of each answer and calculates a total score. Based on the evaluation results, the server generates detailed feedback. This feedback includes evaluations of each question, the test-taker's strengths and areas for improvement, as well as comments on emotional states detected by the emotion engine. This feedback more accurately reflects the test-taker's performance and provides information to help improve next time.

[0530] Performance data analysis and improvement

[0531] The server stores and analyzes test-takers' performance data. This analysis is used to create future test questions and improve feedback. Furthermore, by incorporating emotion engine data as AI learning data, the system can provide guidance and support based on the test-takers' emotions.

[0532] Applications of the Emotion Engine

[0533] The emotion engine also detects the test-taker's stress level and concentration level. For example, if the stress level is high, the server will extend the answer time or provide additional support information. By providing questions and feedback that take emotional states into account, it is possible to help test-takers perform at their best.

[0534] This combination of systems allows for a highly automated qualification examination process, reflecting test-takers' emotions and stress levels in real time, thereby providing more appropriate and effective feedback. This not only reduces human resources and improves the consistency of evaluations, but also alleviates the mental burden on test-takers and maximizes learning effectiveness.

[0535] The following describes the processing flow.

[0536] Step 1:

[0537] The server collects data related to the certification exam from its internal database. This includes past exam questions, test-takers' results, and performance data for top and bottom-ranked candidates. Next, it checks for missing or abnormal data and supplements or corrects it as needed.

[0538] Step 2:

[0539] The server trains a generative AI based on pre-processed data. Specifically, it learns the evaluation criteria for certification exams, past exam question patterns, and the quality of answers. Through this process, the generative AI builds the foundation for creating exam questions and evaluating answers.

[0540] Step 3:

[0541] The server uses generative AI to automatically generate questions for the first and second exams. The first exam generates multiple-choice and short-answer questions to assess basic knowledge, while the second exam generates scenario-based questions and case studies to assess practical skills. Examples of such questions include, "What are the basic steps of the sales process?" and "What is the best way to handle a customer complaint?"

[0542] Step 4:

[0543] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers on their device.

[0544] Step 5:

[0545] During the exam, the server receives test-takers' answers in real time and evaluates them using an emotion engine. For example, it recognizes emotions from the test-taker's facial expressions and voice via camera and microphone, and analyzes their stress and tension levels.

[0546] Step 6:

[0547] Based on the analysis results from the emotion engine, the server generates additional questions and instructions in real time that take into account the test-taker's emotional state. For example, if a test-taker is feeling stressed, the difficulty level of the test questions may be adjusted or the time allotted for answering may be extended.

[0548] Step 7:

[0549] After the first exam is completed, the server immediately delivers the questions for the second exam. Users continue to answer the second exam questions within the same ZOOM session. The answers to the second exam are also evaluated in real time, and the evaluation results are continuously fed back through the sentiment engine.

[0550] Step 8:

[0551] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. An automated scoring system is used to evaluate the accuracy and quality of each answer and calculate the overall score.

[0552] Step 9:

[0553] The server generates detailed feedback based on the evaluation results. This feedback includes an evaluation of each problem, the test-taker's strengths and areas for improvement, as well as comments on the emotional state detected by the emotion engine. The server compiles this feedback for each test-taker and sends it to the user (test-taker) via email.

[0554] Step 10:

[0555] The server stores and analyzes test-takers' performance data. This analysis is used to create future test questions and improve feedback. Furthermore, incorporating emotion engine data as AI training data enables more accurate emotion recognition and feedback.

[0556] Through these steps, an automated certification exam system incorporating an emotion engine can reflect the test-taker's emotions in real time, providing high-quality feedback while reducing mental stress.

[0557] (Example 2)

[0558] Next, we will describe Example 2. 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".

[0559] In the process of certification examinations, there is a demand for automation and efficiency, as well as feedback that reflects the emotions and stress levels of test-takers in real time. However, in conventional systems, the generation and evaluation of exam questions are often done manually, which is inefficient. Furthermore, there has been a lack of technology to provide feedback that reflects the emotional state of test-takers. Therefore, providing accurate, real-time evaluation and feedback, and efficiently managing certification examinations are challenges.

[0560] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0561] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and supplementing / correcting missing or abnormal data, means for training a generative AI model based on the preprocessed data, means for automatically generating exam questions for the first and second exams using prompt sentences, means for receiving the exams and examinees' answers in real time and evaluating them using an emotion engine, means for automatically scoring examinees' answers and generating feedback, means for analyzing examinees' performance data and improving the next exam questions and feedback, and means for detecting the examinee's emotional state and evaluating stress levels and concentration levels. This makes it possible to highly automate the qualification exam process and provide accurate and effective feedback that reflects the examinee's emotions in real time.

[0562] An "internal database" is a centrally managed collection of information within a system, and its role is to store and provide data related to certification exams.

[0563] "Qualification exam data" refers to data containing information necessary for qualification exams, specifically including past exam questions, examinee results, evaluation criteria, and performance data.

[0564] "Data preprocessing" is the process of converting collected raw data into a format that can be used for analysis and machine learning, and includes supplementing and correcting missing or abnormal data.

[0565] A "generative AI model" is a type of artificial intelligence technology that learns from data and automatically generates new test questions based on the results of that learning.

[0566] A "prompt sentence" is an input sentence used by a generative AI model, and it is an instruction sentence that the AI ​​uses to generate answers or test questions.

[0567] The "emotion engine" is a system that recognizes and evaluates the emotional state of test takers in real time, detecting emotions based on data such as facial expressions and voice.

[0568] "Real-time evaluation" is a process that involves immediately analyzing and evaluating the test-taker's answers and emotional state during the exam, allowing for feedback and responses without delay.

[0569] "Automated scoring" is a process that uses algorithms or AI to mechanically evaluate test-takers' answers and calculate the accuracy rate and quality.

[0570] "Feedback" refers to evaluation information and suggestions for improvement generated based on the test taker's exam results and performance, and is intended to support the test taker in achieving better results in the next exam.

[0571] "Performance data" refers to data that includes all performance indicators related to the exam, such as the test-taker's response status during the exam, response speed, accuracy rate, and emotional state.

[0572] "Stress level" is an indicator that shows the degree of mental burden on test takers during the exam, and it quantitatively evaluates emotional states such as tension and anxiety.

[0573] "Concentration level" is an indicator used to evaluate the level of attention and focus of test-takers during an exam, and a high level of concentration is considered to lead to high performance.

[0574] This system is designed to highly automate the qualification examination process and provide real-time feedback that reflects the examinee's emotions and stress levels. The specific implementation procedure is as follows.

[0575] Collection and preprocessing of qualification exam data

[0576] The server collects data related to certification exams from the company's internal database. This data includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. The collected data is checked for missing or abnormal data and supplemented or corrected as needed. For example, the Python Pandas library is used to impute missing values ​​in the data.

[0577] Learning of generative AI

[0578] The server trains a generative AI model based on pre-processed data. Specifically, it uses past exam questions and test-takers' answers as input data and trains the AI ​​model using machine learning libraries such as TensorFlow and PyTorch. This allows the AI ​​to learn the evaluation criteria for the certification exam, patterns in past exam questions, and the quality of the answers.

[0579] Automatic generation of exam questions

[0580] Once the training is complete, the generative AI will automatically generate questions for the first and second exams according to the specified exam format. The first exam will generate multiple-choice and short-answer questions to assess basic knowledge. For example, it will use prompts like the following:

[0581] "What are the basic steps of the sales process?"

[0582] In the second stage of the exam, scenario-based problems and case studies are generated to assess practical skills. For example,

[0583] "What is the best way to handle a customer complaint that arises during customer service?"

[0584] Exam distribution and administration

[0585] The user (examinee) logs into the online conferencing system from their device at the designated exam date and time. The server delivers the first-stage exam question set through the online conferencing system. The user enters their answers using their device and sends them to the server. During the exam, the server uses Web-RTC technology to capture the examinee's facial expressions and collect data in real time.

[0586] Scoring and feedback generation

[0587] After the exam ends, the server compiles all the answers and begins evaluation through an automated scoring system. It assesses the accuracy and quality of each answer and calculates an overall score. Based on the evaluation results, the server generates detailed feedback. This feedback includes evaluations of each question, the test-taker's strengths and areas for improvement, as well as comments on emotional states detected by the emotion engine.

[0588] Performance data analysis and improvement

[0589] The server stores and analyzes test-takers' performance data. This analysis is used to improve future test questions and feedback. Furthermore, data acquired by the emotion engine is incorporated as AI learning data, enabling guidance and support based on the test-takers' emotions.

[0590] Applications of the Emotion Engine

[0591] The emotion engine detects the test-taker's stress level and concentration level in real time. For example, if the stress level is high, the server will provide an extension of the answer time or additional support information. Specific feedback includes:

[0592] "Your stress level appears high. We have extended your response time by 10 minutes."

[0593] This system, designed in this way, can efficiently automate the qualification examination process, reduce the mental burden on test-takers, and provide accurate and effective feedback.

[0594] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0595] Step 1:

[0596] The server collects past exam questions, test-taker results, evaluation criteria, and performance data for top and bottom-ranked test-takers from an internal database. The data to be collected is easily extracted using SQL queries. It receives connection information to the internal database and data retrieval conditions as input, and outputs the necessary exam data based on that. Specifically, it executes SQL queries to extract the required data and stores it on the server.

[0597] Step 2:

[0598] The server preprocesses the collected data. This includes checking for missing or abnormal data and supplementing or correcting them. It receives the raw data collected in step 1 as input and outputs preprocessed, clean data. Specifically, it uses the Python Pandas library to impute missing values ​​and correct invalid data.

[0599] Step 3:

[0600] The server trains a generative AI model based on preprocessed data. It receives preprocessed data as input and uses it as training data. The output is the trained generative AI model. Specifically, it defines and trains the model using machine learning libraries such as TensorFlow and PyTorch.

[0601] Step 4:

[0602] Once the training is complete, the generative AI model automatically generates questions for the first and second exams. It takes prompts for exam format and basic knowledge questions as input and provides generated exam questions as output. An example of a prompt is "What are the basic steps of the sales process?" In terms of operation, the generative AI is given prompts and generates exam questions based on them.

[0603] Step 5:

[0604] The user (examinee) logs into the online conferencing system from their device at the designated exam date and time. The server distributes the exam questions, and the user enters their answers. The input data consists of the examinee's answers and facial expression data captured in real time. The output consists of the examinee's answer data and emotional state data. Specifically, the answers entered from the device and the facial expression data collected using WEB-RTC technology are sent to the server.

[0605] Step 6:

[0606] After the exam ends, the server collects all the answers and performs automatic scoring and evaluation. It receives the test-takers' answer data as input, evaluates the accuracy and quality of each answer as output, and generates an overall score. Specifically, it uses Python's Scikit-learn to calculate the accuracy of each answer and evaluates them using a dedicated algorithm.

[0607] Step 7:

[0608] The server generates feedback based on the aggregated results. It receives evaluation results and sentiment data from the sentiment engine as input, and generates a feedback document as output. Specifically, it creates detailed feedback including evaluations of each problem, the test-taker's strengths, areas for improvement, and comments on the sentiment state detected by the sentiment engine.

[0609] Step 8:

[0610] The server stores and analyzes test-taker performance data. It receives individual test-taker performance data as input and outputs the next exam questions and feedback for improvement. Specifically, it uses the Python Seaborn library to visualize the data and analyze trends.

[0611] Step 9:

[0612] The server uses an emotion engine to detect the test-taker's stress level and concentration level. It receives real-time facial expression and voice data as input and provides stress level and concentration level as output. Specifically, it uses facial expression and voice analysis algorithms to evaluate the emotional state and adjusts the test environment based on that.

[0613] (Application Example 2)

[0614] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server," and the smart glasses 214 will be referred to as the "terminal."

[0615] The current certification examination system suffers from a lack of automation and difficulty in providing feedback that takes into account the emotions and stress levels of test-takers. This increases the mental burden on test-takers, making it difficult for them to perform at their best. Furthermore, the inconsistency in exam evaluation and feedback makes it difficult to accurately identify areas for improvement.

[0616] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and training a generative AI, means for automatically generating exam questions for the first and second exams, means for distributing the exam and evaluating the examinee's answers in real time, means for automatically scoring the examinee's answers and generating feedback, means for analyzing the examinee's performance data and improving the next exam questions and feedback, and means for detecting the examinee's emotions and stress level using an emotion recognition engine and providing feedback that takes this into account. This makes it possible to automate the exam and provide appropriate feedback according to the examinee's emotions and stress level.

[0617] The "internal database" is a digital repository for collecting and storing qualification exam data, and for retrieving and updating it as needed.

[0618] "Preprocessing" is the process of detecting missing or anomaly data in the collected data, performing necessary completion or correction, and preparing it in a format suitable for training data for generative AI.

[0619] "Generative AI" refers to artificial intelligence that uses machine learning algorithms to learn and automatically generate questions for certification exams based on specific patterns and evaluation criteria.

[0620] The "first-stage examination" refers to a type of examination that includes multiple-choice and short-answer questions used to assess basic knowledge.

[0621] The "second examination" is a type of test that includes scenario-based questions and case studies used to assess practical skills.

[0622] "Real-time evaluation" is a process that instantly evaluates the answers entered by test-takers during the exam and provides necessary feedback on the spot.

[0623] "Automated scoring" is a system that uses generative AI or emotion recognition engines to mechanically evaluate test-takers' answers and calculate an overall score.

[0624] "Feedback" refers to information provided regarding the test taker's answers, including evaluations, comments, areas for improvement, and strengths, which is used as reference for future exams and learning.

[0625] "Performance data" refers to data that indicates a test-taker's ability to perform the test, including the accuracy and quality of answers they provided during the test, as well as their emotional state.

[0626] The "emotion recognition engine" is a technology that analyzes the facial expressions and voice of test takers to detect their emotions and stress levels, and is used for real-time feedback.

[0627] The system of the present invention improves efficiency and the working environment by recognizing the emotions and stress levels of workers in factory work in real time and providing appropriate feedback. Specific embodiments are shown below.

[0628] Data Acquisition and Preprocessing

[0629] The server first collects historical work data and environmental data from its internal database. This data includes past work performance, worker performance data, and operating data of the machinery used. The collected data is preprocessed using libraries such as OpenCV and Dlib. Preprocessing involves imputing missing data and correcting anomalous data, preparing the data as a training dataset for generative AI.

[0630] Learning of generative AI

[0631] The server trains a generative AI based on pre-processed data. During this process, the AI ​​learns evaluation criteria for tasks and past performance patterns, building the foundation necessary for improving work efficiency and generating optimal work processes. OpenAI's GPT model is used as the generative AI to generate specific feedback based on prompt text.

[0632] Use of emotion recognition engine

[0633] The emotion recognition engine is used by the server to analyze the worker's facial expressions and voice data in real time. It utilizes OpenCV and Dlib for facial recognition, Google Cloud Speech-to-Text for speech recognition, and Affectiva and Emotion API (Microsoft) for emotion analysis to detect the worker's emotions and stress levels.

[0634] Provide feedback

[0635] Based on data obtained from the emotion recognition engine, the server uses generative AI to provide appropriate feedback to workers in real time. This feedback covers a wide range of topics, including adjusting work pace, suggesting breaks, and offering advice to reduce stress.

[0636] Specific example:

[0637] For example, if "anger" is detected from the worker's facial expression, the OpenAI GPT model generates feedback using the following prompt message.

[0638] Based on the factory worker's sentiment data: 'Angry', generate appropriate feedback.

[0639] In response, an example of the generated feedback is a comment such as, "You appear to be feeling stressed right now. We recommend taking a 5-minute break."

[0640] Performance data analysis and improvement

[0641] The server accumulates worker performance data daily and performs periodic analysis. This data is used to optimize future work processes and improve feedback. Furthermore, by incorporating data from the emotion recognition engine as AI training data, it becomes possible to provide more precise guidance and support based on the worker's emotions.

[0642] The system described above can improve the efficiency of factory operations, provide appropriate feedback tailored to workers' emotions and stress levels, and significantly improve the work environment.

[0643] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0644] Step 1:

[0645] The server collects historical work data and environmental data from its internal database and performs preprocessing. In this step, missing and anomalous data are supplemented and corrected, and the data is prepared in a format suitable for use as a training dataset for generative AI. Specifically, OpenCV and Dlib are used to analyze and cleanse the data. The input is data obtained from the internal database, and the output is a preprocessed dataset.

[0646] Step 2:

[0647] The server trains a generative AI using preprocessed data. During the training process, the AI ​​learns evaluation criteria for tasks and past performance patterns. The OpenAI GPT model is used in this step. The input is a preprocessed dataset, and the output is a trained generative AI model.

[0648] Step 3:

[0649] The server uses an emotion recognition engine to analyze the worker's facial expressions and voice data in real time. It uses OpenCV and Dlib for facial recognition, Google Cloud Speech-to-Text for speech recognition, and Affectiva and Emotion API (Microsoft) for emotion analysis. Inputs include camera video and audio data, and output data indicating emotions and stress levels is generated.

[0650] Step 4:

[0651] The server uses generative AI to provide appropriate feedback to workers in real time, based on data obtained from the emotion recognition engine. It creates prompt messages and generates feedback content using OpenAI's GPT model. The input consists of emotion analysis data and prompt messages, and the output is a feedback message. For example, a prompt message such as "Generate appropriate feedback based on the factory worker's emotion data: 'Angry'" might be used.

[0652] Step 5:

[0653] The server provides workers with real-time generated feedback, including suggestions for adjusting work pace, taking breaks, and stress reduction advice. The input is generated feedback messages, and the output is specific actions or messages provided to the worker.

[0654] Step 6:

[0655] The server stores and analyzes worker performance data. This data is used to optimize future work processes and improve feedback. Real-time and historical performance data are used as input, and analysis results and improved feedback are generated as output. By feeding this analysis result back into the generative AI as new training data, more precise guidance and support become possible.

[0656] 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 user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0657] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0658] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the smart glasses 214.

[0659] [Third Embodiment]

[0660] Figure 5 shows an example of the configuration of the data processing system 310 according to the third embodiment.

[0661] As shown in Figure 5, the data processing system 310 includes a data processing device 12 and a headset terminal 314. An example of the data processing device 12 is a server.

[0662] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0663] The headset terminal 314 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a display 343. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and display 343 are also connected to the bus 52.

[0664] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0665] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0666] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0667] Figure 6 shows an example of the main functions of the data processing device 12 and the headset terminal 314. As shown in Figure 6, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0668] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0669] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0670] In the headset terminal 314, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0671] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the headset terminal 314 will be referred to as the "terminal".

[0672] The system of the present invention aims to automate qualification examinations, with a server at the center executing multiple functions. The specific procedure for implementing the present invention is shown below.

[0673] Collection and preprocessing of qualification exam data

[0674] The server collects data related to certification exams from internal databases within the company. This includes past exam questions, examinee results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. Here, it checks for any missing or abnormal data and supplements or corrects it as needed. This preprocessed data is then used directly as training data for generative AI.

[0675] Learning of generative AI

[0676] The server trains a generative AI based on pre-processed data. During this process, the AI ​​learns the evaluation criteria for certification exams, patterns in past exam questions, and the quality of answers. This creates a foundation for the AI ​​to evaluate test-takers' responses and generate appropriate exam questions.

[0677] Automatic generation of exam questions

[0678] Once the training is complete, the generative AI automatically generates questions for the first and second exams according to the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge. Specific examples include questions such as, "What are the basic steps of the sales process?" The second exam includes scenario-based questions and case studies to assess practical skills. For example, it generates practical questions such as, "What is the best way to handle a customer complaint?"

[0679] Exam distribution and administration

[0680] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers through their device. During the exam, the server receives the examinee's answers in real time and performs automatic evaluation. Additional questions may be automatically generated based on the accuracy rate and quality of the answers. This process continues into the second exam after the first exam is completed.

[0681] Grading and feedback

[0682] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. The automated scoring system evaluates each answer and calculates a total score. Based on the evaluation results, the server generates detailed feedback. This feedback includes an evaluation of each question, the test-taker's strengths, and areas for improvement. The server compiles this feedback for each test-taker and sends it to the user (test-taker) via email.

[0683] Performance data analysis and feedback improvement

[0684] The server collects and analyzes test-taker performance data. This analysis is then used to improve future exam questions and feedback. In addition, the server collects reviews from test-takers and incorporates them as training data for generative AI, thereby improving the overall system quality.

[0685] As described above, the server-centered automated system streamlines the execution of certification exams and enables the rapid provision of high-quality feedback. This system not only saves human resources but also contributes to improved consistency in evaluations and enhanced skills among test-takers.

[0686] The following describes the processing flow.

[0687] Step 1:

[0688] The server collects data related to the certification exam from its internal database. Specifically, it retrieves past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, it checks for any missing or abnormal data and supplements or corrects it as needed. This pre-processed data is then used as training data for generative AI.

[0689] Step 2:

[0690] The server trains the generative AI based on pre-processed data. Specifically, it trains the AI ​​on evaluation criteria for certification exams, patterns of past exam questions, and the quality of answers. Through this learning process, the generative AI builds a foundation for creating appropriate exam questions and evaluating answers.

[0691] Step 3:

[0692] The server uses generative AI to automatically generate questions for the first and second exams. For the first exam, it generates multiple-choice and short-answer questions to assess basic knowledge. For the second exam, it generates scenario-based questions and case studies to assess practical skills. For example, it generates questions such as, "What are the basic steps of the sales process?" and "What is the best way to handle a customer complaint?"

[0693] Step 4:

[0694] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers using their device. During the exam, the server receives the examinee's answers in real time and performs automatic evaluation. Based on this evaluation, additional questions may be automatically generated as needed.

[0695] Step 5:

[0696] Once a user completes the first exam, the server immediately delivers the questions for the second exam. The user then takes the second exam within the same Zoom session. The answers to the second exam are also evaluated in real time, and the evaluation results are continuously fed back to the user.

[0697] Step 6:

[0698] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. Using an automated scoring system, it evaluates the accuracy and quality of each answer and calculates an overall score. Based on the evaluation results, detailed feedback is generated. The feedback includes an evaluation of each question, the test-taker's strengths, and areas for improvement.

[0699] Step 7:

[0700] The server compiles this feedback for each test taker and sends it to the user (test taker) via email. The feedback is also recorded in the company's internal database and used as reference material for future tests and evaluations.

[0701] Step 8:

[0702] The server collects and analyzes test-takers' performance data. This analysis is used to improve future test questions and feedback. Furthermore, the server collects reviews from test-takers and incorporates them as AI training data, thereby improving the overall quality of the system.

[0703] Through the steps outlined above, this system automates the entire certification examination process, reducing human resources and improving the consistency of evaluations. It also enables the rapid provision of high-quality feedback to test-takers.

[0704] (Example 1)

[0705] Next, we will describe Example 1. 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."

[0706] Traditional certification examination systems require significant human resources and are inefficient because the creation of exam questions, evaluation of answers, and provision of feedback are all done manually. Furthermore, consistency in evaluation and quality of feedback are prone to inconsistencies. Additionally, the analysis of test-taker performance data and the improvement of feedback based on that analysis are not adequately performed. This hinders the improvement of test-takers' abilities and fair evaluation.

[0707] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[0708] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data to supplement and correct any deficiencies or anomalies, means for training a generative AI based on the preprocessed data, means for automatically generating exam questions for the first and second exams in a specified format, means for distributing the exams and evaluating examinees' answers in real time, means for automatically scoring examinees' answers, generating detailed feedback, and sending it via email, and means for analyzing examinee performance data to improve the next exam questions and feedback. This enables automation and efficiency of the exam, improved consistency of evaluation, and rapid provision of high-quality feedback.

[0709] An "internal database" refers to a collection of information and data related to certification exams that is maintained within a specific organization or system.

[0710] "Qualification exam data" refers to a variety of data related to the exam, including past exam questions, examinee results, evaluation criteria, and examinee performance data.

[0711] "Preprocessing" refers to the process of performing tasks such as imputing missing values, correcting outliers, and formatting the collected data.

[0712] "Generative AI" refers to artificial intelligence systems that learn from collected and pre-processed data and automatically generate new test questions and answers.

[0713] The term "first-stage examination" refers to the initial stage of an exam that includes multiple-choice and short-answer questions designed to assess basic knowledge.

[0714] The term "second examination" refers to the second stage of the examination, which includes scenario-based problems and case studies designed to assess practical skills.

[0715] "Exam delivery" refers to the process of providing exam questions to test takers and having them take the exam.

[0716] "Real-time evaluation" refers to a process where the answers entered by test-takers during the exam are immediately analyzed and evaluated.

[0717] "Automated scoring" refers to a process that automatically evaluates and scores test-takers' answers to exam questions.

[0718] "Feedback" refers to information generated based on a test-taker's exam results, including evaluation results and suggestions for improvement.

[0719] "Performance data" refers to data related to test-takers' behavior during the exam, the quality of their answers, and the accuracy of their responses.

[0720] "Analysis" refers to the process of examining collected data and extracting meaningful patterns and trends.

[0721] The system of this invention aims to automate qualification examinations and performs various processes centered around a server. This system is implemented in the following specific steps.

[0722] Collection and preprocessing of qualification exam data

[0723] The server first connects to an internal database to collect data related to the certification exam. This includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. Specifically, it checks for missing or abnormal data and then fills in or corrects them. For example, missing values ​​are filled in with the mean, and abnormal values ​​are corrected appropriately. Python's Pandas library and SQL queries are used for this preprocessing.

[0724] Learning of generative AI

[0725] The pre-processed data is used as training data for generative AI. The server uses this data to train the AI ​​model. The AI ​​models used include advanced generative AIs such as GPT-4. In this process, deep learning frameworks such as TensorFlow and PyTorch are utilized, and the model learns patterns and evaluation criteria from past exam questions. As a result, the AI ​​acquires the ability to automatically generate the next exam questions.

[0726] Automatic generation of exam questions

[0727] Once the generative AI has completed its training, it automatically generates questions for the first and second exams according to the specified exam format. For example, the first exam generates multiple-choice and short-answer questions to assess basic knowledge. On the other hand, the second exam includes scenario-based questions and case studies to assess practical skills.

[0728] Example of a prompt

[0729] "What are the basic steps of the sales process?"

[0730] "What is the best way to handle a customer complaint that arises during customer service?"

[0731] Exam distribution and administration

[0732] Next, the server handles the delivery and administration of the exam. Users (examinees) log in to ZOOM from their devices at the designated exam date and time. The server delivers the first exam question set through the ZOOM session, and users answer it. Answers are sent to the server in real time, and the server evaluates them immediately. In some cases, additional questions may be automatically generated based on the accuracy rate and quality of the answers.

[0733] Grading and feedback

[0734] After the exam is completed, the server compiles all the answers and evaluates them using an automated scoring system. Based on the evaluation results, detailed feedback is generated and sent to the user (test taker) via email. This feedback includes an evaluation of each question, the test taker's strengths, and areas for improvement.

[0735] Performance data analysis and feedback improvement

[0736] The server collects and analyzes test-takers' performance data. This analysis is used to improve the creation of future test questions and feedback. Furthermore, reviews from test-takers are collected and incorporated as training data for generative AI, thereby improving the overall quality of the system.

[0737] As described above, the present invention's automated qualification examination system automates a series of processes centered on a server, including data collection, preprocessing, training of a generative AI, automatic generation of examination questions, distribution and administration of the examination, scoring and generation of feedback, and analysis and improvement of performance data, thereby realizing efficient and high-quality examination management.

[0738] The flow of the specific processing in Example 1 will be explained using Figure 11.

[0739] Step 1: Start data collection

[0740] The server establishes a connection to the internal database and executes an SQL query to extract "certification exam data" (past exam questions, examinee results, evaluation criteria, etc.). For example, it might execute a query like "SELECT FROM ExamData WHERE Year BETWEEN 2018 AND 2023".

[0741] Input: Connection information for the internal database

[0742] Output: Extraction results of qualification exam data

[0743] Step 2: Data checking and preprocessing

[0744] The server reads the collected data using Python's Pandas library and checks for missing values ​​and outliers. For example, it checks for missing values ​​using df.isnull().sum() and imputes them with df.fillna(df.mean()). If outliers exist, it corrects them using an appropriate method.

[0745] Input: Certification exam data extracted in Step 1

[0746] Output: Preprocessed data

[0747] Step 3: Preparing training data for generative AI

[0748] The server organizes the preprocessed data as training data for generative AI and converts it into an applicable format. For example, it selects the necessary features and converts the data into numerical vectors.

[0749] Input: Preprocessed data obtained in Step 2

[0750] Output: Training dataset

[0751] Step 4: Training the generative AI model

[0752] The server uses deep learning frameworks such as TensorFlow or PyTorch to train generative AI models. For example, it trains the model using the form model.fit(X_train, y_train, epochs=100).

[0753] Input: Training dataset prepared in Step 3

[0754] Output: Trained generative AI model

[0755] Step 5: Automatic generation of primary exam questions

[0756] The server uses a pre-trained generative AI model to automatically generate questions for the first stage of the exam. For example, it might create a multiple-choice question such as, "What are the basic steps of the sales process?"

[0757] Input: Pre-trained generative AI model

[0758] Output: First Exam Question Set

[0759] Step 6: Automatic generation of secondary examination questions

[0760] The server uses a pre-trained generative AI model to automatically generate questions for the second stage of the exam. For example, it creates scenario questions such as, "What is the best way to handle a customer complaint that arises during customer service?"

[0761] Input: Pre-trained generative AI model

[0762] Output: Secondary Exam Question Set

[0763] Step 7: Notification of exam date and time

[0764] The server will send a notification email to the test taker containing the date, time, and link to join the ZOOM session.

[0765] Input: Candidate information, exam schedule

[0766] Output: Notification email

[0767] Step 8: Exam delivery and real-time collection of test-takers' responses

[0768] The user (examinee) logs into ZOOM from their device at the designated date and time. The server delivers the first-stage exam question set via the ZOOM session, and the user enters their answers. The server collects and evaluates the answers sent from the device in real time.

[0769] Input: First-stage exam question set, examinee's answers

[0770] Output: Real-time evaluation results

[0771] Step 9: Generating additional questions based on accuracy and quality.

[0772] The server evaluates the accuracy rate and quality of answers based on the responses received in real time, and generates additional questions as needed.

[0773] Input: Test-takers' responses, real-time evaluation results

[0774] Output: Additional questions

[0775] Step 10: Aggregation of test results and generation of feedback

[0776] After the exam is completed, the server compiles all the answers and calculates an overall score using an automated scoring system. Next, it generates detailed feedback and sends it via email.

[0777] Input: All responses from test takers

[0778] Output: Overall score, feedback email

[0779] Step 11: Accumulate and analyze performance data

[0780] The server collects and analyzes test-takers' performance data. This analysis is used to create future exam questions and improve feedback.

[0781] Input: Test-taker's performance data

[0782] Output: analysis results, improvement suggestions

[0783] Through the processing steps described above, the system of the present invention can automate and streamline qualification examinations and provide high-quality feedback quickly.

[0784] (Application Example 1)

[0785] Next, we will explain Application Example 1. In the following explanation, 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."

[0786] Traditional certification exams and training programs require significant human resources for tasks such as creating exam questions, administering exams, grading, and providing feedback. This often leads to inefficient exam administration and the risk of biased human evaluation. Furthermore, in on-site employee training, such as in factories, individual evaluation and feedback are difficult to provide, hindering effective training. To address these challenges, an efficient and consistent automated evaluation system is needed.

[0787] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[0788] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and training a generative AI, means for automatically generating exam questions for the first and second exams, means for distributing the exams and evaluating examinees' answers in real time, means for automatically scoring examinees' answers and generating feedback, means for analyzing examinee performance data and improving the next exam questions and feedback, and means for a robot to generate training questions similar to qualification exams and provide real-time feedback during factory employee training. This streamlines the employee training and evaluation process and enables the provision of consistent, high-quality feedback.

[0789] An "internal database" is a database that stores data related to certification exams and allows users to retrieve that data as needed.

[0790] "Qualification exam data" refers to all data related to the administration of qualification exams, including past exam questions, examinee results, evaluation criteria, and examinee performance data.

[0791] "Preprocessing" is the process of supplementing and correcting any deficiencies or anomalies in the collected data and converting it into a format suitable for learning by generative AI.

[0792] "Generative AI" is a type of artificial intelligence that uses large amounts of data to train models and generate test questions.

[0793] The "first-stage examination" is an exam that includes multiple-choice and short-answer questions designed to assess basic knowledge.

[0794] The "second examination" is a test that includes scenario-based problems and case studies, designed to evaluate practical skills.

[0795] "Automatic scoring" is a process that automatically evaluates the test-takers' answers and calculates their scores.

[0796] "Feedback" refers to providing detailed comments on the test-taker's answers, along with an evaluation of their responses, including their strengths and areas for improvement.

[0797] "Performance data" refers to the overall results a test-taker demonstrated during the exam, along with various analytical data based on those results.

[0798] "Factory employee training" refers to the process of educating and providing factory employees with the necessary skills and knowledge to perform their duties.

[0799] "Real-time evaluation" is a process in which, each time a test-taker submits their answer, it is immediately evaluated and the results are provided.

[0800] "Training problems" are questions or tasks presented to assess an employee's abilities.

[0801] To implement this invention, it is first necessary to collect data related to qualification exams from an internal database. The server retrieves data such as past exam questions, examinee results, and evaluation criteria from the database and preprocesses this data. In this preprocessing, any missing or abnormal data is supplemented and corrected, and the data is standardized.

[0802] Next, the server uses the pre-processed data to train a generative AI. A natural language processing (NLP) model, such as GPT-3, is suitable as a generative AI model. Based on the training data, the AI ​​model learns the question patterns and evaluation criteria of certification exams, enabling the automatic generation of exam questions.

[0803] Once the training is complete, the generative AI generates questions for the first and second exams based on the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge, including specific examples such as "What are basic safety procedures?" The second exam generates scenario questions and case studies to assess practical skills, such as "Explain how to respond if an emergency occurs during work."

[0804] The generated test questions are delivered to factory employees' terminals and robots via a server. Employees provide answers to the terminals or robots, and the server evaluates the answers in real time. An AI-powered automated scoring system is used to evaluate the answers, and feedback is generated based on this evaluation. The feedback includes not only whether the answer is correct or incorrect, but also the employee's strengths and areas for improvement.

[0805] Furthermore, the server collects employee performance data and analyzes it to improve the quality of future exam questions and feedback. This analysis utilizes statistical analysis tools (such as Python's Pandas and NumPy) to identify data trends and develop more effective training programs.

[0806] As a concrete example, a factory robot presents an employee with a training question such as, "Explain how to use the emergency stop button," and the employee answers using voice or controls. The AI ​​evaluates the answer in real time and provides feedback such as, "The procedure itself is correct, but safety checks are missing." In this way, employees can learn practical skills on the spot while being efficiently evaluated.

[0807] Examples of prompt messages include the following:

[0808] Please generate the following training questions regarding factory machine operation:

[0809] 1. Explain how to operate the machine according to the safety procedures.

[0810] 2. Please explain how to use the emergency stop button.

[0811] 3. Please describe the procedure for routine maintenance checks.

[0812] The above describes the specific procedures and system configuration necessary for implementing this invention.

[0813] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[0814] Step 1:

[0815] The server collects data related to certification exams from an internal database. The data retrieved from the database includes past exam questions, examinee results, evaluation criteria, and examinee performance data. All data related to the certification exams is retrieved from the database as input. The collected dataset is passed to a preprocessing module as output.

[0816] Step 2:

[0817] The server preprocesses the collected data. Preprocessing includes imputing missing values, correcting anomalous data, and normalization. The input is the data collected in step 1, and the output is the preprocessed dataset. At this stage, data cleansing and normalization are performed using data preprocessing libraries (e.g., Pandas, NumPy).

[0818] Step 3:

[0819] The server trains a generative AI using preprocessed data. Preprocessed data is used as input, and a generative AI model (e.g., GPT-3) is trained. The output is the trained AI model. Here, the training process is executed using an AI model training framework (e.g., TensorFlow, PyTorch).

[0820] Step 4:

[0821] The server automatically generates questions for the first and second exams using a pre-trained generative AI. The input consists of a pre-trained AI model and prompts, and the output is the automatically generated exam questions. Specifically, the following prompts are input to the AI ​​to generate the questions.

[0822] Please generate the following training questions regarding factory machine operation:

[0823] 1. Explain how to operate the machine according to the safety procedures.

[0824] 2. Please explain how to use the emergency stop button.

[0825] 3. Please describe the procedure for routine maintenance checks.

[0826] Step 5:

[0827] The server distributes the generated test questions to terminals and robots. The input is automatically generated test questions, and the output is the questions distributed to the terminals and robots. The terminals and robots then present these questions to factory workers.

[0828] Step 6:

[0829] The user (factory employee) provides answers to the presented problems using voice or controls. The input is the factory employee's answer, and the output is the answer data sent to the server. The answers are input through a voice recognition system and a sensor network.

[0830] Step 7:

[0831] The server evaluates test-takers' answers in real time. The input is user-submitted answer data, and the output is the evaluation result and feedback. Here, an AI-powered automated scoring system evaluates the answers and provides real-time feedback in text and audio formats.

[0832] Step 8:

[0833] The server collects employee performance data and analyzes it to improve the quality of future training programs and feedback. Real-time evaluation results are used as input, and the output is an improved training program and feedback. Statistical analysis tools (e.g., Python's Pandas, NumPy) are used to perform data analysis and incorporate the findings into future improvements.

[0834] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[0835] The system of the present invention combines automation of qualification examinations with an emotion engine to recognize user emotions and reflect them in the feedback. Specific embodiments for carrying out the present invention are shown below.

[0836] Collection and preprocessing of qualification exam data

[0837] The server collects data related to certification exams from the company's internal database. This includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. During this process, it checks for missing or abnormal data and supplements or corrects it as needed. The preprocessed data is then used as training data for generative AI.

[0838] Learning of generative AI

[0839] The server trains a generative AI based on pre-processed data. During the training process, the AI ​​learns the evaluation criteria for the certification exam, patterns of past exam questions, and the quality of answers. This allows the AI ​​to build a foundation for accurately creating exam questions and appropriately evaluating answers.

[0840] Automatic generation of exam questions

[0841] Once the training is complete, the generative AI automatically generates questions for the first and second exams according to the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge. For example, one question might be, "What are the basic steps of the sales process?" The second exam generates scenario questions and case studies to assess practical skills. For example, one might be, "What is the best way to handle a customer complaint that arises during customer service?"

[0842] Exam distribution and administration

[0843] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers using their device. During the exam, the server receives the examinee's answers in real time and evaluates them using an emotion engine. The emotion engine recognizes emotions from the examinee's facial expressions and voice and uses this to perform real-time evaluations.

[0844] Scoring and feedback generation

[0845] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. Through an automated scoring system, it evaluates the accuracy and quality of each answer and calculates a total score. Based on the evaluation results, the server generates detailed feedback. This feedback includes evaluations of each question, the test-taker's strengths and areas for improvement, as well as comments on emotional states detected by the emotion engine. This feedback more accurately reflects the test-taker's performance and provides information to help improve next time.

[0846] Performance data analysis and improvement

[0847] The server stores and analyzes test-takers' performance data. This analysis is used to create future test questions and improve feedback. Furthermore, by incorporating emotion engine data as AI learning data, the system can provide guidance and support based on the test-takers' emotions.

[0848] Applications of the Emotion Engine

[0849] The emotion engine also detects the test-taker's stress level and concentration level. For example, if the stress level is high, the server will extend the answer time or provide additional support information. By providing questions and feedback that take emotional states into account, it is possible to help test-takers perform at their best.

[0850] This combination of systems allows for a highly automated qualification examination process, reflecting test-takers' emotions and stress levels in real time, thereby providing more appropriate and effective feedback. This not only reduces human resources and improves the consistency of evaluations, but also alleviates the mental burden on test-takers and maximizes learning effectiveness.

[0851] The following describes the processing flow.

[0852] Step 1:

[0853] The server collects data related to the certification exam from its internal database. This includes past exam questions, test-takers' results, and performance data for top and bottom-ranked candidates. Next, it checks for missing or abnormal data and supplements or corrects it as needed.

[0854] Step 2:

[0855] The server trains a generative AI based on pre-processed data. Specifically, it learns the evaluation criteria for certification exams, past exam question patterns, and the quality of answers. Through this process, the generative AI builds the foundation for creating exam questions and evaluating answers.

[0856] Step 3:

[0857] The server uses generative AI to automatically generate questions for the first and second exams. The first exam generates multiple-choice and short-answer questions to assess basic knowledge, while the second exam generates scenario-based questions and case studies to assess practical skills. Examples of such questions include, "What are the basic steps of the sales process?" and "What is the best way to handle a customer complaint?"

[0858] Step 4:

[0859] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers on their device.

[0860] Step 5:

[0861] During the exam, the server receives test-takers' answers in real time and evaluates them using an emotion engine. For example, it recognizes emotions from the test-taker's facial expressions and voice via camera and microphone, and analyzes their stress and tension levels.

[0862] Step 6:

[0863] Based on the analysis results from the emotion engine, the server generates additional questions and instructions in real time that take into account the test-taker's emotional state. For example, if a test-taker is feeling stressed, the difficulty level of the test questions may be adjusted or the time allotted for answering may be extended.

[0864] Step 7:

[0865] After the first exam is completed, the server immediately delivers the questions for the second exam. Users continue to answer the second exam questions within the same ZOOM session. The answers to the second exam are also evaluated in real time, and the evaluation results are continuously fed back through the sentiment engine.

[0866] Step 8:

[0867] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. An automated scoring system is used to evaluate the accuracy and quality of each answer and calculate the overall score.

[0868] Step 9:

[0869] The server generates detailed feedback based on the evaluation results. This feedback includes an evaluation of each problem, the test-taker's strengths and areas for improvement, as well as comments on the emotional state detected by the emotion engine. The server compiles this feedback for each test-taker and sends it to the user (test-taker) via email.

[0870] Step 10:

[0871] The server stores and analyzes test-takers' performance data. This analysis is used to create future test questions and improve feedback. Furthermore, incorporating emotion engine data as AI training data enables more accurate emotion recognition and feedback.

[0872] Through these steps, an automated certification exam system incorporating an emotion engine can reflect the test-taker's emotions in real time, providing high-quality feedback while reducing mental stress.

[0873] (Example 2)

[0874] Next, we will describe Example 2. 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."

[0875] In the process of certification examinations, there is a demand for automation and efficiency, as well as feedback that reflects the emotions and stress levels of test-takers in real time. However, in conventional systems, the generation and evaluation of exam questions are often done manually, which is inefficient. Furthermore, there has been a lack of technology to provide feedback that reflects the emotional state of test-takers. Therefore, providing accurate, real-time evaluation and feedback, and efficiently managing certification examinations are challenges.

[0876] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[0877] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and supplementing / correcting missing or abnormal data, means for training a generative AI model based on the preprocessed data, means for automatically generating exam questions for the first and second exams using prompt sentences, means for receiving the exams and examinees' answers in real time and evaluating them using an emotion engine, means for automatically scoring examinees' answers and generating feedback, means for analyzing examinees' performance data and improving the next exam questions and feedback, and means for detecting the examinee's emotional state and evaluating stress levels and concentration levels. This makes it possible to highly automate the qualification exam process and provide accurate and effective feedback that reflects the examinee's emotions in real time.

[0878] An "internal database" is a centrally managed collection of information within a system, and its role is to store and provide data related to certification exams.

[0879] "Qualification exam data" refers to data containing information necessary for qualification exams, specifically including past exam questions, examinee results, evaluation criteria, and performance data.

[0880] "Data preprocessing" is the process of converting collected raw data into a format that can be used for analysis and machine learning, and includes supplementing and correcting missing or abnormal data.

[0881] A "generative AI model" is a type of artificial intelligence technology that learns from data and automatically generates new test questions based on the results of that learning.

[0882] A "prompt sentence" is an input sentence used by a generative AI model, and it is an instruction sentence that the AI ​​uses to generate answers or test questions.

[0883] The "emotion engine" is a system that recognizes and evaluates the emotional state of test takers in real time, detecting emotions based on data such as facial expressions and voice.

[0884] "Real-time evaluation" is a process that involves immediately analyzing and evaluating the test-taker's answers and emotional state during the exam, allowing for feedback and responses without delay.

[0885] "Automated scoring" is a process that uses algorithms or AI to mechanically evaluate test-takers' answers and calculate the accuracy rate and quality.

[0886] "Feedback" refers to evaluation information and suggestions for improvement generated based on the test taker's exam results and performance, and is intended to support the test taker in achieving better results in the next exam.

[0887] "Performance data" refers to data that includes all performance indicators related to the exam, such as the test-taker's response status during the exam, response speed, accuracy rate, and emotional state.

[0888] "Stress level" is an indicator that shows the degree of mental burden on test takers during the exam, and it quantitatively evaluates emotional states such as tension and anxiety.

[0889] "Concentration level" is an indicator used to evaluate the level of attention and focus of test-takers during an exam, and a high level of concentration is considered to lead to high performance.

[0890] This system is designed to highly automate the qualification examination process and provide real-time feedback that reflects the examinee's emotions and stress levels. The specific implementation procedure is as follows.

[0891] Collection and preprocessing of qualification exam data

[0892] The server collects data related to certification exams from the company's internal database. This data includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. The collected data is checked for missing or abnormal data and supplemented or corrected as needed. For example, the Python Pandas library is used to impute missing values ​​in the data.

[0893] Learning of generative AI

[0894] The server trains a generative AI model based on pre-processed data. Specifically, it uses past exam questions and test-takers' answers as input data and trains the AI ​​model using machine learning libraries such as TensorFlow and PyTorch. This allows the AI ​​to learn the evaluation criteria for the certification exam, patterns in past exam questions, and the quality of the answers.

[0895] Automatic generation of exam questions

[0896] Once the training is complete, the generative AI will automatically generate questions for the first and second exams according to the specified exam format. The first exam will generate multiple-choice and short-answer questions to assess basic knowledge. For example, it will use prompts like the following:

[0897] "What are the basic steps of the sales process?"

[0898] In the second stage of the exam, scenario-based problems and case studies are generated to assess practical skills. For example,

[0899] "What is the best way to handle a customer complaint that arises during customer service?"

[0900] Exam distribution and administration

[0901] The user (examinee) logs into the online conferencing system from their device at the designated exam date and time. The server delivers the first-stage exam question set through the online conferencing system. The user enters their answers using their device and sends them to the server. During the exam, the server uses Web-RTC technology to capture the examinee's facial expressions and collect data in real time.

[0902] Scoring and feedback generation

[0903] After the exam ends, the server compiles all the answers and begins evaluation through an automated scoring system. It assesses the accuracy and quality of each answer and calculates an overall score. Based on the evaluation results, the server generates detailed feedback. This feedback includes evaluations of each question, the test-taker's strengths and areas for improvement, as well as comments on emotional states detected by the emotion engine.

[0904] Performance data analysis and improvement

[0905] The server stores and analyzes test-takers' performance data. This analysis is used to improve future test questions and feedback. Furthermore, data acquired by the emotion engine is incorporated as AI learning data, enabling guidance and support based on the test-takers' emotions.

[0906] Applications of the Emotion Engine

[0907] The emotion engine detects the test-taker's stress level and concentration level in real time. For example, if the stress level is high, the server will provide an extension of the answer time or additional support information. Specific feedback includes:

[0908] "Your stress level appears high. We have extended your response time by 10 minutes."

[0909] This system, designed in this way, can efficiently automate the qualification examination process, reduce the mental burden on test-takers, and provide accurate and effective feedback.

[0910] The flow of the specific processing in Example 2 will be explained using Figure 13.

[0911] Step 1:

[0912] The server collects past exam questions, test-taker results, evaluation criteria, and performance data for top and bottom-ranked test-takers from an internal database. The data to be collected is easily extracted using SQL queries. It receives connection information to the internal database and data retrieval conditions as input, and outputs the necessary exam data based on that. Specifically, it executes SQL queries to extract the required data and stores it on the server.

[0913] Step 2:

[0914] The server preprocesses the collected data. This includes checking for missing or abnormal data and supplementing or correcting them. It receives the raw data collected in step 1 as input and outputs preprocessed, clean data. Specifically, it uses the Python Pandas library to impute missing values ​​and correct invalid data.

[0915] Step 3:

[0916] The server trains a generative AI model based on preprocessed data. It receives preprocessed data as input and uses it as training data. The output is the trained generative AI model. Specifically, it defines and trains the model using machine learning libraries such as TensorFlow and PyTorch.

[0917] Step 4:

[0918] Once the training is complete, the generative AI model automatically generates questions for the first and second exams. It takes prompts for exam format and basic knowledge questions as input and provides generated exam questions as output. An example of a prompt is "What are the basic steps of the sales process?" In terms of operation, the generative AI is given prompts and generates exam questions based on them.

[0919] Step 5:

[0920] The user (examinee) logs into the online conferencing system from their device at the designated exam date and time. The server distributes the exam questions, and the user enters their answers. The input data consists of the examinee's answers and facial expression data captured in real time. The output consists of the examinee's answer data and emotional state data. Specifically, the answers entered from the device and the facial expression data collected using WEB-RTC technology are sent to the server.

[0921] Step 6:

[0922] After the exam ends, the server collects all the answers and performs automatic scoring and evaluation. It receives the test-takers' answer data as input, evaluates the accuracy and quality of each answer as output, and generates an overall score. Specifically, it uses Python's Scikit-learn to calculate the accuracy of each answer and evaluates them using a dedicated algorithm.

[0923] Step 7:

[0924] The server generates feedback based on the aggregated results. It receives evaluation results and sentiment data from the sentiment engine as input, and generates a feedback document as output. Specifically, it creates detailed feedback including evaluations of each problem, the test-taker's strengths, areas for improvement, and comments on the sentiment state detected by the sentiment engine.

[0925] Step 8:

[0926] The server stores and analyzes test-taker performance data. It receives individual test-taker performance data as input and outputs the next exam questions and feedback for improvement. Specifically, it uses the Python Seaborn library to visualize the data and analyze trends.

[0927] Step 9:

[0928] The server uses an emotion engine to detect the test-taker's stress level and concentration level. It receives real-time facial expression and voice data as input and provides stress level and concentration level as output. Specifically, it uses facial expression and voice analysis algorithms to evaluate the emotional state and adjusts the test environment based on that.

[0929] (Application Example 2)

[0930] Next, we will explain application example 2. In the following explanation, 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."

[0931] The current certification examination system suffers from a lack of automation and difficulty in providing feedback that takes into account the emotions and stress levels of test-takers. This increases the mental burden on test-takers, making it difficult for them to perform at their best. Furthermore, the inconsistency in exam evaluation and feedback makes it difficult to accurately identify areas for improvement.

[0932] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and training a generative AI, means for automatically generating exam questions for the first and second exams, means for distributing the exam and evaluating the examinee's answers in real time, means for automatically scoring the examinee's answers and generating feedback, means for analyzing the examinee's performance data and improving the next exam questions and feedback, and means for detecting the examinee's emotions and stress level using an emotion recognition engine and providing feedback that takes this into account. This makes it possible to automate the exam and provide appropriate feedback according to the examinee's emotions and stress level.

[0933] The "internal database" is a digital repository for collecting and storing qualification exam data, and for retrieving and updating it as needed.

[0934] "Preprocessing" is the process of detecting missing or anomaly data in the collected data, performing necessary completion or correction, and preparing it in a format suitable for training data for generative AI.

[0935] "Generative AI" refers to artificial intelligence that uses machine learning algorithms to learn and automatically generate questions for certification exams based on specific patterns and evaluation criteria.

[0936] The "first-stage examination" refers to a type of examination that includes multiple-choice and short-answer questions used to assess basic knowledge.

[0937] The "second examination" is a type of test that includes scenario-based questions and case studies used to assess practical skills.

[0938] "Real-time evaluation" is a process that instantly evaluates the answers entered by test-takers during the exam and provides necessary feedback on the spot.

[0939] "Automated scoring" is a system that uses generative AI or emotion recognition engines to mechanically evaluate test-takers' answers and calculate an overall score.

[0940] "Feedback" refers to information provided regarding the test taker's answers, including evaluations, comments, areas for improvement, and strengths, which is used as reference for future exams and learning.

[0941] "Performance data" refers to data that indicates a test-taker's ability to perform the test, including the accuracy and quality of answers they provided during the test, as well as their emotional state.

[0942] The "emotion recognition engine" is a technology that analyzes the facial expressions and voice of test takers to detect their emotions and stress levels, and is used for real-time feedback.

[0943] The system of the present invention improves efficiency and the working environment by recognizing the emotions and stress levels of workers in factory work in real time and providing appropriate feedback. Specific embodiments are shown below.

[0944] Data Acquisition and Preprocessing

[0945] The server first collects historical work data and environmental data from its internal database. This data includes past work performance, worker performance data, and operating data of the machinery used. The collected data is preprocessed using libraries such as OpenCV and Dlib. Preprocessing involves imputing missing data and correcting anomalous data, preparing the data as a training dataset for generative AI.

[0946] Learning of generative AI

[0947] The server trains a generative AI based on pre-processed data. During this process, the AI ​​learns evaluation criteria for tasks and past performance patterns, building the foundation necessary for improving work efficiency and generating optimal work processes. OpenAI's GPT model is used as the generative AI to generate specific feedback based on prompt text.

[0948] Use of emotion recognition engine

[0949] The emotion recognition engine is used by the server to analyze the worker's facial expressions and voice data in real time. It utilizes OpenCV and Dlib for facial recognition, Google Cloud Speech-to-Text for speech recognition, and Affectiva and Emotion API (Microsoft) for emotion analysis to detect the worker's emotions and stress levels.

[0950] Provide feedback

[0951] Based on data obtained from the emotion recognition engine, the server uses generative AI to provide appropriate feedback to workers in real time. This feedback covers a wide range of topics, including adjusting work pace, suggesting breaks, and offering advice to reduce stress.

[0952] Specific example:

[0953] For example, if "anger" is detected from the worker's facial expression, the OpenAI GPT model generates feedback using the following prompt message.

[0954] Based on the factory worker's sentiment data: 'Angry', generate appropriate feedback.

[0955] In response, an example of the generated feedback is a comment such as, "You appear to be feeling stressed right now. We recommend taking a 5-minute break."

[0956] Performance data analysis and improvement

[0957] The server accumulates worker performance data daily and performs periodic analysis. This data is used to optimize future work processes and improve feedback. Furthermore, by incorporating data from the emotion recognition engine as AI training data, it becomes possible to provide more precise guidance and support based on the worker's emotions.

[0958] The system described above can improve the efficiency of factory operations, provide appropriate feedback tailored to workers' emotions and stress levels, and significantly improve the work environment.

[0959] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[0960] Step 1:

[0961] The server collects historical work data and environmental data from its internal database and performs preprocessing. In this step, missing and anomalous data are supplemented and corrected, and the data is prepared in a format suitable for use as a training dataset for generative AI. Specifically, OpenCV and Dlib are used to analyze and cleanse the data. The input is data obtained from the internal database, and the output is a preprocessed dataset.

[0962] Step 2:

[0963] The server trains a generative AI using preprocessed data. During the training process, the AI ​​learns evaluation criteria for tasks and past performance patterns. The OpenAI GPT model is used in this step. The input is a preprocessed dataset, and the output is a trained generative AI model.

[0964] Step 3:

[0965] The server uses an emotion recognition engine to analyze the worker's facial expressions and voice data in real time. It uses OpenCV and Dlib for facial recognition, Google Cloud Speech-to-Text for speech recognition, and Affectiva and Emotion API (Microsoft) for emotion analysis. Inputs include camera video and audio data, and output data indicating emotions and stress levels is generated.

[0966] Step 4:

[0967] The server uses generative AI to provide appropriate feedback to workers in real time, based on data obtained from the emotion recognition engine. It creates prompt messages and generates feedback content using OpenAI's GPT model. The input consists of emotion analysis data and prompt messages, and the output is a feedback message. For example, a prompt message such as "Generate appropriate feedback based on the factory worker's emotion data: 'Angry'" might be used.

[0968] Step 5:

[0969] The server provides workers with real-time generated feedback, including suggestions for adjusting work pace, taking breaks, and stress reduction advice. The input is generated feedback messages, and the output is specific actions or messages provided to the worker.

[0970] Step 6:

[0971] The server stores and analyzes worker performance data. This data is used to optimize future work processes and improve feedback. Real-time and historical performance data are used as input, and analysis results and improved feedback are generated as output. By feeding this analysis result back into the generative AI as new training data, more precise guidance and support become possible.

[0972] The specific processing unit 290 transmits the result of the specific processing to the headset terminal 314. In the headset terminal 314, the control unit 46A causes the speaker 240 and display 343 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[0973] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[0974] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and specific processing may also be performed by the headset terminal 314.

[0975] [Fourth Embodiment]

[0976] Figure 7 shows an example of the configuration of the data processing system 410 according to the fourth embodiment.

[0977] As shown in Figure 7, the 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.

[0978] The data processing device 12 comprises a computer 22, a database 24, and a communication interface 26. The computer 22 is an example of a "computer" related to the technology of this disclosure. The computer 22 comprises a processor 28, RAM 30, and storage 32. The processor 28, RAM 30, and storage 32 are connected to a bus 34. The database 24 and the communication interface 26 are also connected to the bus 34. The communication interface 26 is connected to a network 54. An example of the network 54 is a WAN (Wide Area Network) and / or a LAN (Local Area Network).

[0979] The robot 414 includes a computer 36, a microphone 238, a speaker 240, a camera 42, a communication interface 44, and a controlled object 443. The computer 36 includes a processor 46, RAM 48, and storage 50. The processor 46, RAM 48, and storage 50 are connected to a bus 52. The microphone 238, speaker 240, camera 42, and controlled object 443 are also connected to the bus 52.

[0980] The microphone 238 receives voice signals from the user 20 and receives instructions from the user 20. The microphone 238 captures the voice signals from the user 20, converts the captured voice into audio data, and outputs it to the processor 46. The speaker 240 outputs audio according to the instructions from the processor 46.

[0981] Camera 42 is a small digital camera equipped with an optical system including a lens, aperture, and shutter, and an image sensor such as a CMOS (Complementary Metal-Oxide-Semiconductor) image sensor or a CCD (Charge Coupled Device) image sensor, and captures images of the area around the user 20 (for example, an imaging range defined by a field of view equivalent to the width of a typical healthy person's field of vision).

[0982] Communication interface 44 is connected to network 54. Communication interfaces 44 and 26 are responsible for the exchange of various information between processor 46 and processor 28 via network 54. The exchange of various information between processor 46 and processor 28 using communication interfaces 44 and 26 is performed in a secure manner.

[0983] The controlled object 443 includes a display device, LEDs in the eyes, and motors that drive 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 robot 414's emotions can be expressed by controlling these motors. Furthermore, the robot 414's facial expressions can also be expressed by controlling the illumination state of the LEDs in its eyes.

[0984] Figure 8 shows an example of the main functions of the data processing device 12 and the robot 414. As shown in Figure 8, the data processing device 12 performs specific processing using the processor 28. The storage 32 stores the specific processing program 56.

[0985] The specific processing program 56 is an example of a "program" relating to the technology of this 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.

[0986] The storage 32 stores the data generation model 58 and the emotion identification model 59. The data generation model 58 and the emotion identification model 59 are used by the identification processing unit 290.

[0987] In robot 414, the processor 46 performs the reception output processing. The storage 50 stores the reception output program 60. 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 processing is realized by the processor 46 operating as a control unit 46A according to the reception output program 60 executed on the RAM 48.

[0988] Next, the specific processing performed by the specific processing unit 290 of the data processing device 12 will be described. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[0989] The system of the present invention aims to automate qualification examinations, with a server at the center executing multiple functions. The specific procedure for implementing the present invention is shown below.

[0990] Collection and preprocessing of qualification exam data

[0991] The server collects data related to certification exams from internal databases within the company. This includes past exam questions, examinee results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. Here, it checks for any missing or abnormal data and supplements or corrects it as needed. This preprocessed data is then used directly as training data for generative AI.

[0992] Learning of generative AI

[0993] The server trains a generative AI based on pre-processed data. During this process, the AI ​​learns the evaluation criteria for certification exams, patterns in past exam questions, and the quality of answers. This creates a foundation for the AI ​​to evaluate test-takers' responses and generate appropriate exam questions.

[0994] Automatic generation of exam questions

[0995] Once the training is complete, the generative AI automatically generates questions for the first and second exams according to the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge. Specific examples include questions such as, "What are the basic steps of the sales process?" The second exam includes scenario-based questions and case studies to assess practical skills. For example, it generates practical questions such as, "What is the best way to handle a customer complaint?"

[0996] Exam distribution and administration

[0997] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers through their device. During the exam, the server receives the examinee's answers in real time and performs automatic evaluation. Additional questions may be automatically generated based on the accuracy rate and quality of the answers. This process continues into the second exam after the first exam is completed.

[0998] Grading and feedback

[0999] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. The automated scoring system evaluates each answer and calculates a total score. Based on the evaluation results, the server generates detailed feedback. This feedback includes an evaluation of each question, the test-taker's strengths, and areas for improvement. The server compiles this feedback for each test-taker and sends it to the user (test-taker) via email.

[1000] Performance data analysis and feedback improvement

[1001] The server collects and analyzes test-taker performance data. This analysis is then used to improve future exam questions and feedback. In addition, the server collects reviews from test-takers and incorporates them as training data for generative AI, thereby improving the overall system quality.

[1002] As described above, the server-centered automated system streamlines the execution of certification exams and enables the rapid provision of high-quality feedback. This system not only saves human resources but also contributes to improved consistency in evaluations and enhanced skills among test-takers.

[1003] The following describes the processing flow.

[1004] Step 1:

[1005] The server collects data related to the certification exam from its internal database. Specifically, it retrieves past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, it checks for any missing or abnormal data and supplements or corrects it as needed. This pre-processed data is then used as training data for generative AI.

[1006] Step 2:

[1007] The server trains the generative AI based on pre-processed data. Specifically, it trains the AI ​​on evaluation criteria for certification exams, patterns of past exam questions, and the quality of answers. Through this learning process, the generative AI builds a foundation for creating appropriate exam questions and evaluating answers.

[1008] Step 3:

[1009] The server uses generative AI to automatically generate questions for the first and second exams. For the first exam, it generates multiple-choice and short-answer questions to assess basic knowledge. For the second exam, it generates scenario-based questions and case studies to assess practical skills. For example, it generates questions such as, "What are the basic steps of the sales process?" and "What is the best way to handle a customer complaint?"

[1010] Step 4:

[1011] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers using their device. During the exam, the server receives the examinee's answers in real time and performs automatic evaluation. Based on this evaluation, additional questions may be automatically generated as needed.

[1012] Step 5:

[1013] Once a user completes the first exam, the server immediately delivers the questions for the second exam. The user then takes the second exam within the same Zoom session. The answers to the second exam are also evaluated in real time, and the evaluation results are continuously fed back to the user.

[1014] Step 6:

[1015] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. Using an automated scoring system, it evaluates the accuracy and quality of each answer and calculates an overall score. Based on the evaluation results, detailed feedback is generated. The feedback includes an evaluation of each question, the test-taker's strengths, and areas for improvement.

[1016] Step 7:

[1017] The server compiles this feedback for each test taker and sends it to the user (test taker) via email. The feedback is also recorded in the company's internal database and used as reference material for future tests and evaluations.

[1018] Step 8:

[1019] The server collects and analyzes test-takers' performance data. This analysis is used to improve future test questions and feedback. Furthermore, the server collects reviews from test-takers and incorporates them as AI training data, thereby improving the overall quality of the system.

[1020] Through the steps outlined above, this system automates the entire certification examination process, reducing human resources and improving the consistency of evaluations. It also enables the rapid provision of high-quality feedback to test-takers.

[1021] (Example 1)

[1022] Next, we will describe Example 1. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1023] Traditional certification examination systems require significant human resources and are inefficient because the creation of exam questions, evaluation of answers, and provision of feedback are all done manually. Furthermore, consistency in evaluation and quality of feedback are prone to inconsistencies. Additionally, the analysis of test-taker performance data and the improvement of feedback based on that analysis are not adequately performed. This hinders the improvement of test-takers' abilities and fair evaluation.

[1024] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 1 is realized by the following means.

[1025] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data to supplement and correct any deficiencies or anomalies, means for training a generative AI based on the preprocessed data, means for automatically generating exam questions for the first and second exams in a specified format, means for distributing the exams and evaluating examinees' answers in real time, means for automatically scoring examinees' answers, generating detailed feedback, and sending it via email, and means for analyzing examinee performance data to improve the next exam questions and feedback. This enables automation and efficiency of the exam, improved consistency of evaluation, and rapid provision of high-quality feedback.

[1026] An "internal database" refers to a collection of information and data related to certification exams that is maintained within a specific organization or system.

[1027] "Qualification exam data" refers to a variety of data related to the exam, including past exam questions, examinee results, evaluation criteria, and examinee performance data.

[1028] "Preprocessing" refers to the process of performing tasks such as imputing missing values, correcting outliers, and formatting the collected data.

[1029] "Generative AI" refers to artificial intelligence systems that learn from collected and pre-processed data and automatically generate new test questions and answers.

[1030] The term "first-stage examination" refers to the initial stage of an exam that includes multiple-choice and short-answer questions designed to assess basic knowledge.

[1031] The term "second examination" refers to the second stage of the examination, which includes scenario-based problems and case studies designed to assess practical skills.

[1032] "Exam delivery" refers to the process of providing exam questions to test takers and having them take the exam.

[1033] "Real-time evaluation" refers to a process where the answers entered by test-takers during the exam are immediately analyzed and evaluated.

[1034] "Automated scoring" refers to a process that automatically evaluates and scores test-takers' answers to exam questions.

[1035] "Feedback" refers to information generated based on a test-taker's exam results, including evaluation results and suggestions for improvement.

[1036] "Performance data" refers to data related to test-takers' behavior during the exam, the quality of their answers, and the accuracy of their responses.

[1037] "Analysis" refers to the process of examining collected data and extracting meaningful patterns and trends.

[1038] The system of this invention aims to automate qualification examinations and performs various processes centered around a server. This system is implemented in the following specific steps.

[1039] Collection and preprocessing of qualification exam data

[1040] The server first connects to an internal database to collect data related to the certification exam. This includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. Specifically, it checks for missing or abnormal data and then fills in or corrects them. For example, missing values ​​are filled in with the mean, and abnormal values ​​are corrected appropriately. Python's Pandas library and SQL queries are used for this preprocessing.

[1041] Learning of generative AI

[1042] The pre-processed data is used as training data for generative AI. The server uses this data to train the AI ​​model. The AI ​​models used include advanced generative AIs such as GPT-4. In this process, deep learning frameworks such as TensorFlow and PyTorch are utilized, and the model learns patterns and evaluation criteria from past exam questions. As a result, the AI ​​acquires the ability to automatically generate the next exam questions.

[1043] Automatic generation of exam questions

[1044] Once the generative AI has completed its training, it automatically generates questions for the first and second exams according to the specified exam format. For example, the first exam generates multiple-choice and short-answer questions to assess basic knowledge. On the other hand, the second exam includes scenario-based questions and case studies to assess practical skills.

[1045] Example of a prompt

[1046] "What are the basic steps of the sales process?"

[1047] "What is the best way to handle a customer complaint that arises during customer service?"

[1048] Exam distribution and administration

[1049] Next, the server handles the delivery and administration of the exam. Users (examinees) log in to ZOOM from their devices at the designated exam date and time. The server delivers the first exam question set through the ZOOM session, and users answer it. Answers are sent to the server in real time, and the server evaluates them immediately. In some cases, additional questions may be automatically generated based on the accuracy rate and quality of the answers.

[1050] Grading and feedback

[1051] After the exam is completed, the server compiles all the answers and evaluates them using an automated scoring system. Based on the evaluation results, detailed feedback is generated and sent to the user (test taker) via email. This feedback includes an evaluation of each question, the test taker's strengths, and areas for improvement.

[1052] Performance data analysis and feedback improvement

[1053] The server collects and analyzes test-takers' performance data. This analysis is used to improve the creation of future test questions and feedback. Furthermore, reviews from test-takers are collected and incorporated as training data for generative AI, thereby improving the overall quality of the system.

[1054] As described above, the present invention's automated qualification examination system automates a series of processes centered on a server, including data collection, preprocessing, training of a generative AI, automatic generation of examination questions, distribution and administration of the examination, scoring and generation of feedback, and analysis and improvement of performance data, thereby realizing efficient and high-quality examination management.

[1055] The flow of the specific processing in Example 1 will be explained using Figure 11.

[1056] Step 1: Start data collection

[1057] The server establishes a connection to the internal database and executes an SQL query to extract "certification exam data" (past exam questions, examinee results, evaluation criteria, etc.). For example, it might execute a query like "SELECT FROM ExamData WHERE Year BETWEEN 2018 AND 2023".

[1058] Input: Connection information for the internal database

[1059] Output: Extraction results of qualification exam data

[1060] Step 2: Data checking and preprocessing

[1061] The server reads the collected data using Python's Pandas library and checks for missing values ​​and outliers. For example, it checks for missing values ​​using df.isnull().sum() and imputes them with df.fillna(df.mean()). If outliers exist, it corrects them using an appropriate method.

[1062] Input: Certification exam data extracted in Step 1

[1063] Output: Preprocessed data

[1064] Step 3: Preparing training data for generative AI

[1065] The server organizes the preprocessed data as training data for generative AI and converts it into an applicable format. For example, it selects the necessary features and converts the data into numerical vectors.

[1066] Input: Preprocessed data obtained in Step 2

[1067] Output: Training dataset

[1068] Step 4: Training the generative AI model

[1069] The server uses deep learning frameworks such as TensorFlow or PyTorch to train generative AI models. For example, it trains the model using the form model.fit(X_train, y_train, epochs=100).

[1070] Input: Training dataset prepared in Step 3

[1071] Output: Trained generative AI model

[1072] Step 5: Automatic generation of primary exam questions

[1073] The server uses a pre-trained generative AI model to automatically generate questions for the first stage of the exam. For example, it might create a multiple-choice question such as, "What are the basic steps of the sales process?"

[1074] Input: Pre-trained generative AI model

[1075] Output: First Exam Question Set

[1076] Step 6: Automatic generation of secondary examination questions

[1077] The server uses a pre-trained generative AI model to automatically generate questions for the second stage of the exam. For example, it creates scenario questions such as, "What is the best way to handle a customer complaint that arises during customer service?"

[1078] Input: Pre-trained generative AI model

[1079] Output: Secondary Exam Question Set

[1080] Step 7: Notification of exam date and time

[1081] The server will send a notification email to the test taker containing the date, time, and link to join the ZOOM session.

[1082] Input: Candidate information, exam schedule

[1083] Output: Notification email

[1084] Step 8: Exam delivery and real-time collection of test-takers' responses

[1085] The user (examinee) logs into ZOOM from their device at the designated date and time. The server delivers the first-stage exam question set via the ZOOM session, and the user enters their answers. The server collects and evaluates the answers sent from the device in real time.

[1086] Input: First-stage exam question set, examinee's answers

[1087] Output: Real-time evaluation results

[1088] Step 9: Generating additional questions based on accuracy and quality.

[1089] The server evaluates the accuracy rate and quality of answers based on the responses received in real time, and generates additional questions as needed.

[1090] Input: Test-takers' responses, real-time evaluation results

[1091] Output: Additional questions

[1092] Step 10: Aggregation of test results and generation of feedback

[1093] After the exam is completed, the server compiles all the answers and calculates an overall score using an automated scoring system. Next, it generates detailed feedback and sends it via email.

[1094] Input: All responses from test takers

[1095] Output: Overall score, feedback email

[1096] Step 11: Accumulate and analyze performance data

[1097] The server collects and analyzes test-takers' performance data. This analysis is used to create future exam questions and improve feedback.

[1098] Input: Test-taker's performance data

[1099] Output: analysis results, improvement suggestions

[1100] Through the processing steps described above, the system of the present invention can automate and streamline qualification examinations and provide high-quality feedback quickly.

[1101] (Application Example 1)

[1102] Next, we will explain Application Example 1. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1103] Traditional certification exams and training programs require significant human resources for tasks such as creating exam questions, administering exams, grading, and providing feedback. This often leads to inefficient exam administration and the risk of biased human evaluation. Furthermore, in on-site employee training, such as in factories, individual evaluation and feedback are difficult to provide, hindering effective training. To address these challenges, an efficient and consistent automated evaluation system is needed.

[1104] The specific processing performed by the specific processing unit 290 of the data processing device 12 in Application Example 1 is realized by the following means.

[1105] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and training a generative AI, means for automatically generating exam questions for the first and second exams, means for distributing the exams and evaluating examinees' answers in real time, means for automatically scoring examinees' answers and generating feedback, means for analyzing examinee performance data and improving the next exam questions and feedback, and means for a robot to generate training questions similar to qualification exams and provide real-time feedback during factory employee training. This streamlines the employee training and evaluation process and enables the provision of consistent, high-quality feedback.

[1106] An "internal database" is a database that stores data related to certification exams and allows users to retrieve that data as needed.

[1107] "Qualification exam data" refers to all data related to the administration of qualification exams, including past exam questions, examinee results, evaluation criteria, and examinee performance data.

[1108] "Preprocessing" is the process of supplementing and correcting any deficiencies or anomalies in the collected data and converting it into a format suitable for learning by generative AI.

[1109] "Generative AI" is a type of artificial intelligence that uses large amounts of data to train models and generate test questions.

[1110] The "first-stage examination" is an exam that includes multiple-choice and short-answer questions designed to assess basic knowledge.

[1111] The "second examination" is a test that includes scenario-based problems and case studies, designed to evaluate practical skills.

[1112] "Automatic scoring" is a process that automatically evaluates the test-takers' answers and calculates their scores.

[1113] "Feedback" refers to providing detailed comments on the test-taker's answers, along with an evaluation of their responses, including their strengths and areas for improvement.

[1114] "Performance data" refers to the overall results a test-taker demonstrated during the exam, along with various analytical data based on those results.

[1115] "Factory employee training" refers to the process of educating and providing factory employees with the necessary skills and knowledge to perform their duties.

[1116] "Real-time evaluation" is a process in which, each time a test-taker submits their answer, it is immediately evaluated and the results are provided.

[1117] "Training problems" are questions or tasks presented to assess an employee's abilities.

[1118] To implement this invention, it is first necessary to collect data related to qualification exams from an internal database. The server retrieves data such as past exam questions, examinee results, and evaluation criteria from the database and preprocesses this data. In this preprocessing, any missing or abnormal data is supplemented and corrected, and the data is standardized.

[1119] Next, the server uses the pre-processed data to train a generative AI. A natural language processing (NLP) model, such as GPT-3, is suitable as a generative AI model. Based on the training data, the AI ​​model learns the question patterns and evaluation criteria of certification exams, enabling the automatic generation of exam questions.

[1120] Once the training is complete, the generative AI generates questions for the first and second exams based on the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge, including specific examples such as "What are basic safety procedures?" The second exam generates scenario questions and case studies to assess practical skills, such as "Explain how to respond if an emergency occurs during work."

[1121] The generated test questions are delivered to factory employees' terminals and robots via a server. Employees provide answers to the terminals or robots, and the server evaluates the answers in real time. An AI-powered automated scoring system is used to evaluate the answers, and feedback is generated based on this evaluation. The feedback includes not only whether the answer is correct or incorrect, but also the employee's strengths and areas for improvement.

[1122] Furthermore, the server collects employee performance data and analyzes it to improve the quality of future exam questions and feedback. This analysis utilizes statistical analysis tools (such as Python's Pandas and NumPy) to identify data trends and develop more effective training programs.

[1123] As a concrete example, a factory robot presents an employee with a training question such as, "Explain how to use the emergency stop button," and the employee answers using voice or controls. The AI ​​evaluates the answer in real time and provides feedback such as, "The procedure itself is correct, but safety checks are missing." In this way, employees can learn practical skills on the spot while being efficiently evaluated.

[1124] Examples of prompt messages include the following:

[1125] Please generate the following training questions regarding factory machine operation:

[1126] 1. Explain how to operate the machine according to the safety procedures.

[1127] 2. Please explain how to use the emergency stop button.

[1128] 3. Please describe the procedure for routine maintenance checks.

[1129] The above describes the specific procedures and system configuration necessary for implementing this invention.

[1130] The flow of a specific process in Application Example 1 will be explained using Figure 12.

[1131] Step 1:

[1132] The server collects data related to certification exams from an internal database. The data retrieved from the database includes past exam questions, examinee results, evaluation criteria, and examinee performance data. All data related to the certification exams is retrieved from the database as input. The collected dataset is passed to a preprocessing module as output.

[1133] Step 2:

[1134] The server preprocesses the collected data. Preprocessing includes imputing missing values, correcting anomalous data, and normalization. The input is the data collected in step 1, and the output is the preprocessed dataset. At this stage, data cleansing and normalization are performed using data preprocessing libraries (e.g., Pandas, NumPy).

[1135] Step 3:

[1136] The server trains a generative AI using preprocessed data. Preprocessed data is used as input, and a generative AI model (e.g., GPT-3) is trained. The output is the trained AI model. Here, the training process is executed using an AI model training framework (e.g., TensorFlow, PyTorch).

[1137] Step 4:

[1138] The server automatically generates questions for the first and second exams using a pre-trained generative AI. The input consists of a pre-trained AI model and prompts, and the output is the automatically generated exam questions. Specifically, the following prompts are input to the AI ​​to generate the questions.

[1139] Please generate the following training questions regarding factory machine operation:

[1140] 1. Explain how to operate the machine according to the safety procedures.

[1141] 2. Please explain how to use the emergency stop button.

[1142] 3. Please describe the procedure for routine maintenance checks.

[1143] Step 5:

[1144] The server distributes the generated test questions to terminals and robots. The input is automatically generated test questions, and the output is the questions distributed to the terminals and robots. The terminals and robots then present these questions to factory workers.

[1145] Step 6:

[1146] The user (factory employee) provides answers to the presented problems using voice or controls. The input is the factory employee's answer, and the output is the answer data sent to the server. The answers are input through a voice recognition system and a sensor network.

[1147] Step 7:

[1148] The server evaluates test-takers' answers in real time. The input is user-submitted answer data, and the output is the evaluation result and feedback. Here, an AI-powered automated scoring system evaluates the answers and provides real-time feedback in text and audio formats.

[1149] Step 8:

[1150] The server collects employee performance data and analyzes it to improve the quality of future training programs and feedback. Real-time evaluation results are used as input, and the output is an improved training program and feedback. Statistical analysis tools (e.g., Python's Pandas, NumPy) are used to perform data analysis and incorporate the findings into future improvements.

[1151] Furthermore, an emotion engine that estimates the user's emotions may be incorporated. That is, the identification processing unit 290 may use the emotion identification model 59 to estimate the user's emotions and perform identification processing using the user's emotions.

[1152] The system of the present invention combines automation of qualification examinations with an emotion engine to recognize user emotions and reflect them in the feedback. Specific embodiments for carrying out the present invention are shown below.

[1153] Collection and preprocessing of qualification exam data

[1154] The server collects data related to certification exams from the company's internal database. This includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. Next, the server preprocesses the collected data. During this process, it checks for missing or abnormal data and supplements or corrects it as needed. The preprocessed data is then used as training data for generative AI.

[1155] Learning of generative AI

[1156] The server trains a generative AI based on pre-processed data. During the training process, the AI ​​learns the evaluation criteria for the certification exam, patterns of past exam questions, and the quality of answers. This allows the AI ​​to build a foundation for accurately creating exam questions and appropriately evaluating answers.

[1157] Automatic generation of exam questions

[1158] Once the training is complete, the generative AI automatically generates questions for the first and second exams according to the specified exam format. The first exam generates multiple-choice and short-answer questions to assess basic knowledge. For example, one question might be, "What are the basic steps of the sales process?" The second exam generates scenario questions and case studies to assess practical skills. For example, one might be, "What is the best way to handle a customer complaint that arises during customer service?"

[1159] Exam distribution and administration

[1160] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers using their device. During the exam, the server receives the examinee's answers in real time and evaluates them using an emotion engine. The emotion engine recognizes emotions from the examinee's facial expressions and voice and uses this to perform real-time evaluations.

[1161] Scoring and feedback generation

[1162] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. Through an automated scoring system, it evaluates the accuracy and quality of each answer and calculates a total score. Based on the evaluation results, the server generates detailed feedback. This feedback includes evaluations of each question, the test-taker's strengths and areas for improvement, as well as comments on emotional states detected by the emotion engine. This feedback more accurately reflects the test-taker's performance and provides information to help improve next time.

[1163] Performance data analysis and improvement

[1164] The server stores and analyzes test-takers' performance data. This analysis is used to create future test questions and improve feedback. Furthermore, by incorporating emotion engine data as AI learning data, the system can provide guidance and support based on the test-takers' emotions.

[1165] Applications of the Emotion Engine

[1166] The emotion engine also detects the test-taker's stress level and concentration level. For example, if the stress level is high, the server will extend the answer time or provide additional support information. By providing questions and feedback that take emotional states into account, it is possible to help test-takers perform at their best.

[1167] This combination of systems allows for a highly automated qualification examination process, reflecting test-takers' emotions and stress levels in real time, thereby providing more appropriate and effective feedback. This not only reduces human resources and improves the consistency of evaluations, but also alleviates the mental burden on test-takers and maximizes learning effectiveness.

[1168] The following describes the processing flow.

[1169] Step 1:

[1170] The server collects data related to the certification exam from its internal database. This includes past exam questions, test-takers' results, and performance data for top and bottom-ranked candidates. Next, it checks for missing or abnormal data and supplements or corrects it as needed.

[1171] Step 2:

[1172] The server trains a generative AI based on pre-processed data. Specifically, it learns the evaluation criteria for certification exams, past exam question patterns, and the quality of answers. Through this process, the generative AI builds the foundation for creating exam questions and evaluating answers.

[1173] Step 3:

[1174] The server uses generative AI to automatically generate questions for the first and second exams. The first exam generates multiple-choice and short-answer questions to assess basic knowledge, while the second exam generates scenario-based questions and case studies to assess practical skills. Examples of such questions include, "What are the basic steps of the sales process?" and "What is the best way to handle a customer complaint?"

[1175] Step 4:

[1176] The user (examinee) logs into ZOOM from their device at the designated exam date and time. The server delivers the first exam question set via the ZOOM session. The user enters their answers on their device.

[1177] Step 5:

[1178] During the exam, the server receives test-takers' answers in real time and evaluates them using an emotion engine. For example, it recognizes emotions from the test-taker's facial expressions and voice via camera and microphone, and analyzes their stress and tension levels.

[1179] Step 6:

[1180] Based on the analysis results from the emotion engine, the server generates additional questions and instructions in real time that take into account the test-taker's emotional state. For example, if a test-taker is feeling stressed, the difficulty level of the test questions may be adjusted or the time allotted for answering may be extended.

[1181] Step 7:

[1182] After the first exam is completed, the server immediately delivers the questions for the second exam. Users continue to answer the second exam questions within the same ZOOM session. The answers to the second exam are also evaluated in real time, and the evaluation results are continuously fed back through the sentiment engine.

[1183] Step 8:

[1184] After the exam is completed, the server compiles all the answers and performs an overall evaluation of the first and second exams. An automated scoring system is used to evaluate the accuracy and quality of each answer and calculate the overall score.

[1185] Step 9:

[1186] The server generates detailed feedback based on the evaluation results. This feedback includes an evaluation of each problem, the test-taker's strengths and areas for improvement, as well as comments on the emotional state detected by the emotion engine. The server compiles this feedback for each test-taker and sends it to the user (test-taker) via email.

[1187] Step 10:

[1188] The server stores and analyzes test-takers' performance data. This analysis is used to create future test questions and improve feedback. Furthermore, incorporating emotion engine data as AI training data enables more accurate emotion recognition and feedback.

[1189] Through these steps, an automated certification exam system incorporating an emotion engine can reflect the test-taker's emotions in real time, providing high-quality feedback while reducing mental stress.

[1190] (Example 2)

[1191] Next, we will describe Example 2. In the following description, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1192] In the process of certification examinations, there is a demand for automation and efficiency, as well as feedback that reflects the emotions and stress levels of test-takers in real time. However, in conventional systems, the generation and evaluation of exam questions are often done manually, which is inefficient. Furthermore, there has been a lack of technology to provide feedback that reflects the emotional state of test-takers. Therefore, providing accurate, real-time evaluation and feedback, and efficiently managing certification examinations are challenges.

[1193] The identification process performed by the identification processing unit 290 of the data processing device 12 in Example 2 is realized by the following means.

[1194] In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and supplementing / correcting missing or abnormal data, means for training a generative AI model based on the preprocessed data, means for automatically generating exam questions for the first and second exams using prompt sentences, means for receiving the exams and examinees' answers in real time and evaluating them using an emotion engine, means for automatically scoring examinees' answers and generating feedback, means for analyzing examinees' performance data and improving the next exam questions and feedback, and means for detecting the examinee's emotional state and evaluating stress levels and concentration levels. This makes it possible to highly automate the qualification exam process and provide accurate and effective feedback that reflects the examinee's emotions in real time.

[1195] An "internal database" is a centrally managed collection of information within a system, and its role is to store and provide data related to certification exams.

[1196] "Qualification exam data" refers to data containing information necessary for qualification exams, specifically including past exam questions, examinee results, evaluation criteria, and performance data.

[1197] "Data preprocessing" is the process of converting collected raw data into a format that can be used for analysis and machine learning, and includes supplementing and correcting missing or abnormal data.

[1198] A "generative AI model" is a type of artificial intelligence technology that learns from data and automatically generates new test questions based on the results of that learning.

[1199] A "prompt sentence" is an input sentence used by a generative AI model, and it is an instruction sentence that the AI ​​uses to generate answers or test questions.

[1200] The "emotion engine" is a system that recognizes and evaluates the emotional state of test takers in real time, detecting emotions based on data such as facial expressions and voice.

[1201] "Real-time evaluation" is a process that involves immediately analyzing and evaluating the test-taker's answers and emotional state during the exam, allowing for feedback and responses without delay.

[1202] "Automated scoring" is a process that uses algorithms or AI to mechanically evaluate test-takers' answers and calculate the accuracy rate and quality.

[1203] "Feedback" refers to evaluation information and suggestions for improvement generated based on the test taker's exam results and performance, and is intended to support the test taker in achieving better results in the next exam.

[1204] "Performance data" refers to data that includes all performance indicators related to the exam, such as the test-taker's response status during the exam, response speed, accuracy rate, and emotional state.

[1205] "Stress level" is an indicator that shows the degree of mental burden on test takers during the exam, and it quantitatively evaluates emotional states such as tension and anxiety.

[1206] "Concentration level" is an indicator used to evaluate the level of attention and focus of test-takers during an exam, and a high level of concentration is considered to lead to high performance.

[1207] This system is designed to highly automate the qualification examination process and provide real-time feedback that reflects the examinee's emotions and stress levels. The specific implementation procedure is as follows.

[1208] Collection and preprocessing of qualification exam data

[1209] The server collects data related to certification exams from the company's internal database. This data includes past exam questions, test-takers' results, evaluation criteria, and performance data for top and bottom-ranked candidates. The collected data is checked for missing or abnormal data and supplemented or corrected as needed. For example, the Python Pandas library is used to impute missing values ​​in the data.

[1210] Learning of generative AI

[1211] The server trains a generative AI model based on pre-processed data. Specifically, it uses past exam questions and test-takers' answers as input data and trains the AI ​​model using machine learning libraries such as TensorFlow and PyTorch. This allows the AI ​​to learn the evaluation criteria for the certification exam, patterns in past exam questions, and the quality of the answers.

[1212] Automatic generation of exam questions

[1213] Once the training is complete, the generative AI will automatically generate questions for the first and second exams according to the specified exam format. The first exam will generate multiple-choice and short-answer questions to assess basic knowledge. For example, it will use prompts like the following:

[1214] "What are the basic steps of the sales process?"

[1215] In the second stage of the exam, scenario-based problems and case studies are generated to assess practical skills. For example,

[1216] "What is the best way to handle a customer complaint that arises during customer service?"

[1217] Exam distribution and administration

[1218] The user (examinee) logs into the online conferencing system from their device at the designated exam date and time. The server delivers the first-stage exam question set through the online conferencing system. The user enters their answers using their device and sends them to the server. During the exam, the server uses Web-RTC technology to capture the examinee's facial expressions and collect data in real time.

[1219] Scoring and feedback generation

[1220] After the exam ends, the server compiles all the answers and begins evaluation through an automated scoring system. It assesses the accuracy and quality of each answer and calculates an overall score. Based on the evaluation results, the server generates detailed feedback. This feedback includes evaluations of each question, the test-taker's strengths and areas for improvement, as well as comments on emotional states detected by the emotion engine.

[1221] Performance data analysis and improvement

[1222] The server stores and analyzes test-takers' performance data. This analysis is used to improve future test questions and feedback. Furthermore, data acquired by the emotion engine is incorporated as AI learning data, enabling guidance and support based on the test-takers' emotions.

[1223] Applications of the Emotion Engine

[1224] The emotion engine detects the test-taker's stress level and concentration level in real time. For example, if the stress level is high, the server will provide an extension of the answer time or additional support information. Specific feedback includes:

[1225] "Your stress level appears high. We have extended your response time by 10 minutes."

[1226] This system, designed in this way, can efficiently automate the qualification examination process, reduce the mental burden on test-takers, and provide accurate and effective feedback.

[1227] The flow of the specific processing in Example 2 will be explained using Figure 13.

[1228] Step 1:

[1229] The server collects past exam questions, test-taker results, evaluation criteria, and performance data for top and bottom-ranked test-takers from an internal database. The data to be collected is easily extracted using SQL queries. It receives connection information to the internal database and data retrieval conditions as input, and outputs the necessary exam data based on that. Specifically, it executes SQL queries to extract the required data and stores it on the server.

[1230] Step 2:

[1231] The server preprocesses the collected data. This includes checking for missing or abnormal data and supplementing or correcting them. It receives the raw data collected in step 1 as input and outputs preprocessed, clean data. Specifically, it uses the Python Pandas library to impute missing values ​​and correct invalid data.

[1232] Step 3:

[1233] The server trains a generative AI model based on preprocessed data. It receives preprocessed data as input and uses it as training data. The output is the trained generative AI model. Specifically, it defines and trains the model using machine learning libraries such as TensorFlow and PyTorch.

[1234] Step 4:

[1235] Once the training is complete, the generative AI model automatically generates questions for the first and second exams. It takes prompts for exam format and basic knowledge questions as input and provides generated exam questions as output. An example of a prompt is "What are the basic steps of the sales process?" In terms of operation, the generative AI is given prompts and generates exam questions based on them.

[1236] Step 5:

[1237] The user (examinee) logs into the online conferencing system from their device at the designated exam date and time. The server distributes the exam questions, and the user enters their answers. The input data consists of the examinee's answers and facial expression data captured in real time. The output consists of the examinee's answer data and emotional state data. Specifically, the answers entered from the device and the facial expression data collected using WEB-RTC technology are sent to the server.

[1238] Step 6:

[1239] After the exam ends, the server collects all the answers and performs automatic scoring and evaluation. It receives the test-takers' answer data as input, evaluates the accuracy and quality of each answer as output, and generates an overall score. Specifically, it uses Python's Scikit-learn to calculate the accuracy of each answer and evaluates them using a dedicated algorithm.

[1240] Step 7:

[1241] The server generates feedback based on the aggregated results. It receives evaluation results and sentiment data from the sentiment engine as input, and generates a feedback document as output. Specifically, it creates detailed feedback including evaluations of each problem, the test-taker's strengths, areas for improvement, and comments on the sentiment state detected by the sentiment engine.

[1242] Step 8:

[1243] The server stores and analyzes test-taker performance data. It receives individual test-taker performance data as input and outputs the next exam questions and feedback for improvement. Specifically, it uses the Python Seaborn library to visualize the data and analyze trends.

[1244] Step 9:

[1245] The server uses an emotion engine to detect the test-taker's stress level and concentration level. It receives real-time facial expression and voice data as input and provides stress level and concentration level as output. Specifically, it uses facial expression and voice analysis algorithms to evaluate the emotional state and adjusts the test environment based on that.

[1246] (Application Example 2)

[1247] Next, we will explain application example 2. In the following explanation, the data processing device 12 will be referred to as the "server" and the robot 414 as the "terminal".

[1248] The current certification examination system suffers from a lack of automation and difficulty in providing feedback that takes into account the emotions and stress levels of test-takers. This increases the mental burden on test-takers, making it difficult for them to perform at their best. Furthermore, the inconsistency in exam evaluation and feedback makes it difficult to accurately identify areas for improvement.

[1249] In Application Example 2, the specific processing performed by the specific processing unit 290 of the data processing device 12 is realized by the following means. In this invention, the server includes means for collecting qualification exam data from an internal database, means for preprocessing the collected data and training a generative AI, means for automatically generating exam questions for the first and second exams, means for distributing the exam and evaluating the examinee's answers in real time, means for automatically scoring the examinee's answers and generating feedback, means for analyzing the examinee's performance data and improving the next exam questions and feedback, and means for detecting the examinee's emotions and stress level using an emotion recognition engine and providing feedback that takes this into account. This makes it possible to automate the exam and provide appropriate feedback according to the examinee's emotions and stress level.

[1250] The "internal database" is a digital repository for collecting and storing qualification exam data, and for retrieving and updating it as needed.

[1251] "Preprocessing" is the process of detecting missing or anomaly data in the collected data, performing necessary completion or correction, and preparing it in a format suitable for training data for generative AI.

[1252] "Generative AI" refers to artificial intelligence that uses machine learning algorithms to learn and automatically generate questions for certification exams based on specific patterns and evaluation criteria.

[1253] The "first-stage examination" refers to a type of examination that includes multiple-choice and short-answer questions used to assess basic knowledge.

[1254] The "second examination" is a type of test that includes scenario-based questions and case studies used to assess practical skills.

[1255] "Real-time evaluation" is a process that instantly evaluates the answers entered by test-takers during the exam and provides necessary feedback on the spot.

[1256] "Automated scoring" is a system that uses generative AI or emotion recognition engines to mechanically evaluate test-takers' answers and calculate an overall score.

[1257] "Feedback" refers to information provided regarding the test taker's answers, including evaluations, comments, areas for improvement, and strengths, which is used as reference for future exams and learning.

[1258] "Performance data" refers to data that indicates a test-taker's ability to perform the test, including the accuracy and quality of answers they provided during the test, as well as their emotional state.

[1259] The "emotion recognition engine" is a technology that analyzes the facial expressions and voice of test takers to detect their emotions and stress levels, and is used for real-time feedback.

[1260] The system of the present invention improves efficiency and the working environment by recognizing the emotions and stress levels of workers in factory work in real time and providing appropriate feedback. Specific embodiments are shown below.

[1261] Data Acquisition and Preprocessing

[1262] The server first collects historical work data and environmental data from its internal database. This data includes past work performance, worker performance data, and operating data of the machinery used. The collected data is preprocessed using libraries such as OpenCV and Dlib. Preprocessing involves imputing missing data and correcting anomalous data, preparing the data as a training dataset for generative AI.

[1263] Learning of generative AI

[1264] The server trains a generative AI based on pre-processed data. During this process, the AI ​​learns evaluation criteria for tasks and past performance patterns, building the foundation necessary for improving work efficiency and generating optimal work processes. OpenAI's GPT model is used as the generative AI to generate specific feedback based on prompt text.

[1265] Use of emotion recognition engine

[1266] The emotion recognition engine is used by the server to analyze the worker's facial expressions and voice data in real time. It utilizes OpenCV and Dlib for facial recognition, Google Cloud Speech-to-Text for speech recognition, and Affectiva and Emotion API (Microsoft) for emotion analysis to detect the worker's emotions and stress levels.

[1267] Provide feedback

[1268] Based on data obtained from the emotion recognition engine, the server uses generative AI to provide appropriate feedback to workers in real time. This feedback covers a wide range of topics, including adjusting work pace, suggesting breaks, and offering advice to reduce stress.

[1269] Specific example:

[1270] For example, if "anger" is detected from the worker's facial expression, the OpenAI GPT model generates feedback using the following prompt message.

[1271] Based on the factory worker's sentiment data: 'Angry', generate appropriate feedback.

[1272] In response, an example of the generated feedback is a comment such as, "You appear to be feeling stressed right now. We recommend taking a 5-minute break."

[1273] Performance data analysis and improvement

[1274] The server accumulates worker performance data daily and performs periodic analysis. This data is used to optimize future work processes and improve feedback. Furthermore, by incorporating data from the emotion recognition engine as AI training data, it becomes possible to provide more precise guidance and support based on the worker's emotions.

[1275] The system described above can improve the efficiency of factory operations, provide appropriate feedback tailored to workers' emotions and stress levels, and significantly improve the work environment.

[1276] The flow of a specific process in Application Example 2 will be explained using Figure 14.

[1277] Step 1:

[1278] The server collects historical work data and environmental data from its internal database and performs preprocessing. In this step, missing and anomalous data are supplemented and corrected, and the data is prepared in a format suitable for use as a training dataset for generative AI. Specifically, OpenCV and Dlib are used to analyze and cleanse the data. The input is data obtained from the internal database, and the output is a preprocessed dataset.

[1279] Step 2:

[1280] The server trains a generative AI using preprocessed data. During the training process, the AI ​​learns evaluation criteria for tasks and past performance patterns. The OpenAI GPT model is used in this step. The input is a preprocessed dataset, and the output is a trained generative AI model.

[1281] Step 3:

[1282] The server uses an emotion recognition engine to analyze the worker's facial expressions and voice data in real time. It uses OpenCV and Dlib for facial recognition, Google Cloud Speech-to-Text for speech recognition, and Affectiva and Emotion API (Microsoft) for emotion analysis. Inputs include camera video and audio data, and output data indicating emotions and stress levels is generated.

[1283] Step 4:

[1284] The server uses generative AI to provide appropriate feedback to workers in real time, based on data obtained from the emotion recognition engine. It creates prompt messages and generates feedback content using OpenAI's GPT model. The input consists of emotion analysis data and prompt messages, and the output is a feedback message. For example, a prompt message such as "Generate appropriate feedback based on the factory worker's emotion data: 'Angry'" might be used.

[1285] Step 5:

[1286] The server provides workers with real-time generated feedback, including suggestions for adjusting work pace, taking breaks, and stress reduction advice. The input is generated feedback messages, and the output is specific actions or messages provided to the worker.

[1287] Step 6:

[1288] The server stores and analyzes worker performance data. This data is used to optimize future work processes and improve feedback. Real-time and historical performance data are used as input, and analysis results and improved feedback are generated as output. By feeding this analysis result back into the generative AI as new training data, more precise guidance and support become possible.

[1289] The specific processing unit 290 transmits the result of the specific processing to the robot 414. In the robot 414, the control unit 46A causes the speaker 240 and the controlled object 443 to output the result of the specific processing. The microphone 238 acquires audio indicating user input for the result of the specific processing. The control unit 46A transmits the audio data indicating user input acquired by the microphone 238 to the data processing unit 12. In the data processing unit 12, the specific processing unit 290 acquires the audio data.

[1290] Data generation model 58 is a type of so-called generative AI (Artificial Intelligence). One example of data generation model 58 is ChatGPT (Internet search<URL: https: / / openai.com / blog / chatgpt> ), Gemini (Internet search) <url: https: gemini.google.com ?hl="ja">Examples of generative AI include the following. The data generation model 58 is obtained by performing deep learning on a neural network. The data generation model 58 is input with prompts containing instructions, and with inference data such as audio data representing speech, text data representing text, and image data representing images. The data generation model 58 infers from the input inference data according to the instructions indicated by the prompts, and outputs the inference results in data formats such as audio data and text data. Here, inference refers to, for example, analysis, classification, prediction, and / or summarization.

[1291] In the above embodiment, an example was given in which specific processing is performed by the data processing device 12, but the technology of this disclosure is not limited thereto, and the specific processing may also be performed by the robot 414.

[1292] Furthermore, the emotion identification model 59, acting as an emotion engine, may determine the user's emotion according to a specific mapping. Specifically, the emotion identification model 59 may determine the user's emotion according to a specific mapping, which is an emotion map (see Figure 9). Similarly, the emotion identification model 59 may also determine the robot's emotion, and the identification processing unit 290 may perform identification processing using the robot's emotion.

[1293] Figure 9 shows an emotion map 400 in which multiple emotions are mapped. In the emotion map 400, emotions are arranged in concentric circles radiating from the center. The closer to the center of the concentric circles, the more primitive the emotions are located. Further out of the concentric circles, emotions representing states and actions arising from mental states are located. Emotion is a concept that includes feelings and mental states. On the left side of the concentric circles, emotions that are generally generated from reactions occurring in the brain are located. On the right side of the concentric circles, emotions that are generally induced by situational judgment are located. Above and below the concentric circles, emotions that are generally generated from reactions occurring in the brain and induced by situational judgment are located. In addition, the emotion of "pleasure" is located on the upper side of the concentric circles, and the emotion of "displeasure" is located on the lower side. Thus, in the emotion map 400, multiple emotions are mapped based on the structure in which emotions arise, and emotions that are likely to occur simultaneously are mapped close together.

[1294] These emotions are distributed at the 3 o'clock position on the Emotion Map 400, and usually fluctuate between feelings of security and anxiety. In the right half of the Emotion Map 400, situational awareness takes precedence over internal feelings, resulting in a calm impression.

[1295] The inside of the Emotion Map 400 represents inner thoughts, while the outside represents actions. Therefore, the further you go from the outside of the Emotion Map 400, the more visible (expressed in actions) your emotions become.

[1296] Here, human emotions are based on various balances, such as posture and blood sugar levels. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. Similarly, in robots, cars, motorcycles, etc., emotions can be created based on various balances, such as posture and battery level. When these balances deviate from the ideal, it results in discomfort, and when they approach the ideal, it results in pleasure. The emotion map can be generated, for example, based on Dr. Mitsuyoshi's emotion map (Research on a system for analyzing brain physiological signals of speech emotion recognition and emotion, Tokushima University, doctoral dissertation: https: / / ci.nii.ac.jp / naid / 500000375379). The left half of the emotion map contains emotions belonging to a region called "response," where sensation is dominant. The right half of the emotion map contains emotions belonging to a region called "situation," where situational awareness is dominant.

[1297] The emotion map defines two emotions that promote learning. One is the emotion around the middle of the negative "repentance" and "reflection" on the situation side. In other words, it is when the robot experiences negative emotions such as "I never want to feel this way again" or "I don't want to be scolded again." The other is the emotion around the positive "desire" on the reaction side. In other words, it is when the robot has positive feelings such as "I want more" or "I want to know more."

[1298] The emotion identification model 59 inputs user input into a pre-trained neural network, obtains emotion values ​​representing each emotion shown in the emotion map 400, and determines the user's emotion. This neural network is pre-trained based on multiple training data sets, which are combinations of user input and emotion values ​​representing each emotion shown in the emotion map 400. Furthermore, this neural network is trained so that emotions located close together have similar values, as shown in the emotion map 900 in Figure 10. Figure 10 shows an example where multiple emotions such as "reassured," "calm," and "confident" have similar emotion values.

[1299] The above description primarily focuses on the functions of the data processing device 12 in relation to this disclosure. However, the system related to this disclosure is not necessarily implemented on a server. The system related to this disclosure may be implemented as a general information processing system. This disclosure may be implemented, for example, as a software program that runs on a personal computer or as an application that runs on a smartphone. The method related to this disclosure may be provided to users in SaaS (Software as a Service) format.

[1300] In the above embodiment, an example was given in which a specific process is performed by a single computer 22. However, the technology of this disclosure is not limited thereto, and a distributed processing of the specific process may be performed by multiple computers, including computer 22. For example, a data generation model 58 may be provided in an external device of the data processing device 12, and the external device may generate data according to the input data.

[1301] In the above embodiment, an example was given in which the specific processing program 56 is stored in the storage 32, but the technology of this disclosure is not limited thereto. For example, the specific processing program 56 may be stored in a portable, computer-readable, non-temporary storage medium such as a USB (Universal Serial Bus) memory. The specific processing program 56 stored in the non-temporary storage medium is installed in the computer 22 of the data processing device 12. The processor 28 executes specific processing according to the specific processing program 56.

[1302] Alternatively, the specific processing program 56 may be stored in a storage device such as a server connected to the data processing device 12 via the network 54, and the specific processing program 56 may be downloaded and installed on the computer 22 in response to a request from the data processing device 12.

[1303] Furthermore, it is not necessary to store the entirety of the specific processing program 56 in a storage device such as a server connected to the data processing device 12 via the network 54, or to store the entirety of the specific processing program 56 in the storage 32; it is acceptable to store only a portion of the specific processing program 56.

[1304] The following types of processors can be used as hardware resources to perform specific processing. Examples of processors include a CPU, a general-purpose processor that functions as a hardware resource to perform specific processing by executing software, i.e., a program. Other examples of processors include dedicated electrical circuits, such as FPGAs (Field-Programmable Gate Arrays), PLDs (Programmable Logic Devices), or ASICs (Application Specific Integrated Circuits), which have circuit configurations specifically designed to perform specific processing. All of these processors have built-in or connected memory, and all of them perform specific processing by using memory.

[1305] The hardware resource that performs a specific process may consist of one of these various processors, or it may consist of a combination of two or more processors of the same or different types (for example, a combination of multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, the hardware resource that performs a specific process may consist of a single processor.

[1306] Examples of configurations using a single processor include, firstly, a configuration in which one or more CPUs and software are combined to form a single processor, and this processor functions as a hardware resource that performs a specific process. Secondly, there is a configuration using a processor that realizes the functions of the entire system, including multiple hardware resources that perform a specific process, on a single IC chip, as exemplified by SoCs (System-on-a-chip). In this way, a specific process is realized using one or more of the above types of processors as hardware resources.

[1307] Furthermore, the hardware structure of these various processors can more specifically utilize electrical circuits that combine circuit elements such as semiconductor devices. Also, the specific processing described above is merely an example. Therefore, it goes without saying that unnecessary steps can be deleted, new steps added, or the processing order rearranged, as long as it does not deviate from the main purpose.

[1308] The descriptions and illustrations presented above are detailed explanations of the technical aspects of this disclosure and are merely examples of the technical aspects. For example, the above descriptions of the structure, function, operation, and effect are examples of the structure, function, operation, and effect of the technical aspects of this disclosure. Therefore, it goes without saying that you may delete unnecessary parts, add new elements, or replace elements in the descriptions and illustrations presented above, as long as you do not deviate from the essence of the technical aspects of this disclosure. Furthermore, in order to avoid confusion and facilitate understanding of the technical aspects of this disclosure, explanations of common technical knowledge and the like that do not require special explanation to enable the implementation of the technical aspects of this disclosure have been omitted from the descriptions and illustrations presented above.

[1309] All documents, patent applications, and technical standards described herein are incorporated by reference to the same extent as if each individual document, patent application, and technical standard were specifically and individually noted as being incorporated by reference.

[1310] The following is further disclosed regarding the embodiments described above.

[1311] (Claim 1)

[1312] A means of collecting qualification exam data from an internal database,

[1313] A method for preprocessing collected data and training a generative AI,

[1314] A means for automatically generating examination questions for the first and second examinations,

[1315] A means of distributing the exam and evaluating the examinees' answers in real time,

[1316] A means of automatically grading test-takers' answers and generating feedback,

[1317] Analyzing test-taker performance data to improve future exam questions and feedback,

[1318] A system that includes this.

[1319] (Claim 2)

[1320] The system according to claim 1, which collects qualification exam data from an internal database and supplements and corrects missing or abnormal data.

[1321] (Claim 3)

[1322] The system according to claim 1, which generates questions for the first examination and simultaneously creates explanations for the correct answers to each question.

[1323] "Example 1"

[1324] (Claim 1)

[1325] A means of collecting qualification exam data from an internal database,

[1326] A means of preprocessing the collected data and supplementing or correcting any deficiencies or anomalies,

[1327] A method for training a generative AI based on preprocessed data,

[1328] A means for automatically generating examination questions for the first and second examinations in a specified format,

[1329] A means of distributing the exam and evaluating the examinees' answers in real time,

[1330] A method for automatically grading test-takers' answers, generating detailed feedback, and sending it via email,

[1331] Analyzing test-taker performance data to improve future exam questions and feedback,

[1332] A system that includes this.

[1333] (Claim 2)

[1334] The system according to claim 1, which collects qualification exam data from an internal database and supplements and corrects missing or abnormal data.

[1335] (Claim 3)

[1336] The system according to claim 1, which generates questions for the first and second examinations and simultaneously creates explanations for the corresponding correct answers.

[1337] "Application Example 1"

[1338] (Claim 1)

[1339] A means of collecting qualification exam data from an internal database,

[1340] A method for preprocessing collected data and training a generative AI,

[1341] A means for automatically generating examination questions for the first and second examinations,

[1342] A means of distributing the exam and evaluating the examinees' answers in real time,

[1343] A means of automatically grading test-takers' answers and generating feedback,

[1344] Analyzing test-taker performance data to improve future exam questions and feedback,

[1345] A means by which a robot generates training problems similar to qualification exams and provides real-time feedback during the training of factory workers,

[1346] A system that includes this.

[1347] (Claim 2)

[1348] The system according to claim 1, which collects qualification exam data from an internal database and supplements and corrects missing or abnormal data.

[1349] (Claim 3)

[1350] The system according to claim 1, which generates questions for the first examination and simultaneously creates explanations for the correct answers to each question.

[1351] "Example 2 of combining an emotion engine"

[1352] (Claim 1)

[1353] A means of collecting qualification exam data from an internal database,

[1354] A means of preprocessing the collected data and supplementing or correcting missing or abnormal data,

[1355] A method for training a generative AI model based on preprocessed data,

[1356] A means for automatically generating examination questions for the first and second examinations using prompt statements,

[1357] A means of receiving the exam delivery and examinee answers in real time and evaluating them using an emotion engine,

[1358] A means of automatically grading test-takers' answers and generating feedback,

[1359] Analyzing test-taker performance data to improve future exam questions and feedback,

[1360] A means to detect the emotional state of test takers and evaluate their stress levels and concentration levels,

[1361] A system that includes this.

[1362] (Claim 2)

[1363] The system according to claim 1, which collects qualification exam data from an internal database and supplements and corrects missing or abnormal data.

[1364] (Claim 3)

[1365] The system according to claim 1, which generates questions for the first and second examinations and simultaneously creates explanations for the correct answers to each question.

[1366] "Application example 2 when combining with an emotional engine"

[1367] (Claim 1)

[1368] A means of collecting qualification exam data from an internal database,

[1369] A method for preprocessing collected data and training a generative AI,

[1370] A means for automatically generating examination questions for the first and second examinations,

[1371] A means of distributing the exam and evaluating the examinees' answers in real time,

[1372] A means of automatically grading test-takers' answers and generating feedback,

[1373] Analyzing test-taker performance data to improve future exam questions and feedback,

[1374] A means of detecting the test-taker's emotions and stress level using an emotion recognition engine and providing feedback that takes this into account,

[1375] A system that includes this.

[1376] (Claim 2)

[1377] The system according to claim 1, which collects qualification exam data from an internal database and supplements and corrects missing or abnormal data.

[1378] (Claim 3)

[1379] The system according to claim 1, which generates questions for the first examination and simultaneously creates explanations for the correct answers to each question. [Explanation of Symbols]

[1380] 10, 210, 310, 410 Data Processing Systems 12 Data Processing Devices 14 Smart Devices 214 Smart Glasses 314 Headset-type terminal 414 Robots< / url:> < / url:> < / url:> < / url:>

Claims

1. A means of collecting qualification exam data from an internal database, A method for preprocessing collected data and training a generative AI, A means for automatically generating examination questions for the first and second examinations, A means of distributing the exam and evaluating the examinees' answers in real time, A means of automatically grading test-takers' answers and generating feedback, Analyzing test-taker performance data to improve future exam questions and feedback, A system that includes this.

2. The system according to claim 1, which collects qualification exam data from an internal database and supplements or corrects missing or abnormal data.

3. The system according to claim 1, which generates questions for the first examination and simultaneously creates explanations for the correct answers to each question.

Images